Extraordinary neoteny of synaptic spines in the human
Zdravko Petanjeka, Miloš Judaša, Goran Šimic ´a, Mladen Roko Rašina,b,c, Harry B. M. Uylingsd, Pasko Rakicb,c,1,
and Ivica Kostovic ´a
aCroatian Institute for Brain Research, School of Medicine, University of Zagreb, 10,000 Zagreb, Croatia;bDepartment of Neurobiology andcKavli Institute
for Neuroscience, Yale University, New Haven, CT 06520; anddDepartment of Anatomy and Neuroscience, VU University Medical Center, 1007 MB,
Amsterdam, The Netherlands
Edited* by Jean-Pierre Changeux, Institut Pasteur, Paris Cedex 15, France, and approved June 27, 2011 (received for review March 30, 2011)
The major mechanism for generating diversity of neuronal connec-
tions beyond their genetic determination is the activity-dependent
stabilization and selective elimination of the initially overproduced
synapses [Changeux JP, Danchin A (1976) Nature 264:705–712]. The
cerebral cortex of human and nonhuman primates. It is generally
accepted that synaptic pruning in the cerebral cortex, including pre-
frontal areas, occurs at puberty and is completed during early ado-
present study we analyzed synaptic spine density on the dendrites
oflayerIIICcortico–corticalandlayerV cortico–subcortical projecting
pyramidal neurons in a large sample of human prefrontal cortices in
subjects ranging in age from newborn to 91 y. We confirm that
dendritic spine density in childhood exceeds adult values by two-
to threefold and begins to decrease during puberty. However, we
also obtained evidence that overproduction and developmental
remodeling, including substantial elimination of synaptic spines,
continues beyond adolescence and throughout the third decade
of life before stabilizing at the adult level. Such an extraordinarily
long phase of developmental reorganization of cortical neuronal
circuitry has implications for understanding the effect of environ-
mental impact on the development of human cognitive and emo-
tional capacities as well as the late onset of human-specific neu-
association cortex|critical period|schizophrenia|synaptogenesis
than 4 decades ago (1). This hypothesis gained considerable sup-
port from the discovery that synaptic connections in the cerebral
cortex of human and nonhuman primates initially are over-
produced to about two times the adult number and are then
pruned during puberty to reach the adult level at the onset of
adolescence (2–5). The selective-elimination hypothesis basically
assumes that during a period of overproduction of synapses neu-
ronal activity tunes the molecular structure of individual synapses
and determines which will be retained and which removed from
the neural network (6, 7). Previous EM analyses in nonhuman
primates revealed that synaptic elimination in the monkey pre-
frontal cortex occursmainly by removal of asymmetric synapses on
spines, whereas the number of symmetric synapses on dendritic
shafts remains constant (5, 8).
The tempo of elimination of supranumerary synaptic spines
and the identification of the end of this critical period are ex-
tremely important, because these factors are related to estab-
lishment of cognitive abilities and duration of the window for
optimal acquisition of new language and mathematical skills as
well as personality transformation from the developmental mode
to adult status. In addition, the several leading hypotheses for the
explanation of late-onset neuropsychiatric disorders, such as
schizophrenia and drug- or stress-induced psychoses, implicate
defective pruning of the initially overproduced synapses on den-
dritic spines (8–15). Furthermore, spine dysgenesis is the only
elective stabilization of developing synapses as a mechanism
for specification of neuronal connections was proposed more
detectable anatomical phenotype in some human cognitive dis-
orderssuchasnonsyndromic mental retardation (16).Finally, this
biomedically and socially important concept (9) is also the subject
ofcontinuing dialoguebetweenproponents ofselectionism versus
The end of the critical period of synaptic spine elimination in
the human cortex basically relies on the pioneering study of
Huttenlocher and colleagues (2, 4). Thus, it usually is assumed
tacitly that the period of synaptic overproduction in the human
cerebral cortex is completed by the end of puberty (18), even
though Huttenlocher’s study contains only a single 19-y-old brain
specimen in the age group between 15 and 32 y. In contrast,
more recent studies using electroencephalography (15, 19), PET
(20), and functional MRI (18, 21–26) have suggested that the
dynamic changes in gray matter density and white matter in-
tegrity in the human association neocortex extend into the third
decade of life (21, 23, 26, 27). These changes, observed by MRI,
cannot be explained by a capacious increase in dendritic length,
because the dendritic growth in the human neocortex is limited
mainly to early childhood (28). Therefore, it has been assumed
that functional plasticity probably reflects reorganization of cir-
cuitry, including synaptic elimination (29–32), essential for ac-
quisition of the highest brain functions in humans, including
affective modulation of emotional cues, self-conceptualization,
mentalization, cognitive flexibility, and working memory (33–36).
However, the cellular data supporting this assumption have
To fill this gap in our knowledge, we analyzed the initial over-
production and subsequent elimination of dendritic spines on
layer IIIc and layer V pyramidal neurons in the dorsolateral
prefrontal cortex [Brodmann area 9 (28, 37)]. Our focus was on
the prefrontal cortex because of the relevance of this region for
late-onset, human-specific, neuropsychiatric disorders and the
possible implications in understanding the mechanisms of envi-
ronmental impacts such as education and training on prolonged
development of human cognitive capacities (8–12, 38–41). Be-
cause dendritic spines are impregnated reliably by the rapid Golgi
methods in postmortem brain tissue (Material and Methods), we
decided to examine their development and measure their density
on dendrites in well-preserved human tissue, including ages that
were not analyzed in previous studies (Table S1).
We selected to focus on the large pyramidal cells in layer IIIc
and layer V (28), that form cortico–cortical and subcortical pro-
Author contributions: Z.P., M.J., H.B.M.U., and I.K. designed research; Z.P., G.Š., M.R.R.,
and H.B.M.U. performed research; Z.P., M.J., G.Š., M.R.R., H.B.M.U., P.R., and I.K. analyzed
data; and Z.P., M.J., H.B.M.U., P.R., and I.K. wrote the paper.
The authors declare no conflict of interest.
*This Direct Submission article had a prearranged editor.
1To whom correspondence should be addressed. E-mail: email@example.com.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
www.pnas.org/cgi/doi/10.1073/pnas.1105108108 PNAS Early Edition
| 1 of 6
jections, respectively (Fig. 1A and Figs. S1 and S2). The dendrites
were defined by their division into basal and oblique branches
emanating from the main apical shaft (Fig. 1 A and B). In addi-
tion, the oblique branches were subdivided into proximal and
distal groups, and all types were quantified separately (Fig. 1C).
We found that the vast majority of spines (about 90% after the
neonatal period) belong to the mushroom type characterized by
a neck 1.5–3.5 μm long and up to 0.5 μm thick that expands into
a bulb with a radius of 1–2 μm. The morphology of the head and
neck of mushroom spines remained relatively constant except in
the late adolescent stage (age 16–20 y), when spine heads become
slightly larger. The second type, hair-like thin spines, which lack
an obvious bulb and have terminal expansions slightly wider than
the neck, were observed predominantly during the first postnatal
month, and their percentage diminished afterward. The per-
centage of the third type of dendritic spines, stubby spines with a
broad neck, was low on both oblique and basal dendrites at all
ages analyzed (Fig. 1C).
Statistical analysis of interindividual differences in dendritic
spine density (DSD), using the a posteriori Student–Newman–
Keuls test for multiple comparisons, revealed that in all dendritic
segments of both layer IIIc and layer V pyramidal neurons, the
cortex of a 16-y-old subject. Black arrows indicate basal dendrites, and gray arrows indicate oblique dendrites. (Scale bar: 100 μm.) (B) Neurolucida re-
construction of layer IIIc pyramidal neuron of a 49-y-old subject, illustrating sites selected for counting spines over a 50-μm length of apical distal oblique
dendrites (green), apical proximal oblique dendrites (blue), and basal dendrites (red). (Scale bar: 100 μm.) (C) Representative high-power magnification
images of rapid Golgi-impregnated layer IIIc pyramidal neurons from the dorsolateral prefrontal cortex showing basal dendrites (Left) and distal apical
oblique dendrites (Right) during different stages: an infant 1 mo of age, a 2.5-y-old child, and 16-y-old, 28-y-old, and 49-y-old subjects. (Scale bar: 10 μm.) (D)
Graphs representing number of dendritic spines per 50-μm dendrite segment on basal dendrites after the first bifurcation (red); apical proximal oblique
dendrites originating within 100 μm from the apical main shaft (blue); and apical distal oblique dendrites originating within the second 100-μm segment from
the apical main shaft (green) of layer IIIc (filled symbols) and layer V (open symbols) pyramidal cells in the dorsolateral prefrontal cortex. Squares represent
males; circles represent females. The age in postnatal years is shown on a logarithmic scale. Puberty is marked by a shaded bar. B, birth (fourth postnatal day);
P, puberty. Specification of tissue analyzed is given in Table S1, and the positions of sections on which pyramidal neurons were measured are indicated on the
reconstructed pyramidal neuron shown in B.
(A) Representative low-magnification photographs of the rapid Golgi-impregnated layer IIIc and V pyramidal cells in the dorsolateral prefrontal
2 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1105108108Petanjek et al.
DSD increased significantly during infancy and reached its peak
during childhood, when the DSD was, on average, more than two
times higher than in the adults (Figs. 1D, 2 A–C, and Table S2).
Importantly, although the DSD diminished gradually during late
childhood and adolescence (9–22 y), it remained significantly
higher throughout this period than in the adult (Figs. 1D and 2).
The rate of decrease in DSD varied among dendritic segments.
The highest DSD values for all segments in layer IIIc pyramidal
neurons (Fig. S2) were reached by the age of 2.5–7 y (Fig. 1 C and
D,Fig.S2,andTable S2).Althougha significantdeclineofDSDin
basal and proximal apical oblique dendrites started at age 7–9 y,
the DSD of distal apical oblique dendrites did not decrease sig-
nificantly until age 17 y (Figs. 1 C and D and 2 A–C). In all den-
dritic segments of layer IIIc pyramidal neurons, the DSD
decreased to an adult level by age 30 y and remained stable
thereafter (Figs. 1D and 2 and Table S2). In almost all subjects
older than 30 y (n = 16), DSD values for all segments of layer IIIc
pyramidal neurons were significantly lower than in subjects of
a younger age group (15 m to 28 y; n = 11). Despite the relatively
smaller sample and interindividual variability, the DSD on all
segments of layer V pyramids clearly displayed the highest values
between age 7 and 9 y before beginning to decline. Thus, the
overall developmental course of the DSD was similar in both layer
IIIc and layer V pyramidal neurons, and the DSD in both types
attained the stable adult value around age 30 y (Figs. 1D and 2).
During the peak in number of dendritic spines (i.e., between 2
and 19 y of age), the DSD values were 25–40% higher in layer
IIIc than in layer V pyramidal neurons (Fig. 2); the difference was
larger on the apical proximal oblique dendrites (P = 0.04) than
on apical distal oblique (P = 0.10) or basal dendrites (P = 0.12).
curves fit the distribution of data from the basal dendrites (A), apical proximal oblique dendrites (B), and apical distal oblique dendrites (C) of pyramidal cells
from layer IIIc and V. In all cases the equation of the curves is a double exponential function in the form: y = a*exp(−bt) + c*exp (−dt) + e, where a, b, c, d, and
e are fixed coefficients, and t is time in years (Table S3).
The DSD, as defined in Fig. 1, plotted at the linear scale to illustrate the dynamics of changes occurring during the 100-y human lifespan. Regression
Petanjek et al. PNAS Early Edition
| 3 of 6
Because both types of neurons attain the adult-like total den-
dritic length during the third postnatal year (28), our findings
suggest that the number of overproduced spines is consistently
higher on layer IIIc than on layer V pyramidal neurons. This
difference probably reflects a more recent evolutionary history of
layer IIIc pyramidal neurons, which in primates represent a ma-
jor source of highly expanded ipsi- and contralateral associative
cortico–cortical connections (28, 42–46).
The average DSD value remained relatively constant between
38 and 65 y of age for a particular segment type (Fig. 2). How-
ever, on layer IIIc neurons, the DSD was most prominent in the
distal apical oblique dendrites (Fig. 2 and Table S2); those
neurons display DSD values that are ∼20% higher in the adult
prefrontal cortex than in layer V pyramidal neurons (P = 0.18
for apical distal oblique dendrites, P = 0.05 for apical proximal
oblique dendrites, and P = 0.10 for basal dendrites) (Fig. 2). It
should be noted that both basal and apical oblique dendrites of
layer IIIc pyramid neurons are ∼10% longer than those of layer
V neurons (28). Thus, the total number of dendritic spines
remains greater in layer IIIc than in layer V pyramidal neurons
both during development and in the adult prefrontal cortex.
The present study provides three findings concerning maturation
of the human prefrontal cortex: (i) the period of overproduction
and elimination of dendritic spines on pyramidal neurons in this
area extends to the third decade of life; (ii) the pruning of su-
pernumerary dendritic spines is more pronounced in layer IIIc
cortico–cortical neurons than in comparable segments of layer V
subcortically projecting neurons; and (iii) for layer IIIc neurons,
spine pruning begins earlier in basal and proximal apical oblique
dendrites than in distal apical oblique dendrites.
Previous data from human and nonhuman primates showed
somewhat higher synaptic overproduction in supragranular than
in infragranular layers (8), results that are consistent with the data
obtained in the present study. Furthermore, recent data on
identified neurons obtained from both the rhesus monkey (47)
and human (48) showed marked regional differences in the
number of spines grown and pruned in the basal dendritic tree of
layer III pyramidal neurons. The spine formation and elimination
between sensory, association, and executive cortex in these studies
displayed a similar pattern. However, the number of spines in the
adult cortex was related to functional hierarchy, as was the
number of spines overproduced, being highest in the prefrontal
cortex and lowest in the primary sensory regions (47, 48).
In the present study we have not performed the regional
comparison because of the lack of appropriate material. How-
ever, there is converging evidence that dynamics of synaptic
overproduction and elimination differ among cytoarchitectonic
areas both in humans and in nonhuman primates (2, 8, 47–49).
Most of these studies indicate that the prefrontal cortex under-
goes the largest overproduction and the slowest rate of elimina-
tion of all areas, an observation which is explained by the late
evolutionary emergence of the prefrontal cortex (47). Although
the present study provides data only for the prefrontal cortex, it
reveals a difference in the rate of spine formation and elimination
between two evolutionarily different pyramidal cell populations
and their dendrites situated in the different layers within the same
region. Thus, our findings support the notion that different types
of microcircuitry may have different rates of synaptic formation
and elimination. These findings are in line with the finding that
reorganization of intracortical excitatory synaptic systems in
macaque prefrontal cortex continues after puberty, when the
cortico–cortical synapses reach maturity (49). Therefore, we
propose that the most extensive and protracted overproduction in
humans is related to the associative and intracortical excitatory
network that becomes more represented across the functional
hierarchy and markedly prominent in the areas of highest order,
such as the prefrontal cortex (28, 43, 45, 46, 50).
Taken together, the previous data and the present findings
strongly indicate that anatomical (23, 24, 26–28, 31) and func-
tional changes (15, 18, 19, 21, 22, 25, 35) in the prefrontal cortex
observed in vivo during late adolescence and young adulthood
reflect the dynamic reorganization of synaptic circuitry rather
than solely activity-dependent molecular tuning of the stable
synaptic connections. Experimental studies performed in de-
veloping and adult rodents indicate that dendritic spines in the
cerebral cortex are remarkably plastic initially but gradually be-
come very stable, with the majority lasting throughout the entire
lifespan (51–55). Analysis of the human frontal cortex also shows
significant changes in synapse-associated molecules during the
period of growth and strengthening of synaptic elements in
childhood (32). Therefore, molecular tuning of synaptic strength
during the formative years may be a major mechanism for the
environmental effect on structural reorganization, including
elimination of supernumerary spines and synapses (1, 6, 7). This
hypothesis is supported by the finding that a peak in the ex-
pression of genes regulating neuronal development, including
those that are associated with schizophrenia, occurs between age
15 and 25 y (56). Finally, comparative analysis of mRNA expres-
sion in the prefrontal cortex shows that the dramatic changes in
transcriptome profiles in the human brain are delayed relative to
nonhuman primates (57). Thus, our data on spine overproduction
and elimination are not in contradiction to the molecular changes
in the synaptic membranes. It is likely that molecular changes
occur in each phase, but after the period of synaptic stabilization
molecular tuning becomes the predominant way of interacting
with the environment (51, 52, 54).
Although the molecular mechanisms that regulate prolonged
reorganization of dendritic spines are not well understood, there
are indications that they reflect the changes in dopaminergic in-
nervation. In both human and nonhuman primates, the dopami-
nergic input, together with glutamatergic synapses, terminates
predominantly on thedendrites oflayerIIIc pyramidal neurons of
the prefrontal cortex (8, 29, 42, 58), where it modulates neuronal
activity (59). The magnitude of dopaminergic innervation in
the monkey and human cortex, including gene expression of the
D1-receptor that is essential for the bidirectional modulation of
synaptic plasticity in the medial prefrontal cortex (60), increases
up to young adulthood (61) and reaches its highest level during
adolescence and young adulthood (62–64). These data led to the
hypothesis that an increase in dopaminergic innervation in the
prefrontal cortex is associated with an increase in modulation of
in synaptic stabilization (65). Furthermore, dopamine–glutamate
interaction on dendritic spines of pyramidal neurons in the pre-
protecting vulnerable subjects from developing schizophrenia
(66). It may be significant that the largest postnatal increase in the
extracellular dopamine levels in the prefrontal cortex) occurs be-
tween the second and the fourth decade (67), supporting the hy-
pothesis that the dopamine–glutamate interaction is involved in
regulating synaptic elimination in the human prefrontal cortex.
It may seem paradoxical that the period during which learning
and acquisition of new knowledge are highest in the human
coincides with a net decrease rather than an increase in the
number of synapses. The protracted postadolescent period of
synaptic elimination and increase in dopaminergic innervation of
the prefrontal cortex (61) may be linked to human-specific cog-
nitive functions and circuitry specializations (58) that are a prod-
uctofcooperationbetween genetic endowmentandenvironment,
as postulated by the selective-stabilization hypothesis (1, 3, 8).
The prolonged developmental plasticity in the associative frontal
cortex in human allows an unprecedented opportunity for acqui-
4 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1105108108 Petanjek et al.
sition of the highest level of cognitive abilities (27, 39, 50, 68–70)
but also is susceptible to the formation of abnormal circuitry
that is manifested in late-expressed neuropsychiatric disorders
Materials and Methods
91y,werestudiedquantitatively (Table S1).Noneofthesubjectshadaclinical
historyof neurological disorder or aneuropathologicalalteration detected at
autopsy. All analyzed subjects lived under standard environmental and so-
cioeconomic conditions. Details on subject age, sex, postmortem delay, and
cause of death as described in autopsy and medical records are given in Table
S1. In two subjects (19 and 22 y old) death was caused by suicide, but personal
and clinical records did not point to any specific psychiatric disorders. The
brains were collected with the approval of the Ethical Committee of the
University of Zagreb School of Medicine in compliance with Croatian law. The
prefrontal cortex tissue was studied in sections from the Zagreb Collection
located at the Croatian Institute for Brain Research (75).
The time interval between death and fixation of the tissue (i.e., the
postmortem delay) was <8 h for early postnatal cases, <13 h for infants, <16 h
for children, and <20 h for adults. All analyzed subjects died without pre-
agonal state, so that the postmortem delay actually represents the interval
in which neuron death took place. No staining artifacts caused by post-
mortem delay described for rapid Golgi staining were detected in the cases
that were studied quantitatively.
Tissue Preparation. The parts of the prefrontal cortex examined included the
superior and middle frontal gyrus, mainly defined as the frontal granular and
magnopyramidal Brodmann’s area 9 (76). Blocks of tissue (1 cm3) were sec-
tioned perpendicular to the long axis of the frontal gyrus, from the right
cortical tissue was immersed immediately in rapid Golgi solution (0.3% os-
the dichromate solution was replaced by 1% silver nitrate for 2 d. Then the
tissue was dehydrated and embedded rapidly in 8% celloidin. After embed-
ding, a microtome was used to section the blocks serially into coronal sections
160–200 μm in thickness. This thickness was chosen as a compromise to have
many dendrites and good microscopic clarity. Nissl-stained sections from ad-
jacent blocks were cut at 30 μm to check and additionally ensure that the
neurons quantified were taken from Brodmann’s area 9 (37).
Quantitative Analysis. The criteria for cell selection for quantitative analysis
included clear impregnation of the finest dendrites and dendritic spines
visibleonall partsofthedendritic tree,withoutsmoothsegments andatleast
10 well-impregnated neurons per layer. The entire measured part of the
dendritic segments had to be sharply visible with a 40× objective without
moving the microscope in z direction (depth). The use of these criteria for
quantification resulted in the inclusion of only 32 of 109 subjects analyzed.
The measurements were performed using a 63×-oil immersion objective
with a long working distance (Olympus 1.4 N.A.). The cases were coded, so
that investigators were not aware of the subjects’ age, sex, or medical his-
tory. Large layer IIIc pyramidal neurons were always located within the 200-
μm–wide zone above layer IV. Layer V was detected in counterstained Nissl
sections and in Golgi sections as a 200-μm–thick layer below the transparent
layer IV. Only neurons of typical pyramidal morphology were analyzed (28,
45, 46, 77). Modified pyramidal neurons were not included in the analysis.
Dendritic spine density was analyzed on apical side branches (oblique
dendrites) and basal dendrites (Fig. 1 A and B). Data in Table S2 give spine
numbers on (i) the most proximal 50-μm length of the first-order basal
dendrite, (ii) the first 50-μm length in side branches of apical dendrites
(oblique dendrites), which were divided into two groups: proximal oblique
dendrites, originating from the apical dendrite segment at up to 100-μm
distance from soma, and (iii) distal oblique dendrites, originating in the
segment of apical dendrite at a distance of 100–200 μm from soma (Fig. 1B).
number was tested separately for each layer and each segment with one-way
(28). In the statistical analysis every subject represents a separate age. The
a posteriori Student-Newman-Keuls test for multiple comparisons was ap-
plied to determine which subjects were significantly different. P values <0.05
were considered statistically significant. Statistical analysis with parametric
and nonparametric procedures showed comparable results.
DSD was plotted at the linear scale to illustrate the dynamics of changes
occurring during the 100-y human lifespan (Fig. 2) and to obtain regression
curves fitting the distribution of data from the basal, apical proximal oblique,
and apical distal oblique dendrites of the pyramidal cells from layer IIIc and V.
Inall casestheequationofthecurves was adoubleexponential functioninthe
form: y = a*exp(−bt) + c*exp(−dt) + e, where a, b, c, d, and e are fixed coef-
ficients and t is time in years (Table S3).
ACKNOWLEDGMENTS. We thank J. Arellano for insightful discussions and A.
Bernacchia for help with regression curves in Fig. 2. This work was supported
by grants from the Ministry of Science, Education, and Sports of the Republic
of Croatia (to Z.P., M.J., I.K., and G.Š.), the Unity Through Knowledge Fund
(to I.K.), the National Institutes of Health, and the Kavli Institute for Neuro-
science at Yale University (to P.R.).
1. Changeux JP, Danchin A (1976) Selective stabilisation of developing synapses as
a mechanism for the specification of neuronal networks. Nature 264:705–712.
2. Huttenlocher PR, Dabholkar AS (1997) Regional differences in synaptogenesis in hu-
man cerebral cortex. J Comp Neurol 387:167–178.
3. Rakic P, Bourgeois JP, Eckenhoff MF, Zecevic N, Goldman-Rakic PS (1986) Concurrent
overproduction of synapses in diverse regions of the primate cerebral cortex. Science
4. Huttenlocher PR (1979) Synaptic density in human frontal cortex—developmental
changes and effects of aging. Brain Res 163:195–205.
5. Bourgeois JP, Goldman-Rakic PS, Rakic P (1994) Synaptogenesis in the prefrontal
cortex of rhesus monkeys. Cereb Cortex 4:78–96.
6. Rakic P, Riley KP (1983) Overproduction and elimination of retinal axons in the fetal
rhesus monkey. Science 219:1441–1444.
7. Purves D, Lichtman JW (1980) Elimination of synapses in the developing nervous
system. Science 210:153–157.
8. Rakic P, Bourgeois JP, Goldman-Rakic PS (1994) Synaptic development of the cerebral
cortex: Implications for learning, memory, and mental illness. Prog Brain Res 102:
9. Bizzi E, Scolnick E, Desimone R (2010) Advances in brain sciences: Implication for
therapy. Bull Am Acad Sci 63:15–22.
10. Weinberger DR (1987) Implications of normal brain development for the pathogen-
esis of schizophrenia. Arch Gen Psychiatry 44:660–669.
11. Bennett A O MR (2008) Dual constraints on synapse formation and regression in
schizophrenia: Neuregulin, neuroligin, dysbindin, DISC1, MuSK and agrin. Aust N Z J
12. Garber K (2007) Neuroscience. Autism’s cause may reside in abnormalities at the
synapse. Science 317:190–191.
13. Hutsler JJ, Zhang H (2010) Increased dendritic spine densities on cortical projection
neurons in autism spectrum disorders. Brain Res 1309:83–94.
14. Bossong MG, Niesink RJ (2010) Adolescent brain maturation, the endogenous can-
nabinoid system and the neurobiology of cannabis-induced schizophrenia. Prog
15. Feinberg I, Campbell IG (2010) Sleep EEG changes during adolescence: An index of
a fundamental brain reorganization. Brain Cogn 72:56–65.
16. Dierssen M, Ramakers GJ (2006) Dendritic pathology in mental retardation: From
molecular genetics to neurobiology. Genes Brain Behav 5(Suppl 2):48–60.
17. Quartz SR, Sejnowski TJ (1997) The neural basis of cognitive development: A con-
structivist manifesto. Behav Brain Sci 20:537–556, discussion 556–596.
18. Casey BJ, Jones RM, Hare TA (2008) The adolescent brain. Ann N Y Acad Sci 1124:
19. Whitford TJ, et al. (2007) Brain maturation in adolescence: Concurrent changes in
neuroanatomy and neurophysiology. Hum Brain Mapp 28:228–237.
20. Chugani HT (1998) A critical period of brain development: Studies of cerebral glucose
utilization with PET. Prev Med 27:184–188.
21. Dosenbach NU, et al. (2010) Prediction of individual brain maturity using fMRI. Sci-
22. Luna B, Padmanabhan A, O’Hearn K (2010) What has fMRI told us about the de-
velopment of cognitive control through adolescence? Brain Cogn 72:101–113.
23. Asato MR, Terwilliger R, Woo J, Luna B (2010) White matter development in ado-
lescence: A DTI study. Cereb Cortex 20:2122–2131.
24. Gogtay N, et al. (2004) Dynamic mapping of human cortical development during
childhood through early adulthood. Proc Natl Acad Sci USA 101:8174–8179.
25. Jolles DD, van Buchem MA, Crone EA, Rombouts SA (2011) A comprehensive study of
whole-brain functional connectivity in children and young adults. Cereb Cortex 21:
26. Knickmeyer RC, et al. (2010) Maturational trajectories of cortical brain development
through the pubertal transition: Unique species and sex differences in the monkey
revealed through structural magnetic resonance imaging. Cereb Cortex 20:
27. Shaw P, et al. (2006) Intellectual ability and cortical development in children and
adolescents. Nature 440:676–679.
28. Petanjek Z, Judaš M, Kostovi? c I, Uylings HB (2008) Lifespan alterations of basal den-
dritic trees of pyramidal neurons in the human prefrontal cortex: A layer-specific
pattern. Cereb Cortex 18:915–929.
Petanjek et al.PNAS Early Edition
| 5 of 6
29. Gonzalez-Burgos G, et al. (2008) Functional maturation of excitatory synapses in layer Download full-text
3 pyramidal neurons during postnatal development of the primate prefrontal cortex.
Cereb Cortex 18:626–637.
30. Paus T, Keshavan M, Giedd JN (2008) Why do many psychiatric disorders emerge
during adolescence? Nat Rev Neurosci 9:947–957.
31. Sowell ER, et al. (2003) Mapping cortical change across the human life span. Nat
32. Webster MJ, Elashoff M, Weickert CS (2011) Molecular evidence that cortical synaptic
growth predominates during the first decade of life in humans. Int J Dev Neurosci 29:
33. Leppänen JM, Nelson CA (2009) Tuning the developing brain to social signals of
emotions. Nat Rev Neurosci 10:37–47.
34. Blakemore SJ (2008) The social brain in adolescence. Nat Rev Neurosci 9:267–277.
35. Sebastian C, Viding E, Williams KD, Blakemore SJ (2010) Social brain development and
the affective consequences of ostracism in adolescence. Brain Cogn 72:134–145.
36. Germine LT, Duchaine B, Nakayama K (2011) Where cognitive development and
aging meet: Face learning ability peaks after age 30. Cognition 118:201–210.
37. Rajkowska G, Goldman-Rakic PS (1995) Cytoarchitectonic definition of prefrontal
areas in the normal human cortex: I. Remapping of areas 9 and 46 using quantitative
criteria. Cereb Cortex 5:307–322.
38. Beckman M (2004) Neuroscience. Crime, culpability, and the adolescent brain. Science
39. Kishiyama MM, Boyce WT, Jimenez AM, Perry LM, Knight RT (2009) Socioeconomic
disparities affect prefrontal function in children. J Cogn Neurosci 21:1106–1115.
40. Dehaene S, Changeux JP (1997) A hierarchical neuronal network for planning be-
havior. Proc Natl Acad Sci USA 94:13293–13298.
41. Hill J, et al. (2010) Similar patterns of cortical expansion during human development
and evolution. Proc Natl Acad Sci USA 107:13135–13140.
42. Goldman-Rakic PS (1995) Cellular basis of working memory. Neuron 14:477–485.
43. Elston GN, et al. (2006) Specializations of the granular prefrontal cortex of primates:
Implications for cognitive processing. Anat Rec A Discov Mol Cell Evol Biol 288:26–35.
44. Dehaene S, Kerszberg M, Changeux JP (1998) A neuronal model of a global work-
space in effortful cognitive tasks. Proc Natl Acad Sci USA 95:14529–14534.
45. Elston GN, Benavides-Piccione R, Elston A, Manger PR, Defelipe J (2011) Pyramidal
cells in prefrontal cortex of primates: Marked differences in neuronal structure
among species. Front Neuroanat 5:2.
46. Zeba M, Jovanov-Milosevi? c N, Petanjek Z (2008) Quantitative analysis of basal den-
dritic tree of layer III pyramidal neurons in different areas of adult human frontal
cortex. Coll Antropol 32(Suppl 1):161–169.
47. Elston GN, Oga T, Fujita I (2009) Spinogenesis and pruning scales across functional
hierarchies. J Neurosci 29:3271–3275.
48. Jacobs B, Driscoll L, Schall M (1997) Life-span dendritic and spine changes in areas 10
and 18 of human cortex: A quantitative Golgi study. J Comp Neurol 386:661–680.
49. Woo TU, Pucak ML, Kye CH, Matus CV, Lewis DA (1997) Peripubertal refinement of
the intrinsic and associational circuitry in monkey prefrontal cortex. Neuroscience 80:
50. Jacobs B, et al. (2001) Regional dendritic and spine variation in human cerebral cor-
tex: A quantitative Golgi study. Cereb Cortex 11:558–571.
51. Yang G, Pan F, Gan WB (2009) Stably maintained dendritic spines are associated with
lifelong memories. Nature 462:920–924.
52. Grutzendler J, Kasthuri N, Gan WB (2002) Long-term dendritic spine stability in the
adult cortex. Nature 420:812–816.
53. Bloss EB, et al. (2011) Evidence for reduced experience-dependent dendritic spine
plasticity in the aging prefrontal cortex. J Neurosci 31:7831–7839.
54. Lin YC, Koleske AJ (2010) Mechanisms of synapse and dendrite maintenance and their
disruption in psychiatric and neurodegenerative disorders. Annu Rev Neurosci 33:
55. Arellano JI, Espinosa A, Fairén A, Yuste R, DeFelipe J (2007) Non-synaptic dendritic
spines in neocortex. Neuroscience 145:464–469.
56. Harris LW, et al. (2009) Gene expression in the prefrontal cortex during adolescence:
Implications for the onset of schizophrenia. BMC Med Genomics 2:28.
57. Somel M, et al. (2009) Transcriptional neoteny in the human brain. Proc Natl Acad Sci
58. Raghanti MA, et al. (2008) Cortical dopaminergic innervation among humans, chim-
panzees, and macaque monkeys: A comparative study. Neuroscience 155:203–220.
59. Henze DA, González-Burgos GR, Urban NN, Lewis DA, Barrionuevo G (2000) Dopa-
mine increases excitability of pyramidal neurons in primate prefrontal cortex.
J Neurophysiol 84:2799–2809.
60. Huang YY, Simpson E, Kellendonk C, Kandel ER (2004) Genetic evidence for the bi-
directional modulation of synaptic plasticity in the prefrontal cortex by D1 receptors.
Proc Natl Acad Sci USA 101:3236–3241.
61. Wahlstrom D, White T, Luciana M (2010) Neurobehavioral evidence for changes in
dopamine system activity during adolescence. Neurosci Biobehav Rev 34:631–648.
62. Weickert CS, et al. (2007) Postnatal alterations in dopaminergic markers in the human
prefrontal cortex. Neuroscience 144:1109–1119.
63. Lambe EK, Krimer LS, Goldman-Rakic PS (2000) Differential postnatal development of
catecholamine and serotonin inputs to identified neurons in prefrontal cortex of
rhesus monkey. J Neurosci 20:8780–8787.
64. Rosenberg DR, Lewis DA (1995) Postnatal maturation of the dopaminergic in-
nervation of monkey prefrontal and motor cortices: A tyrosine hydroxylase immu-
nohistochemical analysis. J Comp Neurol 358:383–400.
65. Goldman-Rakic PS, Muly, EC, 3rd, Williams GV (2000) D(1) receptors in prefrontal cells
and circuits. Brain Res Brain Res Rev 31:295–301.
66. Tan HY, Callicott JH, Weinberger DR (2009) Prefrontal cognitive systems in schizophre-
nia: Towards human genetic brain mechanisms. Cogn Neuropsychiatry 14:277–298.
67. Tunbridge EM, et al. (2007) Catechol-o-methyltransferase enzyme activity and protein
expression in human prefrontal cortex across the postnatal lifespan. Cereb Cortex 17:
68. Nelson CA, 3rd, et al. (2007) Cognitive recovery in socially deprived young children:
The Bucharest Early Intervention Project. Science 318:1937–1940.
69. Uylings HBM (2006) Series in Cognitive Neuroscience and Language Learning and
Processing, eds Gullberg M, Indefrey P (Blackwell, Oxford), pp 59–90.
70. Raizada RD, Kishiyama MM (2010) Effects of socioeconomic status on brain de-
velopment, and how cognitive neuroscience may contribute to levelling the playing
field. Front Hum Neurosci 4:3.
71. Eluvathingal TJ, et al. (2006) Abnormal brain connectivity in children after early severe
72. Diamond A, Barnett WS, Thomas J, Munro S (2007) Preschool program improves
cognitive control. Science 318:1387–1388.
73. Changeux JP (2009) Nicotinic receptors and nicotine addiction. C R Biol 332:421–425.
74. Penzes P, Cahill ME, Jones KA, VanLeeuwen JE, Woolfrey KM (2011) Dendritic spine
pathology in neuropsychiatric disorders. Nat Neurosci 14:285–293.
75. Judaš M, et al. (2011) The Zagreb Collection of human brains: A unique, versatile,
but underexploited resource for the neuroscience community. Ann N Y Acad Sci
76. Rajkowska G, Goldman-Rakic PS (1995) Cytoarchitectonic definition of prefrontal
areas in the normal human cortex: II. Variability in locations of areas 9 and 46 and
relationship to the Talairach Coordinate System. Cereb Cortex 5:323–337.
77. DeFelipe J, Fariñas I (1992) The pyramidal neuron of the cerebral cortex: Morpho-
logical and chemical characteristics of the synaptic inputs. Prog Neurobiol 39:563–607.
6 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1105108108 Petanjek et al.