Content uploaded by Peter T. Fox
Author content
All content in this area was uploaded by Peter T. Fox on Jan 11, 2024
Content may be subject to copyright.
Automated Talairach Atlas Labels For
Functional Brain Mapping
Jack L. Lancaster,*Marty G. Woldorff, Lawrence M. Parsons,
Mario Liotti, Catarina S. Freitas, Lacy Rainey, Peter V. Kochunov,
Dan Nickerson, Shawn A. Mikiten, and Peter T. Fox
Research Imaging Center, University of Texas Health Science Center at San Antonio
䉬䉬
Abstract: An automated coordinate-based system to retrieve brain labels from the 1988 Talairach Atlas,
called the Talairach Daemon (TD), was previously introduced [Lancaster et al., 1997]. In the present study,
the TD system and its 3-D database of labels for the 1988 Talairach atlas were tested for labeling of
functional activation foci. TD system labels were compared with author-designated labels of activation
coordinates from over 250 published functional brain-mapping studies and with manual atlas-derived
labels from an expert group using a subset of these activation coordinates. Automated labeling by the TD
system compared well with authors’ labels, with a 70% or greater label match averaged over all locations.
Author-label matching improved to greater than 90% within a search range of ⫾5 mm for most sites. An
adaptive grey matter (GM) range-search utility was evaluated using individual activations from the M1
mouth region (30 subjects, 52 sites). It provided an 87% label match to Brodmann area labels (BA4&BA
6) within a search range of ⫾5 mm. Using the adaptive GM range search, the TD system’s overall match
with authors’ labels (90%) was better than that of the expert group (80%). When used in concert with
authors’ deeper knowledge of an experiment, the TD system provides consistent and comprehensive
labels for brain activation foci. Additional suggested applications of the TD system include interactive
labeling, anatomical grouping of activation foci, lesion-deficit analysis, and neuroanatomy education.
Hum. Brain Mapping 10:120–131, 2000. ©2000 Wiley-Liss, Inc.
Keywords: Talairach Daemon, volume occupancy, Talairach Labels, brain labels
䉬䉬
INTRODUCTION
It is common practice in brain mapping experiments
to report locations of functional and anatomical sites
using standardized x-y-z coordinates [Fox et al., 1985;
Friston et al., 1989, 1991; Fox 1995]. Widespread use of
Talairach coordinates [Talairach et al., 1988] fostered
the development of the BrainMap威database that en-
codes and queries the locations of functional neuroim-
aging findings using these coordinates [Fox et al.,
1994]. The Talairach Daemon (TD) system [Lancaster
et al., 1997] expands this concept by providing easy
Internet access to a 3-dimensional (3-D) database of
brain labels accessed by Talairach coordinates. The
Talairach labels database uses a volume-filling hierar-
chical naming scheme to organize labels for brain
structures ranging from hemispheres to cytoarchitec-
tural regions [Freitas et al., 1996, Lancaster et al., 1997].
This scheme is reflected in the database name, Volume
Edited by: Karl Friston, Associate Editor.
Contract grant sponsor: NIMH; Contract grant number: 5 P01
MH52176-07; Contract grant sponsor: NLM; Contract grant number:
1 RO1 LM06858-01.
*Correspondence to: Jack L. Lancaster, Ph.D., University of Texas
Health Science Center at San Antonio, Research Imaging Center,
7703 Floyd Curl Drive, San Antonio, Texas 78284.
E-mail: jlancaster@uthscsa.edu
Web address for this project: http://nc.uthscsa.edu/projects/
tnlairachdaemon.html
Received 1 November 1999; Accepted 28 April 2000.
䉬Human Brain Mapping 10:120–131(2000) 䉬
©2000 Wiley-Liss, Inc.
Occupancy Talairach Labels (VOTL). The focus of this
report is an evaluation of the accuracy of the TD
system for obtaining Talairach labels for functional
activation sites.
Brain atlases define and catalog rudimentary spatial
features of brain structure using traditional nomencla-
ture. Atlas labels provide a consistent terminology for
qualitative description of regional brain structures.
This is exemplified by the broad use of the 1988 Ta-
lairach atlas by the human brain mapping community
[Steinmetz et al., 1989; Fox 1995]. Subcortical struc-
tures such as thalamus, caudate, and lentiform are
easily identified by visual comparison of atlas sections
with high-resolution 3-D MR images. Unlike the sub-
cortical region, however, visual labeling in the cortex
is highly problematic. Visual identification of gyri in
serial MR section images is tedious, subject to repro-
ducibility problems [Sobel et al., 1993], and subject to
failure for secondary and tertiary sulci, that are not
always present [Ono et al., 1990]. Gyri identification
can be improved by use of surface rendering [Watson
et al., 1993] and/or surface flattening [Dale and Ser-
eno, 1993; Drury and VanEssen, 1997] to create images
with better definition of sulcal and gyral boundaries.
These processing methods, however, are neither time-
efficient nor readily available and provide little sup-
port for standardized labeling of individual brains.
The coordinate-based labeling scheme of the TD sys-
tem, with concise gyral definitions, provides a consis-
tent labeling alternative to visual labeling of gyri.
Visual labeling is hampered by vague boundary
definitions for many brain structures. For example,
complete boundaries of lobes and Brodmann areas are
not found in popular atlases due to lack of consensus
for their definition, including the 1988 Talairach atlas.
Users must therefore select among several possible
labels near non-delineated boundaries, increasing in-
ter-user variability and making labeling susceptible to
observer expectancy-based bias. The VOTL database
provides explicit boundaries for each labeled brain
structure to manage this problem (see Methods).
Visual labeling is problematic even when explicit
atlas labels are available for lobes, gyri and Brodmann
areas. Boundaries for these regions are not easy to
identify in high-resolution MR images, more difficult
in group-average MR images, and practically impos-
sible to identify in low resolution images (PET,
SPECT, and fMRI), whether from individual subjects
or group averages. Numerous automated methods,
based on computerized mapping of atlas region labels
onto medical images, have been developed to facilitate
labeling [Bohm et al., 1983, 1991; Evans et al., 1991;
Roland et al., 1994; Collins et al., 1994, 1995]. However,
none of these atlas-to-image label-mapping methods
fully label the brain volume, nor are they simple to use
or widely distributed. Alternatively, an efficient im-
age-to-atlas mapping method is provided in the TD
system to retrieve 1988 Talairach atlas labels for acti-
vation coordinates [Lancaster et al., 1997].
A coordinate-based automated labeling system for
functional activation sites should provide appropriate
labels regardless of methodological differences. To test
this capability in the TD system, it was compared with
a large set of published functional activations that
reported both coordinates and labels.
METHODS
The methods section is divided into development
and evaluation subsections. The development subsec-
tion describes: (1) creation of a 3-D Talairach Atlas
from the published atlas, (2) formulation of the VOTL
labeling scheme for this 3-D atlas, and (3) features of
the TD system software. The evaluation subsection
focuses on three important areas: (1) labeling by the
TD system vs. Talairach labels from published func-
tional brain studies, (2) labeling by the TD system vs.
knowledgeable users of the Talairach atlas, and (3)
labeling accuracy of the TD system in a PET activation
study.
TD System Development
3-D Talairach Atlas
The 1988 Talairach Atlas contains a series of high-
detail color tracings (coronal, axial, and sagittal sec-
tions) derived from MR images of a 60-year old right-
handed European female. Axial section images were
chosen for developing the 3-D atlas because this sec-
tion format is the most common acquired by tomo-
graphic medical imagers. There are twenty-seven axial
sections ranging from z ⫽⫹65 mm and to z ⫽-40
mm. The section spacing ranges from 2 mm near the
AC to 5 mm for the slices near the top of the brain.
Each section image was digitized with x-y resolution
of 0.43 mm, and each major structure segmented using
a combination of color discrimination (Adobe Photo-
Shop™, San Jose, CA) and thresholding (Alice™,
PAREXEL, Boston, MA). The digitized section images
served as reference planes to interpolate a continuous
3-D brain atlas using pixel (x & y) and section (z)
locations relative to the AC or origin. Section images
were assigned the z-coordinate designated in the atlas.
All images were carefully registered during digitiza-
䉬Automated Talairach Atlas Labels 䉬
䉬121 䉬
tion, to guarantee that the x-y coordinate of the origin
was maintained at a consistent location and that the x-
and y-axes were properly aligned. A contiguous 1-mm
3-D Talairach atlas volume was created from the ref-
erence images using resampling in the x and y direc-
tions and nearest neighbor interpolation in the z di-
rection. Care was taken to guarantee that the brain
dimensions matched that previously published for the
atlas brain [Lancaster et al., 1995] (L-R ⫽136 mm,
A-P ⫽172 mm, and S-I ⫽118 mm). Small regional
differences were seen, but these were usually less than
1 mm. Since most axial sections of the 1988 Talairach
atlas were incomplete on the right, right-side data was
made by reflecting left-side data about the y-axis for
all images.
Volume Occupancy Talairach Labels
Database (VOTL)
The anatomical structure-naming scheme devised
for organizing the numerous 3-D anatomical regions
(volumes of interest - VOIs) from the 3-D Talairach
atlas is based on volume occupancy (Fig. 1). More
specifically a volume-filling, hierarchical, anatomical-
labeling scheme is used, wherein each VOI is defined
using 3-D coordinates (for location) and a unique code
(for anatomical label). The VOIs in the computerized
3-D Talairach atlas were organized into five hierarchi-
cal levels: Hemisphere, Lobe, Gyrus, Tissue type, and
Cell type (Table I). Rules and procedures were
adopted for the VOTL labeling scheme to explicitly
define boundaries for labeled brain structures [Freitas
et al., 1996, Lancaster et al., 1997], and the entire brain
volume fully labeled at each hierarchical level (Fig. 1).
The rules and procedures for the VOTL labeling
scheme are outlined below:
Hemisphere Level. The largest anatomical structures in
the brain (cerebrum, cerebellum, and brainstem) are
assigned to the hemisphere level. The outer borders of
hemisphere structures were extended slightly to in-
clude small invaginations (Fig. 2). The TD system
Figure 1.
A typical set of 3-D data used to create the volume occupancy (VOTL) database around the z ⫽
⫹1 level. Openings in Lobe through Cell levels were provided to emphasize the 3-D nature of data
at each level.
䉬Lancaster et al. 䉬
䉬122 䉬
returns “*” or “interhemispheric” labels for coordi-
nates falling outside hemisphere level structures. The
left- and right-side attribute is assigned at this level.
Lobe Level. For the cerebrum, the lobe level consists of
the four main lobes (Frontal, Temporal, Parietal, and
Occipital), a single lobular equivalent (Limbic lobe)
TABLE I. Volume-occupancy talairach labels
Level Label
Hemisphere
(Level 1)
Cerebrum (R/L) Cerebellum
(R/L)
Brainstem
(R/L)
Lobe
(Level 2)
Lobes Sub-Lobar In Progress In Progress
Gyrus
(Level 3)
Gyri Sub-Gyral Nuclei Extra-Nuclear “ “
Tissue Type
(Level 4)
GM WM WM CSF GM WM WM CSF “ “
Cell Type
(Level 5)
BA — — — Sub-
Nuc.
— — Space “ “
BA - Brodmann Area; WM - White Matter; GM - Gray Matter; CSF - Cerebral Spinal Fluid
Figure 2.
Example of VOTL region labels for a Talairach atlas section image at the z ⫽⫹1 level. Lobe labels
are illustrated with patterned color fills. Several Brodmann areas (cell level) are illustrated on the
top left using solid color fills. Several gyral level structures are illustrated on the bottom left using
bold color outlines.
䉬Automated Talairach Atlas Labels 䉬
䉬123 䉬
deeper within the brain, and a sub-lobar region. Using
this labeling scheme, cingulate is within the limbic
lobe and insula within the sub-lobar region. Outer
boundaries of lobes were defined following cortical
boundaries. Since explicit inner lobe boundaries were
not present in the Talairach atlas, inner boundaries
were defined using lines connecting the deepest extent
of bounding sulci in axial section images (Fig. 2). The
sub-lobar region was assigned to the volume of the
cerebrum not specifically designated as a lobe, com-
pleting the lobe-level volume-occupancy labeling for
the cerebrum. The cerebellum and brainstem labeling
(in progress) will follow this general scheme.
Gyrus Level. This level includes gyri for lobes and
various deep gray-matter structures within the sub-
lobar region. Outer gyral boundaries follow lobar
boundaries, while inner boundaries were defined us-
ing lines drawn between bounding sulcal pits in axial
section images. Since inner lobar boundaries often
extended deeper than inner gyral boundaries a sub-
gyral label was assigned to the region between the
inner boundary of a gyrus and the inner boundary of
its lobe (Fig. 2). Boundaries for sub-lobar gyrus-level
structures (Caudate, Thalamus, Lentiform, etc.) were
well delineated in the Talairach Atlas and used with-
out modification. There are forty-eight different gyrus
level labels.
Tissue Level. This level provides labels for gray mat-
ter, white matter, and cerebral spinal fluid (CSF). The
lateral, third, and fourth ventricles were the only CSF-
designated regions since non-ventricular CSF was not
labeled in the Talairach atlas and is too variable to
label consistently. Brain regions normally associated
with gray matter (cortex and nuclei) were labeled as
gray matter. Brain regions not labeled as gray matter
or CSF were designated as white matter.
Cell Level. The brain was labeled by cell-type in the
cortex using Brodmann’s scheme [Garey, 1994] and
tracts, spaces, and sub-nuclear regions for other por-
tions of the brain. Forty-seven Brodmann areas (BA)
labels were defined. Boundaries between adjacent
Brodmann areas are not defined in the Talairach Atlas,
so explicit BA boundaries were established that con-
formed to their original descriptions as closely as pos-
sible [Garey, 1994]. The BA boundaries were set at a
sulcal pit or gyral crown when possible (Fig. 2). Care
was taken to provide continuity of BA boundaries
between adjacent sections [Freitas et al., 1996]. Other
cell-level labels (e.g., for sub-lobar regions) closely
follow explicit Talairach Atlas labels. Labels for tracts
and spaces have not yet been created. The insular
region had not been assigned a Brodmann area, so it
was designated as BA 13. There are sixty-six different
cell level labels.
Cerebellum Labels
Several sources [Schmahmann et al., 1999;
Courchesne et al., 1989; Press et al., 1989, 1990; Ange-
vine et al., 1961] are being used for development of
cerebellar labels. The cerebellum labels will not come
from a single brain, but rather from a consensus based
on numerous sources. A working version of cerebellar
labels is in place. Since numerous refinements are
anticipated cerebellar labels were not evaluated in the
present study.
Talairach Daemon System Software
The general operating scheme of the TD system is
that a user sends Talairach coordinates to the TD
system server via a TD client. The server looks up
labels in the VOTL database and sends coordinates
with labels back to the user. The TD server is Internet
accessible and supports multiple concurrent requests.
It was implemented as a multi-threaded application,
with network communication using Berkeley Sockets,
with a query/response protocol using ASCII strings,
and a streamlined query/response processing method.
A variety of freely distributed client applications, in-
cluding Java versions, provide Internet access to TD
databases (http://ric.uthscsa.edu/projects/talairach-
daemon.html). The client-server database environ-
ment was named the Talairach Daemon for the atlas
space it references and the fact that the server is a
UNIX daemon process [Freitas et al., 1996; Lancaster et
al., 1997]. A disk-based version of the TD system (cli-
ent & VOTL database) has also been developed to
support high-speed labeling for large groups of coor-
dinates.
Newer TD clients provide an adaptive range-search
utility to find nearby GM labels. Searching is per-
formed in a cubic region centered on the coordinate of
interest. The search is started using a 3x3x3-mm
3
cube,
that is, with a search range of ⫹/- 1-mm along major
coordinate directions. If no GM label is found, the
search range is expanded to 5⫻5⫻5mm
3
(⫹/-2mm).
Expansion of the search range is continued in this
manner to a maximum range of 11⫻11⫻11 mm
3
(⫹/-
5 mm). If a single GM label is found at any search
range, its VOTL label and search range (⫹/- mm) are
returned, and the search terminated. If multiple GM
labels are encountered within a search range, the label
䉬Lancaster et al. 䉬
䉬124 䉬
with the highest frequency of occurrence is used. In
cases of ties the search range is enlarged until the tie is
broken. If no GM label is found, “No GM” is returned.
This GM search algorithm has been successful for all
Talairach coordinates tested to date, and there have
been no cases of ties.
TD System Evaluation
TD vs. Author Labels
A data set containing author-designated labels and
Talairach coordinates for activation sites widely dis-
tributed throughout the brain was compared with la-
bels obtained from the VOTL database by the TD
system. Mostly group-mean data (⬎5000 points; pub-
lished prior to 1996) from numerous laboratories
around the world were retrieved from the BrainMap威
database [Fox et al., 1994; Fox and Lancaster, 1998]. All
data from our research group were removed to elim-
inate potential bias. Additionally, all points with miss-
ing or incompatible label descriptions were removed
resulting in a refined data set of approximately 2500
points. The refined data set (113 authors and ⬎250
studies) provided a wide range of TD system labels for
comparison: all lobes, 33 gyri, 42 Brodmann areas, and
17 sublobar labels (nuclei, ventricles, and corpus cal-
losum).
Multiple author-designated labels were often seen
for the same anatomical region, so an author-to-VOTL
label-mapping scheme was designated to support like-
label comparisons (Table II). This table indicates that
author’s labels paired reasonably well with gyrus and
cell level labels. All label comparisons were therefore
made at the “Gyrus” or “Cell” levels since hemi-
sphere, lobe, and tissue labels can generally be in-
ferred from these. A 670-coordinate subset of labels for
ten commonly referenced brain regions was selected
for a detailed comparison with the TD system (Fig. 3,
Legend). To characterize distances from coordinates to
author-designated structures, label-matching scores
were calculated as a function of search range from 0 to
⫹/- 5mm.
TD vs. Human for Atlas Labels
It was proposed that the TD system would provide
a reliable method to obtain Talairach labels for func-
tional activation studies, and that automated labeling
could be done as well as or better than a user with
atlas in hand. To test these propositions TD-derived
labels were compared with labels looked up in the
1988 Talairach atlas by three knowledgeable users
(ML, LP, and MW). A set of approximately 100 labels
was targeted for testing by this group. The set of 670
author-labeled coordinates (See Previous Section)
were sorted by author label and like labels grouped
using Table II as a guide. A subset of widely distrib-
uted structures was selected including motor, sensory,
association, and limbic areas, encompassing numer-
ous Brodmann areas, as well as several deep gray
structures. Coordinates were equally distributed be-
tween left and right hemispheres, and labels distrib-
uted across all lobes. Data for each structure were
selected to include as many authors as possible to
minimize bias. The resulting test set from 51 authors
contained 106 labels in 16 different gyrus-level and 19
different cell-level structures.
Since the TD system was based on atlas axial section
images, human labelers (testers) were instructed to
initially use axial sections to determine a label for each
point’s Talairach coordinate. They then reviewed
coronal and sagittal section images to determine if the
same label would be obtained. Testers were instructed
to indicate the nearest gray matter label if a coordinate
fell outside gray matter. The testers were provided
with a 1988 Talairach atlas axial section image similar
to Figure 2 as an example of the VOTL scheme for lobe
and gyrus labeling (included lobe, sub-lobar, gyral
and sub-gyral region bounds outlined in color).
Testers were also briefed on how Brodmann area
boundaries were determined for the VOTL scheme.
The method for looking up labels using x-y-z coordi-
nates was left to the testers. Each tester indicated a
level of difficulty (easy, medium, and hard) for each
label and the total labeling time required.
Brodmann Area Labels for Cortical Activations
The accuracy of the TD system was tested using
Talairach coordinates taken from a previously re-
ported O-15 PET water study designed to activate M1
mouth motor areas [Fox et al., 1997, 1999]. The change
distribution analysis (CDA) method [Fox et al., 1988]
was used to determine significant activation foci in
each of 30 subjects [Fox et al., 1999]. All subjects were
healthy, right-handed, native English speakers be-
tween the ages of 21 and 49 (mean ⫽32; SD ⫽7).
Statistically significant activations (uncorrected p ⬍
0.001) for BA 4 or 6 were isolated for testing using
visual inspection (PTF). Fifty-two different x-y-z coor-
dinates were selected, 24 on the right and 28 on the left
side. Talairach coordinates for each activation site
were submitted to the TD system for labeling. If no
䉬Automated Talairach Atlas Labels 䉬
䉬125 䉬
GM label was found, the adaptive range-search utility
was invoked to provide a nearby GM label.
RESULTS
TD vs. Author Labels
Label matching between TD system (VOTL) labels
and author-designated labels for ten regions (670 ac-
tivation sites) is presented in Figure 3. The graph in
this figure indicates the probability of matching the
label designated by authors as a function of search
range. For the Cuneus and Thalamus (gyrus-level la-
bels) the initial match (search range ⫽0 mm) was
about 70%. Label matching rose to 80% for a search
range of ⫾2 mm and to over 90% for a ⫾3-mm range.
Label matching for Lentiform (cell-level) and Amyg-
dala (gyrus-level) labels followed a similar trend. The
primary motor area (BA 4) and the parahippocampal
gyrus had considerably lower initial label matching
TABLE II. Author-to-TD label mapping
Author label Corresponding TD label Hierarchical level
Cuneate Cuneus Gyrus
Cuneus
Cuneus-striate junction
Precuneate cortex Precuneus Gyrus
Precuneus
Calcarine (sulcus) Brodmann area 17 Cell
Primary visual (area or cortex)
Striate
Primary motor Brodmann area 4 Cell
Motor area
Motor cortex
Motor hand area Brodmann areas 4 or 6
Insula Insula (BA 13) Gyrus (cell)
Insular cortex
Insular gyrus
Insular region
Sylvian-insular
Globus pallidus Lentiform Gyrus
Lenticular nucleus
Lenticulate
Lentiform nucleus
Putamen
Pallidum
Caudate Caudate Gyrus
Caudate nucleus
Caudate, head
Caudatum
Caudatus
Anterior cingulate Anterior cingulate (BA 24,32) Gyrus (cell)
Anterior cingulate cortex
Primary auditory Brodmann area 41, 42 Cell
Heschl’s gyrus
Frontal eye fields Brodmann area 8 Cell
Wernicke’s area Brodmann area 22 Cell
Broca’s area Brodmann area 44, 45 Cell
Somatosensory Brodmann area 1, 2, 3 Cell
Thalamus Thalamus Gyrus
Amygdala Amygdala Gyrus
SMA Brodmann area 6 Cell
Parahippocampal gyrus Parahippocampal gyrus Gyrus
Uncus Brodmann area 34 Cell
䉬Lancaster et al. 䉬
䉬126 䉬
scores (32%). Label matching showed little improve-
ment for a ⫾1-mm range but approached 80% for a ⫾
3-mm range. For the four remaining regions (caudate,
insula, anterior cingulate, and primary visual – BA 17),
label matching scores fell between the values for the
worst (primary motor area) and the best (Thalamus)
label matches. Author-TD label matching was above
90% for half of the regions evaluated within a search
range of ⫾4 mm. For eight of ten of the regions, label
matching was greater than 90% using a search range
of ⫾5 mm. These results are for group-mean data, and
individual subject results are presented in the Brod-
mann Area Labels section below.
TD vs. Human for Atlas Labels
Difficulty ratings provided by the human labelers
(testers) showed that similar numbers of sites were of
low (47%) and medium (40%) difficulty. Tester’s im-
pression of difficulty was mirrored by their author
label matching scores (Table III, columns 1 & 2). Cat-
egorization of difficulty by level was not obviously
associated with any specific label, nor did it appear to
be consistent among testers. Interestingly, the TD sys-
tem (Raw – without range search) had its highest
matching score for sites in the tester’s medium-diffi-
culty category. Both the testers and the TD system had
lowest matching scores for sites in the high-difficulty
category. Testers’ author-matching scores for labels
derived from atlas axial sections were similar to TD
raw scores, which were also from axial sections (Table
III, columns1&3).
Testers’ scores improved somewhat when all atlas
sections (axial, sagittal, and coronal) were used to
determine a best label, with the score averaged across
all sites increasing from 73% to 80%. Author-matching
scores for the TD system improved dramatically when
the range-search utility was used to find nearest GM
labels, with scores rising by as much as 37% for the
high difficulty sites. The score averaged across all sites
rose from 71% to 90%. This trend is similar to that
found for the TD-to-author evaluation with 670 points
(Fig. 3). Percentage matching to author’s labels by the
TD system, using the adaptive GM range-search util-
ity, exceeded that by the tester group in all difficulty
rating categories (Table III, columns 4 vs. 2). The time
Figure 3.
Percent match between TD and ten author-designated labels as a
function of the search range in mm. Data was calculated from 670
points reported by numerous authors. The legend data is ordered
by initial percent match (search range ⫽0 mm). The numbers after
each label are organized as follows: (# of different first authors, #
of points for the label).
䉬Automated Talairach Atlas Labels 䉬
䉬127 䉬
to obtain the axial labels for the 106 sites was over
three hours for each tester, while the TD system, with
the GM range-search utility, completed this task in
less than one minute.
Median scores for label agreement between testers
and the TD system (without range search) dropped lev-
el-by-level (hemisphere: 98%, lobe: 83%, gyrus: 63%, and
cell: 55%). Label consistency among testers (all agreed)
dropped similarly (hemisphere: 96%, lobe: 72%, gyrus:
42%, and cell: 53%). Incomplete agreement at the hemi-
sphere level and diminished agreement at the gyrus
level were due to inconsistent nomenclature among us-
ers. Full label matching disparity (all testers disagreed)
was seen only at the gyrus (13%) and cell (10%) levels.
These data demonstrate a potentially high level of vari-
ability among users for manually obtaining labels for
activation sites from the Talairach atlas. There were no
reproducibility problems observed with the TD system.
A consensus (at least 2 agreed) among testers’ best
labels was seen for 94% of the sites. Consensus labels
are less sensitive to task-related user variability and
provide a good data set to evaluate the potential per-
formance for the manual-labeling task. The consensus
labels compared well with labels from the TD range
search (89% agreement) and provided a better overall
match with author designated labels (86%). However,
the TD system (with adaptive range-search utility)
had a slightly higher author label-matching score
(90%) than that of testers’ consensus labeling.
Brodmann Area Labels for Cortical Activations
The TD system returned matching Brodmann labels
(BA 4 or 6) for 40% of the coordinates without using
the range search. The label match increased rapidly to
56%, 79%, and 85% with search ranges of ⫾1, ⫾2, and
⫾3 mm. It rose slightly after that with a final label
match of 87% at ⫾5 mm. Seven errant labels were
seen, all postcentral gyrus (BA 2 and 3), and five of
these occurred at a search range of zero. Similar num-
bers of errant labels were seen for left (4) and right (3)
sides. Label matching as a function of search range
was better than that recorded for the 670 group-mean
coordinates taken from the BrainMap database (Fig. 3,
Primary Motor).
DISCUSSION
The large set of author labels selected for testing
provided reasonable anatomical labeling accuracy, a
diversity of sites distributed throughout the brain, and
numerous methodological challenges, all important
for evaluating automated labeling of functional acti-
vation sites. Authors were not assumed to all correctly
label sites. However, their labels were considered to
be more nearly correct than labels obtained from Ta-
lairach coordinates alone [Roland, et al., 1997]. This
assumption was based on additional information
available to authors (i.e., experimental design, over-
lays of functional maps on MR images, etc.).
Automated labeling by the TD system performed as
well as or better than our experts, or even the consen-
sus of 2 of 3 experts, when compared with author’s
labels. It should be noted, however, that neither
achieved a 100% match with author-designated labels.
The fact that the TD system can achieve an author
match of 90% for many labels indicates that it can be a
valuable asset to brain mappers seeking to standard-
ize labeling of activation foci. However, it is recom-
mended that the TD system be used with caution, and
that users treat labels as candidate labels to be re-
TABLE III. Manual vs. TD system matching to author labels
using talairach coordinates
Difficulty rating
Percentage match to authors’ labels
Testers TD system
Axial
section*
All
sections** Raw
Range
search
All sites (regardless of rating) 73% 80% 71% 90%
Low (47% of sites) 83% 88% 72% 92%
Medium (40% of sites) 72% 76% 77% 89%
High (13% of sites) 40% 64% 47% 84%
* Average scores for three experts manually labeling 106 coordinates using axial section images from
the 1988 Talairach Atlas.
** Average scores of expert’s best label using all (axial, coronal, and sagittal) section images.
䉬Lancaster et al. 䉬
䉬128 䉬
viewed for appropriateness. Questionable labels
should be resolved by inspecting overlays of func-
tional maps on high-resolution MR images. This is
especially important for sites in highly variable re-
gions such as the occipital-cerebellar boundary. For
example, if the TD system reports labels from the
cerebellum that are obviously in the occipital lobe
(based on overlays on MR images), it is useful to
expand the search range to produce an extended list of
label options. This approach is similar to how authors
use the Talairach atlas when such confusion arises, to
review nearby labels and select the most likely candi-
date.
Brain Labeling Accuracy
The degree of feature correspondence between in-
dividual brain images and an atlas varies throughout
the brain [Ono et al., 1990; Mazziotta et al., 1995; Zilles
et al., 1997; Roland et al., 1997]. For visual labeling, a
more informed choice can be made after a survey of
nearby features in both brain image and atlas. A label
is then selected, along with a qualitative estimate of
confidence level. However, there is no guarantee that
multiple users will choose the same label given iden-
tical information or that the label will be correct. Ac-
cordingly, the Talairach atlas labels provided by the
TD system should be treated as candidate labels. In-
trinsic coordinate-based TD-system labeling errors are
expected to be smallest for larger structures seen at the
hemisphere and lobe level and largest for smaller
structures and near boundaries (Fig. 3). However, us-
ing the adaptive GM range-search utility, the TD-
system matching with author’s labels exceeded 90%
for many labels (Fig. 3 and Table III).
Label errors for coordinate-based labeling methods
can come from incomplete anatomical matching by
global spatial normalization as well as methodological
differences [Strother et al., 1994]. While several non-
linear, high degree-of-freedom 3-D regional spatial
normalization algorithms have been developed [Col-
lins et al., 1994; Friston et al., 1995; Woods et al., 1998,
1998a, Kochunov et al., 1999], their accuracy remains
unproved, and they have not been implemented to
regionally match the standard 1988 Talairach atlas. A
design goal of the adaptive GM range-search utility
was to accommodate residual errors following spatial
normalization. The notable improvement in label ac-
curacy (56 to 87%) for M1 hand motor PET activation
studies, using linear affine spatial normalization [Lan-
caster, et al., 1995], shows that this strategy can work
reasonably well. While label errors were reduced,
more accurate spatial normalization (bringing coordi-
nates closer to Talairach labels), coupled with the
range-search utility, should produce even better re-
sults. Another design goal of the GM range-search
utility was to accommodate methodological difference
between different laboratories. The improved author
matching (71 to 90%) for all labels (51 authors) using
the adaptive GM range-search utility (Table III, col-
umn 4 vs. 3) shows that this strategy works reasonably
well.
Common Talairach Daemon System Uses
BrainMap姞
The BrainMap威database provides access to brain
findings from more than 200 research papers, 700
experiments, and 7000 locations in the human brain
[Fox et al., 1994; Fox and Lancaster, 1998]. Data for the
evaluation of the TD system was taken from this da-
tabase. Access to TD labels is provided in the Brain-
Map威Search & View client, where a user can quickly
review candidate Talairach labels for coordinates from
published brain function studies (Fig. 4). A future
enhancement of the BrainMap威database is automatic
entry of VOTL labels of each coordinate to support
searching using VOTL labels.
Anatomical Organization of Coordinate Data
Statistical analyses of brain function studies often
result in a long list of coordinates for activation foci.
For software that provides Talairach coordinates in an
x-y-z list format (SPM; MEDx, Sensor Systems, Ster-
ling, VA) it is helpful to retrieve VOTL database labels
using the TD system and to reorganize the list ana-
tomically. This sorting feature was used to categorize
activation sites for testing in this report.
Lesion-Deficit Analysis
The TD system provides a means to tabulate and
standardize labels for the volume occupied by a brain
lesion in Talairach space. This standardization sup-
ports correlation of standard anatomical nomenclature
with measures of neurological deficit (NIH/NINDS 2
RO1 NS 21889-16).
Education
A Java applet called the Talairach Daemon was
developed as an educational tool (http://ric.uthscsa.
edu/projects/talairachdaemon.html). It contains a
complete hand-detailed version of the 1988 Talairach
䉬Automated Talairach Atlas Labels 䉬
䉬129 䉬
atlas axial sections and overlays for all VOTL structures.
The interactive labeling features of the TD applet pro-
vide excellent examples for teaching brain anatomy.
CONCLUSIONS
The Talairach Daemon system provides rapid access
to Talairach atlas labels for functional activation foci
using Talairach coordinates. Many gyri and cell-level
labels in the volume occupancy Talairach labels
(VOTL) database directly map to common author la-
bels. VOTL labels matching author’s labels were
found within ⫾5 mm of author-designated coordi-
nates for most activation foci. The GM range-search
utility greatly enhanced the TD system, enabling it to
perform better than a group of three knowledgeable
users, when attempting to match published Talairach
coordinates with labels. Using the GM range-search
utility, Brodmann area labels for the M1 hand motor
area activations (BA 4 and 6) were correctly identified
in 87% of the subjects. The TD system, with its numer-
ous Internet distributed client applications, provides
an extensive informatics resource for the human func-
tional brain mapping community.
REFERENCES
Angevine JB, Mancall EL, Yakovlev PI. (1961): The human cerebel-
lum: An atlas of gross topography in serial sections. Little,
Brown, and Co., Boston.
Bohm C, Greitz T, Kingsley D, Berggren, Olsson L. (1983): Adjust-
able computerized stereotaxic brain atlas for transmission and
emission tomography. AJNR 4:731–733.
Bohm C, Greitz T, Seitz R, Eriksson L. (1991): Specification and
selection of regions of interest (TOIs) in a computerized brain
atlas. J Cereb Blood Flow and Metab 11:A64–A68.
Collins DL, Neelin P, Peters TM, Evans AC. (1994): Automatic 3D
intersubject registration of MR volumetric data in standardized
Talairach space. J Comput Assist Tomogr 18:192–205.
Collins DL, Holmes CJ, Peters TM, Evans AE. (1995): Automatic 3-D
model-based neuroanatomical segmentation. Hum Brain Mapp
3:209–223.
Courchesne E, Press GA, Murakami J, Berthoy D, Grafe M, Wiley CA,
Hesselink JR. (1989): The cerebellum in sagittal plane-Anat-
omic-MR correlation: The vermis. Am J of Neurol Res 10:659–665.
Figure 4.
The BrainMap™ Search & View 4.1 application provides direct access to the VOTL (Talairach
labels) database to obtain anatomical labels. A coordinate returned following a search on Brodmann
area 10 was used to automatically retrieve its hierarchical label.
䉬Lancaster et al. 䉬
䉬130 䉬
Drury HA and Van Essen DC. (1997): Functional specilizations in
human cerebral cortex analyzed using the visible man surface-
based atlas. Hum Brain Mapp 5:233–237.
Evans AC, Marrett S, Torrescorozo J, Ju S, Collins L. (1991): MRI-
PET correlation in three dimensions using a volume-of-interest
(VOI) atlas. J Cereb Blood Flow and Metab 11:A69–A78.
Fox PT, Perlmutter JS, Raichle ME. (1985): A stereotactic method of
anatomical localization for positron emission tomography.
J Comput Assist Tomogr 9:141–153.
Fox PT, Mintum MA, Reiman EM, Raichle ME. (1988): Enhanced
detection of focal brain responses using intersubject averaging
and change distribution analysis of subtracted PET images.
J Cereb Blood Flow Metab 8:642–653.
Fox PT. (1995): Spatial Normalization: Origins, objectives, applica-
tions and alternatives. Hum Brain Mapp 3:161–164.
Fox PT, Mikiten S, Davis G, Lancaster JL. (1994): BrainMap: A
database of human functional brain mapping. In: Thatcher RW,
Zeffiro T, Huerta M (eds): Advances in Functional Neuroimag-
ing: Technical Foundations. Orlando: Academic Press, 95–106.
Fox PT, Ingham TJ, Ingham JC, Hirsch TB, Downs JH, Martin C,
Jerabek P, Glass T, Lancaster JL. (1996): A PET study of the
neural systes of stuttering. Nature 382:158–162.
Fox PT, Lancaster JL, Parsons LM, Xiong J, Zamarripa E. (1997):
Functional volumes modeling: Theory and preliminary assess-
ment. Hum Brain Mapp 5(4):306–311.
Fox PT, Lancaster JL. (1998): BrainMap威: Electronic integration of
mind and brain. In: Positron Emisssion Tomography: A critical
assessment of research trends, Gulyas B and Muller-Gartner
HW, eds., Kluwer Academic Publishers.
Fox PT, Huang A, Parsons LM, Xiong J, Zamarripa F, Rainey L,
Lancaster JL. (2000): Location-probabiolity profiles for the mouth
region of human primary motor cortex: Metanalysis and valida-
tion, NeuroImage: in process.
Freitas CS, Summerlin JL, Lancaster JL, Fox PT. (1996): Soc Neurosci
Ann Mtg V–101.
Friston KJ, Passingham RE, Nutt JG, Heather JD, Sawle GV, Frack-
owiak RSJ. (1989): Localization of PET images: Direct fitting of
the intercommissural (AC-PC) line. J Cereb Blood Flow Metab
9:690–695.
Friston KJ, Frith CD, Liddle PF, Frakowiak RSJ. (1991): Comparing
functional (PET) images: The assessment of significant change.
J Cereb Blood Flow Metab 11:690–699.
Friston KJ, Ashburner J, Frith CD, Poline JB, Heather JD, Frakowiak
RSJ. (1995): Spatial registration and normalization of images.
Hum Brain Mapp 3:165–189.
Garey LJ. (1994): Brodmann’s Localisation in the Cerebral Cortex.
London: Smith-Gordon, p 109–129.
Kochunov PV, Lancaster JL, Fox PT. (2000): Accurate high-speed
spatial normalization using an octree method. NeuroImage: in
press.
Lancaster JL, Glass TG, Lankipalli BR, Downs H, Mayberg H, Fox
PT. (1995): A modality-independent approach to spatial normal-
ization. Hum Brain Mapp 3:209–223.
Lancaster JL, Rainey LH, Summerlin JL, Freitas CS, Fox PT, Evans
AE, Toga AW, Mazziotta JC. (1997): Automated labeling of the
human brain: A preliminary report on the development and
evaluation of a forward-transform method. Hum Brain Mapp
5:238–242.
Lancaster JL, Fox PT, Downs H, Nickerson D, Hander T, El Mallah
M, Zamarripa F. (1999): Global spatial normalization of the
human brain using convex hulls, J of Nucl Med 40:942–955.
Mazziotta JC, Toga AW, Evans A, Lancaster JL, Fox PT. (1995): A
probabilistic atlas of the human brain: Theory and rational for its
development. Neuroimage 2:89–101.
Ono M, Kubik S, Abernathey CD. (1990): Atlas of the Cerebral Sulci.
New York: Theime Medical Publishers.
Pelizzari CA, Chen GTY, Spelbring DR, Weichselbaum RR, Chen
CT. (1989): Accurate three-dimensional registration of CT, PET,
and/or MRI images of the brain. J Comput Assist Tomogr 13:
20–26.
Press GA, Murakami J, Courchesne E, Berthoy D, Grafe M, Wiley
CA, Hesselink JR. (1989): The cerebellum in sagittal plane–Ana-
tomic-MR correlation: The cerebellar hemispheres. Am J of Neu-
rol Res 10:667–676.
Press GA, Murakami J, Courchesne E, Grafe M, Hesslink JR. (1990):
The cerebellum: 3. Anatomic-MR correlation in the coronal
plane. Am J Neurol Res 11:41–50.
Roland PE, Graufelds CJ, Wahlin J, Ingelman L, Andersson M,
Ledberg A, Pederson J, Akerman S, Dabringhaus A, Zilles K.
(1994): Human brain atlas: For high-resolution functional and
anatomical mapping. Hum Brain Mapp 1:173–184.
Roland PE, Geyer S, Amunts K, Schormann T, Schleicher A, Malik-
ovic A, Zilles K. (1997): Cytoarchitectural maps of the human
brain in standard anatomical space. Human Brain Mapping
5:222–227.
Schmahmann JD, Doyon J, McDonald D, Holmes C, Lavoie K,
Hurwitz AS, Kabani N, Toga A, Evans A, Petrides M. (1999):
Three-dimensional MRI atlas of the human cerebellum in pro-
portional stereotaxic space. NeuroImage 10:233–260.
Sobel DF, Gallen CC, Schwartz BJ, et al. (1993): Locating the central
sulcus: comparison of MR anatomical and magnetoencephalo-
graphic functional methods. AJNR 14:915–925.
Steinmetz H, Furst G, Freund HJ. (1989): Cerebral Cortical Local-
ization: Application and validation of the proportional grid sys-
tem in MR imaging. J Comput Assist Tomogr 13:10–19.
Stevens WR. (1990): UNIX Network Programming. Englewood
Cliffs, NJ: Prentice Hall.
Strother SC, Anderson JR, Xu XL, Liow JS, Bonar DC, Rottenberg
DA. (1994): Quantitative comparisons of image registration tech-
niques based on high-resolution MRI of the brain. J Comput Asst
Tomogr 18:954–962.
Talairach J, Tournoux P. (1988): Co-Planar Stereotaxic Atlas of the
Human Brain. New York: Thieme Medical Publishers.
Watson JDG, Meyers R, Frackowiak RSJ, Hajnal JV, Woods RP,
Mazziotta JC, Shipp S, Zeki S. (1993): Area V5 of the human
brain: Evidence from a combined study using positron emission
tomography and magnetic resonance imaging. Cereb Cortex
3:79–94.
Woods RP, Mazziotta JC, Cherry SR. (1993): MRI–PET registration
with automated algorithm. J Comput Assist Tomogr 17:536–546.
Woods RP, Grafton ST, Holmes CJ, Cherry SR, Mazziotta JC. (1998):
Automated image registration: I. General methods and intrasu-
bject, intramodality validation. J Comput Asst Tomogr 22:139–
152.
Woods RP, Grafton ST, Watson JDG, Sicotte NL, Mazziotta JC.
(1998a): Automated image registration: II. Intersubject validation
of linear and nonlinear models. J Comput Asst Tomogr 22:153–
165.
Zilles K, Schleicher A, Langemann C, Amunts K, Morosan P, Pal-
omero-Gallagher N, Schormann T, Mohlberg H, Burgel U, Stein-
metz H, Schlaug G, Roland PE. (1997): Quantitative analysis of
sulci in the human cerebral cortex: Development, regional het-
erogeneity, gender difference, asymmetry, intersubject variabil-
ity and cortical architecture. Hum Brain Mapp 5:218–221.
䉬Automated Talairach Atlas Labels 䉬
䉬131 䉬
