ArticlePDF AvailableLiterature Review

The Search for True Numbers of Neurons and Glial Cells in the Human Brain: A Review of 150 Years of Cell Counting: Quantification of Neurons and Glia in Human Brain


Abstract and Figures

For half a century, the human brain was believed to contain about 100 billion neurons and one trillion glial cells, with a glia:neuron ratio of 10:1. A new counting method, the isotropic fractionator, has challenged the notion that glia outnumber neurons and revived a question that was widely thought to have been resolved. The recently validated isotropic fractionator demonstrates a glia:neuron ratio of less than 1:1 and a total number of less than 100 billion glial cells in the human brain. A survey of original evidence shows that histological data always supported a 1:1 ratio of glia to neurons in the entire human brain, and a range of 40-130 billion glial cells. We review how the claim of one trillion glial cells originated, was perpetuated, and eventually refuted. We compile how numbers of neurons and glial cells in the adult human brain were reported and we examine the reasons for an erroneous consensus about the relative abundance of glial cells in human brains that persisted for half a century. Our review includes a brief history of cell counting in human brains, types of counting methods that were and are employed, ranges of previous estimates, and the current status of knowledge about the number of cells. We also discuss implications and consequences of the new insights into true numbers of glial cells in the human brain, and the promise and potential impact of the newly validated isotropic fractionator for reliable quantification of glia and neurons in neurological and psychiatric diseases. This article is protected by copyright. All rights reserved.
Content may be subject to copyright.
The Search for True Numbers of Neurons and Glial Cells in the
Human Brain: A Review of 150 Years of Cell Counting
Christopher S. von Bartheld1,*, Jami Bahney1, and Suzana Herculano-Houzel2
1Department of Physiology and Cell Biology, University of Nevada School of Medicine, Reno, NV
89557, USA
2Instituto de Ciências Biomédicas, Universidade Federal do Rio de Janeiro, and Instituto Nacional
de Neurociência Translacional, CNPq/MCT, Brasil
For half a century, the human brain was believed to contain about 100 billion neurons and one
trillion glial cells, with a glia:neuron ratio of 10:1. A new counting method, the isotropic
fractionator, has challenged the notion that glia outnumber neurons and revived a question that was
widely thought to have been resolved. The recently validated isotropic fractionator demonstrates a
glia:neuron ratio of less than 1:1 and a total number of less than 100 billion glial cells in the
human brain. A survey of original evidence shows that histological data always supported a 1:1
ratio of glia to neurons in the entire human brain, and a range of 40–130 billion glial cells. We
review how the claim of one trillion glial cells originated, was perpetuated, and eventually refuted.
We compile how numbers of neurons and glial cells in the adult human brain were reported and
we examine the reasons for an erroneous consensus about the relative abundance of glial cells in
human brains that persisted for half a century. Our review includes a brief history of cell counting
in human brains, types of counting methods that were and are employed, ranges of previous
estimates, and the current status of knowledge about the number of cells. We also discuss
implications and consequences of the new insights into true numbers of glial cells in the human
brain, and the promise and potential impact of the newly validated isotropic fractionator for
reliable quantification of glia and neurons in neurological and psychiatric diseases.
Glia number; Neuron number; Glia-neuron ratio; Cell counts; Human brain; Quantification;
*CORRESPONDENCE TO: Christopher von Bartheld, Department of Physiology and Cell Biology, Mailstop 352, University of
Nevada School of Medicine, Reno, NV 89557 (USA), Phone: (775) 784-6022, FAX: (775) 784-6903,
There are no conflicts of interest.
All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data
analysis. Study concept and design: CSvB and SHH. Acquisition of data: CSvB and JB. Analysis and interpretation of data: CSvB and
SHH. Drafting of the manuscript: CSvB. Critical revision of the manuscript for important intellectual content: CSvB and SHH.
Obtained funding: CSvB, JB and SHH. Administrative, technical, and material support: JB. Study supervision: CSvB.
HHS Public Access
Author manuscript
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Published in final edited form as:
J Comp Neurol
. 2016 December 15; 524(18): 3865–3895. doi:10.1002/cne.24040.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
“More attention must … be paid to quantitative studies of neuroglia and nerve cells,
as opinions are often conflicting and frequently based on faulty technique.”
Paul Glees, mentor of celebrities Paul Wall and Oliver Sacks (Wall, 2001) in his
foreword to “Neuroglia”, page ix (Glees, 1955)
Quantification of cells and their ratios in the nervous system is considered an important
approach to understand the cellular composition, development, and evolution of the brain,
neurological and psychiatric diseases, and aging (Coggeshall and Lekan, 1996; Morrison
and Hof, 1997; Azevedo et al., 2009; Hilgetag and Barbas, 2009; Lent et al., 2012; Yuhas
and Jabr, 2012; Herculano-Houzel, 2009, 2014; Geuna and Herrera-Rincon, 2015).
Quantification adds an essential, new dimension to the topic of investigation, as famously
expressed by Lord Kelvin (Thomson, 1889; von Bartheld and Wouters, 2015). Recent
studies have shown that the cellular composition of the human brain is very different than
was believed and taught for almost half a century (Azevedo et al., 2009; Hilgetag and
Barbas, 2009; Lent et al., 2012; Yuhas and Jabr, 2012; Herculano-Houzel, 2009, 2014). A
major motivation for our work is to provide a comprehensive analysis of the events and
circumstances that delayed recognition of the true cellular composition of the human brain.
We envision that our review will be utilized in multiple ways. Foremost, our review
examines from a historical perspective the efforts that have been made to estimate and report
cell numbers and ratios in the human brain. As such, it reviews the origin, perpetuation, and
recent refutation of the claim of one trillion glial cells, compares different counting methods,
and emphasizes the importance of proper citation of relevant previous work. We attempt to
provide a comprehensive account of previous studies that quantified cells in the human
brain, to serve as a useful reference for current and future investigations.
Cell counting in the human brain has had a complex history. Cells in the brain can be
quantified and reported in three different ways: Total neuron numbers; total glia numbers;
and the ratio of glia to neurons (“GNR”), which refers not only to astrocytes but to all glial
cells (astrocytes, oligodendrocytes and microglia) in the tissue. Historically, these three ways
of numerical accounting have followed surprisingly distinct trajectories that seemed to co-
exist, on superficial inspection, in agreement. Although they are linked in a simple
mathematical formula (G/N = GNR, where G is the number of glia, N is the number of
neurons, and GNR is the ratio of G/N for any given structure), this relationship was
neglected on multiple occasions.
Brain cell counting can be roughly divided into three historical phases. In the first phase,
data were collected only for parts of the human brain, in particular the cerebral cortex. Some
investigators admitted uncertainty about absolute numbers for the whole brain, while others
calculated or postulated GNRs for the whole brain (Hyden, 1960; Kuffler and Nicholls,
1976; Kandel and Schwartz, 1981). This phase lasted until about the 1970s. A second phase
witnessed the first publications of serious estimates of total numbers, for both glial cells
(40–130 billion: compiled by Blinkov and Glezer, 1968, and Haug, 1986) and neurons (70–
85 billion: compiled by Haug, 1986, also reviewed in Williams and Herrup, 1988). Even
von Bartheld et al. Page 2
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
though these cell density-based estimates supported a total GNR of about 1:1, this was either
not recognized or not effectively communicated, thus allowing statements of a 10:1 or 50:1
GNR in major textbooks and reviews to remain essentially unchallenged from the 1960s
until 2009 (Phase 2, Kandel et al., 1991, 2000; Nicholls et al., 1992; Bear et al., 2001, 2007).
In this phase, most textbooks reiterated the view of a 10:1 abundance of glia, while
neglecting the few, but existing published primary data that conflicted with this notion. The
10:1 GNR had – prematurely and mistakenly – attained the status of “common knowledge.”
The third and most recent phase began with the study by Azevedo et al. (2009) that revealed
the discrepancy with “textbook knowledge” and essentially confirmed the numbers
published by Blinkov and Glezer (1968) and Haug (1986).
There was a disconnect between published reports on actual counts of cells in the human
brain, and how such numbers were reported in review articles and text books.
Inconsistencies in reports of neuron content in the human brain were first documented for
psychology textbooks and reviews in the 1980s (Soper and Rosenthal, 1988). We here
provide a similar analysis for neuroscience reviews and textbooks, but we compile, besides
neuron counts, also reports about glia counts and the GNR, and add trends and insights from
a more longitudinal, long-term perspective over several decades.
We also review the different types of counting methods that have been developed and have
been employed for estimating cell numbers in human brains. Numerical ranges based on
these different methods will be discussed, as well as the advantages and limitations of each
of these methods.
With the benefit of hindsight, we examine the origin of the claim of a 10:1 or 50:1 glia-
neuron ratio (GNR), with a corresponding total number of between 1 and 5 trillion glial cells
in the human brain. We also examine reasons for the longevity (more than half a century) of
the notion of one trillion or more glial cells in human brains. Surprisingly, the main reason
for the origin and persistence of the notion of one trillion glial cells was not the technical
disadvantage of the histological (and other) counting methods for global cell counts in
heterogeneous tissue, but rather the failure to notice that published numbers for all three
components: neuron counts, glia counts and the assumed GNR of 10:1 contradicted each
other, and therefore one or more components had to be false. Major textbooks consistently
presented the notion as a well-established fact, thereby allowing circumvention of the
normal mechanism of peer validation of new claims. Additional sections give examples of
the impact of cell counting and discuss the potential role of the new counting method, the IF,
on obtaining and verifying glial and neuronal cell numbers and their ratios in human
diseases. For reasons of space, we restrict our review primarily to the literature on cell
counts in adult human brains.
It is useful to briefly review the three types of counting methods that have been employed to
quantify cells in the human brain. The unit that is being counted is the cell body with its
nucleus, the building blocks of the brain. For the purpose of this review, we do not take into
account that neurons have different sizes, shapes, or their varying dendritic or axonal
von Bartheld et al. Page 3
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
morphology, or that they belong to different cell types. To determine numbers of glial cells,
most studies have similarly combined astrocytes, oligodendrocytes and microglia. Thus, the
GNR reflects the ratio of numbers of all glial cells to all neurons in a structure, regardless of
their sizes. The three different principal approaches to estimate the number of cells in the
brain are: (1) Either model-based or design-based counting of stained cells, nuclei or
nucleoli or their fragments in histological sections; (2) DNA extraction and measurement of
total DNA content to calculate cell numbers; and (3) “direct enumeration” of cells in
homogenized brain tissue by counting cell nuclei in suspension (a rudimentary precursor of
the isotropic fractionator), and the isotropic fractionator itself.
This is the most often used approach, and it has been detailed in numerous reports
(Abercrombie, 1946; Ebbesson and Tang, 1965; Cragg, 1967; Blinkov and Glezer, 1968;
Konigsmark, 1970; Haug et al., 1984; Haug, 1987; Howard and Reed, 1998; Schmitz and
Hof, 2005; Lyck et al., 2009). Tissues are fixed, usually in a formaldehyde-based fixative,
embedded in a suitable medium, sectioned into thin slices, stained with a dye, and cells or
subcellular particles are counted under the microscope (Fig. 1). There are two major types of
the histology approach: model-based and design based. The traditional model-based
approach (profile counting) relies on analysis of thin sections (of 5–15 microns thickness),
spaced 10 or 20 sections apart. Subcellular particles (usually nuclei or nucleoli) are counted
in those thin sections, then one extrapolates for the sections in between the ones used for
counting, and applies correction factors to account for the fact that larger particles appear in
multiple sections (Abercrombie, 1946; Ebbesson and Tang, 1965; Blinkov and Glezer, 1968;
Konigsmark, 1970; Clarke and Oppenheim, 1995). This requires knowledge or assumptions
about the size and shape of particles. The design-based approach (stereology) uses thicker
sections of 20–100 microns, takes random samples within these sections so that the samples
are representative of the particle density, and applies the random sampling scheme to the
entire reference space (Haug et al., 1984; Haug, 1987; Gundersen et al., 1988; Williams and
Rakic, 1988; Howard and Reed, 1998; Schmitz and Hof, 2005). Such a method is unbiased
in theory, although bias can arise due to tissue deformation and loss of particles during tissue
processing and other errors (von Bartheld, 1999, 2002; Guillery, 2002). For this reason,
investigators have recommended calibration of both methods against the ultimate standard,
i.e. 3-dimensional serial section reconstructions of an entire region or a sample thereof
(Coggeshall et al., 1990; Hatton and von Bartheld, 1999; von Bartheld, 2001; von Bartheld,
2002; Williams et al., 2003; Kaplan et al., 2010).
Major challenges of the histological approach are to make sure that samples are truly
representative of the reference volume, to prevent double counting of particles that appear in
multiple sections, to account for differential shrinkage that changes with age and tissue
composition (white matter vs. grey matter), to distinguish correctly between neurons and
glia (Fig. 1) (discussed in more detail below), to identify the true borders and dimensions of
the reference volume, and to measure the true height of tissue sections (von Bartheld, 2001,
2002; Guillery, 2002; Schmitz and Hof, 2005). The importance of counting absolute
numbers of cells rather than cell densities was underscored by the finding that tissues shrink
differentially with age (Haug et al., 1984). Neglect of the fact of differential shrinkage of
von Bartheld et al. Page 4
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
brain tissue with age led to the false belief that neuron number declines steadily and
significantly in human brains during normal aging (Brody, 1955; Haug et al., 1984;
Morrison and Hof, 1997; Peters et al., 1998; Mouton, 2002; Peters, 2002). It is important to
assess absolute numbers of cells as opposed to cell densities within unclear reference
volumes – densities can be misleading when such volumes change due to confounding
variables – and can give rise to the so-called “reference trap” (Haug et al., 1984; West,
1993a; Mayhew and Gundersen, 1996; Mouton, 2002). The histology/stereology approach is
considered a valuable method for analysis of well-defined regions with precise borders, but
has limitations when large tissues with heterogeneous composition or components and/or
fuzzy borders need quantification (Peters et al., 1998; Benes and Lange, 2001; Lent et al.,
2012; Herculano-Houzel et al., 2015).
DNA extraction
An alternative approach to histology is to extract and measure DNA content and to calculate
cell numbers based on knowledge of DNA content per cell nucleus (Heller and Elliott, 1954;
Hess, 1961; Zamenhof et al., 1964; Margolis, 1969; Bass et al., 1971; Hess and Thalheimer,
1971; Dobbing and Sands, 1973; Zamenhof, 1976; Mares et al., 1985; Jacobson, 1991;
Yuhas and Jabr, 2012). However, this technique also has its drawbacks: complete recovery of
DNA is required; there can be contamination with other nucleic acids; not all cells are
euploid, and only total cell number, but not cell type is revealed.
DNA extraction has been used mostly in the 1950s through the 1970s, primarily to
determine changes or trends, by applying the known amount of DNA per cell nucleus in a
given species and to make relative comparisons rather than to obtain absolute numbers
(Robins et al., 1956; Hess, 1961; Zamenhof et al., 1964; Margolis, 1969; Hess and
Thalheimer, 1971; Dobbing and Sands, 1973; Zamenhof, 1976; Jacobson, 1991). Some of
these studies compared DNA content in primate cortex with glial and neuronal densities as
obtained by histological techniques (Brizzee et al., 1964; Cragg, 1967; Bass et al., 1971;
Ling and Leblond, 1973; Leuba and Garey, 1989), but these comparisons were done in
animal models, and not in the human brain. While theoretically an elegant solution
(Jacobson, 1991), the DNA approach has been criticized for a number of reasons, as recently
compiled (Bahney and von Bartheld, 2014): (1) many initial reports relied on DNA-P
measurement, but P may not necessarily be derived exclusively from DNA (Drasher, 1953);
(2) it requires complete DNA extraction and recovery; (3) there are concerns that the large
and more fragile neuronal nuclei may be preferentially destroyed during the isolation
procedures (Nurnberger and Gordon, 1957); (4) mitochondria also contain a small amount of
DNA (Nass and Nass, 1963); (5) DNA extraction is problematic when lipids and
lipoproteins are abundant in the tissue of interest, as is the case in white matter (Zamenhof et
al., 1964; Penn and Suwalski, 1969; Saldanha et al., 1984); (6) aldehyde fixation causes
DNA denaturation (Srinivasan et al., 2002) and possibly irreversible crosslinking of peptides
to DNA, thereby decreasing the yield of DNA that can be measured by spectrophotometry
(Savioz et al., 1997); (7) euploidy in brain cells is assumed, yet as many as 20% of adult
human neurons are hyperploid (Mosch et al., 2007).
von Bartheld et al. Page 5
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Despite these caveats, some of the reports based on the DNA method were used to support
notions about human cell numbers or GNRs (Nurnberger and Gordon, 1957, Discussion
pages 129–138; Hess, 1961; Hess and Thalheimer, 1971; Yuhas and Jabr, 2012), and
therefore contributed to the development of an apparent consensus about the GNR.
Homogenization and counting cells in suspension (“brain soup”) – also called “direct
enumeration” and more recently “isotropic fractionator”
This approach was originally designed in the 1950s (Nurnberger and Gordon, 1957;
Nurnberger, 1958). Dissected chilled tissue was weighed, homogenized, diluted in a known
volume of medium, stained with methylene blue, mixed, and aliquots of the diluted medium
were counted on a hemocytometer. The original paper suggested that neuronal nuclei could
be distinguished from vascular and glial cell nuclei on the basis of centrally located single
nucleoli as opposed to multiple eccentric nucleoli, and differences in intensity of staining
(Nurnberger and Gordon, 1957). However, these and subsequent investigators (Brizzee et al.,
1964) also stated that nuclei of small neurons (such as cerebellar granule cells) were
misidentified as glial cells (page 112), so that the neuron counts may be too low, in
particular for the cerebellum. The original version of the “direct enumeration” method
suffered from several shortcomings, primarily rapid degradation of unfixed cells and lack of
distinction between cell types, and therefore it was rarely applied. Comparisons with
histological cell counts on rat, monkey and human brains revealed discrepancies, and it
remained unclear how to resolve them (Nurnberger and Gordon, 1957; Brizzee et al., 1964).
Subsequent modifications introduced a formalin fixation step for the dissected tissue, used
disintegration in water, ultrasonication, followed by dilution, resuspension and staining with
thionine (Zamenhof, 1976; Zamenhof and Klimuszko, 1977). These modifications allowed
to easily recognize larger cerebellar neurons, but the distinction between granule cells and
glial cells remained problematic. Comparison with histological counts suggested that
numbers obtained with the “direct enumeration” method were too low, by at least one third
(Clarke and Oppenheim, 1995), possibly due to rupturing of cells during the mechanical
disintegration and sonication steps.
Without knowledge of Zamenhof’s attempts to improve Nurnberger’s method, significant
further modifications of this method were introduced in 2005, and the greatly improved
method was called the “isotropic fractionator” (Herculano-Houzel and Lent, 2005; Zorzetto,
2012) (Fig. 2). The new modifications included fixation of animal brains by perfusion with
buffered 4% paraformaldehyde of tissues or immersion fixation of human brains, followed
by perfusion through the carotid arteries within 24 hours post mortem (Azevedo et al., 2009;
Andrade-Moraes et al., 2013), detergent-assisted mechanical dissociation, centrifugation to
collect nuclei in the pellet, visualization of nuclei with a fluorescent nuclear stain (4,6-
Diamidino-2-Phenylindole, Dihydrochloride, DAPI), and distinction between neuronal and
non-neuronal cell nuclei by use of a neuron-specific antibody, anti-NeuN (Herculano-Houzel
and Lent, 2005). This solved some of the major limitations of previous versions of this
approach. Furthermore, the method has the advantages of being easy, fast, and accurate,
generating estimates of numbers of cells that are independent of tissue volume or cell
density, and overcoming problems of heterogeneity of tissues. However, there are also
von Bartheld et al. Page 6
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
limitations of the IF: the use of antibodies against nuclear proteins (to distinguish neurons
from non-neuronal cells) does not identify cell types among the non-neuronal cells, NeuN
antigens are not expressed by a small number of neuronal populations (Mullen et al., 1992),
and only regions and volumes of tissues that can be dissected macroscopically can be
analyzed (Lent et al., 2012). Automated versions of the IF have been reported, both for the
homogenization procedure (Azevedo et al., 2013) and for the counting procedure, using flow
cytometry (Collins et al., 2010; Young et al., 2012; Herculano-Houzel et al., 2015). Long-
standing concerns about loss of nuclei when using a biochemical homogenization approach
(Brizzee et al., 1964; Hadjiolov et al., 1965; Lovtrup-Rein and McEwen, 1966; Cragg, 1967;
Kato and Kurokawa, 1967; Clarke and Oppenheim, 1995; Yuhas and Jabr, 2012; Carlo and
Stevens, 2013; Verkhratsky and Butt, 2013; Charvet et al., 2015) have recently been
addressed and dispelled in two studies that directly compared the IF, in side-by-side
experiments, with results obtained by stereology (Bahney and von Bartheld, 2014; Miller et
al., 2014). These studies, as well as others (Brautigam et al., 2012; Andrade-Moraes et al.,
2013; Walloe et al., 2014) indicate equivalency between the IF and stereology (Herculano-
Houzel et al., 2015).
There has been considerable interest in quantitative aspects of the human brain for nearly
150 years. Despite the technical limitations of early microscopes’ optical resolution and the
need to develop, refine and optimize fixation and staining methods (Mühlmann, 1936; Glees,
1955; Blinkov and Glezer, 1968; Brodal, 1969; Iniguez et al., 1985; Glees, 1988; Gittins and
Harrison, 2004a), plausible numbers of cells were estimated in the 1900s for animal brains
and for major parts of the human brain, in particular the cerebral cortex. Overall, and
considering that results were obtained by different investigators using different methods,
most of the data are relatively consistent. For example, the majority of studies estimated
total neuron numbers for the entire human cerebral cortex at 10–20 billion (Table 1).
Since the cerebral cortex comprises by volume about 80–85% of the adult human brain
(Stephan et al., 1981; Rilling and Insel, 1999), quantitative data for the cortex was often
equated with or taken to be equivalent to the whole brain. This turned out to be a
consequential over-simplification, because the contribution of the cerebellum (which
contains about 80% of all neurons in the human brain; Azevedo et al. (2009)) was neglected,
and this helped to support the myth of one trillion glial cells in human brains, as discussed in
more detail later in this review. The following sections examine the history of numerical
reports for the three major components of the human brain – cerebral cortex (80–85% of
total brain volume or 1,200 g), the cerebellum (10% of volume or 150 g), the remaining
components, the brainstem, diencephalon and striatum, sometimes called “the rest of brain”
or “remaining regions” (2–8% of volume or 75–110 g; Blinkov and Glezer, 1968; Azevedo
et al. (2009); Andrade-Moraes et al., 2013), and the entire human brain.
von Bartheld et al. Page 7
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Cerebral cortex
We first review the published estimates for neuronal numbers, then the GNR, and finally
glial numbers. Unless indicated otherwise, “cortex” refers to the grey matter only, and
excludes underlying white matter tracts.
Numbers of neurons—Several investigators have estimated numbers of neurons in the
human cerebral cortex, mostly based on histological methods, as compiled in Table 1. The
prevailing approach was to measure cell densities in histological sections, and to cope with
the challenge of differential tissue shrinkage (Nurnberger and Gordon, 1957; Crabb, 1967;
Blinkov and Glezer, 1968). There has been some confusion whether “cerebral cortex” means
only the grey matter, or also includes the underlying white matter. Indeed, the large majority
of studies excluded white matter. The number of neurons in white matter is relatively small –
estimated to be 250–1,000 per mm3 (Garcia-Marin et al., 2010) which is less than 1% of the
number of glial cells, with 20,000–200,000 glial cells per mm3 white matter, see below:
“The number of glial cells.” Therefore, inclusion of white matter does not make a significant
difference for neuron numbers, although it does make a difference for total cell numbers
discussed later. Blinkov and Glezer (1968) and Haug (1986) reviewed the early history of
counting neurons and reporting of numerical estimates in human cerebral cortex, but to our
knowledge there have been no comprehensive reviews of this topic published since that time.
As can be seen in Table 1, the estimates ranged from 1.2–32 billion neurons for the entire
cortex (right and left hemispheres combined), with a majority of studies reporting between
10 and 20 billion neurons. It should be noted that some investigators (e.g., Meynert,
1868/1872; Shariff, 1953) were ambiguous in whether their estimates were applicable to one
or both hemispheres, as mentioned for the Meynert study by von Economo (1926). This type
of confusion explains why Blinkov and Glezer (1968) listed Shariff’s numbers incorrectly
for only one hemisphere, while Haug (1986) correctly listed those numbers for total cortex.
There has been similar confusion whether numerical reports apply to one or both sides in the
1990s (e.g., Mufson and Benzing, 1994; Regeur et al., 1994b; Peters et al., 1998).
Table 1 shows that von Economo’s studies (von Economo and Koskinas, 1925; von
Economo, 1926) were the first to correctly estimate the total number of neurons at about 14
billion. Ironically, their numbers became highly controversial and prompted a harsh rebuttal
by Agduhr (1941). Ultimately, this was one of several controversies where von Economo
and Koskinas were vindicated in history (Triarhou, 2005, 2006).
Table 1 also shows three apparent outliers on the low end by Meynert (the very first report in
1868/1872), Donaldson (1895), and H. Pakkenberg (1966), with estimates between 1.2 and
2.6 billion neurons. On the high end, the group of B. Pakkenberg reported 20–32 billion
neurons (Pakkenberg et al., 1989; Braendgaard et al., 1990; Pakkenberg, 1992, 1993;
Pakkenberg and Gundersen, 1997; Pelvig et al., 2003, 2008). This range appears too high,
based on the previous histological studies and also the results from the isotropic fractionator
(IF) (Azevedo et al. (2009); Andrade-Moraes et al., 2013). There are additional examples
where numbers reported by the group of B. Pakkenberg, one of the pioneers of stereological
counting methods, had to be revised; this is not surprising, given the large biological
variability among human brains and the difficulties of working with human tissues. Another
von Bartheld et al. Page 8
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
potential source of error pertains to sampling issues such as the controversial notion that
counting only 100–200 neurons is sufficient (Gundersen, 1986; Andersen et al., 1992;
Coggeshall and Lekan, 1996), while more recent work employing computer simulations
indicates that considerably more neurons should be counted (Schmitz and Hof, 2000; Geuna
and Herrera-Rincon, 2015). Examples of discrepancies of results include lack of cortical
neuron loss in Alzheimer’s disease (Regeur et al., 1994a; Mufson and Benzing, 1994; Peters
et al., 1998; Andrade-Moraes et al., 2013), numbers of neurons in the cerebellum –
apparently over-estimated by about 50% (Andersen et al., 1992; see below), and the initial
underestimation of the number of neurons in the dorsomedial thalamic nucleus (1.8–7.29
x106 neurons, see “Brainstem, Diencephalon and Striatum,” below).
Nevertheless, it is remarkable that the large majority of the histology-derived estimates
converge at 10–20 billion neurons, which is furthermore supported by estimates obtained by
the IF (Azevedo et al. (2009); Andrade-Moraes et al., 2013). Several studies have
documented the surprisingly wide range of neurons in human cerebral cortex between
individuals (biological variance, Haug, 1986, 1987; Terry et al., 1987; West, 1993a). There
appears to be a normal biological variation in the number of neocortical neurons by a factor
of more than 2; this represents a variance of more than eight times the variance of human
body height (Haug, 1987; Pakkenberg and Gundersen, 1997). The notion that large numbers
of neurons (30–50%) are lost during decades of normal human aging (“Neuronal Fall-Out”,
Brody, 1955; Hanley, 1974; Devaney and Johnson, 1980; Curcio et al., 1982) has been
refuted, primarily through Haug’s pioneering work and others’ (Haug et al., 1984; Haug,
1987; Terry et al., 1987; West, 1993b; Anderton, 1997, see also EXAMPLES SHOWING
IMPACT OF CELL COUNTING). Actual losses appear to be of a much lesser scale and
region-specific (Curcio et al., 1982; West, 1993b; Peters et al., 1998). It still is controversial
whether women have a smaller number of neurons than men and whether neocortex loses a
small amount of neurons (less than 10% over 80 years, Haug, 1987; Pakkenberg and
Gundersen, 1997). Given the large biological variation (over 100%) vs. the small effect size
(West, 1993a), an apparent decrease of less than 10% may be due, at least in part, to secular
(generational) changes in body height, brain size and neuron number (Haug, 1984; Haug,
1987; Pakkenberg, 1989), and furthermore may be functionally insignificant (Peters et al.,
1998). Indeed, recent work indicates that very old women have no reduction in cortical
neuron numbers (Fabricius et al., 2013; Walloe et al., 2014). Overall, excluding the extreme
outliers, the numbers compiled in Table 1 provide a plausible range of estimates for neuronal
numbers in cerebral cortex.
The GNR—The GNR in the human cerebral cortex (grey matter, unless indicated
otherwise) was first established in the 1930s (Mühlmann, 1936; Arutyunova, 1938).
Mühlmann measured densities of glia and neurons in Giemsa-stained samples from the
frontal lobe, and he estimated the GNR to be ~ 1.5 in the adult human cortex (Mühlmann,
1936; Arutyunova, 1938). The GNR of 1.5 in the grey matter of adult human cortex was
confirmed by numerous subsequent investigations as listed in Table 2. Considering well-
established neuronal numbers of 10–20 billion in the human cerebral cortex, this would
place the number of glial cells in the human cortex at about 15–30 billion. The median of
this range is close to the average of 17.4–19.4 billion non-neuronal cells in the human
von Bartheld et al. Page 9
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
cortical grey matter estimated with the isotropic fractionator (IF, Azevedo et al. (2009);
Andrade-Moraes et al., 2013). The number of non-neuronal cells provides a maximum
estimate for the number of glial cells, since non-neuronal cells comprise both glial cells and
endothelial cells. Endothelial cells in the human forebrain and other CNS parts are estimated
to make up about 30% of the non-neuronal cells (equivalent to a ratio of ~2:1
glia:endothelial cells, Nurnberger, 1958; Blinkov and Glezer, 1968; Brasileiro-Filho et al.,
1989; Bjugn and Gundersen, 1993; García-Amado and Prensa, 2012), leaving 70% glial
cells, and reducing the non-neuronal to neuron ratio (nNNR) from 1.48 to a true GNR of
1.04 in Azevedo et al. (2009) and from 1.64 to 1.15 in Andrade-Moraes et al. (2013) (Table
2). It should be noted that endothelial cells in white matter appear to comprise a somewhat
lower percentage (10–20% of non-neuronal cells, Bahney and von Bartheld, 2014) than they
do in cerebral cortex grey matter and other parts of the CNS (about 30%, see below). The
only two discrepancies to the findings of a ~1.5 GNR in human cerebral cortex (Table 2,
with none of these specifying the extent of white matter inclusion) seem to be a 10:1
statement by Hyden and Pigon (1960) and an anecdotal suggestion of a 5:1 ratio made by J.
Olszewski as cited in Heller and Elliott (1954), yet Olszewski published just three years later
a 1.78:1 GNR for human cerebral cortex grey matter (Hawkins and Olszewski, 1957 – see
Table 2). Hyden and Pigon’s claim of a 10:1 ratio in human cortex (unclear whether this
referred to grey matter only) was not backed by any data of their own or other’s original
data. In fact, the discrepancy between Hyden’s 10:1 ratio and those of other investigators
was already noted by Glees (1988).
Taken together, we conclude that based on all available primary data, the GNR of human
(and other primate’s) grey matter of prefrontal cerebral cortex is about 1.5 (Sherwood et al.,
2006; Hilgetag and Barbas, 2009; Ribeiro et al., 2013), and varies locally in the grey matter
between 1.2 in occipital and 3.6 in frontal areas of the human cortical grey matter (Ribeiro et
al., 2013). When white matter is included along with grey matter, then the GNR in cerebral
cortex increases from 1–2 to about 3–4 (Table 2). The average GNR of 1–2 for grey matter
cerebral cortex has been known since 1936 and has to our knowledge never been seriously
disputed (Table 2).
The number of glial cells—Glial cell densities of 200,000 per 1 mm3 in white matter
and about 100,000 per 1 mm3 in grey matter were reported for adult human cortex (Blinkov
and Ivanitskii, 1965), while Schlote (1959) counted 40,000–90,000 per 1 mm3, Hess (1961)
counted 108,000 in white matter, and Blinkov and Glezer (1968) list 48,000 cells (glia and
neurons) per 1 mm3, which is close to Haug’s (1987) report of about 20,000–25,000 glial
cells per 1 mm3, assuming a GNR of between 1 and 2. Applying a total volume of about 250
cm3 per cortical hemisphere grey matter (Blinkov and Glezer, 1968), the number of glial
cells in human cerebral cortex (500 cm3) amounts to 10 billion (Haug, 1987), 20–45 billion
(Schlote, 1959) or 50–100 billion (Blinkov and Ivanitskii, 1965). The number of glial cells
in the grey matter of the human cerebral cortex was more recently reported by using
stereological methods (Pakkenberg et al., 2003; Pelvig et al., 2003, 2008; Karlsen and
Pakkenberg, 2011); these studies estimated between 18.2 and 38.9 billion glial cells (Table
2), while studies using the IF determined the number of non-neuronal (NN) cells at an
von Bartheld et al. Page 10
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
average 17.4–19.1 billion in the grey matter of the cerebral cortex (Azevedo et al. (2009);
Andrade-Moraes et al., 2013) (Table 2).
One of the major – if not most serious – problems in the histology-based counting methods
is the technical difficulty of recognizing glia and distinguishing them from small neurons
(Fig. 1). This problem has a long history (Mühlmann, 1936; Kryspin-Exner, 1943; Glees,
1955; Nurnberger and Gordon, 1957; Braitenberg and Atwood, 1958; Palay, 1958; Schlote,
1959; Iniguez et al., 1985; Andersen et al., 1992; Gittins and Harrison, 2004a), and still
awaits resolution, since immunostaining with the NeuN antibody in tissue sections appears
to be incomplete and variable (Lyck et al., 2009). The difficulty of distinguishing small
neurons from glia may explain some of the conflicting results that have been obtained in
RESEARCH). Therefore, the design of methods that can accurately determine neuronal and
glial cell numbers is important.
Among the glia, numerous investigators have determined the relative contributions of
astrocytes, oligodendrocytes and microglia, mostly in cerebral cortex, as compiled in Table
3. Not surprisingly, oligodendrocytes are more frequent than other glial cell types in white
matter (Table 3). There is also considerable, but not unanimous agreement across primary
sources that in different brain regions, including neocortical grey matter, oligodendrocytes
are the most frequent at 45–75% of glial cells, followed by astrocytes (19–40%), while
microglia contribute 10% or less, although some textbooks and reviews have reported
differently, unfortunately without references (Verkhratsky and Butt, 2007; Pastor and Sola,
2008; Bayraktar et al., 2015). Statements that microglia alone are about as numerous as
neurons (Streit, 1999; Fields, 2009) are incorrect, because they were based on the mistaken
belief of a 10:1 GNR. In conclusion, all three methods: histology, DNA extraction, and the
IF method support numbers of about 10–20 billion neurons and at most a 2-fold larger
number of glial cells (20–40 billion) in the human cerebral cortical grey matter, thus
supporting an average GNR of approximately 1.5. Inclusion of the white matter (that
underlies the grey matter of cerebral cortex) increases the GNR to about 3.0.
The cerebellum is another part of the human brain in which cell numbers were estimated
throughout the last century. Initially, only numbers for the large cerebellar neuronal types
were reported – in particular the easily recognized Purkinje cells with most estimates (8/13)
between 14–26 x 106 (Fig. 1; Table 4). Braitenberg and Atwood (1958) were the first to also
report the number of granule cells (small cerebellar neurons) which alone were estimated to
be “of the order of 10–100 billion.” In 1975, Lange reported the density of neurons in the
human cerebellum as 1,610 cells/0.001 mm3 in the granular layer, with an average 720.8
neurons/0.001 mm3 in cerebellar cortex (Lange, 1975). Applying the reference volume from
other studies, Lange’s neuronal densities in the cerebellum translate to a total of 65–70
billion neurons in the human cerebellar cortex (Williams and Herrup, 1988). In contrast,
Haug estimated about 50 billion neurons in cerebellar cortex (Haug, 1986), based on
Lange’s work and his own counts. A very large number of neurons in the cerebellum had
been suspected by earlier investigators (Elliott in: Nurnberger and Gordon, 1957; Kuffler
von Bartheld et al. Page 11
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
and Nicholls, 1976), as well as a very low number of non-neuronal (glial) cells (Elliott in:
Nurnberger and Gordon, 1957), but the study of Andersen et al. (1992) provided for the first
time direct evidence for a very low number of glial cells in the human cerebellum. Andersen
et al. estimated that among a total of 105 billion cells in the human cerebellum, there were
101 billion granule cells, with most of the remainder, about 3 billion, being glial cells (see
their Figure 8, Andersen et al., 1992). This implied that the GNR of the human cerebellum
had to be extremely low, about 0.03. However, Andersen et al. (1992) did not comment on
how the cellular composition of the cerebellum (GNR of less than 0.1) compared with other
brain structures such as the cerebral cortex (GNR of ~ 2–3, when white and grey matter are
combined). Accordingly, the implications for total neuron and glia numbers in human brains
and the differences between GNRs in the cerebellum and the cerebral cortex remained
hidden. In the meantime, the group of B. Pakkenberg revised their stereological estimates of
the human cerebellum from 101 billion granule cells (Andersen et al., 1992) to about 70
billion granule cells (Andersen et al., 2012), a number that is much closer to the numbers
obtained by using the IF methodology as well as Lange’s and Haug’s estimates, implying
that 50–70 billion is a most plausible range (Table 4).
Based on the study of Andersen et al. (1992), and also taking into account the numbers of
glial cells in the white matter of the cerebellum (Bahney and von Bartheld, 2014), Andersen
and colleagues’ counts of 30,000–40,000 glial cells per mm3 appear plausible, resulting in a
total of about 3 billion glial cells in the cerebellum. Compared with the number of neurons
(about 65 billion), the GNR for the entire human cerebellum appears to be about 0.05.
Studies using the IF have estimated the average number of cells in the cerebellum to be
between 55 and 70 billion (Azevedo et al. (2009); Andrade-Moraes et al., 2013), with
granule cells (granule neurons) constituting the overwhelming majority (Azevedo et al.
(2009)). The same method yields an upper estimate of around 16 billion glial cells; this
counts all non-neuronal cells which comprise the combined total of glial and endothelial
cells in the cerebellum, and therefore amounts to a maximal GNR of 0.23 (Azevedo et al.
(2009); Andrade-Moraes et al., 2013).
When the cerebellum and the cerebral cortex are considered together, the GNR for these two
major parts of the brain amounts to a value of 0.8–0.9, much less than the GNR of the
cerebral cortex alone, without the cerebellum. This difference is so substantial, because the
cerebellum has not only a very large number of neurons, but also a number of glial cells that
is extremely low in comparison. However, the human cerebellum is not an outlier in its GNR
or glial cell composition; if the GNR appears abnormally low, it is because of the very large
density of neurons in this structure (Herculano-Houzel, 2014). The large number of
cerebellar neurons was recognized by early investigators (e.g., Elliott in: Nurnberger and
Gordon, 1957; Kuffler and Nicholls, 1976); Kuffler and Nicholls remarked on the
“staggering numbers of neurons” in the human cerebellum, but the relatively low number of
cerebellar glial cells remained obscure and largely unrecognized even after the report by
Andersen et al. (1992). Without the numbers in the cerebellum, the human brain would have
a GNR of at most 4 (using values from Azevedo et al. (2009)). The unusual cellular
composition of the cerebellum was a key factor in failed attempts to calculate the true GNR
for the total human brain, and a major reason for the persistence of the notion of one trillion
glial cells.
von Bartheld et al. Page 12
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Brainstem, Diencephalon and Striatum
These parts of the brain, primarily the brainstem, have been measured to comprise between 2
and 8% of the volume of the entire brain, but accommodate less than 1% of its neurons
(Azevedo et al. (2009)). The brainstem contains a variety of neuronal nuclei and fiber tracts.
Until 2009 (Azevedo et al. (2009)), there had been no attempts made to estimate the total
number of neurons or glial cells in this part of the brain, although Blinkov (1963) reported
on the glia index for several structures in the human brainstem. A select number of nuclei or
regions was investigated with histological techniques for neuron numbers, including the
reticular formation (5.2 x 106 neurons, Blinkov and Glezer, 1968), corpora quadrigemina
(inferior colliculi: 1.2 x 106 neurons Blinkov and Glezer, 1968), and lateral geniculate
nucleus (on one side: 570,000 neurons, Balado and Franke, 1934; 1.2 x106 neurons, Chacko,
1948; 3.5 x 106, Selemon and Begovic, 2007; 2.0 x 106 neurons, Dorph-Petersen et al.,
2009). The reason for the discrepancies for the lateral geniculate nucleus is unknown, but
both the 2007 and the 2009 studies employed the same stereological method. The supraoptic
nucleus contains about 75,000 neurons and the paraventricular nuclei 85,000 neurons
(various sources, reviewed in Blinkov and Glezer, 1968). The mammillary bodies (medial
nuclei) contain about 800,000 neurons, and there are about 1.3 x106 neurons in the
anteroventral and medial nuclei of the thalamus (Powell et al., 1957). The basal ganglia have
been reported to contain 816 x 106 neurons (Karlsen and Pakkenberg, 2011), with about 100
x 106 small neurons and 570,000–670,000 large neurons in the striatum (Schröder et al.,
1975), 7.8 x 106 neurons in the anterior striatum (Weise et al., 2015), about 700,000 neurons
in the globus pallidus (Thörner et al., 1975), and 300,000 in the subthalamic nucleus (Lange
et al., 1975). The number of glial cells was estimated at 400 x 106 in the striatum (Schröder
et al., 1975) and at 63–82 x 106 in the globus pallidus (Thörner et al., 1975). The number of
neurons in the substantia nigra was reported to be about 450,000 pigmented neurons
(McGeer et al., 1977), 500,000–600,000 neurons (Mann, 1986) and 550,000 pigmented and
260,000 non-pigmented neurons (Pakkenberg et al., 1991; Stark and Pakkenberg, 2004),
while the subthalamic nucleus has 286,000–306,000 neurons (Lange et al., 1975), and the
locus coeruleus contains 32,000–38,000 pigmented neurons (Mouton et al., 1994; Ohm et
al., 1997).
It is in the brainstem and diencephalon where some large GNR values are indeed found. The
superior colliculus has a GNR of about 10 (Blinkov and Glezer, 1968), and the lateral
vestibular nucleus a GNR of about 30–50 (Blinkov, 1963; Ponomarev, 1966; Blinkov and
Glezer, 1968). The GNR was reported near 160 for the globus pallidus (89–114 x106 glial
cells; 688,000–711,000 neurons, Thörner et al., 1975), but is only 3.7 in the striatum (380–
408 x 106 glial cells; 100.7–105.6 x 106 neurons, Schröder et al., 1975). Pakkenberg and
Gundersen (1988) reported neuron and glia numbers for the ventral pallidum (3.97 x 106
neurons; GNR = 12.2) and the dorsomedial thalamic nucleus (1.8 x 106 neurons; GNR =
17). However, the initially reported number of neurons in the dorsomedial thalamic nucleus
turned out to be an underestimate: subsequent studies, also using stereology, reported ~ 3.5 x
106 (Popken et al., 2000), 7.29 x 106 (Dorph-Petersen, 2004), and more recently ~ 6.43 x
106 (Abitz et al., 2007) and ~ 6.4 x 106 (Nielsen et al., 2008). The discrepancies between
studies – even when using the same stereological counting method – illustrate the difficulty
von Bartheld et al. Page 13
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
encountered by efforts to determine the true number of neurons in just one small nucleus in
the brainstem.
Regardless of the precise numbers, it is obvious that the total number of neuronal and glial
cells in the brainstem, diencephalon and striatum does not add up to numbers that are even
close to those in cerebral cortex or cerebellum. Since the volume is small and the total
number of cells is relatively low, this part of the brain contains only about 700 million
neurons and about 6.6–7.7 billion non-neuronal cells, with a GNR of maximal 10:1, as
determined by the isotropic fractionator (Azevedo et al. (2009); Andrade-Moraes et al.,
2013). Therefore, the fluctuations in GNR between specific nuclei or tracts in the brainstem
and diencephalon add little to the overall GNR when compared with the numbers provided
by the cerebral cortex and cerebellum. When the number of neurons in these two structures
together was determined to be between about 80–100 billion, it should have become
apparent that a 10:1 GNR, with the implied 1 trillion or more glial cells, was impossible.
There are not nearly enough glial cells in either the cerebral cortex or in the cerebellum to
arrive at such a number (Azevedo et al. (2009)).
Discrepancies of estimates—Attempts to pinpoint the cause(s) of discrepancies
between studies have proven difficult, not only because most investigators do not provide
sufficiently detailed information (Schmitz and Hof, 2005), but also because a multitude of
potential factors can generate biases. This was shown by studies designed to quantify biases,
by comparison with the gold standard, serial section reconstruction, by changing distinct
variables, and by ultrastructural verification of particle identity (Coggeshall et al., 1990;
Hatton and von Bartheld, 1999; Baryshnikova et al., 2006; Ward et al., 2008; Lyck et al.,
2009; Kaplan et al., 2010). Sources of bias may be in opposite directions, may even cancel
each other, or may skew estimates in the same direction, and then be additive. Without full
access to primary data, to all aspects of tissue processing, and an independent re-
examination of counting, it is impossible to identify sources of bias with any certainty. For
these reasons, it has been recommended, as a practical approach, to calibrate counting
methods against a small sample of serial section reconstructions, still considered the ultimate
standard (Coggeshall et al., 1990; von Bartheld, 2002; Kaplan et al., 2010).
Entire human brain
Based on actual counts of neuronal densities using histological methods, the number of
neurons in the entire human brain was estimated by experts in quantitative neuroscience at
30 billion (Szentagothai, 1983), 70–80 billion (Haug, 1986), and 85 billion (Williams and
Herrup, 1988). Investigators using the isotropic fractionator confirmed these latter neuron
numbers at 67–86 billion neurons (Azevedo et al. (2009); Andrade-Moraes et al., 2013).
Based on glial cell densities, Blinkov and Glezer (1968) estimated the number of glial cells
in the entire human brain to be 100–130 billion, while Haug, using his own densities and
volume measurements, estimated 40–50 billion glial cells for the entire human brain (Haug,
1986). The current estimates of numbers of non-neuronal cells in the entire human brain, as
revealed by the IF, place the total glial numbers well below 85 billion (since these 85 billion
include approximately 20–25 billion endothelial cells), and therefore are closer to the
von Bartheld et al. Page 14
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
estimates of Haug (40–50 billion glial cells) rather than those of Blinkov and Glezer (100–
130 billion glial cells) (Fig. 3).
The GNR or “glia index” is defined as the ratio between the number of glial cells and the
number of neurons in the same volume of brain substance. The GNR and its implications
have attracted interest among investigators for more than a century. The notion of the GNR
was conceived by Nissl (Nissl, 1898), but first applied and studied in a systematic way in the
1930s. While some scientists question the utility of the GNR – or of any cell quantification
(Yuhas and Jabr, 2012), many current investigators conclude that the GNR informs about
brain development, physiology, diseases, aging, and brain evolution (Sherwood et al., 2006;
Hilgetag and Barbas, 2009; Herculano-Houzel, 2014), as detailed below in EXAMPLES
comparative context and when applied to comparable brain regions. Technically, the GNR is
easier to establish than total absolute numbers, especially for distinct brain parts, because no
absolute values are required. Rather, for any given volume, the number of glia and neurons
can be estimated and compared with some certainty in relationship to each other. Thus, the
GNR can be calculated as the ratio between the density of glia and the density of neurons in
any structure or volume, without ever estimating absolute numbers of cells (e.g., Friede,
1954; Hawkins and Olszewski, 1957; Haug 1987; Stolzenburg et al. 1989). Persistent
problems were how to define precise borders between grey and white matter, to clearly
distinguish small neurons from glial cells, and to extrapolate to the whole brain from the
data obtained in spatially restricted samples. Since the GNR was recently discussed in the
context of glial cells and phylogeny (showing a remarkable and evolutionarily conserved
scaling of GNRs with neuronal density between structures and species, Herculano-Houzel,
2014), we focus here on a brief history of the GNR as it relates to human brains and the
claims of glial cell numbers.
Recent work (Fields, 2009; Verkhratsky and Butt, 2013) stated that Fridtjof Nansen was the
first to associate an increasing GNR with increasing intelligence. Unfortunately, this
statement is based on a mis-quotation. Nansen (1886) attributed such increasing mental
abilities to increasing amounts of what he called “dotted substance” which is essentially
neuropil made up of neuronal and glial processes (Table 5). Fields (2009) and Verkhratsky
and Butt (2013) recently adopted Galambos’ (1961) misquote, implying that Nansen was not
referring to the “dotted substance,” but rather to glia exclusively (Table 5; Nansen, 1886,
page 171). The dotted substance was later termed “neuropil” by von Apathy (1897), as
reviewed in detail by Florey (1985).
Accordingly, Franz Nissl was the first to note the prevalence of glial cells in mammalian
cortices (Nissl, 1898; also reviewed in Herculano-Houzel, 2014), while the GNR was first
calculated and reported for a major part of the human brain by Mühlmann (1936).
Mühlmann established that the approximate GNR (“Prozentgehalt der Nerven und der
Gliazellen”) of the grey matter of the human cerebral cortex is about 1.5, a value that since
has been widely confirmed (Table 2). He also conducted a detailed developmental study that
revealed how the GNR in cortex changes from the newborn (GNR = 0.3:1) to the aged adult
von Bartheld et al. Page 15
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
(GNR = 2:1). This showed that the GNR is age-specific and that glia-neuron relations
change as the brain matures. From the 1950s until the 1980s, the GNR was called “glia
index” (Friede, 1953, 1954), glia/neuron index (Brizzee and Jacobs, 1959), or glia/nerve cell
index (Hawkins and Olszewski, 1957). Altman (1967) was the first to use interchangeably
the terms glia index and glia-neuron ratio (GNR), while Bass et al. (1971) and some
subsequent investigators advocated the use of the reciprocal of the GNR: the “neuron/glia
ratio” (Thörner et al., 1975; Diamond et al., 1985; Terry et al., 1987; Leuba and Garey,
1989), the rationale being that the neuronal density varies much more than the glial cell
density (Bass et al., 1971; Reichenbach, 1989). Bass et al. (1971) – incorrectly as it turned
out – assumed that the number of endothelial cells in brains was negligible: “since the
vascular cell fraction is relatively small, the neuron/non-neuron ratio(n) essentially equals
the neuron/glia ratio.” Other’s work showed that as much as one third of non-neuronal cells
were endothelial cells in mammalian, including human, CNS (Blinkov and Glezer, 1968;
Brasileiro-Filho et al., 1989; Bjugn and Gundersen, 1993; Davanlou and Smith, 2004; Lyck
et al., 2009; García-Amado and Prensa, 2012).
Work by Friede and others in the 1950s rapidly confirmed Nissl’s suspicion and revealed
that the GNR differs between species in what appeared to be a “phylogenetic” trend. This
prompted Friede to propose that the GNR serves as an indicator of the “developmental
advancement” of a species – culminating in humans (Friede, 1954; Pfrieger and Barres,
1995; Araque et al., 2001). Brizzee and Jacobs (1959) concluded that brain weight as well as
brain complexity contributed to the GNR. When investigators examined brains larger than
those of humans, they found even larger GNRs (Hawkins and Olszewski, 1957; Tower and
Young, 1973; Haug, 1987; Eriksen and Pakkenberg, 2007). They concluded that the GNR
was associated with brain size rather than with “developmental advancement” or cognitive
abilities. However, the hypothesis originally formulated by Nissl and Friede of glia as being
correlated with increasing intelligence persisted in the literature due to the intuitively
appealing idea that a relatively large GNR in human cerebral cortex compared with other
animals might be related to this species’ cognitive abilities (Jerison, 1973; Diamond et al.,
1985; Witelson et al., 1995; Araque et al., 2001; Fields, 2009; Koob, 2009; Verkhratsky and
Butt, 2013).
It was recognized in the 1960s that differences in GNRs are largely determined by changes
in neuronal densities rather than changes in glial cell densities – glial cell densities remain
remarkably constant between species and even brain structures, at 50,000–130,000 cells per
mm3, while neuronal densities in different parts of the human brain vary between 0 and over
400,000 per mm3 (Blinkov and Glezer, 1968; Bass et al., 1971; Tower and Young, 1973;
Haug, 1987; Herculano-Houzel, 2014). Accordingly, the GNR largely reflects differences in
neuronal density, but not, or only to a very minor extent, differences in glial density (Blinkov
and Glezer, 1968). The GNR was shown not to increase universally with brain mass or
cortical mass, but rather with decreasing neuronal density, which may or may not coincide
with increasing brain mass (Herculano-Houzel, 2014). However, it is still not resolved how
much increasing axon length, dendritic arbor size, and somatic size contribute to increasing
neuronal cell size and thus decreasing neuron density (Friede and van Houten, 1962; Jehee
and Murre, 2008; Herculano-Houzel, 2014). These are crucial questions from an engineering
perspective: how to optimize information processing within finite spaces. While the GNR is
von Bartheld et al. Page 16
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
easier to determine from a sampling standpoint than absolute numbers of glia or neurons,
investigators have to deal with one major technical issue: how to distinguish glia from small
How to best distinguish glia from small neurons
Small neurons are difficult to distinguish from glial cells (Fig. 1). Virtually all investigators
using histology encountered and recognized this as a major problem, especially in the
granular layer of the cerebellum (Kryspin-Exner, 1943; Glees, 1955; Nurnberger and
Gordon, 1957; Braitenberg and Atwood, 1958; Andersen et al., 1992; Lyck et al., 2009).
Mühlmann tested several different stains and recommended the Giemsa stain as the best way
to distinguish glia and neurons (in paraffin sections, Mühlmann, 1936). Kryspin-Exner
(1943) and Schlote (1959) preferred to study glia in Nissl-stained material. Glees (1955)
routinely used silver impregnation and Nissl stain in adjacent sections to confirm cell types.
Braitenberg and Atwood (1958) were “not fully satisfied with any of the methods available”
and acknowledge the “serious difficulty presented by the small size of the granular cells.”
Even at the ultrastructural level, glial cells can be difficult to identify and classify (Palay,
The Giemsa stain is a mixture of dyes (methylene blue and eosin yellow) with the capacity
to stain not only ribonucleic acid in the cytoplasm (neurons), but also nuclear chromatin
(glia), in a temperature- and pH-dependent manner (Iniguez et al., 1985). The utility of the
Giemsa stain and long tradition in distinguishing neurons and glia is often overlooked
(Mufson and Benzing, 1994), and it has been stated that the Giemsa stain was introduced in
neurohistology only in the 1970s (e.g., Scheff and Baldwin, 1996), even though Mühlmann
described in the 1930s in considerable detail the use of the Giemsa stain to distinguish glia
and neurons (Mühlmann, 1936). A method paper devoted to the Giemsa stain in brain
sections further confirmed that this stain is well suited to visualize both neurons and glia
(Iniguez et al., 1985). Thus, utilization of the Giemsa stain predates the adoption of this stain
by Gundersen, West and Pakkenberg for their resin sections in the 1980s and 1990s (e.g.,
Gundersen et al., 1988; West and Gundersen, 1990; Regeur et al., 1994a). Mufson and
Benzig (1994) discuss in their commentary the importance of distinguishing neurons and
glia, and types of stains that have been used to reach this goal.
A breakthrough seemed to have been achieved by utilizing an antibody against a neuron-
specific nuclear antigen (NeuN; Mullen et al., 1992). This was first applied in histology to
distinguish small neurons from non-neuronal cells in tissue sections (Gittins and Harrison,
2004a). A side-by-side analysis of NeuN and Nissl stains in the cerebral cortex showed that
cell counts using Nissl stain underestimated numbers of neurons, apparently because small
interneurons can be confused with glial cells (Gittins and Harrison, 2004a), while another
study found that only a fraction (18–57%) of neurons were identified as NeuN-positive in
histological sections from human cortex, and a panel of neuron-specific antibodies was
recommended for future work (Lyck et al., 2009). On the other hand, the NeuN antibody
was proven a highly efficient tool to separate neuronal from non-neuronal cell nuclei in the
isotropic fractionator method (Herculano-Houzel and Lent, 2005). Additional suitable
antibodies are now becoming available that can be used to further classify neurons into
von Bartheld et al. Page 17
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
subtypes, and to separate the non-neuronal cells unambiguously into glial cell types and
endothelium. Currently, however, the IF does not distinguish glia from endothelial cells, but
rather pools both types together as non-neuronal cells. The ratio obtained with the IF is
therefore not a GNR, but a “nN-NR” ratio (non-neuronal cells/neurons ratio) – which,
however, serves as a useful upper limit to the GNR. Given that the vasculature represents a
constant (and small, 1–5%) fraction of brain tissue (and cerebral cortex in particular;
Buchweitz and Weiss, 1986; Lawers et al., 2008; Tsai et al., 2009; Karbowski, 2011), values
of nN-NR likely translate into GNR by the same proportion across species.
Reports of the GNR and estimates of absolute numbers
In the context of the notion of one trillion glial cells, Table 6 compiles reports of the GNR as
well as estimates of absolute cell numbers in the entire human brain from 1895 until 2015
(see also Fig. 3). We attempted to include all major reviews and textbooks. It is interesting
that in the late 1950s through the 1970s, qualifiers such as “perhaps” and “about” were often
associated with the numbers given, but in the 1980s and beyond, such caution was largely
replaced by an assertiveness that seemed to convey knowledge and evidence rather than a
“best guess” or possible range. Several scientists reported wide ranges in the 1970s, e.g.,
Hubel (1979) and Nauta and Feirtag (1979). “The number of nerve cells, or neurons, that
make up man’s three pounds or so of brain is on the order of 1011 (a hundred billion) give or
take a factor of 10” (Hubel, 1979); and Nauta and Feirtag (1979) wrote: “… there are classes
of neurons so small and densely crowded that it is difficult to judge their number … There
are so many granule cells … that the estimate of 1010 neurons in the entire central nervous
system becomes suspect. The total could easily be an order of magnitude, perhaps two
orders of magnitude, higher.”
As can be seen in Table 6, nearly all authors surveyed endorse a 5:1 – 50:1 abundance of glia
over neurons, with very few exceptions. The exceptions are authors who actually did the
counting (shaded in grey in Table 6: Blinkov and Glezer, 1968; Szentagothai, 1983; Haug,
1986; Azevedo et al. (2009); Andrade-Moraes et al., 2013) or authors who were intimately
familiar with the relevant primary literature (e.g., Jacobson, 1991). Only five publications
report a much lower GNR of 0.7:1 – 1:1 for the whole brain (Haug, 1986; Azevedo et al.
(2009); Andrade-Moraes et al., 2013; Streit, 2013; Verkhratsky and Butt, 2013). Table 6
shows that the range of
numbers in the human brain is by and large within one
order of magnitude, with 20/23 authors giving numbers or a median between 10 and 100
billion. Two texts say one trillion (Kandel and Schwartz, 1981, 1985), and the authors did
not correct this mistake for neuron numbers until subsequent editions of their textbook
(Kandel et al., 1991, 2000, 2013). Remarkably, such errors, in neuron number, glia number
and GNR, were contained in the most prestigious textbook of its generation (Darlington,
2009). For example, the 2000 edition of Kandel et al was praised: “The bible of
neuroscience and the singular source for all things brain. It is 1500 pages of facts,
information, data, theory, and on a level of scholarship unparalleled. Ever since its first
edition came out in the early 1980s, this book has set the standard for erudition in the
sciences and is probably on the bookshelf of almost every neuroscientist in the world …”
(Lambert and Kinsley, 2004).
von Bartheld et al. Page 18
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Several authors implicitly postulate a number of 1–50 trillion glial cells in human brains,
because they provide the GNR as well as estimates of the total number of neurons (Kandel
and Schwarz, 1981, 1985; Kandel et al., 1991, 2000; Bear et al., 2001; Hatton and Parpura,
2004). We calculated those implicit numbers and indicated them in brackets in Table 6
(“[…]”). Accordingly, two editions of the Principles of Neural Science (Kandel and
Schwarz, 1981 and 1985) suggested that human brains contain as many as 50 trillion glial
cells, despite the fact that the largest number ever suggested in the primary literature was
130 billion (Blinkov and Glezer, 1968). The overwhelming number of claims of a 10:1 or
higher GNR (the origin of which will be examined next) outweighed the few original reports
showing a 1:1 GNR (only three publications prior to 2009, Table 6).
None of the textbooks or reviews listed in Table 6 provides a primary reference – or any
valid reference – for the claim of a 10:1 GNR. The lack of citations for the notion of a 10:1
GNR over a 50 year period is an example of a major failure in the scientific process that is
supposed to self-correct invalid claims or reports (Committee on the Conduct of Science,
1989; Neville, 2007; Firestein, 2012; Ioannidis, 2012), as explained in more detail below.
Not surprisingly, the first response of many brain scientists to the “maverick” report by
Azevedo et al. (2009) was disbelief (see below and Yuhas and Jabr, 2012), and it has taken
several years for the new evidence to become accepted (Table 6). The refutation of the
notion of one trillion glial cells is also an example where a new (or substantially improved)
technique, the IF, initiated a paradigm shift, but subsequent scrutiny showed – surprisingly –
that the new paradigm had been supported all along, for decades, by traditional (histological)
techniques. The problem appears to have been disregard of conflicting primary data and a
failure to recognize the lack of supporting data for the prevailing consensus. The false belief
was enabled and facilitated by presenting the 10:1 GNR as a “fact” and as “common
knowledge” not requiring citations (Committee on the Conduct of Science, 1989; Neville,
There are numerous examples of how cell counting has informed and impacted progress in
the field, with classical studies documenting the loss of neurons in degenerative diseases, for
example correlating the extent of neuron loss with disease severity (Damier et al., 1999;
Stark and Pakkenberg, 2004; Kordower et al., 2013). However, cell quantification is
fundamentally important not only in pathology and in the clinical area. We selected here
three examples that illustrate how cell counting has had a significant impact in areas beyond
clinical medicine. The first example is from the aging human brain, the second is from the
evolution of the brain, and the last is from developmental neuroscience.
Is there a significant loss of neurons in the normal aging brain?
Based on studies in the 1950s to 1980s, it was reported and generally believed that the
normal aging brain loses large numbers of neurons each day after 30 years of age (Brody,
1955; Devaney and Johnson, 1980), so that “a 60-year span of adulthood would mean loss of
half the cerebral neurons” (Hanley, 1974; see also Curcio et al., 1982; Kausler et al., 2007).
Reports of this “neuronal fall-out” with normal aging provided a depressing outlook for
von Bartheld et al. Page 19
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
octogenarians: loss of neurons was thought to be the cause of senile dementia, and senile
dementia was thought to be an inevitable part of growing old. Thus, the above-mentioned
cell counting studies may have contributed to the fear of dementia among the elderly
(“greatest cause of distress,” Jorm, 1987; Pitt, 1998), a segment of the population with high
rates of suicides (Meehan et al., 1991; McKeown et al., 2006; Schmutte et al., 2009). In this
context, the innovative and diligent quantitative work of Haug and colleagues (1984)
demonstrated that the studies indicating a constant and significant loss of neurons in the
normal aging brain were flawed. The shrinkage of brains after fixation depends on the
person’s age, and accordingly the reference volumes of brains from older people differ from
those of younger brains. When this was taken into account, there was very little if any
normal loss of neurons in most parts of the brain (Haug et al., 1984; West, 1993b; Morrison
and Hof, 1997; Stark and Pakkenberg, 2004; Fabricius et al., 2013). The new view, that
mental decline is not an imminent or inevitable fate, changed the elderly’s outlook on their
remaining life span rather dramatically, even though the old dogma of continuous age-
dependent neuronal death can still be found in recent literature (Rodriguez-Arellano et al.,
2015; see also Verkhratsky et al., 2004; Kausler et al., 2007). The misconception of the
extent of neuron loss in normal aging brains had profound implications beyond the quality of
life for octogenarians: it complicated and delayed research into the causes of the real
problem: the pathological loss of neurons in Alzheimer’s and related dementias. It took
major efforts to correct this view (Morrison and Hof, 1997; Hof and Mobbs, 2009). As
revealed in our review, once a myth has found its way into textbooks, curricula and common
knowledge, it becomes difficult to rectify.
Evolution of the human brain – insights from the GNR
Throughout much of the 20th century the notion prevailed that the cellular composition of
the human brain was exceptional among species and likely responsible for the superior
cognitive abilities of humans (Gazzaniga, 2008). Previous work had suggested that the
human brain and in particular the human neocortex showed an abnormally high GNR when
compared with other mammals with lesser cognitive abilities (Friede, 1954; Jerison, 1973;
Araque et al., 2001; Fields, 2009; Koob, 2009; Verkhratsky and Butt, 2013). Examination of
Albert Einstein’s post-mortem brain, showing an increased GNR in some regions of his
cortex, appeared to support this idea (Diamond et al., 1985; Witelson et al., 1995; Fields,
2009; Koob, 2009). The development of a more efficient cell counting method, the IF, made
it possible to re-examine GNRs and to survey a much larger number of species, and in more
detail (Azevedo et al. (2009); Herculano-Houzel, 2009; Herculano-Houzel, 2011;
Herculano-Houzel, 2012; Herculano-Houzel, 2014). The results of these studies, comparing
cell numbers and GNRs among a wide range of species, has shown that brain size does not
scale universally with neuron number, that different mammalian species such as primates
and rodents scale differently, that cell numbers in cerebral cortex and cerebellum evolve in a
coordinated fashion, and that glia density and sizes vary much less than neuronal density and
sizes. The GNR is highly conserved between structures and species, pointing to an important
and close regulation of glia numbers (scaling) in response to, or regulated by, neuron density
and neuron sizes (Herculano-Houzel, 2012; Herculano-Houzel, 2014; Mota and Herculano-
Houzel, 2014). Most importantly, a GNR of 10 would indeed have made the human brain
extraordinary – but that is not the case: The human ratio of non-neuronal to neuronal cells of
von Bartheld et al. Page 20
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
1 is similar to that of other primates, firmly establishing humans as non-outliers (Herculano-
Houzel, 2012). Thus, the new studies comparing GNRs of different primate brains have
shown that the human brain and its neocortex have “hardware” and cellular contents that are
expected for its body size and are not extraordinary in their cellular composition.
Accordingly, efforts to explain underlying mechanisms of humans’ cognitive abilities must
look elsewhere (Dicke and Roth, 2016). Such new insights and new directions depended
upon the development and implementation of accurate and efficient counting methods.
How are glia and neuron numbers controlled during development?
Neuron and glia numbers and their ratios fluctuate within relatively narrow ranges even in
different species and different adult brain structures, emphasizing the importance of optimal
quantitative relations between cell types. How these ratios are accomplished during
development has been unclear, although it has been shown that the GNR increases markedly
during early postnatal development (Mühlmann, 1936; Brizee et al., 1964; Bandeira et al.,
2009). Using a combination of the DNA extraction and stereological axon counting
methods, Martin Raff’s group counted retinal ganglion cell axons and quantified glial cells
in the optic nerve and tract; they made significant advances by showing that mice with
genetically increased numbers of retinal ganglion cells and axons caused corresponding glial
cells to increase their numbers proportionally (Burne et al., 1996). These results implied that
the neurons (retinal ganglion cells) communicated signals either to glial cell precursors to
proliferate or to existing glia to allow more of the already produced glial cells to survive, so
that a constant (presumably optimal) ratio between neurons or axons and supporting glial
cells was maintained in the mice with increased neuron numbers (Burne et al., 1996). Thus,
cell counting studies helped to advance a new field of study: neuron-glia interactions and
signaling between these two types of cells in the brain, leading to a better understanding of
how neurons and glia interact, communicate, and depend on each other during normal
development of brains, as well as during abnormal development and disorders of the brain
(Araque et al., 2001; Kettenmann and Ransom, 2013).
The notion of a 10:1 GNR dates back to the 1950s, as can be seen in Table 6. We found that
the earliest published accounts by brain and glia scientists – Glees (1958), Pope (1958) and
Galambos (1961) included qualifiers (such as “perhaps”) in their estimates of a 10-fold
abundance of glia over neurons, or they used vague terms such as “glia cells … in higher
animals are extremely numerous” (Bullock, 1967). On the other hand, Hyden, a glia
researcher (Hyden, 1960, 1961, 1967a) was more assertive and proclaimed: “The glial cells
outnumber the nerve cells by a factor of around 10,” and this was stated in the context of
“the central nervous tissue,” quoted from the chapter “The Neuron” in the influential series
“The Cell” (Hyden, 1960), among other texts (e.g., “the glia are by far the most numerous
cells”, Hyden, 1967b). This makes it sound as if a 10:1 GNR was a known fact. How does a
new finding become a “scientific fact”? This process has been described as follows: “At
each stage, researchers submit their work to be examined by others with the hope that it will
be accepted. This process of public, systematic skepticism is critical in science.” …
“Bypassing the standard routes of validation can short-circuit the self-correcting
von Bartheld et al. Page 21
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
mechanisms of science.” (Committee on the Conduct of Science, 1989). In the case of the
GNR, the normal scientific process of peer review and gradual validation was essentially
“short-circuited.” One researcher or a small group of researchers convinced their
contemporaries and their successors by making a claim (that should have been worded as a
testable hypothesis) sound as if it was common knowledge and therefore did not need a
primary reference or other citations. Neuroscientists then, with very few exceptions, copied
it from review to review and from textbook to textbook for over half a century (Table 6),
before it was exposed as one of the most persistent scientific myths of recent history
(Firestein, 2012).
What made Hyden so convinced about a global 10:1 GNR? The key to understanding this
conviction may lie in the context of Hyden’s own research area, which were the large
Deiters neurons in the lateral vestibular nucleus (Hyden and Pigon, 1960). These brainstem
nuclei indeed have a very large GNR – later determined and verified to be about 30–50
(Blinkov, 1963; Ponomarev, 1966; Blinkov and Glezer, 1968). Accordingly, Hyden was used
to seeing neurons surrounded by a large number of glial cells, and it is likely that this
contributed to his and others’ belief that such an arrangement was representative for the
entire mammalian and human brain (Nicholls, 1991). As revealed by the work of Blinkov
and Glezer (1968) as well as Thörner and colleagues (1975), the GNR can vary substantially
among different brainstem nuclei – thus, the assumption that the distribution in one small
nucleus of the brain was representative for the whole brain likely contributed to the widely
overstated GNR in reviews and textbooks (Table 6). Unfortunately, we cannot ask Hyden
what made him believe in the 10fold GNR – he died in 2000 (Hertz et al., 2001; Delgado
and Estanol, 2013). Although Hyden appears to have been the driving force behind the
initial formulation of the myth, he was not the only one who propagated the abundance of
glial cells. In the late 1950s and early 1960s, it was widely believed that glial cells, and in
particular oligodendrocytes, were the most numerous among the cell types in the human
brain (Pope, 1958), although Schlote (1959) found fewer oligodendrocytes than neurons in
most layers of human cortex. It is obvious that there seemed to be a general consensus in the
late 1950s and early 1960s that glia far outnumbered neurons, as also stated in a
memorandum of the RAND corporation (Maron, 1963), as well as in the popular book first
published in 1963 by Isaac Asimov, a science fiction writer and professor at Columbia
University (Asimov, 1963) (Table 6). The notion appears to have originated as an inadvertent
mistake, with no evidence of deliberate manipulation, as in other instances of mis-
information in science (Proctor and Schiebinger, 2008).
In the previous section, we examined how the claim originated. Here we examine the
question “how did that first, wrong number become so widespread?” (Firestein, 2012). Is it
true, as Firestein surmises, that “textbook writers … just picked it up from one another and
kept passing it around?” Once the notion of an overabundance of glia relative to neurons had
formed and had been incorporated into early influential textbooks (Hyden, 1960; Kuffler and
Nicholls, 1976; Kandel and Schwartz, 1981), the notion was treated as a fact, and the
abundance itself was rarely questioned; rather, it became largely reduced to the question of
by exactly how much glia outnumbered neurons, whether it was 5:1, 10:1, or 50:1.
von Bartheld et al. Page 22
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
There is a certain irony in that the perpetuation of the claim was to a large part due to errors
in Kandel’s textbook editions (Table 6), which helped glia biologists to advance their
arguments of glial neglect (e.g., Fields, 2009; Koob, 2009), yet at the same time, Kandel was
criticized for promoting the “neural dogma” and ignoring the importance of glial cells
(Merrill, 2009).
Our review of textbooks and other published reports on the GNR and neuron and glia
numbers shows that reports can be divided into three types. (1) A few authors remained
cautious and stated a wide range, used non-specific terms (“large number”) or said that
numbers or ratios were unknown (Hubel, 1979; Nauta and Feirtag, 1979; Jacobson, 1991).
(2) Other authors reported numbers based on particular studies and data sets (their own or
others) and properly cited the original reference(s) – this was also relatively rare (Blinkov
and Glezer, 1968; Szentagothai, 1983; Haug, 1986; Williams and Herrup, 1988; Azevedo et
al. (2009); Andrade-Moraes et al., 2013; Verkhratsky and Butt, 2013, Table 6). (3) A large
majority of reports cited a specific number or small range, making it sound as if the exact
number or ratio was known, but did not provide any reference (Table 6).
The claim of an overabundance of glial cells spread beyond quantitative brain science and
reached diverse areas of society: the policies of federal funding agencies that decide on brain
research funding, the National Institute of Neurological Disorders & Stroke (NINDS);
public educational databases ( established by major neuroscience societies
and foundations (Society for Neuroscience, The Kavli Foundation, GATSBY); the curricula
of medical, graduate and undergraduate students, and the media such as National Public
Radio (NPR).
For example, the director of NINDS stated during an NPR interview that was nationally
broadcast in the USA in 2013 that the human brain contained “trillions” of nerve cells.
NINDS publishes an annual narrative for justification of neuroscience research funding to
the legislature. These narratives mirror the misleading statements about glia-neuron ratios in
the textbooks, and are factually wrong, but reflect the “textbook knowledge” of their times:
“glial cells far outnumber nerve cells in the brain” (NINDS, 2001); “non-neuronal cells …
far outnumber nerve cells in the brain” (NINDS, 2011); “non-nerve cells, called glial cells,
outnumber nerve cells in the brain” (NINDS, 2015) (years indicate the fiscal year of the
narratives). Several BrainFacts articles, some as recent as 2012 (
Brain-Basics/Neuroanatomy/Articles/2012/The-Neuron) and 2013 (http://
Brain) repeat the old, incorrect information as “brain facts.”
Missed opportunities to refute the notion
Blinkov and Glezer (1968) compiled a wealth of data, but failed to realize that the GNR, the
neuron numbers, and their own glial cell numbers for the whole brain “did not add up.”
Maximally 130 billion glia with a 10:1 GNR could not be true, because that would cap the
total number of neurons at 13 billion, but there were 10–16 billion neurons in cerebral cortex
alone, plus at least 10 billion neurons in cerebellar cortex (granule neurons alone), and
possibly up to a total of 100 billion neurons in cerebellar cortex. This should have been a
warning signal – that the GNR could not exceed 6.5:1 (130:20 = 6.5), and probably was
von Bartheld et al. Page 23
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
much lower, possibly as low as a GNR of 100:116 which equals 0.86:1. Haug (1986) had
calculated the numbers of glia and neurons to be equally low, with a low GNR – with a
remarkable accuracy, as revealed in hindsight. He was a founding member of the
International Society of Stereology and a prolific worker with more than 160 publications
(Kühnel, 2003), yet he does not appear to have made any efforts to refute the prevailing
numbers. Most influential for the propagation of those numbers appear to have been the
textbooks of prominent neuroscientists such as Kuffler and Nicholls and the textbook by
Nobel laureate Kandel and his colleagues. While at Harvard, Kuffler was a mentor not only
to Kandel, but also to Hubel and Wiesel (both of them also Nobel laureates). Kuffler, the
founder of the department of neurobiology at Harvard, is admired as the “father of modern
neuroscience” and the “most dominant figure in experimental neuroscience in the 1960s and
‘70s” (McMahan, 1990), and Kandel’s textbook editions have been praised as the “bible of
neuroscience” (Lambert and Kinsley, 2004; Darlington, 2009). Endorsement of the
prevailing numbers by the most accomplished neuroscientists thus was a formidable
influence. We conclude that there is not one single predominant reason, but a combination of
factors that contributed to the notion of one trillion glia and its perpetuation: These factors
include failure to realize that the numbers did not add up; focus on parts of the human brain
that were not representative; neglecting the role of the cerebellum; missing primary
literature; copying information from previous reviews without scrutiny; inaccurate quoting
of others’ work, and reluctance to challenge the prevailing dogma (Ioannidis, 2012; Nuzzo,
The first challenge to the statements of one trillion cells in the brain and a GNR of 10:1
came with the estimation of a total number of no more than 130 billion glial cells (Blinkov
and Glezer, 1968). However, the significance was not realized. The second challenge came
with Haug’s estimate of less than 50 billion glial cells and a GNR of less than 1 (Haug,
1986). Again, these estimates were not placed in context, numbers were not compared with
those taught in textbooks, and discrepancies therefore remained hidden. The third challenge
was based on data obtained by the isotropic fractionator which showed that the cellular
composition of the human brain comprised an average of 86 billion neurons, 85 billion non-
neuronal cells, and thus rendered a GNR of less than 1:1 (Azevedo et al. (2009)). This time,
the significance of the findings was realized, and the authors drew attention to the
discrepancies and made considerable efforts to locate the source of the prevailing erroneous
estimates (Herculano-Houzel, 2009; Hilgetag and Barbas, 2009; Firestein, 2012; Yuhas and
Jabr, 2012) (Fig. 3).
Some gliabiologists and neuroscientists, however, disagreed: the IF had not yet been
validated against the current standard in the field, stereology (Yuhas and Jabr, 2012; Carlo
and Stevens, 2013; Verkhratsky and Butt, 2013; Charvet et al., 2015), although the estimates
obtained with the IF for the human brain were very close to those in the literature, where
they existed (Azevedo et al. (2009)). Indeed, the possibility that a significant fraction of cell
nuclei were not recovered or were damaged in the isotropic fractionator method was not
addressed in the initial publication, despite concerns that the dissociation, isolation and
purification methods might damage glial or neuronal cell nuclei preferentially. In fact,
von Bartheld et al. Page 24
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
notions have been controversial whether destruction of nuclei may affect primarily the
larger, neuronal nuclei (Lovtrup-Rein and McEwen, 1966; Clarke and Oppenheim, 1995) or
the smaller, glial cell nuclei (Hadjiolov et al., 1965; Kato and Kurokawa, 1967). Initial
concerns were for unfixed tissues and nuclei (Hadjiolov et al., 1965; Lovtrup-Rein and
McEwen, 1966; Kato and Kurokawa, 1967), but also for fixed tissues (Clarke and
Oppenheim, 1995). For example, Hadjiolov et al. (1965) wrote that “According to an
analysis of the nuclear size distribution, a considerable loss of smaller nuclei (10 to 20μ2),
mainly from glial cells, occurs during the purification procedure” (Hadjiolov et al., 1965),
and that “the purification procedure results in a considerable loss of smaller nuclei (10 to
20/~2) which most probably originate from oligodendroglial and microglial
cells“ (Hadjiolov et al., 1965). “The greater number of smaller nuclei were lost during the
ordinary isolation procedure” (Kato and Kurokawa, 1967). Other researchers, however, were
concerned that the larger, neuronal nuclei were more fragile: “Because of their extreme
fragility, … brain nuclei, mainly those from neurons and astrocytes, are easily disrupted
during the homogenization procedure“ (Lovtrup-Rein and McEwen, 1966). And: “large
numbers of cells might be ruptured by the dissociation procedure … this concern is
supported by the fact that Zamenhof’s total large cells … was 1.75 x 105 on average,
whereas counts in histological sections of only the Purkinje cells … came to 2.62 x 105
(Clarke and Oppenheim, 1995). More recent concerns stated: “This ‘isotropic fractionation’
technique can not be considered flawless, of course. We do not know how many nuclei are
lost in the process …” (Verkhratsky and Butt, 2013, p. 95; see also: Yuhas and Jabr, 2012).
Recent calibration studies have dispelled these concerns and validated the IF against other
counting methods, including stereology (Bahney and von Bartheld, 2014; Miller et al.,
2014). Approaches used were to examine adjacent samples of white matter to directly
compare the methods of IF, histology/stereology, and DNA retrieval. In addition, the two
cerebral hemispheres of the same non-human primate were examined with IF and
stereology, again showing equivalency between methods (Miller et al., 2014). Furthermore,
original data based on histological sections are consistent with a 1:1 ratio.
Still, some researchers remain unconvinced. Barres (cited in Yuhas and Jabr, 2012)
maintains that glia make up at least 80 percent of cells in the human brain, because growing
numbers of glia in the forebrain explain the increase in total forebrain DNA, based on a
report of Dobbing and Sands (1973). Yet these DNA data are entirely consistent with the
finding that the human forebrain (cerebral cortex, including white matter) has a ratio of
about 4:1 between non-neuronal:neuronal cells (3.72:1, Azevedo et al. (2009)). The problem
in the Barres argument is that the data are from the
(containing only 19% of the
brain’s neurons), but he makes conclusions about the entire brain. The newest (5th) edition
of Kandel’s textbook has revised the chapter on the cellular composition of the brain, which
is now co-authored by Barres, from the original “10–50 more glia” statement to “2 to 10
times more glia than neurons” (Kandel et al., 2013). This is an improvement, but still
incorrect, as is the claim of an abundance of glia over neurons in the human brain (4:1
according to Barres et al., 2015). Nevertheless, as shown in Table 6, there now is gradual
acceptance of the IF and its conclusions by many neuroscientists and also glia biologists
(Brautigam et al., 2012; Devinsky et al., 2013; Streit, 2013; Verkhratsky and Butt, 2013).
von Bartheld et al. Page 25
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
We wish to emphasize that just because glia are less numerous in the brain than neurons, and
far less numerous than previously thought, this does not mean in any way that glia are less
important. On the contrary, glial cells perform a long list of essential functions (Baumann
and Pham-Dinh, 2001; Haydon, 2001; Ullian et al., 2001; Doetsch, 2003; Nedergaard et al.,
2003; Allen and Barres, 2009; Fields, 2010; Han et al., 2013; Kettenmann and Ransom,
2013; Verkhratsky and Butt, 2013). The extraordinarily conserved numerical relationship
between glia and neurons over at least 90 million years of evolution alone indicates that glia
cells and their relation with neurons and brain function must be extremely important
(Herculano-Houzel, 2014). A precise balance of glia to neurons in human brain regions
seems essential for normal function and this balance is disturbed in disease and trauma (see
below). The isotropic fractionator may prove to be a reliable and efficient tool to not only
provide insights into brain evolution (Herculano-Houzel, 2009), but also to probe suspected
changes in glia and neuron numbers within dissectable regions of the human brain of
patients with neurological and psychiatric diseases (Andrade-Moraes et al., 2013;
Herculano-Houzel et al., 2015).
A large number of neurological and psychiatric diseases have been implicated with
abnormal glia numbers or GNRs. The earliest such reports originated in the 19th century
(Hammarberg, 1895; Ferrero, 1947; Friede, 1953; Hempel and Treff, 1959; Schlote, 1959).
While the degree and localization of abnormalities differed considerably between studies
(Ferrero, 1947; Rowland and Mettler, 1949; Hempel and Treff, 1959; Benes, 1993; Ongür et
al., 1998; Harrison, 1999; Vawter et al., 2000; Todtenkopf et al., 2005; Bernstein et al., 2015;
Elsayed and Magistretti, 2015), changes in glial cell number, densities or GNRs in discrete
brain regions have been confirmed in more recent studies employing stereological or IF
methods for diseases including autism spectrum disorders, mood disorder, depression,
schizophrenia, and Alzheimer’s disease (Rajkowska, 2000; Cotter et al., 2001; Hof et al.,
2003; Vostrikov et al., 2007; Morgan et al., 2010; Karlsen and Pakkenberg, 2011; Andrade-
Moraes et al., 2013; Verkhratsky et al., 2014).
Early research into glia abnormalities (as described above) was much forgotten – so much
so, that the significance of glial changes in psychiatric diseases had to be re-discovered in
2000 (Coyle and Schwarcz, 2000). “… for too long, glial cells have been grossly neglected
when thinking about the neurobiological features of psychiatric disorders”; “… in the next
century glia will no longer remain the silent majority of brain cells but will assume a major
focus of interest in the study of the causes and treatment of neuropsychiatric disorders.”
However, the lack of reliability, validity and therefore trust in quantitative data has been a
major impediment to progress in defining the potential roles of numerical glia abnormalities
in neurological and psychiatric diseases. As already mentioned in previous sections of this
review, there have been multiple examples where initial reports of numbers or ratios of glial
cells and neurons could not be replicated or had to be substantially revised, even within the
same group of investigators or when using the same type of counting technique (Pakkenberg
and Gundersen, 1988; Guillery and Herrup, 1997; Schmitz et al., 2001; von Bartheld, 2001;
Dorph-Petersen, 2004; Abitz et al., 2007; Nielsen et al., 2008; Dorph-Petersen et al., 2009;
von Bartheld et al. Page 26
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Herculano-Houzel et al., 2015). Therefore, meta-analyses have become common to explore
the status and validity of previously published quantitative studies (Harrison, 1999;
Rajkowska, 2000, 2002; Hof et al., 2003; Lyness et al., 2003; Palmen et al., 2004;
Todtenkopf et al., 2005; Courchesne et al., 2007; Amaral et al., 2008). Unfortunately, meta-
analyses have not been able to resolve all controversies about glia numbers and ratios in
human neurological and psychiatric diseases, especially when the primary data was based on
densities or ratios, rather than absolute numbers, and using profile counting or even
stereology. Use of design-based stereology does not, unfortunately, guarantee unbiased
results – there can be significant numerical differences between studies, indicating that these
techniques are not infallible (Herculano-Houzel et al., 2015).
For this reason, there is hope that the recently developed alternative to histological counting
methods, the isotropic fractionator, may emerge as a more robust option to obtain and
validate quantitative data about glia and neuron numbers and their ratios in deceased
patient’s brains. Isotropic fractionator technology is a relatively fast and simple procedure,
and compatible with a large range of fixatives, which makes this approach more versatile
than histological approaches (Bahney and von Bartheld, 2014). It is yet too early to tell, but
this alternative counting technique may provide a much-needed verification and validation of
previously reported numerical abnormalities in glia and neurons in various neurological and
psychiatric diseases.
It is important that quantitative studies of glia and neuron composition refer to the whole
body of published information, take into account all relevant studies, and compare new data
with previously published work. Too often in the history of cell quantification have
discrepancies between investigators, studies, and techniques remained hidden. We hope that
our review will help to facilitate comparison with previous work. More careful scrutiny of
relevant studies, including primary sources, would increase transparency to better compare
studies, data, and techniques, and would contribute to resolve conflicting opinions and
uncover faulty techniques, as Paul Glees forewarned more than 60 years ago (Glees, 1955).
GRANT SUPPORT: NIH grants NS079884 (CSvB, JB), GM104944 and GM103554 (CSvB); CNPq, Faperj, and
the James S. McDonnell Foundation (SHH).
We thank Rob Williams and Glenn Rosen for sharing unpublished data. Our work was supported by NIH grants
NS079884, GM104944 and GM103554 (Center of Biomedical Research Excellence, funded by the National
Institute for General Medical Science, CSvB). One of the authors (CSvB) acknowledges collaborations on
quantitation of cells with Stefano Geuna, Suleyman Kaplan, Jon Kaas, Ron Oppenheim, Glenn Rosen, Oliver von
Bohlen und Halbach, and Rob Williams. SHH thanks Roberto Lent for collaboration in developing the IF and in
early work on the human brain. Her work was supported by CNPq, Faperj, and the James S. McDonnell
DAPI 4,6-Diamidino-2-Phenylindole, Dihydrochloride
DNA-P desoxyribonucleic acid-phosphorus
GNR glia-neuron ratio
von Bartheld et al. Page 27
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
IF isotropic fractionator
Abercrombie M. Estimation of nuclear populations from microtome sections. Anat Rec. 1946; 94:239–
247. [PubMed: 21015608]
Abitz M, Nielsen RD, Jones EG, Laursen H, Graem N, Pakkenberg B. Excess of neurons in the human
newborn mediodorsal thalamus compared with that of the adult. Cereb Cortex. 2007; 17:2573–2578.
[PubMed: 17218480]
Agashiwala RM, Louis ED, Hof PR, Perl DP. A novel approach to non-biased systematic random
sampling: a stereologic estimate of Purkinje cells in the human cerebellum. Brain Res. 2008;
1236:73–78. [PubMed: 18725208]
Agduhr E. A contribution to the technique of determining the number of nerve cells per volume unit of
the tissue. Anat Rec. 1941; 80:191–202.
Allen NJ, Barres BA. Neuroscience: Glia - more than just brain glue. Nature. 2009; 457:675–677.
[PubMed: 19194443]
Altman, J. Postnatal growth and differentiation of the mammalian brain, with implication for a
morphological theory of memory. In: Qarton, GQ.; Melnechuk, T.; Schmitt, FO., editors. The
Neurosciences: A Study Program. New York: The Rockefeller Univ Press; 1967. p. 721-743.
Amaral DG, Schumann CM, Nordahl CW. Neuroanatomy of autism. Trends Neurosci. 2008; 31:137–
145. [PubMed: 18258309]
Andersen BB, Korbo L, Pakkenberg B. A quantitative study of the human cerebellum with unbiased
stereological techniques. J Comp Neurol. 1992; 326:549–560. [PubMed: 1484123]
Andersen BB, Gundersen HJ, Pakkenberg B. Aging of the human cerebellum: a stereological study. J
Comp Neurol. 2003; 466:356–365. [PubMed: 14556293]
Andersen BB, Pakkenberg B. Stereological quantitation in cerebella from people with schizophrenia.
Br J Psychiatry. 2003; 182:354–361. [PubMed: 12668413]
Andersen K, Andersen BB, Pakkenberg B. Stereological quantification of the cerebellum in patients
with Alzheimer’s disease. Neurobiol Aging. 2012; 33:197, e11–20.
Anderton BH. Changes in the ageing brain in health and disease. Philos Trans R Soc Lond B Biol Sci.
1997; 352:1781–1792. [PubMed: 9460061]
Andrade-Moraes CH, Oliveira-Pinto AV, Castro-Fonseca E, da Silva CG, Guimarães DM, Szczupak D,
Parente-Bruno DR, Carvalho LR, Polichiso L, Gomes BV, Oliveira LM, Rodriguez RD, Leite RE,
Ferretti-Rebustini RE, Jacob-Filho W, Pasqualucci CA, Grinberg LT, Lent R. Cell number changes
in Alzheimer’s disease relate to dementia, not to plaques and tangles. Brain. 2013; 136:3738–
3752. [PubMed: 24136825]
Araque A, Carmignoto G, Haydon PG. Dynamic signaling between astrocytes and neurons. Annu Rev
Physiol. 2001; 63:795–813. [PubMed: 11181976]
Armstrong RC, Clarke PG. Neuronal death and the development of the pontine nuclei and inferior
olive in the chick. Neuroscience. 1979; 4:1635–1647. [PubMed: 514511]
Arutyunova AY. Age changes in the number of nerve and neuroglial cells in man. Azerb Med Zh.
1938; 1:150.
Asimov, I. The Human Brain: Its Capacities and Functions. The University of Michigan; Houghton
Mifflin: 1963. p. 357
Azevedo FAC, Carvalho LRB, Grinberg LT, Farfel JM, Ferretti REL, Leite REP, Jacob Filho W, Lent
R, Herculano-Houzel S. Equal numbers of neuronal and nonneuronal cells make the human brain
an isometrically scaled-up primate brain. J Comp Neurol. 2009; 513:532–541. [PubMed:
Azevedo FAC, Andrade-Moraes CH, Curado MR, Oliveira-Pinto AV, Guimarães DM, Szczupak D,
Gomes BV, Alho AT, Polichiso L, Tampellini E, Lima L, de Lima DO, da Silva HA, Lent R.
Automatic isotropic fractionation for large-scale quantitative cell analysis of nervous tissue. J
Neurosci Methods. 2013; 212:72–78. [PubMed: 23017980]
von Bartheld et al. Page 28
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Bahney J, von Bartheld CS. Validation of the isotropic fractionator: Comparison with unbiased
stereology and DNA extraction for quantification of glial cells. J Neurosci Methods. 2014;
222:165–174. [PubMed: 24239779]
Balado, M.; Franke, E. Das Corpus Geniculatum Externum. In: Foerster, O.; Rudin, E.; Spatz, H.,
editors. Monographien aus dem Gesamtgebiete der Neurologie und Psychiatrie. Berlin: Julius
Springer; 1937. p. 1-118.
Bandeira FC, Lent R, Herculano-Houzel S. Changing numbers of neuronal and non-neuronal cells
underlie postnatal brain growth in the rat. Proc Natl Acad Sci USA. 2009; 106:14108–14113.
[PubMed: 19666520]
Barres, BA.; Freeman, MR.; Stevens, B. Cold Spring Harbor Perspectives in Biology. Cold Spring
Harbor: Cold Spring Harbor Laboratory Press; 2015. Glia.
Baryshnikova LM, von Bohlen und Halbach O, Kaplan S, von Bartheld CS. Two distinct events,
section compression and loss of particles (“lost caps”), contribute to z-axis distortion and bias in
optical disector counting. Microsc Res Tech. 2006; 69:738–756. [PubMed: 16845675]
Bass NH, Hess HH, Pope A, Thalheimer C. Quantitative cytoarchitectonic distribution of neurons,
glia, and DNA in rat cerebral cortex. J Comp Neurol. 1971; 143:481–490. [PubMed: 4945394]
Baumann N, Pham-Dinh D. Biology of oligodendrocyte and myelin in the mammalian central nervous
system. Physiol Rev. 2001; 81:871–927. [PubMed: 11274346]
Bayraktar, OE.; Fuentealba, LC.; Alvarez-Buylla, A.; Rowitch, DH. Astrocyte development and
heterogeneity. In: Barres, BA.; Freeman, MR.; Stevens, B., editors. Glia. Cold Spring Harbor: Cold
Spring Harbor Laboratory Press; 2015. p. 1-16.
Bear, MF.; Connors, BW.; Paradiso, MA. Exploring the Brain. 2. Philadelphia: Lippincott Williams &
Wilkins; 2001. Neuroscience.
Bear, MF.; Connors, BW.; Paradiso, MA. Exploring the Brain. 3. Philadelphia: Lippincott Williams &
Wilkins; 2007. Neuroscience.
Benes FM. Neurobiological investigations in cingulate cortex of schizophrenic brain. Schizophr Bull.
1993; 19:537–549. [PubMed: 8235456]
Benes FM, Lange N. Two-dimensional versus three-dimensional cell counting: a practical perspective.
Trends Neurosci. 2001; 24:11–17. [PubMed: 11163882]
Berger H. Untersuchungen uber den Zellgehalt der menschlichen Grosshirnrinde. Z Ges Neurol
Psychiat. 1921; 69:45–69.
Bernstein HG, Steiner J, Guest PC, Dobrowolny H, Bogerts B. Glial cells as key players in
schizophrenia pathology: recent insights and concepts of therapy. Schizophr Res. 2015; 161:4–18.
[PubMed: 24948484]
Bignami A. Glial cells in the central nervous system. Discussions Neurosci. 1991; 8:7–45.
Bjugn R, Gundersen HJ. Estimate of the total number of neurons and glial and endothelial cells in the
rat spinal cord by means of the optical disector. J Comp Neurol. 1993; 328:406–414. [PubMed:
Black, JA.; Ransom, BR. Glial cells: Morphology, development and function in the CNS. In: Adelman,
G.; Smith, BH., editors. Encyclopedia of Neuroscience. 2. Vol. 1. Elsevier; Amsterdam: 1999. p.
Blinkov SM. The glial index and distribution density of glial cells in the human brain stem. Arkh Anat
Gistol Embriol. 1963; 45:42–47. [PubMed: 14070855]
Blinkov, SM.; Glezer, II. A Quantitative Handbook. New York: Plenum Press; 1968. The Human Brain
in Figures and Tables; p. 482
Blinkov SM, Ivanitski GR. On the number of glial cells in the human brain. (Calculations on the
computer). Biofizika. 1965; 10:817–825. [PubMed: 5870851]
Braendgaard H, Evans SM, Howard CV, Gundersen HJ. The total number of neurons in the human
neocortex unbiasedly estimated using optical disectors. J Microsc. 1990; 157:285–304. [PubMed:
BrainFacts. A Primer on the brain and nervous system. A companion to The Kavli
Foundation, GATSBY, and the Society for Neuroscience.
von Bartheld et al. Page 29
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Braitenberg V, Atwood RP. Morphological observations on the cerebellar cortex. J Comp Neurol.
1958; 109:1–33. [PubMed: 13563670]
Brasileiro-Filho G, Guimaraes RC, Pittella JE. Quantitation and karyometry of cerebral neuroglia and
endothelial cells in liver cirrhosis and in the hepatosplenic schistosomiasis mansoni. Acta
Neuropathol. 1989; 77:582–590. [PubMed: 2501965]
Brautigam H, Steele JW, Westaway D, Fraser PE, St George-Hyslop PH, Gandy S, Hof PR, Dickstein
DL. The isotropic fractionator provides evidence for differential loss of hippocampal neurons in
two mouse models of Alzheimer’s disease. Mol Neurodegener. 2012; 7:58. [PubMed: 23173713]
Brizzee KR, Jacobs LA. The glia/neuron index in the submolecular layers of the motor cortex in the
cat. Anat Rec. 1959; 134:97–105. [PubMed: 13804526]
Brizzee KR, Vogt J, Kharetchko X. Postnatal changes in glia/neuron index with a comparison of
methods of cell enumeration in the white rat. Progr Brain Res. 1964; 4:136–149.
Brodal, A. Neurological Anatomy. 2. New York: Oxford Univ. Press; 1969.
Brodal, P. The Central Nervous System. New York: Oxford Univ. Press; 1992.
Brodal, P. Structure and Function. New York: Oxford Univ. Press; 2010. The Central Nervous System.
Brody H. Organization of the cerebral cortex. III. A study of aging in the human cerebral cortex. J
Comp Neurol. 1955; 102:511–556. [PubMed: 14381544]
Brody H. Structural changes in the ageing nervous system. Interdiscipl Top Gerontol. 1970; 7:9–21.
Brownson RH. Perineuronal satellite cells in the motor cortex of aging brains. J Neuropathol Exp
Neurol. 1956; 15:190–195. [PubMed: 13307281]
Buchweitz E, Weiss HR. Alterations in perfused capillary morphometry in awake vs anesthetized
brain. Brain Res. 1986; 377:105–111. [PubMed: 3730847]
Bullock, TH. Signals and Neuronal Coding. In: Qarton, GQ.; Melnechuk, T.; Schmitt, FO., editors.
The Neurosciences: A Study Program. New York: The Rockefeller Univ Press; 1967. p. 347-352.
Burne JF, Staple JK, Raff MC. Glial cells are increased proportionally in transgenic optic nerves with
increased numbers of axons. J Neurosci. 1996; 16:2064–2073. [PubMed: 8604051]
Carlo CN, Stevens CF. Structural uniformity of neocortex, revisited. Proc Natl Acad Sci USA. 2013;
110:1488–1493. [PubMed: 23297199]
Chacko LW. The laminar pattern of the lateral geniculate body in the primates. J Neurol Neurosurg
Psychiat. 1948; 11:211–224. [PubMed: 18878026]
Charvet CJ, Cahalane DJ, Finlay BL. Systematic, cross-cortex variation in neuron numbers in rodents
and primates. Cereb Cortex. 2015; 25:147–160. [PubMed: 23960207]
Clarke PG, Oppenheim RW. Neuron death in vertebrate development: in vitro methods. Methods Cell
Biol. 1995; 46:277–321. [PubMed: 7609654]
Coggeshall RE, Lekan HA. Methods for determining numbers of cells and synapses: a case for more
uniform standards of review. J Comp Neurol. 1996; 364:6–15. [PubMed: 8789272]
Coggeshall RE, La Forte R, Klein CM. Calibration of methods for determining numbers of dorsal root
ganglion cells. J Neurosci Methods. 1990; 35:187–194. [PubMed: 2084389]
Collins CE, Young NA, Flaherty DH, Airey DC, Kaas JH. A rapid and reliable method of counting
neurons and other cells in brain tissue: a comparison of flow cytometry and manual counting
methods. Front Neuroanat. 2010; 4:5. [PubMed: 20300202]
Committee on the Conduct of Science, National Academy of Sciences of the United States of America.
On being a scientist. Proc Natl Acad Sci USA. 1989; 86:9053–9074. [PubMed: 2594750]
Cotter DR, Pariante CM, Everall IP. Glial cell abnormalities in major psychiatric disorders: the
evidence and implications. Brain Res Bull. 2001; 55:585–595. [PubMed: 11576755]
Courchesne E, Pierce K, Schumann CM, Redcay E, Buckwalter JA, Kennedy DP, Morgan J. Mapping
early brain development in autism. Neuron. 2007; 56:399–413. [PubMed: 17964254]
Coyle JT, Schwarcz R. Mind glue: implications of glial cell biology for psychiatry. Arch Gen
Psychiatry. 2000; 57:90–93. [PubMed: 10632240]
Cragg BG. The density of synapses and neurones in the motor and visual areas of the cerebral cortex. J
Anat. 1967; 101:639–654. [PubMed: 4964696]
Cragg, BG. Gross, microscopical and ultramicroscopical anatomy of the adult nervous system. In:
Davison, AN.; Dobbing, J., editors. Applied Neurochemistry. Oxford: Blackwell; 1968. p. 3-47.
von Bartheld et al. Page 30
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Curcio, CA.; Buell, SJ.; Coleman, PD. Morphology of the aging central nervous system: Not all
downhill. In: Mortimer, JA.; Pirozzolo, FJ.; Maletta, GJ., editors. The Aging Motor System.
Advances in Neurogerontology. Vol. 3. New York: Praeger; 1982. p. 7-35.
Dahlstrom A, Hamberger A. Holger Hyden 1917–2000. Neurochem Res. 2001; 26:335–336.
Damier P, Hirsch EC, Agid Y, Graybiel AM. The substantia nigra of the human brain. II. Patterns of
loss of dopamine-containing neurons in Parkinson’s disease. Brain. 1999; 122:1437–1448.
[PubMed: 10430830]
Darlington, CL. The Female Brain. 2. Boca Raton: CRC Press; 2009.
Davanlou M, Smith DF. Unbiased stereological estimation of different cell types in rat cerebral cortex.
Image Anal Stereol. 2004; 23:1–11.
Delgado G, Estañol B. Acido, ergo sum: Holger Hydén--the neuroscientist in Cortázar’s Hopscotch.
Arq Neuropsiquiatr. 2013; 71:408–410. [PubMed: 23828537]
Devaney KO, Johnson HA. Neuron loss in the aging visual cortex of man. J Gerontol. 1980; 35:836–
841. [PubMed: 7440924]
Devinsky O, Vezzani A, Najjar S, De Lanerolle NC, Rogawski MA. Glia and epilepsy: excitability and
inflammation. Trends Neurosci. 2013; 36:174–184. [PubMed: 23298414]
Diamond MC, Scheibel AB, Murphy GM Jr, Harvey T. On the brain of a scientist: Albert Einstein.
Exp Neurol. 1985; 88:198–204. [PubMed: 3979509]
Dicke U, Roth G. Neuronal factors determining high intelligence. Philos Trans R Soc Lond B Biol Sci.
2016; 371:20150180. [PubMed: 26598734]
Dobbing J, Sands J. Timing of neuroblast multiplication in developing human brain. Nature. 1970;
226:639–640. [PubMed: 5444924]
Dobbing J, Sands J. Quantitative growth and development of human brain. Arch Dis Child. 1973;
48:757–767. [PubMed: 4796010]
Doetsch F. The glial identity of neural stem cells. Nat Neurosci. 2003; 6:1127–1134. [PubMed:
Donaldson, HH. The Growth of the Brain. Chicago, New York: Scribner; 1895.
Dorph-Petersen KA, Pierri JN, Sun Z, Sampson AR, Lewis DA. Stereological analysis of the
mediodorsal thalamic nucleus in schizophrenia: volume, neuron number, and cell types. J Comp
Neurol. 2004; 472:449–462. [PubMed: 15065119]
Dorph-Petersen KA, Caric D, Saghafi R, Zhang W, Sampson AR, Lewis DA. Volume and neuron
number of the lateral geniculate nucleus in schizophrenia and mood disorders. Acta Neuropathol.
2009; 117:369–384. [PubMed: 18642008]
Drasher ML. A criticism of the indiscriminate use of the Schmidt-Thannhauser method for the
fractionation of nucleic acids in biological material. Science. 1953; 118:181–183. [PubMed:
Ebbesson SOE, Tang D. A method for estimating the number of cells in histological sections. J Royal
Microsc Soc. 1965; 84:449–464.
Edelman, GM.; Mountcastle, V. Cortical Organization and the Group-Selective Theory of Higher brain
Function. Cambridge: MIT Press; 1978. The Mindful Brain.
Elsayed M, Magistretti PJ. A new outlook on mental illnesses: Glial involvement beyond the glue.
Front Cell Neurosci. 2015; 9:468. [PubMed: 26733803]
Eriksen N, Pakkenberg B. Total neocortical cell number in the mysticete brain. Anat Rec (Hoboken).
2007; 290:83–95. [PubMed: 17441201]
Fabricius K, Pakkenberg H, Pakkenberg B. No changes in neocortical cell volumes or glial cell
numbers in chronic alcoholic subjects compared to control subjects. Alcohol Alcohol. 2007;
42:400–406. [PubMed: 17341513]
Fabricius K, Jacobsen JS, Pakkenberg B. Effect of age on neocortical brain cells in 90+ year old
human females - a cell counting study. Neurobiol Aging. 2013; 34:91–99. [PubMed: 22878165]
Ferrero C. Le champ frontal granulaire magnocellulaire de l’ecorce cerebrale. Schweiz Arch Neurol
Psychiatr. 1947; 59:41–70.
Fields, RD. The Other Brain. New York: Simon & Schuster; 2009.
von Bartheld et al. Page 31
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Fields RD. Neuroscience. Change in the brain’s white matter. Science. 2010; 330:768–769. [PubMed:
Florey E. The zoological station at Naples and the neuron: personalities and encounters in a unique
institution. Biol Bull. 1985; 168(suppl):137–152.
Friede RL. Gliaindex und Hirnstoffwechsel. Wien Zeitschr Nervenh. 1953; 7:143–152.
Friede R. Der quantitative Anteil der Glia and der Cortexentwicklung. Acta Anat. 1954; 20:290–296.
[PubMed: 13137775]
Friede RL, van Houten WH. Neuronal extension and glial supply: functional significance of glia. Proc
Natl Acad Sci USA. 1962; 48:817–821. [PubMed: 13895192]
Friede RL. The relationship of body size, nerve cell size, axon length, and glial density in the
cerebellum. Proc Natl Acad Sci USA. 1963; 49:187–193. [PubMed: 13959737]
Firestein, S. Ignorance: How it drives Science. New York: Oxford Univ Press; 2012.
Galambos R. A glia-neural theory of brain function. Proc Natl Acad Sci USA. 1961; 47:129–136.
[PubMed: 13703008]
Gallatz K, Palkovits M, Szentágothai J. Ten billion neurons in the human cerebral cortex. Folia
Morphol (Praha). 1982; 30:133–135. [PubMed: 7117989]
Ganong, WF. The Nervous System. Los Altos: Lange medical Publications; 1977.
Ganong, WF. The Nervous System. 2. Los Altos: Lange medical Publications; 1979.
García-Amado M, Prensa L. Stereological analysis of neuron, glial and endothelial cell numbers in the
human amygdaloid complex. PLoS One. 2012; 7:e38692. [PubMed: 22719923]
García-Marín V, Blazquez-Llorca L, Rodriguez JR, Gonzalez-Soriano J, DeFelipe J. Differential
distribution of neurons in the gyral white matter of the human cerebral cortex. J Comp Neurol.
2010; 518:4740–4759. [PubMed: 20963826]
Gazzaniga, MS. Human: The Science Behind What Makes Your Brain Unique. New York: Harper
Collins; 2008.
Geuna S, Herrera-Rincon C. Update on stereology for light microscopy. Cell Tissue Res. 2015; 360:5–
12. [PubMed: 25743692]
Gittins R, Harrison PJ. Neuronal density, size and shape in the human anterior cingulated cortex: a
comparison of Nissl and NeuN staining. Brain Res Bull. 2004a; 63:155–160. [PubMed:
Gittins R, Harrison PJ. A quantitative morphometric study of the human anterior cingulate cortex.
Brain Res. 2004b; 1013:212–222. [PubMed: 15193531]
Glees, P. Morphology and Function. Oxford: Blackwell Scientific Publications; 1955. Neuroglia.
Glees, P. The biology of the neuroglia: a summary. In: Windle, WF., editor. Biology of Neuroglia.
Springfield: Charles C Thomas; 1958. p. 234-242.
Glees, P. The Human Brain. New York: Cambridge Univ Press; 1988.
Gredal O, Pakkenberg H, Karlsborg M, Pakkenberg B. Unchanged total number of neurons in motor
cortex and neocortex in amyotrophic lateral sclerosis: a stereological study. J Neurosci Methods.
2000; 95:171–176. [PubMed: 10752488]
Guillery RW, Herrup K. Quantification without pontification: choosing a method for counting objects
in sectioned tissues. J Comp Neurol. 1997; 386:2–7. [PubMed: 9303520]
Guillery RW. On counting and counting errors. J Comp Neurol. 2002; 447:1–7. [PubMed: 11967890]
Gundersen HJG, Bagger P, Bendtsen TF, Evans SM, Korbo L, Marcussen N, Møller A, Nielsen K,
Nyengaard JR, Pakkenberg B, Sørensen FB, Vesterby A, West MJ. The new stereological tools:
Disector, fractionator, nucleator and point sampled intercepts and their use in pathological
research and diagnosis. Acta Pathol Microbiol Immunol Scand. 1988; 96:857–881.
Gundersen V, Storm-Mathisen J, Bergersen LH. Neuroglial transmission. Physiol Rev. 2015; 95:695–
726. [PubMed: 26084688]
Hadjiolov AA, Tencheva ZS, Bojadjieva-Mikhailova AG. Isolation and some characteristics of cell
nuclei from brain cortex of adult cat. J Cell Biol. 1965; 26:383–393. [PubMed: 5325391]
Haines, D. Fundamental Neuroscience. 2. Philadelphia: Churchill Livingstone; 2002.
von Bartheld et al. Page 32
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Hammarberg, RA. Studien uber Klinik und Pathologie der Idiotie, nebst Untersuchungen uber die
normale Anatomie der Hirnrinde. Upsala: E.D.v. Berling; 1895.
Han X, Chen M, Wang F, Windrem M, Wang S, Shanz S, Xu Q, Oberheim NA, Bekar L, Betstadt S,
Silva AJ, Takano T, Goldman SA, Nedergaard M. Forebrain engraftment by human glial
progenitor cells enhances synaptic plasticity and learning in adult mice. Cell Stem Cell. 2013;
12:342–353. [PubMed: 23472873]
Hanley T. ‘Neuronal fall-out’ in the ageing brain: a critical review of the quantitative data. Age
Ageing. 1974; 3:133–151. [PubMed: 4219098]
Harrison PJ. The neuropathology of schizophrenia. A critical review of the data and their
interpretation. Brain. 1999; 122:593–624. [PubMed: 10219775]
Hatton, GI.; Parpura, V. Glia-Neuronal Signaling. Boston: Kluwer Acad. Publishers; 2004.
Hatton WJ, von Bartheld CS. Analysis of cell death in the trochlear nucleus of the chick embryo:
calibration of the optical disector counting method reveals systematic bias. J Comp Neurol 1999.
1999; 409:169–186.
Haug H. Der Einfluss der sakularen Akzeleration auf das Hirngewicht des Menschen und dessen
Anderung wahrend der Alterung. Gegenbaurs Morphol Jahrb. 1984; 130:481–500. [PubMed:
Haug H. Gibt es Nervenzellverluste wahrend der biologischen Alterung in der menschlichen
Hirnrinde? Ein morphometrischer Beitrag zu dieser Frage. Nervenheilkunde. 1985; 4:103–109.
Haug H. History of neuromorphometry. J Neurosci Methods. 1986; 18:1–17. [PubMed: 3540464]
Haug H. Brain sizes, surfaces, and neuronal sizes of the cortex cerebri: a stereological investigation of
man and his variability and a comparison with some mammals (primates, whales, marsupials,
insectivores, and one elephant). Am J Anat. 1987; 180:126–142. [PubMed: 3673918]
Haug H, Kühl S, Mecke E, Sass NL, Wasner K. The significance of morphometric procedures in the
investigation of age changes in cytoarchitectonic structures of human brain. J Hirnforsch. 1984;
25:353–374. [PubMed: 6481152]
Haug H, Rebhan I. Der Grauzellkoeffizient der menschlichen Hirnrinde. Berechnungen nach dem
Zahlenmaterial v. Economos. II. Acta Anat. 1956; 28:259–287. [PubMed: 13381330]
Hawkins A, Olszewski J. Glia/nerve cell index for cortex of the whale. Science. 1957; 126:76–77.
[PubMed: 13442648]
Haydon PG. GLIA: listening and talking to the synapse. Nat Rev Neurosci. 2001; 2:185–193.
[PubMed: 11256079]
Heller IH, Elliott KA. Desoxyribonucleic acid content and cell density in brain and human brain
tumors. Can J Biochem Physiol. 1954; 32:584–592. [PubMed: 13190467]
Hempel KJ, Treff WM. The glia cell density in clinically normal persons and schizophrenics. A
quantitative-morphological study with reference to glia cell-nerve cell coefficients. J Hirnforsch.
1959; 4:371–411. [PubMed: 14400963]
Herculano-Houzel S. The human brain in numbers: a linearly scaled-up primate brain. Front Hum
Neurosci. 2009; 3:31. [PubMed: 19915731]
Herculano-Houzel S. Not all brains are made the same: new views on brain scaling in evolution. Brain
Behav Evol. 2011; 78:22–36. [PubMed: 21691045]
Herculano-Houzel S. The remarkable, yet not extraordinary, human brain as a scaled-up primate brain
and its associated cost. Proc Natl Acad Sci USA. 2012; 109(Suppl 1):10661–10668. [PubMed:
Herculano-Houzel S. The glia/neuron ratio: How it varies uniformly across brain structures and species
and what that means for brain physiology and evolution. Glia. 2014; 62:1377–1391. [PubMed:
Herculano-Houzel S, Messeder DJ, Fonseca-Azevedo K, Pantoja NA. When larger brains do not have
more neurons: increased numbers of cells are compensated by decreased average cell size across
mouse individuals. Front Neuroanat. 2015; 9:64. [PubMed: 26082686]
Herculano-Houzel S, Lent R. Isotropic fractionator: a simple, rapid method for the quantification of
total cell and neuron numbers in the brain. J Neurosci. 2005; 25:2518–2521. [PubMed:
von Bartheld et al. Page 33
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Herculano-Houzel S, von Bartheld CS, Miller DJ, Kaas JH. How to count cells: the advantages and
disadvantages of the isotropic fractionator compared with stereology. Cell Tissue Res. 2015;
360:29–42. [PubMed: 25740200]
Hertz L, Hansson E, Rönnbäck L. Signaling and gene expression in the neuron-glia unit during brain
function and dysfunction: Holger Hydén in memoriam. Neurochem Int. 2001; 39:227–252.
[PubMed: 11434981]
Hess, HH. The rates of respiration of neurons and neuroglia in human cerebrum. In: Kety, SS.; Elkes,
J., editors. Regional Neurochemistry. Oxford: Pergamon Press; 1961. p. 200-212.
Hess HH, Thalheimer C. DNA and RNA and the cytoarchitecture of human frontal cortex. J
Neurochem. 1971; 18:1281–1290. [PubMed: 5118884]
Hilgetag CC, Barbas H. Are there ten times more glia than neurons in the brain? Brain Struct Funct.
2009; 213:365–366. [PubMed: 19198876]
Hof PR, Haroutunian V, Friedrich VL Jr, Byne W, Buitron C, Perl DP, Davis KL. Loss and altered
spatial distribution of oligodendrocytes in the superior frontal gyrus in schizophrenia. Biol
Psychiatry. 2003; 53:1075–1085. [PubMed: 12814859]
Hof, PR.; Mobbs, CV. Preface. London: Elsevier; 2009. Handbook of the Neuroscience of Aging; p.
Howard, CV.; Reed, MG. Three–Dimensional Measurement in Microscopy. New York: Springer; 1998.
Unbiased Stereology.
Hubel, DH. The Brain. A Scientific American Book. Freeman and Co, San Francisco: Freeman and
Co; 1979. The brain; p. 2-12.[Sci. Am. 241 (3)]
Hyden, H. The Neuron. In: Brachet, J.; Mirsky, AE., editors. The Cell: Biochemistry, Physiology,
Morphology. Vol. 4. New York: Acad Press; 1960. p. 215-323.
Hyden H. Satellite cells in the nervous system. Sci Am. 1961; 205:62–70.
Hyden, H. Dynamic aspects of the neuron-glia relationship - a study with microchemical methods. In:
Hyden, H., editor. The Neuron. Amsterdam: Elsevier; 1967a. p. 179-217.
Hyden H. RNA in brain cells. The Neurosciences: A Study Program. 1967b:248–266.
Hyden H, Pigon A. A cytophysiological study of the functional relationship between oligodendroglial
cells and nerve cells of Deiters’ nucleus. J Neurochem. 1960; 6:57–72. [PubMed: 13717152]
Iñiguez C, Gayoso MJ, Carreres J. A versatile and simple method for staining nervous tissue using
Giemsa dye. J Neurosci Methods. 1985; 13:77–86. [PubMed: 3887046]
Ioannidis JP. Why science is not necessarily self-correcting. Perspect Psychol Sci. 2012; 7:645–654.
[PubMed: 26168125]
Jacobson, M. Developmental Neurobiology. 3. New York: Plenum Press; 1991.
Jehee JF, Murre JM. The scalable mammalian brain: emergent distributions of glia and neurons. Biol
Cybern. 2008; 98:439–445. [PubMed: 18414892]
Jensen, D. The Human Nervous System. New York: Appleton-Century-Crofts; 1980.
Jensen GB, Pakkenberg B. Do alcoholics drink their neurons away? Lancet. 1993; 342:1201–1204.
[PubMed: 7901529]
Jerison, H. Evolution of the brain and intelligence. New York: Academic Press; 1973.
Jorm, AF. Understanding Senile Dementia. Melbourne: Chapman & Hall; 1987. p. 158
Kandel, ER.; Schwartz, JH. Principles of Neural Science. 1. New York: Elsevier; 1981.
Kandel, ER.; Schwartz, JH. Principles of Neural Science. 2. New York: Elsevier; 1985.
Kandel, ER.; Schwartz, JH.; Jessell, TM. Principles of Neural Science. 3. Norwalk, Connecticut:
Appleton & Lange; 1991.
Kandel, ER.; Schwartz, JH.; Jessell, TM. Principles of Neural Science. 4. New York: McGrawhill;
Kandel, ER.; Schwartz, JH.; Jessell, TM.; Siegelbaum, SA.; Hudspeth, AJ. Principles of Neural
Science. 5. New York: McGraw-Hill; 2013.
Kaplan S, Geuna S, Ronchi G, Ulkay MB, von Bartheld CS. Calibration of the stereological estimation
of the number of myelinated axons in the rat sciatic nerve: a multicenter study. J Neurosci
Methods. 2010; 187:90–99. [PubMed: 20064555]
von Bartheld et al. Page 34
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Karbowski J. Scaling of brain metabolism and blood flow in relation to capillary and neural scaling.
PLoS One. 2011; 6:e26709. [PubMed: 22053202]
Karlsen AS, Pakkenberg B. Total numbers of neurons and glial cells in cortex and basal ganglia of
aged brains with Down syndrome - a stereological study. Cereb Cortex. 2011; 21:2519–2524.
[PubMed: 21427166]
Kato T, Kurokawa M. Isolation of cell nuclei from the mammalian cerebral cortex and their assortment
on a morphological basis. J Cell Biol. 1967; 32:649–662. [PubMed: 4226774]
Kausler, DH.; Kausler, BC.; Krupsaw, JA. The Essential Guide to Aging in the Twenty-First Century:
Mind, Body, and Behavior. St Louis: University of Missouri Press; 2007.
Kettenmann, H.; Ransom, BH. Neuroglia. Oxford: Oxford Univ. Press; 1995.
Kettenmann, H.; Ransom, BR. Neuroglia. 3. Oxford: Oxford Univ. Press; 2013.
Kettenmann, H.; Verkhratsky, A. Glial Cells. In: Pfaff, DW., editor. Neuroscience in the 21st Century.
New York: Springer; 2013. p. 475-506.
Koch, C. A Neurobiological Approach. Englewood: Roberts & Company; 2004. The Quest for
Konigsmark, BW. Methods for the counting of neurons. In: Nauta, WJH.; Ebbesson, SOE., editors.
Contemporary Research Methods in Neuroscience. New York: Springer; 1970. p. 315-340.
Konigsmark BW, Murphy EA. Neuronal populations in the human brain. Nature. 1970; 228:1335–
1336. [PubMed: 5488113]
Koob, A. The Root of Thought: Unlocking Glia – the Brain Cell That Will Help Us Sharpen Our Wits,
Heal Injury, and Treat Brain Disease. Upper Saddle River: Pearson Education; 2009.
Kordower JH, Olanow CW, Dodiya HB, Chu Y, Beach TG, Adler CH, Halliday GM, Bartus RT.
Disease duration and the integrity of the nigrostriatal system in Parkinson’s disease. Brain. 2013;
136:2419–2431. [PubMed: 23884810]
Kreuzfuchs S. Die Groesse der Oberflaeche des Kleinhirns. Arb Neur Inst Univ Wien. 1902; 9:274–
Kryspin-Exner W. Beitraege zur Morphologie der Glia im Nissl-Bilde. Z Anat Entw Gesch. 1943;
Kryspin-Exner W. Über die Architektonik der Glia im Zentralnervensystem des Menschen und der
Säugetiere. Proc Ist Internat Congr Neuropath, Rome. 1952; 3:504–510.
Kühnel W. In memoriam des Anatomen Herbert W. Haug (1920–2002). Ann Anat. 2003; 185:293–
Kuffler SW, Nicholls JG. The physiology of neuroglial cells. Ergeb Physiol. 1966; 57:1–90. [PubMed:
Kuffler, SW.; Nicholls, JG. From Neuron to Brain. Sunderland: Sinauer Associates; 1976.
Kuffler, SW.; Nicholls, JG. From Neuron to Brain. Sunderland: Sinauer Associates; 1985.
Lambert, K.; Kinsley, CH. Clinical Neuroscience. New York: Worth Publishers; 2005.
Landolt R, Hess HH, Thalheimer C. Regional distribution of some chemical structural components of
the human nervous system. I. DNA, RNA and ganglioside sialic acid. J Neurochem. 1966;
13:1441–1452. [PubMed: 5964647]
Lange W. Cell number and cell density in the cerebellar cortex of man and some other mammals. Cell
Tissue Res. 1975; 157:115–124. [PubMed: 804353]
Lange H, Thorner G, Hopf A. Morphometric-statistical structure analysis of human striatum, pallidum
and nucleus su-thalamicus. III. Nucleus subthalamicus. J Hirnforsch. 1976; 17:31–41. [PubMed:
Lauwers F, Cassot F, Lauwers-Cances V, Puwanarajah P, Duvernoy H. Morphometry of the human
cerebral cortex microcirculation: General characteristics and space-related profiles. Neuroimage.
2008; 39:936–948. [PubMed: 17997329]
Lee GM, Thornthwaite JT, Rasch EM. Picogram per cell determination of DNA by flow
cytofluorometry. Anal Biochem. 1984; 137:221–226. [PubMed: 6731800]
Lemke G. Glial control of neuronal development. Annu Rev Neurosci. 2001; 24:87–105. [PubMed:
von Bartheld et al. Page 35
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Lent R, Azevedo FA, Andrade-Moraes CH, Pinto AV. How many neurons do you have? Some dogmas
of quantitative neuroscience under revision. Eur J Neurosci. 2012; 35:1–9. [PubMed: 22151227]
Leuba G, Garey LJ. Evolution of neuronal numerical density in the developing and aging human visual
cortex. Hum Neurobiol. 1987; 6:11–8. [PubMed: 3583841]
Leuba G, Garey LJ. Comparison of neuronal and glial numerical density in primary and secondary
visual cortex of man. Exp Brain Res. 1989; 77:31–38. [PubMed: 2792267]
Ling EA, Leblond CP. Investigation of glial cells in semithin sections. II. Variation with age in the
numbers of the various glial cell types in rat cortex and corpus callosum. J Comp Neurol. 1973;
149:73–81. [PubMed: 4121706]
Lojda, Zd. Quantitative study of the Purkinje cells in the brain [Czech. Cs Morfol. 1955; 3:66–78.
Lovtrup-Rein H, McEwen BS. Isolation and fractionation of rat brain nuclei. J Cell Biol. 1966;
30:405–415. [PubMed: 5968977]
Lyck L, Santamaria ID, Pakkenberg B, Chemnitz J, Schrøder HD, Finsen B, Gundersen HJ. An
empirical analysis of the precision of estimating the numbers of neurons and glia in human
neocortex using a fractionator-design with sub-sampling. J Neurosci Methods. 2009; 182:143–
156. [PubMed: 19520115]
Lyness SA, Zarow C, Chui HC. Neuron loss in key cholinergic and aminergic nuclei in Alzheimer
disease: a meta-analysis. Neurobiol Aging. 2003; 24:1–23. [PubMed: 12493547]
Mann, DMA. Dopamine neurons of the vertebrate brain: some aspects of anatomy and pathology. In:
Winlow, W.; Markestein, R., editors. The Neurobiology of Dopamine Systems. Manchester:
Manchester Univ. Press; 1986. p. 87-103.
Mares V, Crkovská J, Marsak TL, Stípek S. DNA content in nerve-cell nucleus. A biochemical and
cytophotometric study of the rat cerebrum. Neuroscience. 1985; 16:45–47. [PubMed: 2423917]
Margolis FL. DNA and DNA-polymerase activity in chicken brain regions during ontogeny. J
Neurochem. 1969; 16:447–456. [PubMed: 5795595]
Maron, ME. The RAND Corporation, Memorandum RM-3522-PR. 1963. Artificial Intelligence and
Brain Mechanisms; p. 1-35.
Mayhew TM, Gundersen HJ. ‘If you assume, you can make an ass out of u and me’: A decade of the
disector for stereological counting of particles in 3D space. J Anat. 1996; 188:1–15. [PubMed:
Mayhew TM, MacLaren R, Henery CC. Fractionator studies on Purkinje cells in the human
cerebellum: numbers in right and left halves of male and female brains. J Anat. 1990; 169:63–70.
[PubMed: 2200769]
McGeer PL, McGeer EG, Suzuki JS. Aging and extrapyramidal function. Arch Neurol. 1977; 34:33–
35. [PubMed: 12731]
McKeown RE, Cuffe SP, Schulz RM. US suicide rates by age group, 1970–2002: an examination of
recent trends. Am J Public Health. 2006; 96:1744–1751. [PubMed: 17008567]
McMahan, UJ. Steve: Remembrances of Stephen W. Kuffler. Sunderland: Sinauer Associates; 1990. p.
Meehan PJ, Saltzman LE, Sattin RW. Suicides among older United States residents: epidemiologic
characteristics and trends. Am J Public Health. 1991; 81:1198–1200. [PubMed: 1951834]
Merril, BC. A Master’s Thesis in Philosophy. University of Nevada; Reno: 2009. The Neural Dogma;
p. 124Thesis
Meynert T. Der Bau der Grosshirnrinde und seine ortlichen Verschiedenheiten, nebst einem
pathologisch-anatomischen Corollarium. Vierteljahresschr Psych. 1868:88–113.
Meynert T. Der Bau der Grosshirnrinde und seine örtlichen Verschiedenheiten. 1872:18. Reprint in
Vierteljahrschrift fur Psychiatrie etc.
Miller DJ, Balaram P, Young NA, Kaas JH. Three counting methods agree on cell and neuron numbers
in chimpanzee primary visual cortex. Front Neuroanat. 2014; 8:36. [PubMed: 24904305]
Morgan JT, Chana G, Pardo CA, Achim C, Semendeferi K, Buckwalter J, Courchesne E, Everall IP.
Microglial activation and increased microglial density observed in the dorsolateral prefrontal
cortex in autism. Biol Psychiatry. 2010; 68:368–376. [PubMed: 20674603]
von Bartheld et al. Page 36
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Morrison JH, Hof PR. Life and death of neurons in the aging brain. Science. 1997; 278:412–419.
[PubMed: 9334292]
Mosch B, Morawski M, Mittag A, Lenz D, Tarnok A, Arendt T. Aneuploidy and DNA replication in
the normal human brain and Alzheimer’s disease. J Neurosci. 2007; 27:6859–6867. [PubMed:
Mota B, Herculano-Houzel S. All brains are made of this: a fundamental building block of brain matter
with matching neuronal and glial masses. Front Neuroanat. 2014; 8:127. [PubMed: 25429260]
Mountjoy CQ, Roth M, Evans NJ, Evans HM. Cortical neuronal counts in normal elderly controls and
demented patients. Neurobiol Aging. 1983; 4:1–11. [PubMed: 6877482]
Mouton, PR. An introduction for bioscientists. Baltimore: Johns Hopkins University Press; 2002.
Principles and practices of unbiased stereology.
Mouton PR, Pakkenberg B, Gundersen HJ, Price DL. Absolute number and size of pigmented locus
coeruleus neurons in young and aged individuals. J Chem Neuroanat. 1994; 7:185–190.
[PubMed: 7848573]
Mufson EJ, Benzing WC. Lack of neocortical nerve cell loss in Alzheimer’s disease: reality or
methodological artifact. Neurobiol Aging. 1994; 15:361–362. discussion: 379–380. [PubMed:
Mühlmann M. About the question of glia formation. Zur Neurogliabildungsfrage Beitr Pathol Anat
Allg Pathol. 1936; 96:361–374.
Mullen RJ, Buck CR, Smith AM. NeuN, a neuronal specific nuclear protein in vertebrates.
Development. 1992; 116:201–211. [PubMed: 1483388]
Nairn JG, Bedi KS, Mayhew TM, Campbell LF. On the number of Purkinje cells in the human
cerebellum: unbiased estimates obtained by using the “fractionator”. J Comp Neurol. 1989;
290:527–532. [PubMed: 2613942]
Nansen F. The Structure and Combination of the Histological Elements in the Central Nervous System.
Bergens Museums Aarsberetning for 1886. 1887:29–195.
Nass S, Nass MM. Intramitochondrial fibers with DNA characteristics. II. Enzymatic and other
hydrolytic treatments. J Cell Biol. 1963; 19:613–629. [PubMed: 14086139]
National Institute of Neurological Disorders and Stroke. FY2001 NINDS Congressional Justification
National Institute of Neurological Disorders and Stroke. FY2011 NINDS Congressional Justification
National Institute of Neurological Disorders and Stroke. FY2015 NINDS Congressional Justification
Nauta, WJH.; Feirtag, M. The Brain. A Scientific American Book. San Francisco: Freeman and Co;
1979. The organization of the brain; p. 41-55.[Sci Am 241 (3)]
Nedergaard M, Ransom B, Goldman SA. New roles for astrocytes: redefining the functional
architecture of the brain. Trends Neurosci. 2003; 26:523–530. [PubMed: 14522144]
Neville, C. The Complete Guide to Referencing and Avoiding Plagiarism. 2. Maidenhead: Open UP;
Nicholls JG. Closing Remarks. Ann New York Acad Sci. 1991; 633:633–636.
Nicholls, JG. Biographical Memoirs. Vol. 74. Natl Acad Press; 1998. Stephen W. Kuffler, August 24,
1913 – October 11, 1980; p. 193-208.
Nicholls, JG.; Martin, AR.; Wallace, BG. From Neuron to Brain. 2. Sunderland: Sinauer; 1985.
Nicholls, JG.; Martin, AR.; Wallace, BG. From Neuron to Brain. 3. Sunderland: Sinauer; 1992.
Nicholls, JG.; Martin, AR.; Fuchs, PA.; Brown, DA.; Diamond, ME.; Weisblat, DA. From Neuron to
Brain. 5. Sunderland: Sinauer; 2012.
Nielsen RD, Abitz M, Pakkenberg B. Neuron and glial cell numbers in the mediodorsal thalamic
nucleus in brains of schizophrenic subjects. Image Anal Stereol. 2008; 27:133–141.
Nishiyama A, Yang Z, Butt A. What’s in a name? J Anat. 2005; 207:687–693. [PubMed: 16367796]
von Bartheld et al. Page 37
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Nissl F. Nervenzellen und graue Substanz. Munch Med Wochenschr. 1898; 45:988–992. 1023–1029,
NPR. New project would map the human brain. 2013.
Noback, CR.; Demarest, RJ. Introduction and Review. New York: McGrawHill; 1972. The Human
Nervous System.
Noback, CR.; Demarest, RJ. Introduction and Review. 2. New York: McGrawHill; 1975. The Human
Nervous System.
Noback, CR.; Strominger, NL.; Demarest, RJ. Introduction and Review. 4. Philadelphia: LEA &
FEBIGER; 1991. The Human Nervous System.
Nolte, J. The Human Brain. St. Louis: The C.V. Mosby Company; 1981.
Nolte, J. The Human Brain. 3. St. Louis: The C.V. Mosby Company; 1993.
Nurnberger JI, Gordon MW. The cell density of neural tissues: direct counting method and possible
applications as a biologic referent. Prog Neurobiol. 1957; 2:100–128. [PubMed: 13477746]
Nurnberger, JI. Direct enumeration of cells of the brain. In: Windle, WF., editor. Biology of Neuroglia.
Springfield: Charles C Thomas; 1958. p. 193-202.
Nuzzo R. Fooling ourselves. Nature. 2015; 526:182–185. [PubMed: 26450039]
Ohm TG, Busch C, Bohl J. Unbiased estimation of neuronal numbers in the human nucleus coeruleus
during aging. Neurobiol Aging. 1997; 18:393–399. [PubMed: 9330970]
Ongür D, Drevets WC, Price JL. Glial reduction in the subgenual prefrontal cortex in mood disorders.
Proc Natl Acad Sci USA. 1998; 95:13290–13295. [PubMed: 9789081]
Palmen SJ, van Engeland H, Hof PR, Schmitz C. Neuropathological findings in autism. Brain. 2004;
127:2572–2583. [PubMed: 15329353]
Pakkenberg B. Pronounced reduction of total neuron number in mediodorsal thalamic nucleus and
nucleus accumbens in schizophrenics. Arch Gen Psychiatry. 1990; 47:1023–1028. [PubMed:
Pakkenberg B. Stereological quantitation of human brains from normal and schizophrenic individuals.
Acta Neurol Scand Suppl. 1992; 137:20–33. [PubMed: 1414258]
Pakkenberg B. Leucotomized schizophrenics lose neurons in the mediodorsal thalamic nucleus.
Neuropathol Appl Neurobiol. 1993; 19:373–380. [PubMed: 8278019]
Pakkenberg B. Total nerve cell number in neocortex in chronic schizophrenics and controls estimated
using optical disectors. Biol Psychiatry. 1993; 34:768–772. [PubMed: 8292680]
Pakkenberg B, Gundersen HJ. Total number of neurons and glial cells in human brain nuclei estimated
by the disector and the fractionator. J Microsc. 1988; 150:1–20. [PubMed: 3043005]
Pakkenberg B, Gundersen HJ. New stereological method for obtaining unbiased and efficient estimates
of total nerve cell number in human brain areas. Exemplified by the mediodorsal thalamic
nucleus in schizophrenics. Acta Pathol Microbiol Immunol Scand. 1989; 97:677–681.
Pakkenberg B, Gundersen HJ. Neocortical neuron number in humans: effect of sex and age. J Comp
Neurol. 1997; 384:312–320. [PubMed: 9215725]
Pakkenberg B, Evans SM, Møller A, Brændgaard H, Gundersen HJG. Total number of neurons in
human neocortex related to age and sex estimated by way of optical disectors. Acta Stereologica.
1989; 8:251–256.
Pakkenberg B, Møller A, Gundersen HJ, Mouritzen Dam A, Pakkenberg H. The absolute number of
nerve cells in substantia nigra in normal subjects and in patients with Parkinson’s disease
estimated with an unbiased stereological method. J Neurol Neurosurg Psychiatry. 1991; 54:30–
33. [PubMed: 2010756]
Pakkenberg B, Pelvig D, Marner L, Bundgaard MJ, Gundersen HJ, Nyengaard JR, Regeur L. Aging
and the human neocortex. Exp Gerontol. 2003; 38:95–99. [PubMed: 12543266]
Pakkenberg H. The number of nerve cells in the cerebral cortex of man. J Comp Neurol. 1966; 128:17–
20. [PubMed: 4165856]
Palay, SL. An electron-microscopical study of neuroglia. In: Windle, WF., editor. Biology of
Neuroglia. Springfield: Charles C. Thomas; 1958. p. 24-38.
von Bartheld et al. Page 38
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Pastor, J.; Sola, RG. The role of astrocytes in epileptogenesis. In: Kaneez, FS., editor. Underlying
Mechanisms of Epilepsy. Rijeka: InTech; 2008. p. 19-44.
Pedersen KM, Marner L, Pakkenberg H, Pakkenberg B. No global loss of neocortical neurons in
Parkinson’s disease: a quantitative stereological study. Mov Disord. 2005; 20:164–171.
[PubMed: 15468109]
Pelvig DP, Pakkenberg H, Regeur L, Oster S, Pakkenberg B. Neocortical glial cell numbers in
Alzheimer’s disease. A stereological study. Dement Geriatr Cogn Disord. 2003; 16:212–219.
[PubMed: 14512716]
Pelvig DP, Pakkenberg H, Stark AK, Pakkenberg B. Neocortical glial cell numbers in human brains.
Neurobiol Aging. 2008; 29:1754–1762. [PubMed: 17544173]
Penn NW, Suwalski R. Quantitative determination of deoxyribonucleic acid in rat brain. Biochem J.
1969; 115:563–568. [PubMed: 5353529]
Peters A, Morrison JH, Rosene DL, Hyman BT. Feature article: are neurons lost from the primate
cerebral cortex during normal aging? Cereb Cortex. 1998; 8:295–300. [PubMed: 9651126]
Peters A. The effects of normal aging on myelin and nerve fibers: a review. J Neurocytol. 2002;
31:581–593. [PubMed: 14501200]
Pfrieger FW, Barres BA. What the fly’s glia tell the fly’s brain. Cell. 1995; 83:671–674. [PubMed:
Pitt B. Loss in late life. BMJ. 1998; 316:1452–1454. [PubMed: 9572763]
Ponomarev VS. Glial index in vestibular nuclei of humans, macacos, and dogs. Arkh Anat Gistol
Embriol. 1966; 51:100–104. [PubMed: 4971939]
Pope, A. Implications of histochemical studies for metabolism of the neuroglia. In: Windle, WF.,
editor. Biology of Neuroglia. Springfield: Charles C Thomas; 1958. p. 211-233.
Pope A. The intralaminar distribution of dipeptidase activity in human frontal isocortex. J Neurochem.
1959; 4:31–41. [PubMed: 13655090]
Popken GJ, Bunney WE Jr, Potkin SG, Jones EG. Subnucleus-specific loss of neurons in medial
thalamus of schizophrenics. Proc Natl Acad Sci USA. 2000; 97:9276–9280. [PubMed:
Powell TPS, Guillery RW, Cowan WM. A quantitative study of the fornix mamillo-thalamic system. J
Anat. 1957; 91:419–435. [PubMed: 13475143]
Proctor, R.; Schiebinger, LL. Agnotology: The Making and Unmaking of Ignorance. Stanford:
Stanford Univ Press; 2008.
Purves, D. How they work and what that tells us about who we are. Upper Saddle River: FT Press;
2010. Brains.
Rajkowska G. Postmortem studies in mood disorders indicate altered numbers of neurons and glial
cells. Biol Psychiatry. 2000; 48:766–777. [PubMed: 11063973]
Rajkowska G. Cell pathology in mood disorders. Semin Clin Neuropsychiatry. 2002; 7:281–292.
[PubMed: 12382210]
Regeur L, Jensen GB, Pakkenberg H, Evans SM, Pakkenberg B. No global neocortical nerve cell loss
in brains from patients with senile dementia of Alzheimer’s type. Neurobiol Aging. 1994a;
15:347–352. [PubMed: 7936059]
Regeur L, Jensen GB, Pakkenberg H, Evans SM, Pakkenberg B. Author’s response to commentaries.
Neurobiol Aging. 1994b; 15:379–380.
Reichenbach A. Glia:neuron index: review and hypothesis to account for different values in various
mammals. Glia. 1989; 2:71–77. [PubMed: 2524444]
Ribeiro PFM, Ventura-Antunes L, Gabi M, Mota B, Grinberg LT, Farfel JM, Ferretti-Rebustini REL,
Leite REP, Jacob Filho W, Herculano-Houzel S. The human cerebral cortex is neither one nor
many: neuronal distribution reveals two quantitatively different zones in the gray matter, three in
the white matter, and explains local variations in cortical folding. Front Neuroanat. 2013; 7:28.
[PubMed: 24032005]
Rilling JK, Insel TR. The primate neocortex in comparative perspective using magnetic resonance
imaging. J Hum Evol. 1999; 37:191–223. [PubMed: 10444351]
von Bartheld et al. Page 39
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Robins E, Smith DE, Eydt KM. The quantitative histochemistry of cerebral cortex. J Neurochem.
1956; 1:54–67. [PubMed: 13346374]
Rodríguez-Arellano JJ, Parpura V, Zorec R, Verkhratsky A. Astrocytes in physiological aging and
Alzheimer’s disease. Neuroscience. 2016 in press.
Rowland LP, Mettler FA. Cell concentration and laminar thickness in the frontal cortex of psychotic
patients; studies on cortes removed at operation. J Comp Neurol. 1949; 90:255–280. [PubMed:
Saldanha J, Gannicliffe A, Itzhaki RF. An improved method for preparing DNA from human brain. J
Neurosci Methods. 1984; 11:275–279. [PubMed: 6096638]
Savioz A, Blouin JL, Guidi S, Antonarakis SE, Bouras C. A method for the extraction of genomic
DNA from human brain tissue fixed and stored in formalin for many years. Acta Neuropathol.
1997; 93:408–413. [PubMed: 9113206]
Scheff SW, Baldwin SA. Preparation and staining of tissue for counting neurons. Kopf Carrier. 1996;
Schlote W. Zur Gliaarchitektonik der menschlichen Grosshirnrinde im Niss-Bild. Arch Psychiatr
Nervenkr Z Gesamte Neurol Psychiatr. 1959; 199:573–595. [PubMed: 14442925]
Schmitz C, Hof PR. Recommendations for straightforward and rigorous methods of counting neurons
based on a computer simulation approach. J Chem Neuroanat. 2000; 20:93–114. [PubMed:
Schmitz C, Hof PR. Design-based stereology in neuroscience. Neuroscience. 2005; 130:813–831.
[PubMed: 15652981]
Schmutte T, O’Connell M, Weiland M, Lawless S, Davidson L. Stemming the tide of suicide in older
white men: a call to action. Am J Mens Health. 2009; 3:189–200. [PubMed: 19477749]
Schröder KF, Hopf A, Lange H, Thörner G. Morphometrical-statistical structure analysis of human
striatum, pallidum and subthalamic nucleus. J Hirnforsch. 1975; 16:333–350. [PubMed:
Selemon LD, Begovic A. Stereologic analysis of the lateral geniculate nucleus of the thalamus in
normal and schizophrenic subjects. Psychiatry Res. 2007; 151:1–10. [PubMed: 17383740]
Shariff GA. Cell counts in the primate cerebral cortex. J Comp Neurol. 1953; 98:381–400. [PubMed:
Sherwood CC, Stimpson CD, Raghanti MA, Wildman DE, Uddin M, Grossman LI, Goodman M,
Redmond JC, Bonar CJ, Erwin JM, Hof PR. Evolution of increased glia-neuron ratios in the
human frontal cortex. Proc Natl Acad Sci USA. 2006; 103:13606–13611. [PubMed: 16938869]
Sholl, DA. The Organization of the Cerebral Cortex. Methuen, London, New York: Wiley and Sons;
Sholl DA. The measurable parameters of the cerebral cortex and their significance in its organization.
Prog Neurobiol. 1956; (2):324–333. [PubMed: 13441807]
Sholl DA. A comparative study of the neuronal packing density in the cerebral cortex. J Anat. 1959;
93:143–158. [PubMed: 13641114]
Snell, RS. Clinical Anatomy for Medical Students. Boston: Little, Brown + Co; 1980.
Snell, RS. Clinical Anatomy for Medical Students. 6. Philadelphia: Lippincott Williams & Wilkins;
Smith K. Settling the great glia debate. Nature. 2010; 468:160–162. [PubMed: 21068805]
Soper B, Rosenthal G. The number of neurons in the brain: how we report what we do not know. Tech
Psychol. 1988; 15:153–156.
Srinivasan M, Sedmak D, Jewell S. Effect of fixatives and tissue processing on the content and
integrity of nucleic acids. Am J Pathol. 2002; 161:1961–1971. [PubMed: 12466110]
Stark AK, Pakkenberg B. Histological changes of the dopaminergic nigrostriatal system in aging. Cell
Tissue Res. 2004; 318:81–92. [PubMed: 15365813]
Stephan H, Frahm H, Baron G. New and revised data on volumes of brain structures in insectivores
and primates. Folia Primatol. 1981; 35:1–29. [PubMed: 7014398]
Steward, O. Principles of cellular, molecular and developmental neuroscience. New York: Springer;
von Bartheld et al. Page 40
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Steward, O. Functional Neuroscience. New York: Springer; 2000.
Streit, WJ. Role of macrophages and microglia in the injured CNS. In: Berry, M.; Logan, A., editors.
CNS Injuries: Cellular Responses and Pharmacological Strategies. Boca Raton, FL: CRC Press;
1999. p. 81-97.
Streit, W. Microglial cells. In: Kettenmann, H.; Ransom, BR., editors. Neuroglia. 3. Oxford: Oxford
Univ Press; 2013. p. 86-97.
Stolzenburg JU, Reichenbach A, Neumann M. Size and density of glial and neuronal cells within the
cerebral neocortex of various insectivoran species. Glia. 1989; 2:78–84. [PubMed: 2524445]
Szentágothai J. The modular architectonic principle of neural centers. Rev Physiol Biochem
Pharmacol. 1983; 98:11–61. [PubMed: 6361964]
Terry RD, DeTeresa R, Hansen LA. Neocortical cell counts in normal human adult aging. Ann Neurol.
1987; 21:530–539. [PubMed: 3606042]
Thörner G, Lange H, Hopf A. Morphometrical-statistical structure analysis of human striatum, pallidus
and subthalamic nucleus. II. Globus pallidus. J Hirnforsch. 1975; 16:401–413. [PubMed:
Thompson H. The total number of functional cells in the cerebral cortex of man, and the percentage of
the total volume of the cortex composed of nerve cell bodies, calculated from Karl
Hammarberg’s data; together with a comparison of the number of giant cells with the number of
pyramidal fibers. J Comp Neurol. 1899; 9:113–140.
Thomson, W. Nature Series. Popular Lectures and Addresses. Vol. 1. London: Macmillan and Co;
1889. The practical applications of electricity. Electrical units of measurement; p. 73-136.(A
lecture delivered at the Institution of Civil Engineers, May 3, 1883)
Thörner G, Lange H, Hopf A. Morphometrical-statistical structure analysis of human striatum, pallidus
and subthalamic nucleus. II. Globus pallidus. J Hirnforsch. 1975; 16:401–413. [PubMed:
Thune JJ, Pakkenberg B. Stereological studies of the schizophrenic brain. Brain Res Rev. 2000;
31:200–204. [PubMed: 10719149]
Todtenkopf MS, Vincent SL, Benes FM. A cross-study meta-analysis and three-dimensional
comparison of cell counting in the anterior cingulate cortex of schizophrenic and bipolar brain.
Schizophr Res. 2005; 73:79–89. [PubMed: 15567080]
Tower DB. Structural and functional organization of mammalian cerebral cortex; the correlation of
neurone density with brain size; cortical neurone density in the fin whale (Balaenoptera physalus
L.) with a note on the cortical neurone density in the Indian elephant. J Comp Neurol. 1954;
101:19–52. [PubMed: 13211853]
Tower DB, Elliott KAC. Activity of acetylcholine system in cerebral cortex of various unanesthetized
mammals. Amer J Physiol. 1952; 168:747–759. [PubMed: 14914888]
Tower DB, Young OM. The activities of butyrylcholinesterase and carbonic anhydrase, the rate of
anaerobic glycolysis, and the question of a constant density of glial cells in cerebral cortices of
various mammalian species from mouse to whale. J Neurochem. 1973; 20:269–278. [PubMed:
Triarhou LC. Georg N. Koskinas (1885–1975) and his scientific contributions to the normal and
pathological anatomy of the human brain. Brain Res Bull. 2005; 68:121–139. [PubMed:
Triarhou LC. The signalling contributions of Constantin von Economo to basic, clinical and
evolutionary neuroscience. Brain Res Bull. 2006; 69:223–243. [PubMed: 16564418]
Tsai PS, Kaufhold JP, Blinder P, Friedman B, Drew PJ, Karten HJ, Lyden PD, Kleinfeld D.
Correlations of neuronal and microvascular densities in murine cortex revealed by direct counting
and colocalization of nuclei and vessels. J Neurosci. 2009; 29:14553–14570. [PubMed:
Ullian EM, Sapperstein SK, Christopherson KS, Barres BA. Control of synapse number by glia.
Science. 2001; 291:657–660. [PubMed: 11158678]
Vawter MP, Freed WJ, Kleinman JE. Neuropathology of bipolar disorder. Biol Psychiatry. 2000;
48:486–504. [PubMed: 11018222]
von Bartheld et al. Page 41
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Verkhratsky A, Mattson MP, Toescu EC. Aging in the mind. Trends Neurosci. 2004; 27:577–578.
[PubMed: 15374665]
Verkhratsky, A.; Butt, A. Glial Neurobiology: A Text Book. Chichester, UK: John Wiley & Sons;
Verkhratsky, A.; Butt, A. Neuroglia: Definition, Classification, Evolution, Numbers, Development.
Vol. Chapter 3. Oxford, UK: Wiley-Blackwell; 2013. Glial Physiology and Pathophysiology; p.
Verkhratsky A, Rodríguez JJ, Parpura V. Neuroglia in ageing and disease. Cell Tissue Res. 2014;
357:493–503. [PubMed: 24652503]
von Apathy S. Das leitende Element des Nervensystems und seine topographischen Beziehungen zu
den Zellen. Erste Mitth Zool Stat Neapel. 1897; 12:495–748.
von Economo C. Ein Koeffizient fur die Organisationshohe der Grosshirnrinde (Zellanzahl derselben
und einige andere Cortex-Masse). Klin Wochenschr. 1926; 5:593–595.
von Economo, C.; Koskinas, GN. Die Cytoarchitektonik der Grosshirnrinde des erwachsenen
Menschen. Wien/Berlin: Springer; 1925.
von Bartheld CS. Systematic bias in an “unbiased” neuronal counting technique. Anat Rec. 1999;
257:119–120. [PubMed: 10467241]
von Bartheld CS. Comparison of 2-D and 3-D counting: the need for calibration and common sense.
Trends Neurosci. 2001; 24:504–506. [PubMed: 11550662]
von Bartheld CS. Counting particles in tissue sections: choices of methods and importance of
calibration to minimize biases. Histol Histopathol. 2002; 17:639–648. [PubMed: 11962763]
von Bartheld CS, Wouters FS. Quantitative techniques for imaging cells and tissues. Cell Tissue Res.
2015; 360:1–4. [PubMed: 25773453]
von Bonin G. The frontal lobes of primates: cytoarchitectural studies. Res Publ Assoc Res Nerv Ment
Dis. 1948; 27:67–83. [PubMed: 18106851]
Vostrikov VM, Uranova NA, Orlovskaya DD. Deficit of perineuronal oligodendrocytes in the
prefrontal cortex in schizophrenia and mood disorders. Schizophr Res. 2007; 94:273–280.
[PubMed: 17566708]
Wall, PD. The History of Neuroscience in Autobiography. Squire, LR., editor. San Diego: Academic
Press; 2001. p. 472-500.Patrick D. Wall
Walloe S, Pakkenberg B, Fabricius K. Stereological estimation of total cell numbers in the human
cerebral and cerebellar cortex. Front Hum Neurosci. 2014; 8:508. [PubMed: 25076882]
Ward TS, Rosen GD, von Bartheld CS. Optical disector counting in cryosections and vibratome
sections underestimates particle numbers: effects of tissue quality. Microsc Res Tech. 2008;
71:60–68. [PubMed: 17868132]
Weise CM, Mouton PR, Eschbacher J, Coons SW, Krakoff J. A post-mortem stereological study of
striatal cell number in human obesity. Obesity (Silver Spring). 2015; 23:100–104. [PubMed:
West MJ. New stereological methods for counting neurons. Neurobiol Aging. 1993a; 14:275–285.
[PubMed: 8367009]
West MJ. Regionally specific loss of neurons in the aging human hippocampus. Neurobiol Aging.
1993b; 14:287–293. [PubMed: 8367010]
West MJ, Gundersen HJ. Unbiased stereological estimation of the number of neurons in the human
hippocampus. J Comp Neurol. 1990; 296:1–22. [PubMed: 2358525]
Williams RW, Herrup K. The control of neuron number. Annu Rev Neurosci. 1988; 11:423–453.
[PubMed: 3284447]
Williams RW, Rakic P. Three-dimensional counting: accurate and direct method to estimate numbers
of cells in sectioned material. J Comp Neurol. 1988; 278:344–352. [PubMed: 3216047]
Williams RW, von Bartheld CS, Rosen GD. Counting cells in sectioned material: a suite of techniques,
tools, and tips. Curr Protoc Neurosci. 2003; Chapter 1(Unit 1.11)
Windle, WF. Biology of Neuroglia. Springfield, Ill: C.C. Thomas; 1958.
Winick M. Changes in nucleic acid and protein content of the human brain during growth. Pediatr Res.
1968; 2:352–355. [PubMed: 5672695]
von Bartheld et al. Page 42
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Witelson SF, Kigar DL, Harvey T. The exceptional brain of Albert Einstein. Lancet. 1999; 35:2149–
Wittrock, MC. The Human Brain. Englewood Cliffs, NJ: Prentice Hall, Inc; 1977.
Young NA, Flaherty DK, Airey DC, Varlan P, Aworunse F, Kaas JH, Collins CE. Use of flow
cytometry for high-throughput cell population estimates in fixed brain tissue. Front Neuroanat.
2012; 6:27. [PubMed: 22798947]
Yuhas, D.; Jabr, F. [Accessed October 30, 2013] Know your neurons Chapter 4: What is the ratio of
glia to neurons in the brain?. 2012.
Zamenhof S, Bursztyn H, Rich K, Zamenhof PJ. The determination of deoxyribonucleic acid and of
cell number in brain. J Neurochem. 1964; 11:505–509. [PubMed: 14200773]
Zamenhof S. Final number of Purkinje and other large cells in the chick cerebellum influenced by
incubation temperatures during their proliferation. Brain Res. 1976; 109:392–394. [PubMed:
Zamenhof S, Klimuszko D. Final number of Purkinje and other large cells in the chick cerebellum
increased by oxygen and by glucose. Brain Res. 1977; 128:385–388. [PubMed: 871920]
Zorzetto, R. [Accessed on 4/21/2015] How many neurons do we really have? (And why we may have
10x fewer glia than we previously thought). 2012.
von Bartheld et al. Page 43
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Fig. 1.
A–B Photomicrographs of Nissl-stained neurons and glial cells. A. Purkinje cells (P) and
granule cells (arrow) in the cerebellum of an adult mouse brain. B. Motoneuron (M),
interneuron (I) and glial cells (arrows) in the trochlear nucleus of an adult mouse brain. Note
that the neuronal granule cell (arrow) in panel A is virtually indistinguishable in size and
shape from glial cells (arrows) in panel B. Thionin stains of 40 μm paraffin sections. Digital
images obtained on a Nikon Eclipse E600 microscope, with no digital adjustments or
manipulations of the images. Scale bar = 10 μm. Histological sections kindly provided by
Dr. Larisa M. Wiggins.
von Bartheld et al. Page 44
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Fig. 2.
A–D Flow chart of the isotropic fractionator (IF) cell counting method. The major steps of
the procedure are illustrated. A. Example of fixed brain tissue. Scale bar = 1 cm. B.
Tenbroek glass homogenizer. C. Appearance of DAPI-stained nuclei (left) and two nuclei
double-labeled with DAPI (upper panel) and NeuN (lower panel). Scale bar = 20 μm. D.
Neubauer counting chamber. DAPI, 4,6-diamidino-2-phenylindole; Fr, fraction; NeuN+,
neuronal nuclear antigen positive; PBS, phosphate-buffered saline; Vol, volume; Modified
from Herculano-Houzel and Lent (2005), Bahney and von Bartheld (2014), and Herculano-
Houzel et al. (2015).
von Bartheld et al. Page 45
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Fig. 3.
This graph summarizes the essence of Table 6. From the 1960s until 2009, the number of
glial cells in human brains was reported to be about one trillion, 10 times more than neurons
(100 billion), as detailed in Table 6. The number of glia, based on published data, is in fact
lower than the number of neurons, resulting in a glia-neuron ratio of less than 1 rather than
von Bartheld et al. Page 46
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
von Bartheld et al. Page 47
Estimates of numbers of neurons (N), non-neurons (nN) and glia (G) in human cerebral cortex (in billion) – Cortex comprises only grey matter, but does
not include white matter (WM), unless specifically indicated [ ].
Author Year N One side N Total nN Total G Total
Meynert 1868/1872 0.612 1.224
Donaldson 1895 1.200
Thompson 1899 9.282
Berger 1921 5.512
von Economo & Koskinas 1925 14
von Economo 1926 13.653
Agduhr 1941 5.0
Shariff 1953 6.9
Sholl 1956 5.000 10.000
Haug & Rebhan 1956 16.5
Haug 1959 8.200 16.400
Pakkenberg 1966 2.6
Gallatz et al. 1982 10.030
Haug 1985 13.8 ± 2.4
Haug 1987 10–19
Pakkenberg et al. 1989 ~20
Braendgaard et al. 1990 13.7 [27.4]
Pakkenberg 1992 25.1
Jensen & Pakkenberg 1993 23.2
Pakkenberg 1993 22.1
Regeur et al. 1994 18.1
Pakkenberg & Gundersen 1997 19.3–22.8 [range: 14.7 – 32.0]
Gredal et al. 2000 22.3
Pakkenberg et al. 2003 19.3–22.8 39
Pelvig et al. 2003 21.2 29.1
Koch 2004 20
Pedersen et al. 2005 18.8
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
von Bartheld et al. Page 48
Author Year N One side N Total nN Total G Total
Pelvig et al. 2008 21.4 – 26.3 27.9 – 38.9
Azevedo et al. 2009 6.18 12.36
Azevedo et al. 2009 [16.34] [60.84]
Lyck et al. 2009 [15–19.7] [35.4–40.6] [18.5–20.3]
Karlsen & Pakkenberg 2011 17.9 18.2
Andrade-Moraes et al. 2013 [12.7] [54.9]
[ ] includes white matter (WM)
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
von Bartheld et al. Page 49
Reports of glia-neuron ratios (GNRs) and non-neuron-neuron ratios (nNNRs) in human cerebral cortex, grey
matter (GM) only, unless indicated.
GNR nNNR Comments Author Year
~2 adult: 1.04 – 2.3, newborn: 0.14 – 0.2 Mühlmann 1936
1.2 – 2.1 Adult Superior Frontal Gyrus, all layers Arutyunova 1938
1.24 – 1.98 Human Cortex Friede 1953
1.24 – 1.98 Human Cortex, layers II – VI Friede 1954
1.78 Human cortex, layers II – VI Hawkins & Olszewski 1957
2.9 – 3.5 4.4–5.2 Striate cortex, GM+WM Nurnberger & Gordon 1957
0.74 – 6.6 tabulated by Blinkov and Glezer, 1968 (p. 416) Schlote 1959
Hyden & Pigon 1960
2 “Human Cortex” Cragg 1968
2.3 Frontal cortex Hess & Thalheimer 1971
0.49 – 0.57 Frontal/parietal cortex (control) Diamond et al. 1985
0.86 – 1.09 Frontal/parietal cortex (Albert Einstein) Diamond et al. 1985
1 – 1.5 Visual cortex Leuba & Garey 1989
1.56 – 2.02 Males and Females, 18–98 years old Pakkenberg et al. 2003
1.37 Neocortex without archicortex, 60–98 years old Pelvig et al. 2003
1.65 Frontal cortex, layers II/III Sherwood et al. 2006
1.32 – 1.49 Females – Males, 18–93 years old Pelvig et al. 2008
1.48 – 1.05
3.72 in GM only Azevedo et al. 2009
2.48 in GM+WM Azevedo et al. 2009
4.31 in GM only Andrade-Moraes et al. 2013
3.01 in GM+WM Andrade-Moraes et al. 2013
1.2–3.6 for GM, not including WM Ribeiro et al. 2013
GM, grey matter; GNR, glia-neuron ratio; nN, non-neuronal cells; N, neurons; nNNr, non-neuronal-neuron ratio; WM, white matter.
Based on a 2:1 ratio of glia to endothelial cells (References: 27–30%: Nurnberger, 1958; Blinkov and Glezer, 1968; Brasileiro-Filho et al., 1989;
Lyck et al., 2009; García-Amado and Prensa, 2012).
No primary data or reference provided
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
von Bartheld et al. Page 50
Types of Glial Cells contributing to the Total Number of Glia in the Human Brain.
Oligodendrocytes Astrocytes Microglia Comments Authors Year
29% 61.5% 9.5% Visual Cortex Kryspin-Exner 1952
40% 54% Caudatum Kryspin-Exner 1952
57% Pallidum Kryspin-Exner 1952
52–74% 30–40% 6–8% Thalamus Kryspin-Exner 1952
77.5% Nucleus ruber Kryspin-Exner 1952
62% Substantia nigra, pc Kryspin-Exner 1952
29–77.5% 30–61.5% 6–9.5% various regions Glees 1955
Review of Kryspin-Exner’s work
51% 40% 9% Motor Cortex, layer V Brownson 1956
45% 45% 10% GM Pope 1958
<67% >23% 10% WM Pope 1958
52% 39% 9% Motor Cortex Windle (Brownson) 1958
45% 45% 10% GM Windle (Pope) 1958
67% WM Windle (Pope) 1958
36.6% 46.5% 16.8% Frontal Cortex GM Pope 1959
69% 24% 6.9% Frontal Cortex WM Pope 1959
Frontal Cortex Schlote 1959
45% 45% 10% Cortex Blinkov & Glezer 1968
Data: Schlote, 1959 Hess & Thalheimer 1971
75% 19% 6% Neocortex GM Pelvig et al. 2003
5% 80% 10–15% CNS Verkhratsky & Butt 2007
74.6–75.6% 17.3–20.2% 5.2–6.5% Males, females Neocortex (GM) Pelvig et al. 2008
15–18% Males, Neocortex Lyck et al. 2009
75% 20% 5% Neocortex (GM) Verkhratsky & Butt 2013
GM, grey matter; WM, white matter; studies reporting primary data are shaded in .
Note: These numbers from Schlote’s 1959 data are compiled according to Hess and Thalheimer (1971), and adjusted for the percentages among glial cells (microglia and endothelial cells are assumed at a
1:1 ratio). As pointed out by Hess and Thalheimer (1971), the figure legends in Schlote (1959) erroneously switched the symbols for astroglia and oligodendroglia. This may explain some text books
reporting of an abundance of astroglia vs. oligodendrocytes (e.g., Verkhratsky and Butt, 2007).
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
von Bartheld et al. Page 51
Note: This percentage includes microglia plus endothelial cells.
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
von Bartheld et al. Page 52
Estimates of cell numbers in the human cerebellum (both sides together).
Number Method Author and Year
Purkinje cells
14 × 106Histology Kreuzfuchs, 1902
25–26 × 106Histology Lojda, 1955
15 × 106Histology Braitenberg & Atwood, 1958
15.4 × 106Stereology Nairn et al., 1989
0.88 × 106Histology Riedel et al., 1989
15.6 × 106Stereology Mayhew et al., 1990
30.5 × 106Stereology Andersen et al., 1992
30.5 × 106Stereology Korbo and Andersen, 1995
28.5 × 106Stereology Andersen & Pakkenberg, 2003
28 × 106Stereology Andersen et al., 2003
22.3 × 106Stereology Agashiwala et al., 2008
26 × 106Stereology Andersen et al., 2012
26 × 106
Stereology Kiessling et al., 2014
Granule cells (granule neurons)
10–100 × 109Histology Braitenberg & Atwood, 1958
19.8 ×109Histology Riedel et al., 1989
101 × 109Stereology Andersen et al., 1992
112.3 × 109Stereology Andersen & Pakkenberg, 2003
109 × 109Stereology Andersen et al., 2003
70 × 109Stereology Andersen et al., 2012
75.2 × 109Stereology Kiessling et al., 2014
Total neurons
65–70 × 109Histology Lange, 1975; Williams & Herrup, 1988
50 × 109Stereology Haug, 1986
105 × 109Stereology Andersen et al., 1992
69 × 109IF Azevedo et al., 2009
54 × 109IF Andrade-Moraes et al., 2013
Glial cells
3 × 109Stereology Andersen et al., 1992
Non-neuronal cells
16 × 109IF Azevedo et al., 2009
15.4 × 109IF Andrade-Moraes et al., 2013
Abbreviations: IF, isotropic fractionator
Data from 10–11 month old infants
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
von Bartheld et al. Page 53
Mis-quotations of Nansen’s original opinion about neuropil [“Leydig’s dotted substance]” being the seat of
intelligence to claims of glia or neuroglia being the seat of intelligence and increasing during evolution in size
or number.
Nansen, 1886: “… the more complicated the structure of dotted substance [neuropil consisting of neuronal and glial processes
] is – the more
highly is the animal mentally developed; in other words, we may conclude that
the more the inteligence of an animal is developed – the more
intricate becomes the web of plaiting of nerve-tubes and fibrillae in its dotted substance
… and this web is probably the principal seat of
inteligence.” (page 171, Nansen, 1886, his italics).
Glees, 1955: “It is worth mentioning Nansen’s opinion … that this substance [Leydig’s dotted substance = ‘plaiting of nerve-tubes and
fibrillae’] was the seat of intelligence as it increases in size from the lower to the higher forms of animal.” (cites Nansen, 1886)
Galambos, 1961: “Nansen … said neuroglia was ‘the seat of intelligence, as it increases in size from the lower to the higher forms of animal.’
“ (cites Glees’ 1955 footnote)
Fields, 2009: “Nansen … observed in 1886 that glia might be ‘the seat of intelligence, as [their number] increase in size from the lower to the
higher forms of animal.’ ” (cites Galambos, 1961)
Verkhratsky and Butt, 2013: “Nansen … postulated that neuroglia was ‘the seat of intelligence, as it increases in size from the lower to the
higher forms of animal’ “ (cites Galambos, 1961).
“Nerve-tubes are …present in great plenty in the dotted substance” (Nansen, 1886, page 124)
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
von Bartheld et al. Page 54
Reports of glia-neuron ratios (GNRs), numbers of total cells, glia, and neurons in human brain
GNR Total cell number Glia # Neuron # Method Reference
3 bn Donaldson, 1895
10:1 (“perhaps”) Glees, 1958
10:1 (“perhaps”) Pope, 1958
10:1 (“around”) Hyden, 1960
10:1 (“perhaps”) Galambos, 1961
10:1 (“about”) Hyden, 1961
10:1 110 bn 100 bn 10 bn Asimov, 1963
10:1 Maron, 1963
(glia “more abundant” than neurons) Kuffler & Nicholls, 1966
100–130 bn Histology Blinkov & Glezer, 1968
“glia …outnumber neurones by several fold” Dobbing & Sands, 1970
5:1 – 10:1 Noback & Demarest, 1975
10:1 (“at least”) >10 bn Kuffler & Nicholls, 1976
~10:1 Ganong, 1977
“Glia … far outnumber(s) neurons” 20–200 bn Wittrock, 1977
50 bn Edelman & Mountcastle 1978
~10:1 Ganong, 1979
10bn – 1 trn Hubel, 1979
10 bn – 100 bn or 1 trn Nauta & Feirtag, 1979
100 bn Stevens, 1979
5:1 Jensen, 1980
5:1 – 10:1 Snell, 1980
9:1 [10 trn] [~9 trn] ~1 trn Kandel & Schwartz, 1981
10:1 (“perhaps”) [~1 trn]100 bn Nolte, 1981
30 bn (“roughly”) Szentagothai, 1983
10:1 Damask & Swenberg, 1984
10:1 – 50:1 [11–51 trn] [10–50 trn]1 trn (“best estimate”) Kandel & Schwartz, 1985
10:1 Nicholls et al., 1985 2nd ed
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
von Bartheld et al. Page 55
GNR Total cell number Glia # Neuron # Method Reference
0.7:1 40–50 bn 70–80 bn Histology Haug, 1986
85 bn Williams & Herrup, 1988
10:1 Steward, 1989
10:1 Bignami, 1991
1:1 or 10:1 Jacobson, 1991
10:1 – 50:1 1.1–5.1 trn [1–5 trn]100 bn (“best estimate”) Kandel et al.,
5:1 – 10:1 (“depending on region”) Noback et al., 1991
GN (glia # “much higher” than neuron #) Brodal, 1992
10:1 (“at least”) 10 bn – 1 trn Nicholls et al., 1992
(“several times that many glial cells”) 100 bn Nolte, 1993
10:1 (glial cells … outnumber neurons 10 to 1”) Black and Ransom, 1999
10:1 – 50:1 [1.1–5.1 trn] [1–5 trn]100 bn (“on the order of …”) Kandel et al., 2000
10:1 Steward, 2000
10:1 [~1 trn]~100 bn Bear et al., 2001, 2nd ed
(“glial cells … vastly outnumber neurons”) Lemke, 2001
10:1 (“thought to be at least ten glia per neuron”) Haydon, 2001
10:1 (“or more”) Levitan & Kaczmarek, 2002
100 bn Haines, 2002
(“glia are the most numerous cells in the brain”) Doetsch, 2003
10:1 – 50:1 [1.1 – 5.1 trn] [1–5 trn]100 bn Hatton & Parpura, 2004
10:1 Bear et al., 2007, 3rd ed
10:1 > “several trillions” (probably) Verkhratsky & Butt, 2007
1:1 170 bn <85 bn 86 bn IF Azevedo et al., 2009
6:1 [767 bn] [667 bn] (85%) 100 bn (15%) Fields, 2009
GN (glia # “much higher” than neuron #) Brodal, 2010
3:1 or 4:1 100 bn Purves, 2010
~1:1 “human brain contains roughly equal numbers of glia and neurons” Smith, 2010
10:1 (“at least”) Nicholls et al., 2012, 5th ed
2:1 – 10:1 [200 bn −1 trn]100 bn Kandel et al., 2013, 5th ed
1:1 Verkhratsky & Butt, 2013
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
von Bartheld et al. Page 56
GNR Total cell number Glia # Neuron # Method Reference
~1:1 <78.6 bn 67.3 bn IF Andrade-Moraes et al., 2013
~1:1 “glial cells are as abundant as neurons” Kettenmann and Verkhratsky, 2013
~1:1 “roughly … equal numbers of neurons and glia” Streit, 2013, p. 86
4:1 “glia … constitute … the majority of cells …, 80% in the human brain” Barres et al., 2015
~1:1 “neuroglial cells … are about as numerous as neurons in the brain as a whole” Gundersen et al., 2015
[ ], implied numbers - not stated explicitly; bn, billion; ed, edition; GNR, glia-neuron ratio; IF, isotropic fractionator; trn, trillion;
J Comp Neurol
. Author manuscript; available in PMC 2017 December 15.
... Notably, a single human astrocyte domain has been estimated to contact up to 2 million synapses, underscoring their important role in human brain function. Astrocytes are very abundant in the CNS, accounting for between one and two out of every five cells, but the CNS may possess fewer astrocytes than it does neurons [40], in stark contrast to the ratio between SGCs and neurons in the peripheral ganglia. ...
Full-text available
Preclinical studies have identified glial cells as pivotal players in the genesis and maintenance of neuropathic pain after nerve injury associated with diabetes, chemotherapy, major surgeries, and virus infections. Satellite glial cells (SGCs) in the dorsal root and trigeminal ganglia of the peripheral nervous system (PNS) and astrocytes in the central nervous system (CNS) express similar molecular markers and are protective under physiological conditions. They also serve similar functions in the genesis and maintenance of neuropathic pain, downregulating some of their homeostatic functions and driving pro-inflammatory neuro-glial interactions in the PNS and CNS, i.e., “gliopathy”. However, the role of SGCs in neuropathic pain is not simply as “peripheral astrocytes”. We delineate how these peripheral and central glia participate in neuropathic pain by producing different mediators, engaging different parts of neurons, and becoming active at different stages following nerve injury. Finally, we highlight the recent findings that SGCs are enriched with proteins related to fatty acid metabolism and signaling such as Apo-E, FABP7, and LPAR1. Targeting SGCs and astrocytes may lead to novel therapeutics for the treatment of neuropathic pain.
... Microglia are resident macrophages of CNS in mammals, derived from yolk sac myeloid progenitors during neurodevelopmental stages, colonize the brain early in brain development and establish in the brain parenchyma. In the adult stage, they represent about 10% of the CNS population (Lawson et al. 1990;Jurga et al. 2020) and 5%-20% of glial cells (Polazzi and Monti 2010;von Bartheld et al. 2016;Zhang et al. 2018). These resident brain immune cells are divergent from other peripherally immunocompetent cell populations, such as infiltrated, perivascular, and bordersassociated macrophages due to their ontogeny from the yolk sac and autorenewal capacity, about 28% per year, with a lifespan of 4.2 years, totally independent from bone marrow (Réu et al. 2017). ...
Full-text available
Microglia, the resident macrophages of the central nervous system, are essential players during physiological and pathological processes. Although they participate in synaptic pruning and maintenance of neuronal circuits, microglia are mainly studied by their activity modulating inflammatory environment and adapting their phenotype and mechanisms to insults detected in the brain parenchyma. Changes in microglial phenotypes are reflected in their morphology, membrane markers, and secreted substances, stimulating neighbor glia and leading their responses to control stimuli. Understanding how microglia react in various microenvironments, such as chronic inflammation, made it possible to establish therapeutic windows and identify synergic interactions with acute damage events like stroke. Obesity is a low-grade chronic inflammatory state that gradually affects the central nervous system, promoting neuroinflammation development. Obese patients have the worst prognosis when they suffer a cerebral infarction due to basal neuroinflammation, then obesity-induced neuroinflammation could promote the priming of microglial cells and favor its neurotoxic response, potentially worsening patients’ prognosis. This review discusses the main microglia findings in the obesity context during the course and resolution of cerebral infarction, involving the temporality of the phenotype changes and balance of pro- and anti-inflammatory responses, which is lost in the swollen brain of an obese subject. Graphical Abstract Obesity enhances proinflammatory responses during a stroke. Obesity-induced systemic inflammation promotes microglial M1 polarization and priming, which enhances stroke-associated damage, increasing M1 and decreasing M2 responses.
... The human brain has about a 100 billion neurons and glia (von Bartheld, Bahney, & Herculano-Houzel, 2016). Glia associate closely with neurons to regulate neuron shape and function, thereby 48 impacting animal behavior (Barres, 2008). ...
Full-text available
Each glia interacts with multiple neurons, but the fundamental logic of whether it interacts with all equally remains unclear. We find that a single sense-organ glia modulates different contacting neurons distinctly. To do so, it partitions regulatory cues into molecular microdomains at specific neuron contact-sites, at its delimited apical membrane. For one glial cue, K/Cl transporter KCC-3, microdomain-localization occurs through a two-step, neuron-dependent process. First, KCC-3 shuttles to glial apical membranes. Second, some contacting neuron cilia repel it, rendering it microdomain-localized around one distal neuron-ending. KCC-3 localization tracks animal aging, and while apical localization is sufficient for contacting neuron function, microdomain-restriction is required for distal neuron properties. Finally, we find the glia regulates its microdomains largely independently. Together, this uncovers that glia modulate cross-modal sensor processing by compartmentalizing regulatory cues into microdomains. Glia across species contact multiple neurons and localize disease-relevant cues like KCC-3. Thus, analogous compartmentalization may broadly drive how glia regulate information processing across neural circuits.
... Taken together, these findings underline the strong susceptibility and malleability of ECs, which are directly exposed to secreted factors in both brain parenchyma and blood, to adapt to changes in their microenvironment, which is consistent with pervious observations from our lab 25,26 and others 30,43,59,69 . Therefore, ECs, despite comprising <5% of the total number of brain cells 103 , are a promising and accessible target for the treatment of aging and its associated diseases 104 . ...
Full-text available
Aging is a complex process involving transcriptomic changes associated with deterioration across multiple tissues and organs, including the brain. Recent studies using heterochronic parabiosis have shown that various aspects of aging-associated decline are modifiable or even reversible. To better understand how this occurs, we performed single-cell transcriptomic profiling of young and old mouse brains after parabiosis. For each cell type, we cataloged alterations in gene expression, molecular pathways, transcriptional networks, ligand–receptor interactions and senescence status. Our analyses identified gene signatures, demonstrating that heterochronic parabiosis regulates several hallmarks of aging in a cell-type-specific manner. Brain endothelial cells were found to be especially malleable to this intervention, exhibiting dynamic transcriptional changes that affect vascular structure and function. These findings suggest new strategies for slowing deterioration and driving regeneration in the aging brain through approaches that do not rely on disease-specific mechanisms or actions of individual circulating factors.
... A recent study using a similar triculture showed a reversed trend, with microglia appearing to reduce spike frequency and other electrophysiological features [61]. However, in that paper, the amount of microglia was increased to 15-25% of the total cell population, which is more than double the number of microglia found in our tri-culture model ( Figure S4) [32] and the proportions found in vivo [62]. Furthermore, the authors found that increased microglia reactivity corresponds with increasing microglia density, suggesting that the decrease in spike frequency may be a function of reactive microglia as opposed to more homeostatic microglia and would be in line with the results we obtained from our LPS-treated cultures. ...
Full-text available
Neuroinflammation plays a central role in many neurological disorders, ranging from traumatic brain injuries to neurodegeneration. Electrophysiological activity is an essential measure of neuronal function, which is influenced by neuroinflammation. In order to study neuroinflammation and its electrophysiological fingerprints, there is a need for in vitro models that accurately capture the in vivo phenomena. In this study, we employed a new tri-culture of primary rat neurons, astrocytes, and microglia in combination with extracellular electrophysiological recording techniques using multiple electrode arrays (MEAs) to determine the effect of microglia on neural function and the response to neuroinflammatory stimuli. Specifically, we established the tri-culture and its corresponding neuron-astrocyte co-culture (lacking microglia) counterpart on custom MEAs and monitored their electrophysiological activity for 21 days to assess culture maturation and network formation. As a complementary assessment, we quantified synaptic puncta and averaged spike waveforms to determine the difference in excitatory to inhibitory neuron ratio (E/I ratio) of the neurons. The results demonstrate that the microglia in the tri-culture do not disrupt neural network formation and stability and may be a better representation of the in vivo rat cortex due to its more similar E/I ratio as compared to more traditional isolated neuron and neuron-astrocyte co-cultures. In addition, only the tri-culture displayed a significant decrease in both the number of active channels and spike frequency following pro-inflammatory lipopolysaccharide exposure, highlighting the critical role of microglia in capturing electrophysiological manifestations of a representative neuroinflammatory insult. We expect the demonstrated technology to assist in studying various brain disease mechanisms.
... Microglial colonization of the CNS occurs well before the formation of astrocytes and oligodendrocytes, which are derived from the neuroectoderm prior to hematopoiesis [20]. However, accurate measure of the true numbers of neuroglial cells has been a challenging target that is the subject of prolonged debate [21]. During early developmental stages, microglia disseminate throughout the CNS in a somewhat homogeneous manner whilst simultaneously undergoing an adaptive determination of their phenotype. ...
Full-text available
Microglia are the primary immunocompetent cells of the central nervous system (CNS). Their ability to survey, assess and respond to perturbations in their local environment is critical in their role of maintaining CNS homeostasis in health and disease. Microglia also have the capability of functioning in a heterogeneous manner depending on the nature of their local cues, as they can become activated on a spectrum from pro-inflammatory neurotoxic responses to anti-inflammatory protective responses. This review seeks to define the developmental and environmental cues that support microglial polarization towards these phenotypes, as well as discuss sexually dimorphic factors that can influence this process. Further, we describe a variety of CNS disorders including autoimmune disease, infection, and cancer that demonstrate disparities in disease severity or diagnosis rates between males and females, and posit that microglial sexual dimorphism underlies these differences. Understanding the mechanism behind differential CNS disease outcomes between men and women is crucial in the development of more effective targeted therapies.
Astrocytes are active participants in the performance of the Central Nervous System (CNS) in both health and disease. During aging, astrocytes are susceptible to reactive astrogliosis, a molecular state characterized by functional changes in response to pathological situations, and cellular senescence, characterized by loss of cell division, apoptosis resistance, and gain of proinflammatory functions. This results in two different states of astrocytes, which can produce proinflammatory phenotypes with harmful consequences in chronic conditions. Reactive astrocytes and senescent astrocytes share morpho-functional features that are dependent on the organization of the cytoskeleton. However, such changes in the cytoskeleton have yet to receive the necessary attention to explain their role in the alterations of astrocytes that are associated with aging and pathologies. In this review, we summarize all the available findings that connect changes in the cytoskeleton of the astrocytes with aging. In addition, we discuss future avenues that we believe will guide such a novel topic.
Full-text available
The human brain comprises heterogeneous cell subtypes whose composition can be altered with physiological and pathological conditions. New approaches to discern the diversity and distribution of brain cells associated with neurological conditions would significantly advance the study of brain-related pathophysiology and neuroscience. We demonstrate that DNA-based cell-type deconvolution achieves an accurate resolution of seven major cell types. Unlike single-nuclei approaches, DNA methylation-based deconvolution does not require special sample handling or processing, is cost-effective, and easily scales to large study designs. Current methods for brain cell deconvolution are limited only to neuronal and non-neuronal cells. Using DNA methylation profiles of the top cell-type-specific differentially methylated CpGs, we employed a hierarchical modeling approach to deconvolve GABAergic neurons, glutamatergic neurons, astrocytes, microglial cells, oligodendrocytes, endothelial cells, and stromal cells. We demonstrate the utility of our method by applying it to data on normal tissues from various brain regions and in aging and diseased tissues, including Alzheimer's disease, autism, Huntington’s disease, epilepsy, and schizophrenia. We expect that the ability to determine the cellular composition in the brain using only DNA from bulk samples will accelerate understanding brain cell type composition and cell-type-specific epigenetic states in normal and diseased brain tissues.
The epilepsies are a diverse spectrum of disease states characterized by spontaneous seizures and associated comorbidities. Neuron-focused perspectives have yielded an array of widely used anti-seizure medications and are able to explain some, but not all, of the imbalance of excitation and inhibition which manifests itself as spontaneous seizures. Furthermore, the rate of pharmacoresistant epilepsy remain high despite the regular approval of novel anti-seizure medications. Gaining a more complete understanding of the processes that turn a healthy brain into an epileptic brain (epileptogenesis) as well as the processes which generate individual seizures (ictogenesis) may necessitate broadening our focus to other cell types. As will be detailed in this review, astrocytes augment neuronal activity at the level of individual neurons in the form of gliotransmission and the tripartite synapse. Under normal conditions, astrocytes are essential to the maintenance of blood-brain barrier integrity and remediation of inflammation and oxidative stress, but in epilepsy these functions are impaired. Epilepsy results in disruptions in the way astrocytes relate to each other by gap junctions which has important implications for ion and water homeostasis. In their activated state, astrocytes contribute to imbalances in neuronal excitability due their decreased capacity to take up and metabolize glutamate and an increased capacity to metabolize adenosine. Furthermore, due to their increased adenosine metabolism, activated astrocytes may contribute to DNA hypermethylation and other epigenetic changes that underly epileptogenesis. Lastly, we will explore the potential explanatory power of these changes in astrocyte function in detail in the specific context of the comorbid occurrence of epilepsy and Alzheimer's disease and the disruption in sleep-wake regulation associated with both conditions.
This book seeks to present, through a combination of morphological data and physiological and neurological studies, a comprehensive survey of our knowledge of the human brain. The major emphasis is upon structural organisation, based upon the evolution of this most complex of organs. However, functional aspects, including experimental research and clinical findings, have also been incorporated, broadening the interest for students of neurobiology and clinical medicine.
Developmental Neurobiology provides an up-to-date survey of the cellular events and the molecular contributors that contribute to the assembly of the vertebrate nervous system. The text will serve as a readily tractable source for advanced undergraduate neuroscience majors and beginning graduate students who will benefit from a single source to begin their study of a more detailed understanding of neural development. Each chapter is peppered with a sound mixture of historical context and descriptions from both the vertebrate and invertebrate literature that best illustrate specific aspects of development. The liberal use of simple diagrams and tables, which readily illustrate complex issues, is a welcome addition for instructor and student alike. While classic topics of neural development, including axial patterning, cell proliferation, migration, cell death and synapse formation are covered, of particular interest are subjects that oftentimes received superficial coverage in texts, including separate, detailed chapters on oligodendrocyte and astrocyte development, and developmental mechanisms that relate to the process of aging. Multi-authored texts are often tricky to assemble for consistency, but Developmental Neurobiology succeeds in providing a sound introduction to the most exciting questions that neuroscientists will address experimentally for years to come. Pat Levitt, Ph.D., Director, Vanderbilt Kennedy Center for Research on Human Development, Professor of Pharmacology, Vanderbilt University, Nashville, TN When the inaugural edition of Developmental Neurobiology appeared in 1970, it was the first attempt to comprehensively assess our understanding of neuronal development since the publication of S.R. Detwiler’s book Neuroembryology in 1936. Although progress had been made in the intervening 34 years, the author, Marcus Jacobson, was correct in noting that in 1970 "most aspects of neural ontogeny could be surveyed at a glance". In contrast, by the time the 3rd edition appeared in 1991, the size of the book had increased by 40% and the number of references cited went from around 2,000 to over 8,000. Since 1991, however, the field has grown at an even more rapid rate and has now reached a point that makes it virtually impossible for a single individual to comprehensively and authoritatively assess the entire gamut of neural ontogeny. Before his death in 2001, Jacobson together with the co-editor of the present edition, M. Rao, conceived the plan for a 4th edition that would be multiauthored with each chapter being written by experts in a sub-field. Although one of the joys of reading the previous three editions was the consistency of Jacobson’s inimitable prose style, in the present book there remains a smoothness and consistency of style that is unusual and refreshing in a multiauthored text. In 14 chapters that begins with neural induction and ends with developmental mechanisms of aging, virtually all of the major topics of neural development are discussed in a clear and coherent fashion and with the aid of ample illustrations. The inclusion of historical antecedents , past and present controversies, technical and conceptual advances together with a comprehensive discussion of each topic all add up to an excellent assessment of the field as it enters the 21st century. Although one misses the Jacobsonian idiosyncrasies of previous editions, the 4th edition is a fitting legacy of Marcus Jacobson’s four decades of empirical and pedagogical contributions to developmental neurobiology. Ronald W. Oppenheim, Ph.D., Neuroscience Program, Wake Forest University School of Medicine, Winston-Salem, NC
Glial Neuronal Signaling fills a need for a monograph/textbook to be used in advanced courses or graduate seminars aimed at exploring glial-neuronal interactions. Even experts in the field will find useful the authoritative summaries of evidence on ion channels and transporters in glia, genes involved in signaling during development, metabolic cross talk and cooperation between astrocytes and neurons, to mention but a few of the timely summaries of a wide range of glial-neuronal interactions. The chapters are written by the top researchers in the field of glial-neuronal signaling, and cover the most current advances in this field. The book will also be of value to the workers in the field of cell biology in general. When we think about the brain we usually think about neurons. Although there are 100 billion neurons in mammalian brain, these cells do not constitute a majority. Quite the contrary, glial cells and other non-neuronal cells are 10-50 times more numerous than neurons. This book is meant to integrate the emerging body of information that has been accumulating, revealing the interactive nature of the brain's two major neural cell types, neurons and glia, in brain function.
The desoxyribonucleic acid (DNA) content per nucleus and per unit weight of tissue have been determined chemically in normal cerebral cortex, cerebellar cortex, and corpus callosum of man, dog, and cat and in various human brain tumors. Nuclear densities have been calculated from these determinations. Corpus callosum contains approximately the same total number of cells as does cerebral cortex; cerebellar cortex contains several times this number. The nuclear density in tumors is usually higher than in cerebral cortex or corpus callosum. The amount of DNA per nucleus in primary brain tumors is considerably higher than in normal tissue. The average DNA per nucleus in the more primitive and malignant tumors appears to be higher than in the more differentiated tumors. Calculations indicate that the increase in the DNA per nucleus in brain tumors is more likely to be due to polyploidy than to increased mitotic activity.
Only a short time after the neuron became identified as the essential unit of the nervous system, the first attempts were made to estimate the number of neurons in different parts of the nervous system. During the past century, a great number of methods have been used in making such estimates. Although the most widely used and accepted method is that of direct counting in the microscope, various other techniques, including photographic, projection, homogenate, automatic, and ocular methods have been designed. Brief descriptions of these techniques will be given in the following account.