Download Disproportion of cerebral surface areas and volumes in

braini0202
Brain (1997), 120, 271–281
Disproportion of cerebral surface areas and
volumes in cerebral dysgenesis
MRI-based evidence for connectional abnormalities
S. M. Sisodiya and S. L. Free
Epilepsy Research Group, Institute of Neurology, National
Hospital for Neurology and Neurosurgery, and National
Society for Epilepsy, London, UK
Correspondence to: Dr S. M. Sisodiya, Department of
Clinical Neurology, Radcliffe Infirmary, Woodstock Road,
Oxford OX2 6HE, UK
Summary
In the normal adult human brain, there are quantitative
relationships between a surface area measure of the grey
matter and the volume of hemispheric grey matter, the volume
of the hemispheric subcortical matter and the cross-sectional
area of the corpus callosum as revealed by analysis of high
resolution MRI data. These relationships reflect structural
order in, and biological features of, normal human cerebral
hemispheres. Cerebral dysgenesis (CD) is associated with
disruption of the normal organization of the hemispheres to
a greater or lesser extent and is often manifest as refractory
epilepsy. We have examined structural proportions and their
disruption in the brains of patients with epilepsy and CD.
We found that structural measures were abnormal in 60% of
patients with CD, with abnormalities in 64% of hemispheres
that, on visual inspection alone, appeared completely normal.
We showed that the disruptions found are compatible with
expected histopathology in cases where histopathology may
be predictable, and that extensive abnormalities may be due
to abnormal patterns of connections within the hemispheres.
In some cases, it may be possible to predict histopathology
on the basis of quantitative analyses of high resolution
MRI data, when such prediction is not possible on visual
inspection alone.
Keywords: brain measurement; cerebral dysgenesis; surface area; connectivity
Abbreviations: CCA 5 corpus callosal cross-sectional area; CD 5 cerebral dysgenesis; EA 5 extra subcortical matter
surface area; ECC 5 EA:CCA ratio; EGM 5 EA:GMV ratio; ESM 5 EA:SMV ratio; GM 5 grey matter; GMV 5 GM
volume; SM 5 subcortical matter; SMA 5 SM surface area; SMV 5 SM volume
Introduction
The development and organization of the vast number of
neurons (Braendgaard et al., 1990) and synapses (Cherniak,
1990) in the human brain is governed by a comparatively
minute amount of genetic information, suggesting that general
rules must exist for the establishment of cerebral structure,
almost certainly with environmental influence producing local
modulation (Purves et al., 1992). General rules are reflected
in quantitative measures of normal cerebral structure: the
mean number of neurons per unit volume and per unit surface
area is relatively constant (Haug, 1987; Braendgaard et al.,
1990); the variation in neocortical thickness is limited, both
within and between species (Welker, 1990; Rakic, 1995); the
proportions of neurons allocated to specific functional groups,
e.g. projectional or nonprojectional, are relatively constant
(Rockel et al., 1980; Winfield et al., 1980). The existence
© Oxford University Press 1997
of general rules is highlighted by the finding that single
gene defects can result in dramatic, extensive and
histologically homogeneous structural disorder in human
development (Reiner et al., 1993). Because of these rules,
useful macroscopic parameters of cerebral structure exist
(Hofman 1985, 1991; Haug, 1987; Filipek et al., 1994;
Sisodiya et al., 1995, 1996). Also, reliable determination
of cortical surface area or volume might provide a means of
estimating total (Mountcastle, 1978), regional and functionspecific neuronal numbers in normal brains. A knowledge of
the number of neurons, and the proportion allocated to a
certain laminar or functional group, would be of importance
in the further investigation of normal cortical organization
and its disruption. The availability and usefulness of such
information would be enhanced if it could be gathered in vivo.
272
S. M. Sisodiya and S. L. Free
Fig. 1 Schematic drawing of the derivation of EA. The
corrugated circle represents the real SM surface in cross-section;
the plain circle encompasses the same area but is unfolded; the
difference in the perimeters of the two regions is the extra length
(area in three dimensions) generated by the folding of the surface
of the SM. This extra area is termed EA, and is defined by:
EA 5 SMA – 4kπ(3SMV/4π)2/3 where SMA is the measured SM
surface area and SMV the measured SM volume, and k is a
constant converting volume to voxel number (for details, see
Sisodiya et al., 1996).
MRI allows quantitative data on human cerebral structure
to be acquired in vivo: it permits the study of macroscopic
cerebral parameters, such as volumes and surface areas.
Using MRI, we have examined cerebral surface areas and
volumes in normal adult human brains and shown that a
number of quantitative relationships do indeed exist between
various parameters, reflecting general organizational rules
(Sisodiya et al., 1996). Thus, in the normal brain, total
hemispheric grey matter volume and (inner) grey matter
surface area correlate linearly, implying that variation in the
range of cortical thickness across normal subjects is limited,
and reflecting the histological finding of a relatively constant
mean neuronal surface density. In addition, the corpus
callosum cross-sectional area, a measure of the number of
fibres passing through it (Aboitiz et al., 1992), also correlates
with a surface area measure of the grey matter, supporting
the suggestion that a fixed proportion of neurons belong to
a given functional class.
Whilst all previously published data on cortical areas
concern the free surfaces (grey–CSF interfaces; e.g.
Schlaug et al., 1995), usually of post-mortem brains (Blinkov
and Glezer, 1968; Henery and Mayhew, 1989), the inner
surface (grey–white matter boundary) can be defined with
greater precision globally on MRI. This inner surface is also
the surface of the subcortical matter (SM, defined as the
subcortical white matter and subcortical nuclei excluding
the caudate) and is created by the spatial dominance of
projectional (extracortical) fibres; these may form intergyral
arcuate (or ‘U’) fibres or fibres travelling further afield. Such
projectional fibres are most plentiful in the walls and at the
crowns of gyri and least numerous in the depths of the sulci
(Welker, 1990). The variable EA, defined as a measure of
the extra surface area generated by the folding of the SM
surface (see Fig. 1), may also be taken as an estimate of the
surface area of the white matter cores of gyri, that is the area
of SM directly overlain by cortical grey matter in gyral
crowns and walls rather than in sulcal depths (Sisodiya et al.,
1996). Folding in normal brains is necessitated by the number
of projectional axons and the amount of neuropil (intracortical
synapses, dendrites and axons) covering a limited volume of
white matter (Prothero and Sundsten, 1984; Ruppin et al.,
1993). EA is thus a better average measure of projectional
(afferent and efferent) axon numbers and neuropil quantity
than is the total SM area (SMA); EA may be considered to
be more biologically relevant.
Abnormal organization of the cortex defines cerebral
cortical dysgenesis (CD) and may occur at any level, from
the synaptic to the macroscopic (Becker, 1991; Sarnat, 1992;
Raymond et al., 1995). The nomenclature of dysgenesis is
not yet universally agreed; we use the system proposed by
Raymond et al. (1995), based on the appearance of dysgenesis
as seen on routine inspection of high-resolution MRI. We
propose that in dysgenesis quantitative measures and normal
proportions of cerebral structure might be altered and that
examination of quantitative parameters in brains with dysgenesis might shed light, in vivo, on the nature, distribution and extent of the dysgenetic process.
Based on the outcome from surgery, it has been proposed
that in human CD, structural abnormalities may extend
beyond the visualized boundaries of a lesion (Taylor et al.,
1971; Awad et al., 1991; Andermann, 1994). It has
subsequently been shown that there is indeed both local
lesional extension (Palmini et al., 1995) and distant alteration
in cerebral structure (Sisodiya et al., 1995) in cerebrum
appearing normal on inspection. It has been suggested that
abnormalities spatially removed from a lesion may be due
to abnormal connectivity associated with apparently focal
dysgenesis, given evidence from animal experiments
(Goldman-Rakic, 1980; Rakic, 1988; Loopuijt et al., 1995)
and the high degree of connectivity in the normal human
brain (Cherniak, 1990). Abnormal connectivity has been
considered in dysgenesis directly using MRI through the
analysis of developmental anomalies of the corpus callosum,
known to be associated with the presence of abnormal
cerebral structure (Friede, 1975; Barkovich and Norman,
1988) and function (Hynd et al., 1995).
In this report, we develop previously reported parameters
of cerebral structure further and examine their use in
investigating structural disturbances in vivo in patients with
CD. Specifically, we attempt (i) to determine if there is
disruption of surface structural measures in CD; (ii) to
correlate any changes with known or presumed
histopathology; (iii) to look for further evidence that CD
may extend beyond visualized abnormalities and (iv) to
Cerebral surface areas and volumes in cerebral dysgenesis
273
Table 1 Patient clinical and neuroimaging data
Patient no./sex/
age at scan (years)/
age at seizure onset
MRI diagnosis
Seizure types
1/F/36/24 years
2/M/17/8 years
3/F/35/21 years
4/M/18/1 month
5/F/36/6 months
6/M/23/12 years
7/F/15/10 years
8/F/18/5 years
9/F/34/23 years
10/F/29/2 years
11/F/27/17 years
12/F/18/13 months
R frontoparietal polymicrogyria
Extensive R occipital dysgenesis (Fig. 2)
Extensive R CH gyral abnormality, polymicrogyria
R medial occipital and R lateral parietal clefts (partial thickness)
L frontal medial and lateral gyral abnormality, L frontal SEH
Hypothalamic hamartoma
Bilateral posterior macrogyria (Fig. 4)
L occipitotemporal macrogyria
Bilateral full thickness clefts
Bilateral SEH and R posterior localised macrogyria
Bilateral small parietal subcortical heterotopic nodules
Agenesis of corpus callosum and R temporal, anterior
parietal CD, full thickness cleft, bilateral SEH (Fig. 3)
Bilateral SEH (occipital and temporal horns; diffuse)
R parietal macrogyria
L occipital, posterior parieto-temporal full thickness cleft
Bilateral subependymal heterotopia, parietal full thickness clefts
Agenesis of corpus callosum
Bilateral subcortical band heterotopia, frontal
Bilateral subcortical band heterotopia.
R posterior parieto-temporal full thickness cleft
L hemisensory, SGS 1–2/month
CPS, tonic extension L leg; SGS once only
SPS, L hemisensory; no CPS for 10 years; no SGS
L arm aura then SGS, 3/month
Dropping head, raising arms, falling down; no SGS
Gelastic seizures; drop attacks
SPS; drop attacks; rare SGS
CPS; visual aura; no SGS
Focal motor, L arm; no SGS
CPS with buzzing; no SGS
L arm focal motor seizures
CPS with jerking of all limbs; no SGS
13/F/31/17 years
14/F/28/12 years
15/F/15/10 months
16/M/26/18 months
17/F/23/15 years
18/M/29/12 years
19/M/33/18 months
20/M/27/26 years
CPS; three SGS only
Nightly SGS
SPS (aura); CPS. R focal motor; no SGS
CPS; myoclonic jerks; four SGS only
CPS only with hypothermia
R focal motor and two SGS
Absences; 2–3 GS/year
CPS; no SGS
CH 5 cerebral hemisphere; CPS 5 complex partial seizure; SPS 5 simple partial seizure; GS 5 generalized seizures; SGS 5 secondary
generalized seizures; SEH 5 subependymal heterotopia.
examine the possibility that such extension is associated with
abnormalities of interneuronal connections.
Methods
Subjects
Thirty-three neurologically normal volunteers, 11 female and
22 male, were scanned using MRI. Their age range was
19–52 years. All gave informed consent for the scanning
procedure; scanning was approved by the ethics committee
of the National Hospital for Neurology and Neurosurgery.
Twenty patients, with definite CD identified on routine
visual inspection of MRI scans by an experienced neuroradiologist, were selected because the grey–white interface
on postprocessing was of a certain degree of complexity and
was neither thicker than normal nor blurred (see below in
‘Area measurements’). All patients attended clinics at the
National Hospital or the Chalfont Centre for Epilepsy. Patient
details are given in Table 1.
There was no significant difference between the ages of
the controls and the ages of the patients (Mann–Whitney,
two-tailed, P . 0.2).
Imaging, segmentation and block analysis
The imaging and segmentation protocols have been described
in detail previously (Sisodiya et al., 1995). Briefly, all subjects
were scanned using sagittal T1-weighted, coronal T1-weighted
volumetric, axial T2-weighted and axial proton density
sequences. All images had been reviewed by an experienced
neuroradiologist who specifically examined for subtle
dysgenesis. In no case was there any evidence of either
dysgenesis other than that reported in Table 1, or of gliosis
on T2-weighted imaging. On each coronal image for each
subject, the cortical grey matter (GM) and SM were isolated
using a dedicated image processing workstation (Allegro,
ISG Technologies, Toronto, Canada) and these regions-ofinterest reconstructed into three-dimensional images. The
volumes of the hemispheric GM and SM (GMV and SMV)
reconstructions were automatically calculated. The symmetry
and regional distributions of GMV and SMV in these
reconstructions were also assessed (see Sisodiya et al., 1995).
In essence, the method quantifies the amount of cortical grey
matter or subcortical matter within prescribed proportions
(blocks each extending one-tenth of the anterior–posterior
axis) of each individual hemisphere. Eighty measures are
thus generated (10 GMV and 10 SMV per hemisphere, 20
individual block ratios of GMV:SMV, 10 ratios each of GMV
and SMV in a block in the left hemisphere compared with
the volume in the homologous block in the right hemisphere).
Normal ranges, lying ,3 SDs from each of these 80 means
from the 33 control subjects, can identify and locate
abnormalities of the distribution of a particular tissue within
the hemispheres. None of the 33 control subjects had more
than one abnormal value out of the total of eighty. On this
basis, the presence of more than one abnormal variable (of
80 in total) in a particular subject was considered abnormal;
this finding has previously been given a biological
interpretation (Sisodiya et al.,1995).
274
S. M. Sisodiya and S. L. Free
Area measurements
The surface area of the SM reconstructions (SMA) and the
cross-sectional area of the corpus callosum (CCA) were
obtained as summarized below and described in full detail
elsewhere (Sisodiya et al., 1996).
SM surface area
The surface area of a reconstruction was defined as the
number of voxels in its surface contour. For the SM, this
contour is effectively the grey–white interface, except for the
small region where the SM surface is formed by the corpus
callosum, where the contour is not overlain by GM.
This surrogate measure is only valid under certain
circumstances. The method is model-based and therefore not
unbiased (Mayhew, 1992): orientation of the brain in the
scanner could alter the final voxel count obtained. In practice,
however, because of the complexity of the SM surface, which
is highly convoluted in three dimensions, the actual measured
effect of deliberately altering intrascanner orientation in a
limited number of brains is small with respect to biological
variation and the effect of segmentation. This finding of
orientation-independence can be applied to other brains
without formal testing, provided that the SM surface
complexity is similar. The surface complexity was measured
by estimation of its fractal dimension (Mandelbrot, 1982),
and for controls falls within a very narrow range (right SM
surface 2.27–2.31; left SM surface 2.28–2.31; Free et al.,
1996). The selection of subjects for this study was limited
to those in whom the fractal dimension lay within the normal
range. Voxel counting also estimates surface area directly
only for those scans in which the grey–white interface is not
abnormally thickened. So all the individuals studied were
also selected by ensuring that no extensive grey–white
interface blurring was seen on visual inspection of their
MRIs by an experienced neuroradiologist. Thus the surface
voxel count is, for these brains, a valid estimate of the area
SMA, effectively independent of brain orientation and not
complicated by blurring of the grey–white interface.
Corpus callosal cross-sectional area
The coronally acquired volumetric data set was reformatted
in three mutually perpendicular planes using the proprietary
image-analysis software, and on these data an optimal
interhemispheric plane of section was chosen to produce an
approximately sagittal image through the smallest crosssectional area of the corpus callosum, from which CCA was
measured by segmentation (Sisodiya et al., 1996).
Surface area derivatives
From the measurements made above, the following were
calculated for each hemisphere (see Sisodiya et al., 1996,
for details). First, the extra SM surface area, EA, produced
by the folds in the SM, as detailed in Fig. 1.
It has been shown that EA correlates highly, in controls,
with CCA and with right and left hemispheric GMV
separately (Pearson correlation coefficients, P , 0.005).
EA also correlates highly in both hemispheres with SMV
(P,0.0005). On this basis, the following three ratios were
calculated for right and left hemispheres separately for
all subjects:
(i) for the relationship between GMV (in cm3) and EA
(in voxels),
EGM 5 EA/GMV voxels/cm3;
(ii) for the relationship between SMV (in cm3) and EA,
ESM 5 EA/SMV voxels/cm3;
(iii) for the relationship between EA and CCA (in cm2),
ECC 5 EA/CCA voxels/cm2.
For controls, there was no significant sex difference for these
ratios or CCA, so that all control subjects were grouped to
form sex-independent normal ranges for these variables.
Individual hemispheric GMV and SMV normal ranges were
segregated for sex. Abnormal values of the ratios are defined
as those lying .3 SDs from the mean for control subjects.
Abnormal values are categorized as either high or low,
allowing interpretation of abnormalities; the actual values
are not given as this does not currently add to their
interpretation (see below).
Reliability
The inter- and intra-rater reliabilities of the segmentation
process and the measures of hemispheric constituent volumes
and surface areas have all been published previously and
shown to be high (Sisodiya et al., 1995, 1996).
Statistics
Analysis was performed using SPSS for Windows, Version
6.1 (SPSS Inc, Chicago, Ill., USA). The statistical test used
is indicated in each case. Means6SDs were calculated for
each of the derived ratios; the normal range was defined as
being within 63 SDs of the mean. Significance was taken
at P , 0.01 level.
Results
Control subjects
The mean6SD values for right and left hemispheric volumes
GMV and SMV, EA and surface area derivative ratios (EGM,
ESM and ECC) are given in Table 2. No control subjects
had values for any of these parameters outside the normal
range for that parameter. The mean6SD of the CCA across
the control group was 696673 mm2. No control subject had
more than one abnormal block variable in the total 80
Cerebral surface areas and volumes in cerebral dysgenesis
275
Table 2 Results for normal subjects
Variable
Right hemisphere
(mean6SD)
Grey matter volume (GMV) in cm3
Male
263628
Female
225621
Subcortical matter volume (SMV) in cm3
Male
261625
Female
229636
Extra subcortical-matter surface-area and derived ratios
EA (voxels)
299 000631 000
EGM (voxels/cm3)
12006140
1200678.3
ESM (voxels/cm3)
ECC (voxels/cm2)
435644.7
Left hemisphere
(mean6SD)
272628
235624
245621
222631
291 000628 200
11206104
1230681.2
424644.8
EA 5 extra subcortical matter surface area; ECC 5 EA:CCA ratio; EGM 5 EA:GMV ratio; ESM 5
EA:SMV ratio; GMV 5 grey matter volume; SMV 5 subcortical matter volume
Table 3 Abnormal values in patients
Patient
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Hemispheric volume or corpus
callosal area (CCA) abnormality
SMV-R, CCA low
SMV-R, GMV-R high
CCA low
SMV-L low
SMV-R, SMV-L, GMV-R low
CCA absent
SMV-L, CCA low
CCA low
Surface area derivatives†
Abnormal
blocks (n)
Abnormal blocks
in normal CH‡ (n)
Normal
ECC-R, ECC-L, ESM-R high
ECC-L high
Normal
ECC-R, ECC-L high
Normal
EGM-R, EGM-L, ECC-R, ECC-L low
Normal
ESM-R, ESM-L high
ECC-R, ECC-L high
Normal
ECC-R, ECC-L, ESM-R, ESM-L high
ECC-R high
ECC-R, ECC-L high
ECC-R, ECC-L, ESM-R, ESM-L high
Normal
ECC-R, ECC-L, ESM-L high
Normal
Normal
ECC-R, ESM-R, ESM-L high
2
4
17
4
2
8
31
10
4
0
0
16
0
9
*
6
16
3
3
4
0
2
4
1
0
8
NA
0
NA
NA
NA
NA
NA
3
*
NA
16
NA
NA
1
*Block analysis precluded by stereological considerations (see text). †-R or -L 5 right or left; low or high 5 low or high compared with
normal ranges. ‡For patients with midline or unilateral CD only (CH 5 cerebral hemisphere); NA 5 not appropriate as CD is bilateral.
variables per subject; i.e. no control subject had an abnormal
regional distribution of GM or SM, as defined above.
Patients
Values calculated for the patients were compared with the
normal ranges: abnormal results (except for EA) are given
in Table 3. EA was abnormal in only one patient (Patient 7;
EA low bilaterally). Abnormal surface area measures were
found in 12 out of 20 (60%) of patients. The ESM ratio was
high in 10 hemispheres (six patients), associated with a low
SMV in three hemispheres (two patients). The ECC ratio
was high in 10 out of 20 patients (17 out of 40 hemispheres)
and low in one patient (both hemispheres); in six out of 11
of the patients with an abnormal ECC ratio, the area of the
callosum itself was within the normal range. The EGM ratio
was altered in only one patient (bilaterally low in Patient 7).
Sixteen of twenty patients had a significant number of
abnormalities of the measures of the regional distribution
of volume (block abnormalities); block analysis was not
possible in one other patient because excessive head tilt
confounded measurements. Of 11 patients with visually
discernable unilateral or midline CD, seven had abnormality
of surface area derivatives in the contralateral hemisphere,
276
S. M. Sisodiya and S. L. Free
whilst seven also had at least one abnormal block in the
contralateral hemisphere (block measurements were not
possible in Patient 15).
Discussion
Interpretation of surface area derivatives
The surface area derivatives quantify mean structural
properties of entire hemispheres. The parameters assess the
relationship between the extra SM area, EA, and GMV, SMV
and CCA. These ratios are independent of brain size and
may identify changes in structural proportions even though the
underlying variables (surface areas and volumes) fall within
the normal range. For control subjects, values all fall within
3 SDs of the mean. In patients, small areas of disproportion,
due to structural abnormality, may be averaged out by larger
areas of normal proportions (structural normality) within the
same hemisphere. If abnormal values for these parameters
are found, then the pathological process present in the
hemisphere must be quantitatively dominant.
The ratio ECC is perhaps most simply interpreted. In
control subjects, EA is a measure of the number of
projectional axons and the amount of neuropil. CCA is a
measure of the number of interhemispheric fibres (Tomasch,
1954; Aboitiz et al., 1992). ECC is a function of these
quantities. An increase in ECC implies that either (i) the
proportion of (noninterhemispheric) projectional axons is
increased and/or the proportion of interhemispheric fibres
is reduced (or that they are thinner on average) or (ii) the
amount of neuropil is increased (or some combination of
these findings). In any case, there must be an alteration in
the normal pattern of interneuronal connectivity (as mediated
by axons, synapses and dendrites), given that the proportion
of neurons belonging to a given functional class is thought
to be fixed (Winfield et al., 1980) and that the mean amount
of neuropil/neuron in normal brains is also relatively constant
(see above). The specific nature of average changes in cortical
structure cannot be determined solely from a change in ECC,
but a clearer impression might be formed if changes in EGM,
ESM and hemispheric volumes were also taken into account.
A reduction in ECC has the converse implications with
respect to changes in cerebral structure, and also implies
extensive alteration in averaged interneuronal connectivity.
Geometrically, GMV is a function of grey matter surface
area and thickness. Changes in grey matter thickness are
very limited in normal brains (Rakic, 1995). It has been
suggested that this is because of biophysical constraints
acting on the apical dendrites of pyramidal cells (Prothero
and Sundsten, 1984). An equally valid reason may be that if
an abnormal number of neurons are stacked vertically and
yet maintain normal dendritic expansions, then the tangential
area available for the passage of the increased number of
afferent and efferent axons required to keep the increased
number of neurons connected is reduced, limiting the number
of neurons that can be supported in such a stack. This is
essentially a smaller scale and more superficial (pial)
application of the gyral window concept of Prothero and
Sundsten (1984). Therefore in abnormally thick cortex, if the
number of neurons is increased, then the proportion of axons
that cross the grey–white interface must be reduced; if the
number of neurons is unchanged or reduced, then in the
absence of gross gliosis, the mean amount of neuropil per
neuron must increase.
EGM is a function of EA and GMV. GMV depends on
neuronal numbers and, predominantly, on neuropil volume
(Haug, 1956), gross gliosis being excluded in this study by
the negative T2 findings. In both focal cortical dysplasia and
polymicrogyria, increased astrocyte numbers may be seen
histologically (Taylor et al., 1971; Williams et al., 1976),
but, to our knowledge, gross gliosis occupying a large volume
of the cortex has not been reported in the types of dysgenesis
of the varieties under discussion: gliosis per se is therefore
unlikely to explain our findings. Any change in neuropil is
equivalent to a change in connectivity; changes in neuronal
numbers, given the finite volume of GM and that all neurons
have neuropil, must be accompanied by neuropil volume
change and therefore alteration in connectivity. A reduction
in EGM therefore implies (some combination of) either (i) a
reduction in the amount of neuropil or in the proportion of
projectional axons or (ii) a disproportionate increase in the
GMV, that again necessitates a fall in projectional axon
number. In pathologically thickened cortex (e.g. as a result
of pachygyria), EA may be a better measure of projectional
axonal numbers than either SMA or GMV.
The ESM ratio is a function of EA and SMV. The biological
correlate of SMV is the product of the number of all
extracortical projectional axons of various diameters and the
mean volume of axons of a given diameter. If ESM is
increased, then either (i) EA must be disproportionately large
(i.e. there are more projectional axons, demanding a reduced
mean volume per axon in the SM, or more total neuropil—
but this cannot be associated with increased extracortical
projection as then SMV would rise—so that local connectivity
must rise disproportionately, without axons entering and
contributing to the SM) and/or (ii) there is simply a reduced
volume per axon in the SM (with or without a reduction in
their number). In either case (or with some combination of
the two possibilities), the reduced mean volume per
projectional axon implies altered connectivity, as axons
must be shorter and/or thinner than normal. Thinner axons
cannot support as great a terminal arbor as thicker axons
(Mitchison, 1991), therefore any alteration in axon diameter
confounding ESM (or ECC) measures is likely a priori to
be accompanied by an alteration in connectivity. Shorter axons
would suggest a tendency to increased local connectivity at
the expense of more distant connectivity. In this group of
patients, when the ESM ratio was abnormal, it was always
increased. In no case was an increase in the ESM ratio
associated with an abnormally high value of EA itself.
Cerebral surface areas and volumes in cerebral dysgenesis
277
Surface area derivatives in patients
Polymicrogyria
The interpretation of surface area derivative abnormalities is
exemplified by consideration of changes in patients with
known pathological abnormalities. Patients with full thickness
clefts of the neopallium commonly have polymicrogyria in the
surrounding and distant neocortex, possibly with additional
subependymal heterotopia (Dekaban, 1965) and reduction in
the subcortical white matter and cross-sectional area of the
corpus callosum (Levine et al., 1974). In cases of layered
polymicrogyria, as seen in association with pallial clefts,
very few axons or dendrites traverse the astroglial-rich layer
seen in the deeper layers of the neocortex (Williams et al.,
1976). Extracortical projection from neurons pial (superficial)
to this cell-sparse layer is reduced, leading to reduced
subcortical white matter volume and callosal area, as reported
by Levine et al. (1974). Histological analysis (Richman et al.,
1974), including Golgi (Williams et al., 1976), has shown,
however, that the number of neurons in the remaining layers,
and the amount of neuropil, need not be attenuated. In
patients with full thickness clefts revealed by MRI in this
study, similar structural abnormalities might be expected.
In two patients with full thickness clefts, there is indeed
reduction of SMV (Patients 9 and 15), with a reduction
in CCA in one (Patient 15). Additional abnormalities of
connections are revealed by the surface area derivatives.
One might predict that in these patients, interhemispheric
connections would be fewer than expected for a given brain
size, and that extracortical projection would be disproportionately reduced, as axons and dendrites would not
cross the astroglial layer in the polymicrogyric cortex. However, the remaining neuropil is normal (see above), so that
this contribution to EA may be normal. As a result of these
changes, from the interpretation of surface area derivatives
given above, the ECC and ESM ratios should both be
abnormally high. In four out of six patients with cleft(s),
there are precisely such abnormalities of the surface area
derivatives. The ESM ratio is increased bilaterally in a patient
with bilateral full thickness clefts (Patient 9), and also
bilaterally in Patients 12, 15 and 20, who have only unilateral
clefts on routine inspection of the MRI. The ECC ratio is
also abnormally high in three of these cases.
It should be noted that in these cases perilesional polymicrogyria was not obvious or definite on routine inspection,
but is presumed to be present. Quantification shows that
structural disproportion compatible with polymicrogyria is
present in 66% of such cases and may be found in hemispheres
that are completely normal on routine inspection. This lends
further support to the hypothesis that structural abnormalities
in CD may extend beyond visualized boundaries, and be
associated with abnormalities of connections (e.g. interhemispheric axons).
Therefore in some cases, quantitative changes in postprocessed MRI data and expected histology can be directly
correlated, raising the possibility of predicting histopathology
Fig. 2 Coronal MRI image of Patient 2. Note the extensive
dysgenesis affecting the right hemisphere; the underlying
histopathology cannot be uniquely determined on visual
inspection of the MRI alone.
when this is not directly apparent from routine inspection of
the MRI alone. The nature of the dysgenesis in the right
hemisphere of Patient 2 is unclear (see Fig. 2). However,
analysis of the surface area derivatives shows a pattern
similar to that in patients with probable polymicrogyria, with
elevation of ECC and ESM in the dysgenetic hemisphere,
suggesting that the underlying pathology in this case may also
be polymicrogyria. Abnormality is found on quantification in
the visually normal left hemisphere, in further support of the
extensive and possibly occult nature of dysgenesis.
Two patients with clefts, Patient 16 with bilateral full
thickness clefts and Patient 4 with unilateral partial thickness
clefts, do not have any abnormalities of the surface area
derivatives. This is possibly because the extent of presumed
structural abnormality (polymicrogyria) around a cleft is
variable (Dekaban, 1965; Levine et al., 1974), such that
quantitative normality of the rest of the hemisphere swamps
localized quantitative abnormality. This may also be the case
for Patient 1 who is known to have polymicrogyria. These
cases illustrate the need for regionalization of area measures.
278
S. M. Sisodiya and S. L. Free
Fig. 3 Surface rendering of the MRI scan of brain of Patient 12
viewed from a right anterior oblique projection. Note the
macrogyric region in the temporal lobe.
Agenesis of the corpus callosum
Agenesis of the corpus callosum may be associated with
hemispheric dysgeneses other than the radial rearrangement
of gyri on the mesial surfaces (Friede, 1975; Barkovich and
Norman, 1988), such as polymicrogyria (Billette de Villemeur
et al., 1992) and pyramidal tract hypoplasia (Parrish et al.,
1979). The ratio ECC should be, and is, abnormal bilaterally
in our cases (Patients 12 and 17), as interhemispheric projection is obviously reduced. In both cases, additional abnormalities of the ESM ratio are found, bilaterally in Patient 12 and
on the left only in Patient 17. As interpreted above on the
basis of analysis of brains with clefts, this implies some
combination of increased (or unchanged) neuropil and
reduced mean extracortical projection, associated with the
presence, though not necessarily exclusively, of polymicrogyria. It is therefore possible that the gyral abnormality
(see Fig. 3) seen in Patient 12, who has a cleft, callosal
agenesis and specific surface area derivative abnormalities,
is polymicrogyria, though this is not the only diagnosis that
can be made on visual inspection only of the reconstruction.
Patient 17 has the expected abnormalities of gyral disposition
on the mesial surfaces of the hemispheres; the quantitative
results suggest that there may also be extensive polymicrogyria in the left hemisphere, though this was not seen
on routine inspection. Polymicrogyria may be occult on
macroscopic inspection or imaging and be revealed only
on histology (Galaburda and Kemper, 1979; Kuzniecky
et al., 1994).
Macrogyria
Bilateral posterior macrogyria is seen in Patient 7, with
cortical thickening in the affected area seen on routine
inspection in multiple planes. Cortical thickening may be due
to several different histopathologies (Friede, 1975; Raymond
et al., 1995). A priori, truly thickened cortex implies either
reduced projection out of and into the cortex (if the surface
Fig. 4 View from the vertex of surface rendering of the MRI scan
of cerebral hemispheres in Patient 7. Macrogyria is obvious
posteriorly, especially in comparison with the gyral widths
anteriorly: note the extent of the obvious gyral abnormality.
density of neurons increases), supported by the low EA
bilaterally in this case, or increased mean neuronal neuropil
(if the surface density of neurons does not increase). Increased
total neuropil and reduced projection are associated with an
increased ESM ratio in four out of six cases of probable
polymicrogyria; the normality of the ESM ratio in this case,
despite the extent of visible gyral abnormality (see Fig. 4)
and the bilaterally low EA, argue against an increase in
neuropil. Instead, extracortical axonal projection is likely to
be reduced, irrespective of any alteration in the number of
neurons. A reduction in projection is supported by the
abnormally low SMV in the left hemisphere. The normal
CCA suggests that, although total extracortical projection is
reduced, interhemispheric projection seems to be maintained.
This suggests lamina 3 neurons have maintained their
projectional integrity (Jones, 1984), whether or not they are
correctly positioned in the thicker cortex, and necessitates a
reduction in intrahemispheric extracortical projection. This
combination is therefore unlike polymicrogyria, but
compatible with pachygyria, in which reduction in neuropil
locally has been reported (Takada et al., 1994). Other
histopathologies are possible; confirmation of the interpretation of the changes in surface area derivatives must
await analysis, including stereological cell counts, of cases
of known agyria (e.g. with familial lissencephaly). It is
possible, nevertheless, that abnormality of ESM is seen only
in association with polymicrogyria.
Cerebral surface areas and volumes in cerebral dysgenesis
Alteration in the total number of neurons may complicate
these analyses; the effect of changes in cell density has been
discussed for Patient 7 above. In all the other cases of
dysgenesis, the effect of altered neuronal numbers can be
similarly addressed; neurons probably need to be connected
to survive in the adult brain (Cowan et al., 1984), so that
any change in neuron number must be accompanied by
appropriate changes in neuropil and therefore in connections.
Alteration in the number of neurons in lesional areas has
been reported based on qualitative analysis in both pachygyria
(Stewart et al., 1975) and polymicrogyria (Dvorak et al.,
1978; Sarnat, 1992), but, to our knowledge, there are no
stereological quantitative analysis of cell numbers in CD to
confirm this. Indeed, experimental alteration in the number
of neurons during various stages of development does not
seem to lead to abnormalities of gyration (Langman and
Shimada, 1971).
There remain patients with abnormal surface area derivatives in whom analysis is limited. Patient 3, for example, has
a visually abnormal right hemisphere, with reduced distant
connectivity suggested by the low SMV, and reduced interhemispheric connectivity suggested by the low CCA. The
ECC ratio is high in the contralateral hemisphere, detecting the
extensive connectional disturbance that affects the apparently
normal hemisphere if interhemispheric transfer is reduced.
Patients 5, 10 and 13 all have subependymal heterotopia,
with additional dysgenesis in Patients 5 (subcortical heterotopia) and 13 (unilateral limited posterior macrogyria). All
three of these patients appear to have connectional abnormalities as they all have an abnormal ECC ratio, two in visually
normal hemispheres, including one patient who does not
have any block abnormalities. Patient 14 also has an abnormal
ECC ratio bilaterally, although the dysgenesis is limited, on
inspection, to the right parietal region. However, further
interpretation is not possible from the available results. In a
further four patients, the surface area derivatives are normal,
though in three there are block abnormalities. Only one
patient has no abnormalities of either blocks or surface area
derivatives. This suggests that the two methods of quantitative
analysis examine different aspects of cerebral structure.
The nature of CD
The analysis of the surface area derivatives in the patients
has revealed structural abnormalities in seven out of 11
hemispheres that, on routine inspection, appeared completely
normal, in the same fashion that regional volumetric analysis
has revealed changes in apparently normal hemispheres
(Sisodiya et al., 1995). This provides further evidence for the
hypothesis that CD may extend beyond visualized boundaries.
The actual form of the disruption in visually normal
regions of the brain may not necessarily fall into a currently
recognized histopathological variety of dysgenesis, such as
polymicrogyria, for example. In its broadest sense, dysgenesis
means abnormal formation and whether this includes
qualitative disruption of anatomy, or a quantitative change,
279
is a matter of semantics and the level at which disruption is
considered. Purpura (1974) and Huttenlocher (1974), who
studied mentally retarded patients with and without epilepsy,
respectively, used Golgi staining to demonstrate abnormal
synaptic development in hemispheres judged normal on
inspection and routine microscopy. These changes may be
considered dysgenetic because they are likely to be due to
abnormal development. Additionally, their work supports the
connectional basis of some forms of structural abnormality,
with alteration in neuropil morphology and dendritic spine
number; neuronal architectural changes, suggestive of
connectional abnormalities in dysgenesis, are also seen in
pachygyria (Bordarier et al., 1986; Takada et al., 1994) and
polymicrogyria (Ferrer, 1984).
It might be argued that the changes found are due to
epilepsy rather than associated with its cause. However, there
is no significant difference within the patient group between
those with surface area derivative abnormalities and those
without, in terms of the duration of epilepsy or the history
of secondarily generalized seizures. Block-volume
abnormalities, it has been shown, are also unlikely to be due
to epilepsy (Sisodiya et al., 1995). The changes seen are
unlikely to be epiphenomena and are likely to be associated
with pathogenic structural abnormalities.
Conclusions
The results show that order, present in normal cerebral
hemispheres and manifest as proportions of quantitative
parameters, may be disrupted extensively in patients with
cerebral dysgenesis. The changes in surface area parameters
predicted from the known histopathology of polymicrogyria
are seen in patients presumed to have this condition. The
interpretation of the quantitative results in polymicrogyria
differs from that in non-polymicrogyric pathology on
theoretical grounds and may be useful in identifying cortical
thickening that is non-polymicrogyric in nature. In 64% of
cases, abnormalities of surface area derivatives were found
in visually normal hemispheres, supporting the hypothesis
that cerebral dysgenesis is an extensive disorder. Our results
also support the hypothesis that in humans this extension of
dysgenesis is associated with connectional abnormalities, as
seen in 60% of our cases.
The regionalization of surface area measures may allow
further analysis, possibly revealing localized (and possibly
multifocal) abnormalities that are swamped by quantitatively
dominant normal cortex (as seen with regional measures of
GMV and SMV), and may thus help in the analysis of the
eight out of 20 patients who have no surface area derivative
abnormalities, though the existence of abnormal connectivity
patterns is suggested in seven of the eight by abnormal
regionalized volume measures (abnormal blocks). It may also
allow better correlation of area changes with predicted
pathology: thus regional analysis of patients with bilateral
perisylvian syndrome (known to be polymicrogyric in nature)
or familial lissencephaly may allow confirmation of the
280
S. M. Sisodiya and S. L. Free
interpretation of the changes in surface-area derivatives,
without recourse to post-mortem studies. Regional analysis
would also allow correlation in surgical cases; local surfacearea based predictions of pathology could be tested in
resection specimens.
The methods presented here could also be applied to other
cerebral conditions in which connectional abnormalities may
be present; regionalization may increase the utility of the
technique and might help determine, in vivo, the nature of
cerebral changes in, for example, patients with partial epilepsy
and visually completely normal scans.
Acknowledgements
We wish to thank Dr J. M. Stevens for reporting the MRI
scans, and Professor S. D. Shorvon and Dr D. R. Fish for
helpful comments. The work was supported by the Wellcome
Trust and the National Society for Epilepsy.
References
Aboitiz F, Scheibel AB, Fisher RS, Zaidel E. Fiber composition of
the human corpus callosum. Brain Res 1992; 598: 143–53.
Andermann F. Brain structure in epilepsy. In: Shorvon SD, Fish
DR, Andermann F, Bydder GM, Stefan H, editors. Magnetic
resonance scanning and epilepsy. New York: Plenum Press, 1994:
21–7.
Awad IA, Rosenfeld J, Ahl J, Hahn JF, Luders H. Intractable
epilepsy and structural lesions of the brain: mapping, resection
strategies, and seizure outcome. Epilepsia 1991; 32: 179–86.
microgyria and status verrucosus deformis in rats—pathogenesis
and interrelation. Histological and autoradiographical study. Acta
Neuropathol (Berl) 1978; 44: 121–9.
Ferrer I. A Golgi analysis of unlayered polymicrogyria. Acta
Neuropathol (Berl) 1984; 65: 69–76.
Filipek PA, Richelme C, Kennedy DN, Caviness VS Jr. The young
adult human brain, an MRI-based morphometric analysis. Cereb
Cortex 1994; 4: 344–60.
Free SL, Sisodiya SM, Cook MJ, Fish DR, Shorvon SD. Three
dimensional fractal analysis of the white matter surface from MR
images of the human brain. Cereb Cortex 1996; 6: 830–6.
Friede RL. Developmental neuropathology. Wien: SpringerVerlag, 1975.
Galaburda AM, Kemper TL. Cytoarchitectonic abnormalities in
developmental dyslexia: a case study. Ann Neurol 1979; 6: 94–100.
Goldman-Rakic PS. Morphological consequences of prenatal injury
to the primate brain. [Review]. Progr Brain Res 1980; 53: 3–19.
Haug H. Remarks on the determination and significance of the gray
cell coefficient. J Comp Neurol 1956; 104: 473–92.
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: 126–42.
Henery CC, Mayhew TM. The cerebrum and cerebellum of the
fixed human brain: efficient and unbiased estimates of volumes and
cortical surface areas. J Anat 1989; 167: 167–80.
Hofman MA. Size and shape of the cerebral cortex in mammals.
Brain Behav Evol 1985; 27: 28–40.
Barkovich AJ, Norman D. Anomalies of the corpus callosum:
correlation with further anomalies of the brain. AJNR AM J
Neuroradiol 1988; 9: 493–501.
Hofman MA. The fractal geometry of convoluted brains. J Hirnforsch
1991; 32: 103–11.
Becker LE. Synaptic dysgenesis. [Review]. Can J Neurol Sci 1991;
18: 170–80.
Huttenlocher PR. Dendritic development in neocortex of children
with mental defect and infantile spasms. Neurology 1974; 24:
203–10.
Billette de Villemeur T, Chiron C, Robain O. Unlayered
polymicrogyria and agenesis of the corpus callosum: a relevant
association? Acta Neuropathol (Berl) 1992; 83: 265–70.
Blinkov SM, Glezer II. The human brain in figures and tables. A
quantitative handbook. New York: Basic Books, 1968.
Bordarier C, Robain O, Rethore MO, Dulac O, Dhellemes C.
Inverted neurons in agyria. Hum Genet 1986; 73: 374–8.
Braendgaard H, Evans SM, Howard CV, Gundersen HJG. The total
number of neurons in the human neocortex unbiasedly estimated
using optical disectors. J Microsc 1990; 157: 285–304.
Cherniak C. The bounded brain: toward quantitative neuroanatomy.
J Cogn Neurosci 1990; 2: 58–68.
Cowan WM, Fawcett JW, O’Leary DDM, Stanfield BB. Regressive
events in neurogenesis. Science 1984; 225: 1258–65.
Dekaban AS. Large defects in the cerebral hemispheres associated
with cortical dysgenesis. J Neuropathol Exp Neurol 1965; 24:
512–30.
Dvorak K, Feit J, Jurankova Z. Experimentally induced focal
Hynd GW, Hall J, Novey ES, Eliopulos D, Black, Gonzalez JJ,
et al. Dyslexia and corpus callosum morphology. Arch Neurol 1995;
52: 32–8.
Jones EG. Laminar distribution of cortical efferent cells. In: Jones
EG, Peters A, editors. Cerebral Cortex, Volume 1. New York:
Plenum Press, 1984: 521–54.
Kuzniecky R, Andermann F, Guerrini R. The epileptic spectrum in
the congenital bilateral perisylvian syndrome. Neurology 1994; 44:
379–85.
Langman J, Shimada M. Cerebral cortex of the mouse after prenatal
chemical insult. Am J Anat 1971; 132: 355–74.
Levine DN, Fisher MA, Caviness VS Jr. Porencephaly with
microgyria: a pathologic study. Acta Neuropathol (Berl) 1974; 29:
99–113.
Loopuijt LD, Villablanca JR, Hovda DA. Morphological changes
in the thalamus and neocortex of the cat brain after a restricted
unilateral fetal neocortical lesion. Brain Res Dev Brain Res 1995;
85: 259–72.
Cerebral surface areas and volumes in cerebral dysgenesis
281
Mandlebrot BB. The fractal geometry of nature. San Francisco:
W.H.Freeman, 1982.
a 27-week fetus: an architectonic and topographic analysis. J
Neuropathol Exp Neurol 1974; 33: 374–84.
Mayhew TM. A review of recent advances in stereology for
quantifying neural structure. [Review]. J Neurocytol 1992; 21:
313–28.
Rockel AJ, Hiorns RW, Powell TPS. The basic uniformity in
structure of the neocortex. Brain 1980; 103: 221–44.
Mitchison G. Neuronal branching patterns and the economy of
cortical wiring. Proc R Soc Lond B Biol Sci 1991; 245: 151–8.
Mountcastle VB. An organizing principle for cerebral function: the
unit module and the distributed system. In: Edelman GM,
Mountcastle VB, editors. The mindful brain: cortical organisation
and the group-selective theory of higher brain function. Cambridge
(MA): MIT Press, 1978: 7–50.
Palmini A, Gambardella A, Andermann F, Dubeau F, da Costa JC,
Olivier A, et al. Intrinsic epileptogenicity of human dysplastic
cortex as suggested by corticography and surgical results. Ann
Neurol 1995; 37: 476–87.
Ruppin E, Schwartz EL, Yeshurun Y. Examining the volume
efficiency of the cortical architecture in a multi-processor network
model. Biol Cybern 1993; 70: 89–94.
Sarnat HB. Cerebral dysgenesis. New York: Oxford University
Press, 1992.
Schlaug G, Jäncke L, Huang Y, Steinmetz H. In vivo evidence of
structural brain asymmetry in musicians [see comments]. Science
1995; 267: 699–701. Comment in: Science 1995; 267: 616, comment
in: Science 1995; 268: 621–2.
Sisodiya SM, Free SL, Stevens JM, Fish DR, Shorvon SD.
Widespread cerebral structural changes in patients with cortical
dysgenesis and epilepsy. Brain 1995; 118: 1039–50.
Parrish ML, Roessmann U, Levinsohn MW. Agenesis of the corpus
callosum: a study of the frequency of associated malformations.
Ann Neurol 1979; 6: 349–54.
Sisodiya S, Free S, Fish D, Shorvon S. MRI-based surface area
estimates in the normal adult human brain: evidence for structural
organisation. J Anat 1996; 188: 425–38.
Prothero JW, Sundsten JW. Folding of the cerebral cortex in
mammals: a scaling model. Brain Behav Evol 1984; 24: 152–67.
Stewart RM, Richman DP, Caviness VS Jr. Lissencephaly and
pachygyria: an architectonic and topographical analysis. Acta
Neuropathol (Berl) 1975; 31: 1–12.
Purpura DP. Dendritic spine ‘dysgenesis’ and mental retardation.
Science 1974; 186: 1126–8.
Purves D, Riddle DR, LaMantia A-S. Iterated patterns of brain
circuitry (or how the cortex gets its spots) [see comments]. [Review].
Trends Neurosci 1992; 15: 362–8. Comment in: Trends Neurosci
1993; 16: 178–80.
Rakic P. Specification of cerebral cortical areas. [Review]. Science
1988; 241: 170–6.
Rakic P. A small step for the cell, a giant leap for mankind: a
hypothesis of neocortical expansion during evolution. [Review].
Trends Neurosci 1995; 18: 383–8.
Raymond AA, Fish DR, Sisodiya SM, Alsanjari N, Stevens JM,
Shorvon SD. Abnormalities of gyration, heterotopias, tuberous
sclerosis, focal cortical dysplasia, microdysgenesis, dysembryoplastic neuroepithelial tumour and dysgenesis of the
archicortex in epilepsy. Clinical, EEG and neuroimaging features
in 100 adult patients. [Review]. Brain 1995; 118: 629–60.
Reiner O, Carrozzo R, Shen Y, Wehnert M, Faustinella F, Dobyns
WB, et al. Isolation of a Miller-Dieker lissencephaly gene containing
G protein β-subunit-like repeats. Nature 1993; 364: 717–21.
Richman DP, Stewart RM, Caviness VS Jr. Cerebral microgyria in
Takada K, Becker LE, Chan F. Aberrant dendritic development in
the human agyric cortex: a quantitative and qualitative Golgi study
of two cases. Clin Neuropathol 1994; 7: 111–9.
Taylor DC, Falconer MA, Bruton CJ, Corsellis JA. Focal dysplasia
of the cerebral cortex in epilepsy. J Neurol Neurosurg Psychiatry
1971; 34: 369–87.
Tomasch J. Size, distribution, and number of fibres in the human
corpus callosum. Anat Rec 1954; 119: 119–35.
Welker W. Why does cerebral cortex fissure and fold? A review of
determinants of gyri and sulci. In: Jones EG, Peters A, editors.
Cerebral cortex, Vol. 8B. New York: Plenum Press, 1990: 3–136.
Williams RS, Ferrante RJ, Caviness VS Jr. The cellular pathology
of microgyria. A Golgi analysis. Acta Neuropathol (Berl) 1976; 36:
269–83.
Winfield DA, Gatter KC, Powell TPS. An electron microscopic
study of the types and proportions of neurons in the cortex of the
motor and visual areas of the cat and rat. Brain 1980; 103: 245–58.
Received December 20, 1995. Revised September 18, 1996.
Accepted 11 October, 1996