Morphological Variation of Layer III
Pyramidal Neurones in the
Occipitotemporal Pathway of the Macaque
Monkey Visual Cortex
Guy N. Elston and Marcello G.P. Rosa
We compared the morphological characteristics of layer III
pyramidal neurones in different visual areas of the occipitotemporal
cortical ‘stream’, which processes information related to object
recognition in the visual field (including shape, colour and texture).
Pyramidal cells were intracellularly injected with Lucifer Yellow in
cortical slices cut tangential to the cortical layers, allowing
quantitative comparisons of dendritic field morphology, spine density
and cell body size between the blobs and interblobs of the primary
visual area (V1), the interstripe compartments of the second visual
area (V2), the fourth visual area (V4) and cytoarchitectonic area TEO.
We found that the tangential dimension of basal dendritic fields of
layer III pyramidal neurones increases from caudal to rostral visual
areas in the occipitotemporal pathway, such that TEO cells have, on
average, dendritic fields spanning an area 5–6 times larger than V1
cells. In addition, the data indicate that V1 cells located within blobs
have significantly larger dendritic fields than those of interblob cells.
Sholl analysis of dendritic fields demonstrated that pyramidal cells in
V4 and TEO are more complex (i.e. exhibit a larger number of
branches at comparable distances from the cell body) than cells in
V1 or V2. Moreover, this analysis demonstrated that the dendrites of
many cells in V1 cluster along specific axes, while this tendency is
less marked in extrastriate areas. Most notably, there is a relatively
large proportion of neurones with ‘morphologically orientationbiased’ dendritic fields (i.e. branches tend to cluster along two
diametrically opposed directions from the cell body) in the interblobs
in V1, as compared with the blobs in V1 and extrastriate areas.
Finally, counts of dendritic spines along the length of basal dendrites
revealed similar peak spine densities in the blobs and the interblobs
of V1 and in the V2 interstripes, but markedly higher spine densities
in V4 and TEO. Estimates of the number of dendritic spines on the
basal dendritic fields of layer III pyramidal cells indicate that cells in
V2 have on average twice as many spines as V1 cells, that V4 cells
have 3.8 times as many spines as V1 cells, and that TEO cells have
7.5 times as many spines as V1 cells. These findings suggest the
possibility that the complex response properties of neurones in
rostral stations in the occipitotemporal pathway may, in part, be
attributed to their larger and more complex basal dendritic fields,
and to the increase in both number and density of spines on their
basal dendrites.
strict (e.g. Krubitzer and Kaas, 1989; Merigan and Maunsell,
1993; Webster et al., 1994; Gegenfurtner and Hawken, 1996),
important functional differences seem to exist between the
areas comprising each pathway (e.g. Bullier et al., 1994). In
addition to this aspect of parallel processing, a hierarchical
organization of areas has been proposed, based primarily on
laminar patterns of connections between areas (Rockland and
Pandya, 1979; Tigges et al., 1981; Maunsell and Van Essen, 1983;
Boussaoud et al., 1990). However, the notion of a cortical
hierarchy, and specifically whether or not the hierarchical levels
revealed by anatomical tracing correspond precisely to
physiological levels of serial processing, has been the subject of
considerable debate (for reviews, see Felleman and Van Essen,
1991; Bullier and Nowak, 1995; Rockland, 1997).
In recent years, some anatomical characteristics of the
intrinsic cortical circuitry have been observed to vary among
cortical visual areas. For example, the cross-sectional area,
centre-to-centre spacing and spread of intrinsic connections
increase between caudal and rostral visual areas (Amir et al.,
1993; Lund et al., 1993; Fujita and Fujita, 1996), a result which
has been interpreted as ref lecting the ‘more global computational roles proposed for higher-order areas’ (Amir et al.,
1993). In addition, it has been established that the tangential area
of dendritic fields of supragranular pyramidal neurones, and
their complexity and spine density vary approximately following
a caudal-to-rostral gradient. Cells in V1 and V2 are small in
comparison with cells in the fourth visual area (V4) and the
middle temporal area (MT; Lund et al., 1993; Elston et al., 1996;
Elston and Rosa, 1997). These cyto- and histological differences
between visual areas may be important clues to the type of
neuronal operations performed therein; for example, cells with
complex and spine-dense dendritic trees are likely to be able to
integrate excitatory projections from more sources than cells
with simple, sparsely spined dendritic trees (Elston and Rosa,
1997).
In order to explore further the extent of morphological
differences between visual cortical areas, we tested for variation
in dendritic field complexity of layer III pyramidal neurones
among four of the visual areas forming the occipitotemporal
pathway: V1, V2, V4 and area TEO, a cytoarchitectural subdivision of the temporal lobe (von Bonin and Bailey, 1947). The
results demonstrate that the basal dendritic fields of layer III
pyramidal neurones become larger, more complex and have
more spines with rostral progression along the occipitotemporal
pathway.
Introduction
In primates, cortical processing of visual information is carried
out by two parallel pathways which connect the first and second
visual areas (V1 and V2), in the occipital lobe, to areas in the
temporal and parietal lobes. While neurones in the areas of the
occipitotemporal pathway appear to be specialized for
processing of object shape, colour and texture, those in areas of
the occipitoparietal pathway appear to extract information
about object location, including motion (Ungerleider and
Mishkin, 1982; Kaas, 1987; Maunsell and Newsome, 1987;
DeYoe and Van Essen, 1988; Desimone and Ungerleider, 1989;
Felleman and Van Essen, 1991). Although the anatomical and
functional segregation between visual cortical pathways is not
Vision, Touch and Hearing Research Centre, Department of
Physiology and Pharmacology, The University of Queensland,
Queensland 4072, Australia
Materials and Methods
Data were collected from three adult macaque monkeys (Macaca
fascicularis; Table 1). One of these monkeys (AM1) was also used in a
parallel study of the occipitoparietal pathway (Elston and Rosa, 1997). In
all animals one hemisphere was used for electrophysiological recording
or optical imaging experiments, under barbiturate anaesthesia. Upon
Cerebral Cortex XX X 1998;8:278–294; 1047–3211/98/$4.00
Figure 1. Schematics showing the parts of the brain from which blocks of tissue were excised (grey, in the inserts) and the relative location of individually injected neurones (dots)
in flat-mounted slices obtained from areas V1, V2, V4 and TEO. In the V2 diagram, the occipital operculum was everted, revealing the posterior banks of the lunate and inferior occipital
sulci. The magnified view of the posterior bank of the inferior occipital sulcus indicates the location of CO-rich thick stripes (dark grey), thin stripes (light grey) and CO-poor interstripes
(white). The scale bar (under the TEO diagram) corresponds to 5 mm.
Table 1
Animals and visual areas in which neurones were injected
Animal
Age
(months)
Sex
Weight
(kg)
Areas studied
RM12
SM1
AM1
28
23
14
male
male
male
1.7
1.5
1.3
V1
V4
V2, TEO
completion of the recordings the animals were administered an overdose
of sodium pentobarbitone (200 mg/kg) and intracardially perfused, first
with 0.95% saline, then with 4% paraformaldehyde in phosphate buffer
(0.1 mol/l). Blocks of tissue were taken from the non-experimental
hemispheres (Fig. 1). In all areas, sections were taken from the region of
cortex corresponding to the central 5° of the visuotopic representation,
as determined from previously published mapping studies (Daniel and
Whitteridge, 1961; Gattass et al., 1981, 1988; Boussaoud et al., 1991).
The V1 block was excised from the exposed portion of the occipital lobe,
the V2 block from the posterior bank of the inferior occipital sulcus, the
Cerebral Cortex Apr/May 1998, V 8 N 3 279
Figure 2. Alternate 50 µm and 250 µm sections were cut tangential to the cortical
layers in V1. The 50 µm sections (A,C) were processed for CO, thus revealing features
specific to different layers within V1. The 250 µm section in which neurones were
injected (B) includes the sublaminae immediately above the honeycomb pattern
characteristic of sublamina IIIbβ (C). Thus, neurones were injected in sublaminae IIIa
and IIIbα. The blobs and interblobs characteristic of supragranular cortex in V1 were
revealed in the 50 µm section (A) immediately above the 250 µm section (B) in which
we injected neurones. By aligning blood vessels (e.g. arrows) we were able to establish
whether each injected neurone was located in a blob or an interblob. Scale bar =
500 µm.
V4 block from the middle third of the prelunate gyrus, and the TEO block
from the ventrolateral portion of the occipitotemporal transition. The
tissue was f lattened between glass slides as described previously (e.g.
Rosa et al., 1988) and left overnight in paraformaldehyde (4% in 0.1 mol/l
280 Supragranular Pyramidal Neurones in Macaque Ventral Stream • Elston and Rosa
phosphate buffer) at 4°C. The following day the tissue was sectioned on a
vibratome.
Throughout this paper, the nomenclature suggested by Hassler (1966)
for the layers of striate cortex is used in preference to that of Brodmann
(1909), for reasons outlined in Casagrande and Kaas (1994) and Elston
and Rosa (1997). In the context of the present study it is important to note
that layers IIIbβ and IIIc of Hassler correspond to layers IVa and IVb of
Brodmann respectively. In V1 and V2, alternate 50 and 250 µm sections
were cut, and the 50 µm sections were processed for cytochrome oxidase
(CO) reactivity (Wong-Riley, 1979). In V1, the 250 µm section containing
sublaminae IIIa and IIIbα could be identified because the underlying
50 µm section contained the CO ‘honeycomb’ pattern characteristic of
sublamina IIIbβ (Brodmann’s layer IVa; Hendrickson et al., 1978).
Furthermore, by aligning blood vessels between the 250 µm section
in which neurones were injected and the alternate 50 µm sections
processed for CO (Fig. 2), we were able to establish whether each
neurone was centred within a blob or an interblob (Horton and Hubel,
1981). Only cells that were clearly within the CO-rich blobs were
analysed as belonging to the blobs; cells that were located near the
borders of the blobs/interblobs were not included for analysis. By virtue
of the manner in which neurones were injected (grid pattern spaced ∼500
µm apart, Fig. 2B) a larger number of cells were injected within the blobs
than within interblobs. In V2, V4 and TEO, the basal limit of layer III was
determined by the distinct cytoarchitecture of layers III and IV, as
revealed by DAPI (4,6-diamidino-2-phenylindole; Sigma D9542) staining
(Elston and Rosa, 1997; see below). In these areas, we targeted neurones
located within the lower half of layer III. In addition, in V2 we aligned
injected neurones with the CO stripes seen in the alternate 50 µm
sections (Fig. 1), thus establishing whether each neurone was centred in
a CO-rich thick or thin stripe or a CO-poor interstripe (Tootell et al.,
1983; Abel et al., 1997). Although the CO stripes can be somewhat
difficult to determine, in the present study the analysis was restricted to
the interstripes on either side of a well-labelled thick stripe.
Methods of cell injection, immunohistochemical processing, and
morphological and statistical analysis employed in this study have been
described in detail elsewhere (Elston et al., 1996, 1997; Elston and Rosa,
1997; see also Einstein, 1988; Voigt et al., 1988; Buhl and Lübke, 1989).
Prior to injection, tissue was incubated in DAPI, revealing the
cytoarchitecture of cortical layers. In areas V2, V4 and TEO the transition
between layer III and the granular layer was determined by focusing
through the thickness of the slices, and determining the transition
between the larger, more scattered cell bodies of layer III and the small,
densely packed cells of layer IV (see Elston and Rosa, 1997). Labelling the
tissue with DAPI also allowed for the visualization of individual somata,
thus facilitating electrode penetration of a large number of neurones.
Pyramidal neurones were injected in 250 µm slices with the
naphthalamide tracer Lucifer Yellow (Sigma L0259). Tissue in which
neurones had been injected was immunohistologically processed, using
3,3′-diaminobenzidine (DA B; Sigma D8001) as the chromagen (Fig. 3),
and wet-mounted (intact) in 50% glycerol in phosphate buffer. Spiny cells
were categorized as pyramidal or multipolar (see Lund, 1984; DeFelipe
and Fariñas, 1992). In order to be included for analysis, each neurone had
to: (i) be of the pyramidal type, (ii) be located at the base of layer III,
(iii) have the complete basal dendritic arbor contained within the section,
and (iv) be completely filled. The applicability of the last two criteria
to individual cells was judged by tracing each terminal dendrite to an
abrupt, well-defined tip. Cells in which the distal branches had
an uneven, or beaded appearance were excluded. All neurones with an
unambiguous apical dendrite were classified as pyramidal cells; thus,
the various forms of modified pyramidal neurones (Lorente de Nó,
1938; Lund, 1973, 1984; Jones, 1975; DeFelipe and Fariñas, 1992;
Nieuwenhuys, 1994) were included for analysis. The random nature of
the slice boundaries relative to the cortical layers, and the desire to inject
cells that were both away from the upper limit of the slice and above the
limit of layer IV meant that we were not able to obtain samples from each
area in every monkey. Instead, we concentrated the analyses in the best
slice obtained from each area across cases, as judged by the appropriate
laminar position of cells and the quality of the cell filling. While some
interindividual variability is expected, it is unlikely that this factor alone
can explain the variations of cell morphology we report, as shown by
comparison of data from two areas in single monkeys (e.g. case AM1,
Figure 3. Photomicrograph of layer III pyramidal neurones in area TEO injected with Lucifer Yellow and processed for a DAB (3,3′-diaminobenzidine) reaction product. Our experimental paradigm allowed for the injection of a large number of neurones in each slice (A) and allowed for detailed analysis of spine densities (B). Scale bars: A, 250 µm; B, 20 µm.
Table 1) and by our previous work comparing cells from four different
areas of single animals (Elston et al., 1996). Individual variability,
whenever we had a chance to assess it (e.g. Elston and Rosa, 1997),
proved small in comparison with inter-areal differences.
All cells that satisfied the four criteria outlined above were drawn with
the aid of a camera lucida attached to a Zeiss Axioplan microscope. It
should be noted that, as we injected DAPI-labelled neurones in a
pseudorandom manner, the data are most likely to include different
populations of projection neurones, such as those that project
intrinsically, to subcortical targets, and to other ipsi- and contralateral
cortical areas (Tigges et al., 1981; Gilbert and Wiesel, 1983; Diamond et
al., 1985; Rockland, 1985; Voigt et al., 1988; Buhl and Singer, 1989; Shipp
and Zeki, 1989). Basal dendritic field areas were calculated by drawing a
polygon (hull) joining the distal tips of the outermost dendrites, and using
standard features of NIH-IMAGE (NIH Research Services, Bethesda, MD)
for the Macintosh. In order to calculate spine densities, 20 tangentially
projecting basal dendrites of layer III pyramidal neurones from each area
were drawn at high power (using a ×100 Zeiss oil-immersion objective).
Cell bodies were also drawn at this magnification to test for a difference
between cross-sectional areas (in the dimension parallel to the cortical
layers), and for correlation between cell body and basal dendritic field
areas. Modified Sholl analysis (Fig. 4) was performed to study the
complexity of dendritic fields at different distances from the cell body and
to detect any possible directional biases and clustering of dendrites. In
order to quantify bias and clustering in the dendritic field we calculated
the angle (θ) subtended by the polar plot and a circle of radius equal to the
half-maximal value (Fig. 4); cells with tightly clustered dendritic trees
have low values of θ, and cells with no apparent bias have θ values near
360°. Statistical comparisons were made using GB-Stat for the Macintosh
(Dynamic Microsystems, Silver Spring, MD).
Results
Five hundred and ninety-eight DAPI-labelled neurones were
injected in f lat-mounted preparations containing layer III of
areas V1, V2, V4 and TEO. Two hundred and forty-six of these
neurones were included in this analysis by virtue of being
pyramidal, having their complete basal arborization contained
within the section, and being completely filled (Table 2). These
included pyramidal neurones in sublaminae IIIa and IIIbα of V1,
and those in the lower half of layer III in extrastriate areas V2, V4
and TEO. In V2, only neurones in the CO-poor interstripes
were included for analysis. A few pyramidal neurones were
successfully injected in the CO-rich thin stripes (Table 2), but
the sample was considered too small for meaningful comparisons, and these cells were not included for further analysis.
Basal Dendritic Field Areas of Layer III Pyramidal Cells
Basal dendritic field areas of layer III pyramidal neurones differed
between visual areas, increasing markedly with rostral progression along the occipitotemporal pathway. To demonstrate
this qualitatively we ranked pyramidal neurones in each visual
area in ascending order of basal dendritic field area, as illustrated
in Figure 5. The smallest cell in each visual area is pictured on
the left of each row and the largest cell on the right; the other
cells represent the 20th, 40th, 60th and 80th percentiles in each
sample. Statistical analysis revealed the difference in basal
dendritic field area between visual areas to be highly significant
Cerebral Cortex Apr/May 1998, V 8 N 3 281
Figure 4. Analysis of dendritic field complexity. The basal dendritic field of each cell was drawn (A,D) and a modified Sholl analysis (B,E) was performed in order to generate polar
plots (C,F) that describe the number of intersections between the dendritic tree and the circles forming the sampling grid, as a function of direction form the cell body. The orientation
of the dendritic fields is arbitrary. In order to quantify the degree of dendritic bias, we calculated the sum of the angles (θ) subtended between a circle (radius equal to the half-maximal
value of the polar plot) and the polar plot, for each cell. The polar plot in C illustrates an example of a non-biased neurone (θ = 360°), whereas the polar plot in F illustrates an example
of a ‘morphologically orientation-biased’ cell (θ = 163°).
Table 2
Basal dendritic field areas of layer III pyramidal neurones
Table 3
Statistical comparison between basal dendritic field areas
Visual area
n
SD
SEM
Minimum
Maximum
Mean
(× 103 µm2) (× 103 µm2) (× 103 µm2) (× 103 µm2) (× 103 µm2)
V1i
V1b
V2 interstripe
V2 thin stripe
V4
TEO
43
92
33
5
34
44
20.06
26.99
42.38
46.34
74.78
122.26
10.00
10.59
11.78
5.75
14.14
19.57
1.53
1.10
2.05
2.57
2.43
2.95
5.64
5.28
22.22
40.62
47.34
82.58
51.49
48.64
63.76
53.63
102.29
158.68
Kruskal–Wallis test: df = 4, n = 246, H = 187.392, P < 0.0001
Visual area
Mean rank
V1i
V1b
V2
V4
TEO
52.698
82.185
133.330
185.588
223.727
Mann–Whitney U-tests
(separate comparisons were made for populations of neurones in
the blobs and interblobs in V1 and prestriate areas V2, V4 and
TEO: Fig. 6 and Tables 2 and 3). In addition, a comparison
between basal dendritic fields of pyramidal neurones located
within blobs and interblobs of V1 revealed a significant
difference, with blob cells having more extensive dendritic fields
(Tables 2 and 3).
Complexity of Basal Dendritic Fields
As well as an overall increase in area, the basal dendritic fields of
layer III pyramidal neurones became more complex with rostral
282 Supragranular Pyramidal Neurones in Macaque Ventral Stream • Elston and Rosa
Comparison
U′
z
V1i vs. V1b
V1i vs. V2
V1i vs. V4
V1i vs. TEO
V1b vs. V2
V1b vs. V4
V1b vs. TEO
V2 vs. V4
V2 vs. TEO
V4 vs. TEO
2778
1302
1461
1892
2508
3127
4048
1093
1452
1462
–3.643*
–6.209*
–7.488*
–8.031*
–5.455*
–8.590*
–9.415*
–6.672*
–7.473*
–7.195*
*P < 0.001.
Figure 5. Basal layer III pyramidal neurones from the primary (V1), second (V2; CO-poor interstripes), and fourth (V4) visual areas and cytoarchitectonic area TEO, viewed in the plane
of section tangential to the cortical layers. The neurone with the smallest basal dendritic field area from each visual area is illustrated on the left of each row, and the largest on the
right of each row. The other neurones in each row represent 20% rank increments along the distribution of dendritic field area (i.e. the 20th, 40th, 60th and 80 th percentiles).
Table 4
Two-way ANOVA (repeated measures) on Sholl analysis (5 × 6 design) of ‘Area’ (V1(i,b), V2, V4,
TEO) by ‘Distance’ (from cell body, 25 µm intervals)
Source
Sum of
squares
df
Mean
squares
F-ratio
Probability
Area
Distance
Area × Distance
Within cells
25029.13
15246.74
9822.83
26692.09
4
5
20
1134
6257.28
3049.35
491.14
23.54
265.84
129.55
20.87
P < 0.01
P < 0.01
P < 0.01
Post-hoc comparison between pairs of areas (Newman–Keuls)
V1b
V2
V4
TEO
V1i vs.
V1b vs.
V2 vs.
V4 vs.
P > 0.05
P < 0.05
P < 0.05
P < 0.05
P < 0.05
P < 0.05
P < 0.05
P < 0.05
P < 0.05
P < 0.05
progression through the visual areas of the occipitotemporal
pathway. While Figure 5 allows a qualitative impression of
changes between areas, these are better illustrated by Figure 7,
which shows the results of a Sholl analysis of neurones in the
different subdivisions in V1, and in extrastriate areas. For all
areas and for both blobs and interblobs in V1, the complexity of
the basal dendritic fields (estimated by the number of
intersections between dendrites and imaginary concentric
circles centred on the cell body) was greatest within the
proximal third of the dendritic tree. For example, in V1, the
greatest complexity occurred ∼50 µm from the cell body (17.6 ±
5.5 intersections), whereas in area TEO the greatest complexity
occurred between 75 and 100 µm from the cell body (30.4 ± 4.4
intersections). Furthermore, there was an increase in the
complexity of the basal dendritic fields of layer III pyramidal
neurones with rostral progression along areas of the occipitotemporal stream (Table 4). For example, with the exception of
the ring 25 µm around the cell body, cells in area TEO were more
complex than those in V1 and V2 at all distances from the cell
body. The significant interaction between the ANOVA factors
‘Area’ and ‘Distance’ (Table 4) also demonstrates that visual areas
differ in the way dendritic field complexity varies as a function
of distance from the cell body. No difference in complexity was
found between cells located within the blobs and interblobs of
V1 (Table 4).
Dendritic Biases
As illustrated in Figure 4, the analysis of directional dendritic bias
was done by subdividing the imaginary circles drawn for Sholl
analysis in arbitrary 30° sectors, representing different
directions away from the cell body, and then drawing polar plots
Cerebral Cortex Apr/May 1998, V 8 N 3 283
Figure 6. Frequency histograms of basal dendritic field areas of layer III pyramidal
neurones in V1, V2, V4 and area TEO in the macaque monkey. Basal dendritic fields of
layer III pyramidal neurones in the CO-rich blobs of V1 were larger than those in the
CO-poor interblobs. The average size of the basal dendritic field of layer III pyramidal
neurones increased with rostral progression through the visual areas.
representing the number of intersections (between circles and
dendritic tree) as a function of direction from the cell body (Fig.
4). While most cells had circular or gently elliptical dendritic
trees, in many cases it was found that the dendrites tended to
cluster in specific sectors, resulting in a biased distribution of
complexity as a function of direction from the cell body. As
detailed previously (see Elston and Rosa, 1997), to facilitate the
comparison between areas, dendritic fields were classified into
four categories. Cells in which the dendritic branches were
homogeneously distributed around the cell body were classified
as ‘non-biased’. Both the ‘morphologically direction-biased’ and
284 Supragranular Pyramidal Neurones in Macaque Ventral Stream • Elston and Rosa
Figure 7. Sholl analysis of dendritic trees of layer III pyramidal neurones in four visual
areas. This analysis yields a graphical representation of the complexity of the basal
dendritic fields of layer III pyramidal neurones. Neurones in the blobs and interblobs in
V1 were less complex than those in the CO-poor interbands in V2, neurones in V4 had
more complex basal dendritic fields than those in V1 or V2, and neurones in area TEO
had the most complex basal dendritic fields.
the ‘morphologically orientation-biased’ cells had their dendrites
primarily clustered along a single axis. ‘Morphologically
direction-biased’ cells had dendrites that projected primarily in
one direction from the cell body, such that the projection in one
direction along the preferred axis was more than twice that of
the diametrically opposed direction. The dendrites of
‘morphologically orientation-biased’ cells clustered in two
diametrically opposed directions from the cell body. Cells which
displayed other patterns (e.g. three clusters of dendrites, or two
clusters along different axes) were classified as ‘morphologically
complex’. Moreover, because cells in the different visual areas
varied in the ‘tightness’ of their clustering along the preferred
Figure 8. Graphical representation of the proportion of different types of dendritic bias
found in the basal dendritic fields of layer III pyramidal neurones in the different visual
areas of the occipitotemporal pathway. There was a higher proportion of neurones that
exhibited morphological direction or orientation bias in V1 and V2 than in V4 or TEO. The
majority of neurones in V4 and TEO were non-biased.
axis, we quantified the half-peak bandwidth of the polar plots
(e.g. Fig. 4), as an independent measure of dendritic clustering.
The results indicated that the proportions of cells in each of
the four categories change according to the visual area (Figs
8–10). Within the interblob compartments of V1 and the
interstripes of V2, we found that morphologically orientationand direction-biased neurones accounted for about two-thirds of
the pyramidal cell population, whereas in the V1 blobs the
proportion of such biased cells was lower (51%). In contrast with
V1 and V2, non-biased cells were the most common type in both
V4 and TEO, comprising 46.7% and 42% of pyramidal neurones
respectively (Fig. 8). Morphologically direction-biased neurones
were the second most common type in areas V4 and TEO (33.3%
and 26.3% respectively), while morphologically orientationbiased cells (13.3% and 21.1%) formed the third most numerous
population in these areas. Morphologically complex neurones
were present in all areas, in relatively small proportions (Fig. 8).
From the analysis of the half-peak bandwidth of the polar plots
(θ; see Materials and Methods), illustrated in Figure 9, it can be
seen that there were considerable differences between visual
areas. Most notably, neurones in the interblobs of V1 included
the smallest values for θ (corresponding to morphologically
orientation- and direction-biased cells in the above classification), whereas there was a greater proportion of neurones
with large values of θ in the blobs (i.e. fewer morphologically
orientation- and direction-biased cells). A two-tailed unpaired
t-test revealed a significant difference between values of θ for
neurones in the blobs and interblobs in V1 (df = 87, t = –3.15, P
= 0.002). A ll extrastriate areas had right-skewed frequency
distributions of θ, a result that confirms the existence of a large
proportion of non-biased neurones in these areas.
Spine Densities on Basal Dendrites
There was a marked variation in the number of spines present on
the basal dendrites of layer III pyramidal neurones between V1
and the extrastriate visual areas of the occipitotemporal
pathway. This change was a result of increases in both the length
of basal dendrites and the spine density along these dendrites. In
all areas studied there was variability in spine density, such that
spine density was low close to the cell body and greatest in the
proximal half of the dendrites, decreasing slightly toward the
distal tip of the dendrites. Quantitative analysis of peak spine
Figure 9. Frequency distributions of θ (the angle subtended between the polar plots
and a circle of radius equal to the half-maximal value) for layer III pyramidal neurones in
visual areas of the occipitotemporal pathway. Arrow heads point to the median value of
the distributions. Note the different distributions of θ between blobs (V1b) and
interblobs (V1i), and the large proportion of cells with high bandwidths in V4 and TEO.
density (defined as the average density along the five consecutive
10 µm segments yielding the highest spine count; see Fig. 11) in
different areas revealed that peak spine densities were similar in
V1 (blobs and interblobs) and V2. Spine densities of V4 neurones
were significantly greater than those of V1 and V2 cells, and
spine densities of cells in area TEO were even greater than those
of V4 cells (Fig. 11 and Table 5).
Cerebral Cortex Apr/May 1998, V 8 N 3 285
Figure 10. Polar plots of the basal dendritic fields of layer III pyramidal neurones revealed a difference in the relative proportion of neurones with specific types of dendritic bias. For
each area, several plots of the most common type of dendritic bias have been superimposed on the one set of axes illustrated on the left, and the second most common type of bias
is illustrated on the right. The radial dimension in the polar plots was computed by adding the number of intersections between the dendritic tree and concentric circles drawn 25,
50, 75, 100, 125 and 150 µm from the cell body, at 30° polar angle intervals. Each division along the axes equals 10 intersections. The directions relative to the cell body are arbitrary.
286 Supragranular Pyramidal Neurones in Macaque Ventral Stream • Elston and Rosa
Table 5
Statistical comparison of peak spine densities
Kruskal–Wallis test: df = 4, n = 484, H = 292.17, P < 0.0001
Visual area
Mean rank
V1i
V1b
V2
V4
TEO
156.90
145.45
154.93
348.71
392.02
Mann-Whitney U-tests
Comparison
U′
z
V1i vs. V1b
V1i vs. V2
V1i vs. V4
V1i vs. TEO
V1b vs. V2
V1b vs. V4
V1b vs. TEO
V2 vs. V4
V2 vs. TEO
V4 vs. TEO
4455.5
4572
8373
8762.5
4984
8715.5
9052
9343.5
9726
6611.5
–0.63
–0.06
–10.06*
–11.07*
–0.873
–10.53*
–11.39*
–10.66*
–11.58*
–3.96*
*P < 0.001.
‘average’ neurone in each visual area by multiplying the average
number of dendritic segments between concentric circles (25
µm increments), from the Sholl analysis, by the average spine
density at a corresponding distance from the cell body. This
estimate makes apparent the fact that the combined increases in
area, complexity and spine density of the dendritic field lead to a
dramatic increase in the number of spines present on neurones
in the rostral areas of the occipitotemporal pathway, as
compared with those in V1 (Fig. 12). While the number of spines
on the basal dendritic field of the ‘average’ neurone located in
the interblobs in V1 was similar to that in the blobs (597 and 688
respectively), the average V2 neurone has nearly twice as many
spines (1139). The number of spines along the basal dendritic
tree roughly doubles at each successive stage of the
occipitotemporal pathway: in V4 the average neurone has 2429
spines, and in area TEO 4812 spines. Overall, there is a 7.5-fold
difference in the total number of basal dendritic spines between
the ‘average’ layer III pyramidal neurones in V1 and TEO.
Figure 11. Plots of spine densities along basal dendrites of pyramidal neurones in the
interblobs and blobs of upper and middle layer III of V1, and at the base of layer III in V2,
V4, and area TEO in the macaque monkey. Twenty basal dendrites were randomly
selected from neurones in each region to determine if there was any difference in peak
spine density. Spine density, which was calculated per 10 µm segment of dendrite,
varied as a function of distance from the soma. Peak spine density was relatively low for
neurones in V1 and V2, higher in V4, and greatest in area TEO. Error bars = standard
deviations; arrows indicate the limits of the segments yielding the ‘peak spine density’.
The total number of spines on the average basal dendritic
field increased substantially from V1 to TEO. The number of
spines on the basal dendritic field was calculated for the
Somal Cross-sectional Areas
The cross-sectional areas of the cell bodies of pyramidal
neurones in V1 (blob and interblob), V2, V4 and area TEO were
compared (Tables 6 and 7). Two hundred and thirty-four of the
246 neurones were included for this analysis of somal area, 12
cells being excluded because leakage of dye obscured the outline
of the soma. Somal areas of layer III pyramidal neurones did not
differ significantly between CO-rich blobs and CO-poor
interblobs in V1 (Figure 13, Tables 6 and 7). Layer III pyramidal
neurones in CO-pale interstripes of V2 were larger than those in
either blobs or interblobs, although there was a great deal of
overlap. Cells in V4 were larger than those in V2, but there was
no significant difference in somal area of layer III pyramidal
neurones between V4 and area TEO (see Fig. 13, Tables 6 and 7).
In V4, somal areas had a bimodal distribution, suggesting there
may be two anatomically distinct subclasses of pyramidal cells
(Fig. 13).
Somal area was compared with basal dendritic field area of
individual neurones to establish whether there is any correlation
Cerebral Cortex Apr/May 1998, V 8 N 3 287
Figure 12. Number of spines in the ‘average’ basal dendritic field of layer III pyramidal
neurones in visual areas of the occipitotemporal pathway.
Table 6
Somal cross-sectional areas of layer III pyramidal neurones in different visual areas
Visual area
n
mean
(µm2)
SD
SEM
Minimum
(µm2)
Maximum
(µm2)
V1i
V1b
V2
V4
TEO
43
92
33
22
44
99.52
102.41
130.92
238.74
229.46
18.68
18.19
27.48
52.96
37.61
2.85
1.90
4.78
11.29
5.67
71.68
74.84
85.46
150.38
155.25
154.43
157.58
192.01
346.82
338.04
Table 7
Statistical comparison between cell body cross-sectional area
Kruskal–Wallis test: df = 4, n = 234, H = 157.13, P < 0.0001
Visual area
Mean rank
V1i
V1b
V2
V4
TEO
68.98
76.51
127.94
203.14
199.98
Mann–Whitney U-tests
Comparison
U′
z
V1i vs. V1b
V1i vs. V2
V1i vs. V4
V1i vs. TEO
V1b vs. V2
V1b vs. V4
V1b vs. TEO
V2 vs. V4
V2 vs. TEO
V4 vs. TEO
2175
1181
945
1892
2452
2023
4047
712
1438
536
–0.93
–4.94*
–6.54*
–8.03*
–5.23*
–7.26*
–9.41*
–5.99*
–7.33*
–0.07
between these variables. A plot of somal area vs. dendritic field
area of layer III pyramidal neurones revealed a correlated
increase (r2 = 0.72) of these two variables in the occipitotemporal pathway areas (Fig. 14). However, the specific
relationship seems to vary between areas: for example, for a
given cell body size, V4 cells (triangles) tended to have smaller
dendritic trees than did TEO cells (crosses).
288 Supragranular Pyramidal Neurones in Macaque Ventral Stream • Elston and Rosa
Figure 13. Frequency histograms of somal areas of layer III pyramidal neurones in V1
(blobs and interblobs), V2, V4 and area TEO in the macaque monkey. There was an
overall increase in somal area with rostral progression of the visual areas.
Discussion
The objective of this study was to establish the degree of
morphological variation of layer III pyramidal neurones in
different areas of the occipitotemporal pathway (V1, V2, V4 and
TEO) of the macaque monkey. We found that the basal dendritic
field area of layer III pyramidal neurones increased in a caudal to
rostral progression, such that neurones in area TEO were much
larger than those in V1. Furthermore, the complexity of basal
dendritic fields, the peak spine density and the total number of
spines also increased with rostral progression along the
occipitotemporal pathway. The differences in morphology, and
particularly spine density, may allow for sampling of a greater
Figure 14. Dendritic field areas were plotted against cross-sectional somal areas for pyramidal neurones in V1, V2, V4 and area TEO. Linear regression analysis revealed a significant
correlation (r2 = 0.72).
diversity of inputs to layers III and IV in rostral visual areas,
including ‘feedforward’ (Rock land and Pandya, 1979, 1981;
Rock land, 1989; Felleman and van Essen, 1991), intrinsic
(Yoshioka et al., 1992; Amir et al., 1993; Lund et al., 1993; Fujita
and Fujita, 1996), ‘feedback’ (Boussaoud et al., 1990; Rockland
et al., 1994), and callosal projections (Van Essen et al., 1982;
Cusick et al., 1984; Kennedy et al., 1986; Beck and Kaas, 1994).
It would be interesting to know whether the relative proportions
of inputs coming from these different sources vary among visual
areas. The stronger modulation of activity in rostral visual areas
by memory and attention (Moran and Desimone, 1985; Li et al.,
1993; Miyashita et al., 1993a,b; Sobotka and R ingo, 1993;
Nakamura and Kubota, 1995) suggests that neurones in these
areas may receive a larger proportion of non-visual or ‘feedback’
projections than do V1 or V2 cells.
Does the change in average neuronal size result only from
changes in the proportions of specific populations of cells? One
possible explanation for the increase in the average size and
complexity of the dendritic fields of pyramidal neurones along
the occipitotemporal stream is that these trends are based on
changes of proportions of cells belonging to specific
subpopulations. Pyramidal neurones in layer III visual cortex
have been shown to form subpopulations of cells that project to
different targets, including intrinsic (Gilbert and Wiesel, 1983;
Rockland, 1985; Boyd and Matsubara, 1991; Kisvarday and Eysel,
1992; Thejomayen and Mastsubara, 1993; Levitt et al., 1994),
callosal (Voigt et al., 1988; Buhl and Singer, 1989; Diao and So,
1991; Weisskopf and Innocenti, 1991; Martinez-Garcia et al.,
1994), and ipsilateral inter-areal projections (Tigges et al., 1981;
Diamond et al., 1985; Shipp and Zeki, 1989; Spatz et al., 1991;
Rosa et al., 1993; Vogt-Weisenhorn et al., 1995; Elston and Rosa,
1997). In a recent study, Matsubara et al. (1996) have shown
that, within a given visual area, these different subpopulations of
neurones are characterized by different morphologies: for
example, intrinsic projection pyramidal neurones are smaller
and less complex than those forming inter-areal and callosal
projections. Therefore, the differences in cellular morphology
reported here could be a ref lection of a greater proportion of
callosal and inter-areal projecting neurones in rostral visual
areas. Although this may account, partially, for the increase in
cellular complexity in rostral visual areas, this factor alone can
not account for the present results, as the data indicate that the
largest cells injected in V1 and V2 are much smaller than the
smallest cells found in TEO. Moreover, an increase in the
percentage of neurones forming one subpopulation must result
in a decrease in the percentage of cells forming another
population; for example, a small average dendritic field area
could be generated if intrinsically projecting cells, which form
the smallest and least complex subpopulation of pyramidal
neurones in layer III (Matsubara et al., 1996), were abundant in
V1 but rare in V4 and TEO. Yet, these cells have been observed
to form a rich network of intrinsic projections in rostral visual
areas (e.g. Fujita and Fujita, 1996), as in V1 (e.g. Amir et al.,
1993; Lund et al., 1993).
Increase in the Size of Basal Dendritic Fields Allows for
Integration Across Larger Areas of the Visual Field
We found evidence for a serial increase in basal dendritic field
area with rostral progression through visual areas. Because each
of the occipitotemporal pathway areas forms a visuotopic
representation (for review, see Rosa, 1997), the net effect of this
serial increase in basal dendritic field area, combined with the
decrease in magnification factor as one progresses from V1 to
TEO, is that each layer III pyramidal neurone covers an
increasingly larger fraction of the visuotopic map (see also Elston
et al., 1996; Elston and Rosa, 1997). It is therefore natural to ask
whether or not this increase in the ‘sampling area’ of pyramidal
neurones parallels the increase in the excitatory receptive field
Cerebral Cortex Apr/May 1998, V 8 N 3 289
Table 8
Comparison of pyramidal cell sampling area and receptive field area in V1, V2, V4 and TEO
Visual area
Surface area
(mm2)
Dendritic field area
(× 103 µm2)
Percentage of visual area
sampled by the average
neurone
(× 10–2%)
Pyramidal neurone
sampling area
(relative to V1)
Receptive field area for
eccentricities <5°
(deg2)
Receptive field area
relative to V1
(RFarea/RFV1)
V1
V2
V4
TEO
1142a
1203c
479e
275g
24.80
42.38
74.78
122.26
0.22
0.35
1.56
4.45
1.00
1.62
7.19
20.51
0.10b
1.20d
11.09f
89.30h
1.00
12.12
111.65
899.18
a
Average of values given by Van Essen et al. (1984), Le Vay et al. (1985), Felleman and Van Essen (1991), Florence and Kaas (1992) and Horton and Hocking (1996).
b
c
Average of values given by Hubel and Wiesel (1974), Dow et al. (1981) and Van Essen et al. (1984).
Average of value given by Felleman and Van Essen (1991) and measurement of illustration by Distler et al. (1993).
d
Average of measurements taken from Burkhalter and Van Essen (1986), Felleman and Van Essen (1987) and Roe and T’so (1995).
e
Average of value given by Felleman and Van Essen (1991) and measurement of illustrations by Boussaoud et al. (1991) and Distler et al. (1993).
f
Average of values given by Schein and Desimone (1990) and Boussaoud et al. (1991).
g
Average of values measured from illustrations by Boussaoud et al. (1991) and Distler et al. (1993)
h
Boussaoud et al. (1991).
area along the occipitotemporal pathway. In order to explore
this question quantitatively, we list in Table 8 current estimates
of surface area of macaque V1, V2, V4 and TEO, as well as the
dendritic field area and receptive field area for neurones in the
central representation of these visual areas. Receptive field area
increases ∼10 times for each step in the hierarchy (e.g. from V1
to V2), leading to a nearly 900 times cumulative change of
receptive field size between the central (<5°) representations of
V1 and TEO (Table 8). In contrast, the anatomically defined
pyramidal neurone sampling area (i.e. the fraction of the surface
of a visual area covered by the dendrites of a single cell) varies 20
times between V1 and TEO. Thus, the results demonstrate that
the changes in receptive field size cannot be solely explained
in geometrical terms (i.e. assuming a similar amount of
convergence and divergence at each step of the pathway). Other
factors that may contribute to the increase in receptive field size
along the occipitotemporal pathway are the more widespread
intrinsic connections (Yoshioka et al., 1992; Amir et al., 1993;
Fujita and Fujita, 1996) and the potential for neurones in rostral
areas, which have a greater density of spines, to integrate
converging and diverging inputs from more cortical areas (see
also Felleman and Van Essen, 1991).
Correlation between Dendritic Territories of Layer III
Pyramidal Neurones and Intrinsic Axonal Clusters in
Visual Areas Comprising the Occipitotemporal Pathway
Cells in layer III of primate cortex are interconnected by a
network of intrinsic, intra-areal connections that form regularly
spaced patches (see Yoshioka et al., 1992; Lund et al., 1993;
Fujita and Fujita, 1995 for recent studies). Usually, the tangential
dimension of intrinsic axonal clusters has been documented in
terms of the width of the patches. In order to compare the
cross-sectional area of the basal dendritic fields of layer III
pyramidal neurones with that of intrinsic axonal clusters we
estimated the area of clusters from the average widths, assuming
circular symmetry. In addition, direct measurements of the areas
of intrinsic axonal clusters were obtained from A mir et al.
(1993). The comparison of our estimates of basal dendritic field
areas with estimates of intrinsic axonal clusters in the
occipitotemporal pathway of the macaque reveals a highly
significant relationship (r2 = 0.967; Fig. 15). This comparison
supports the observations of Lund et al. (1993), who proposed
that a constant relationship between these variables exists in
290 Supragranular Pyramidal Neurones in Macaque Ventral Stream • Elston and Rosa
Figure 15. Correlation between dendritic field area and cross-sectional area of
patches of intrinsic connections in V1, V2, V4 and TEO. The means and standard
deviations of data on intrinsic connection patches reflect the variability between the
estimates of different studies (Amir et al., 1993; Lund et al., 1993; Yoshioka et al., 1994;
Fujita and Fujita, 1996).
cortical areas. Thus, while it is possible that the cortex is formed
by modular, repeating units that have similar components
throughout (Lorente de Nó, 1938; Szentagothai, 1975; Eccles,
1984), it is now clear that the dimensions of these modules may
vary considerably in different cortical areas.
Morphological Variation of Layer III Pyramidal
Neurones in V1
Our finding of a significant difference in basal dendritic field
area of layer III pyramidal neurones in CO-rich blobs and
CO-poor interblobs in V1 differs from the observations of a
previous cell injection study (Hübener and Bolz, 1992) in which
the authors found no difference in basal dendritic field structure
between these two histochemically defined regions. Several
factors may account for this discrepancy. Firstly, Hübener and
Bolz (1992) included layer II pyramidal neurones in their
analysis, cells which have different inputs and outputs (Blasdel et
al., 1985; Fitzpatrick et al., 1985; Lachica et al., 1993), and
which have smaller dendritic trees than layer III pyramidal cells
(unpublished observations; see also Fig. 3 of Hübener and Bolz,
1992). Therefore, inclusion of layer II cells may have masked any
difference between the blob and interblob cells of middle layer
III. The second factor may have been non-specific variance in the
samples. Hübener and Bolz (1992) had a smaller sample of
injected neurones (n = 92), and their cells were collected from a
number of different monkeys (blob vs. interblob comparisons
were drawn from three monkeys), whereas our sample (n = 161)
was drawn from a single section in one monkey. Therefore,
factors such as possible variation with different eccentricity,
histochemical processing (e.g. different intensity of CO-labelling
and degrees of shrinkage) and interindividual differences in
columnar organization (Horton and Hocking, 1996) may have
introduced variance in their samples. Thus, while it has been
suggested that morphology of supragranular pyramidal neurones
does not contribute to functional variation between the blobs
and interblobs (Hübener and Bolz, 1992), our results on both
dendritic field size and dendritic bias (see below) indicate that
cell morphology may be part of a distinction between these
compartments of V1. This finding ties in logically with the
results of Kim et al. (1997), who demonstrated a lower cell
density in the blobs, as compared with the interblobs. These
authors have suggested that cells within the blobs may be further
apart than those in the interblobs due to the increased density of
neuropil. Perhaps their result and our finding of longer dendrites
of blob cells will ref lect the connectivity of direct subcortical
projections to the blobs (Fitzpatrick et al., 1983; Lachica and
Casagrande, 1992), or the integration of parvo-, magno- and
koniocellular projections within blobs (Lachica et al., 1992).
Dendritic Bias
The existence of dendritic bias was investigated and compared
between visual areas and V1 compartments. In V1, morphologically direction-biased and non-biased neurones with tightly
clustered dendritic trees were more common in the interblobs
than in the blobs. Electrophysiological studies (e.g. Livingstone
and Hubel, 1984; T’so and Gilbert, 1988; Sato et al., 1996)
suggest that neurones in the interblobs tend to have their
responses more sharply tuned for orientation than cells in the
blobs, paralleling the morphological differences revealed by the
present analysis. Furthermore, in a previous study (Elston and
Rosa, 1997), we found that 85% of the pyramidal cells had
morphologically orientation- or direction-biased dendritic trees
in layer IIIc (Brodmann’s layer IVb), which is also characterized
by sharp orientation selectivity (Blasdel and Fitzpatrick, 1984;
Hawken and Parker, 1984; Livingstone and Hubel, 1984; Sato et
al., 1996). These morphological data agree with the suggestion
(Colonnier, 1964) that biased sampling along specific axes of the
fine-grained layer IV visuotopic map may be one of the factors
contributing to orientation selectivity, in addition to excitatory
and inhibitory intrinsic lateral connections (Kisvarday et al.,
1994; Crook et al., 1996, 1997; Sato et al., 1996; Vidyasagar et
al., 1996; Bosking et al., 1997). However, the morphological
distinctions between cells in blobs and interblobs are subtle,
only becoming obvious through the analysis of a large
population of cells. The large overlap between the morphological characteristics of the two cell populations is reminiscent of
results of recent studies, which suggest a more gradual change in
response properties between neurones belonging to these two
histochemically defined V1 compartments (Leventhal et al.,
1995). In extrastriate areas beyond V2 most cells have nonbiased or slightly biased dendritic fields, paralleling our previous
observations on the occipitoparietal pathway (Elston and Rosa,
1997). Because these areas lack a precise visuotopy in the input
layer, such as that imposed by geniculate projections to V1, this
difference is not suprising.
Comparison with the Occipitoparietal Pathway
In a previous study we demonstrated that the average basal
dendritic field area of pyramidal cells located in layer III
increases serially between the striate cortex, V2, and MT, but
that there is little difference between neurones in MT, the
ventral portion of the lateral intraparietal area (LIPv) and area 7a.
Comparison of the data of the present study with those obtained
from areas forming the occipitoparietal pathway reveals that the
degree of morphological variation of pyramidal cells appears to
be more pronounced in the occipitotemporal pathway than in
the occipitoparietal pathway. For example, TEO neurones have
7.5 times more spines than V1 neurones, whereas neurones in
area 7a have 3.2 times as many spines as cells in V1.
Furthermore, the present data demonstrate a consistent trend for
an increase in cellular complexity with rostral progression
through the occipitotemporal pathway, whereas this was not the
case for the occipitoparietal pathway. In summary, the present
results reinforce the notion that pyramidal neurones with larger,
more complex and more densely spined basal dendritic fields
tend to occur in visual areas associated with more ‘global’ levels
of processing of visual information, rather than in V1 or V2.
However, the results of Elston and Rosa (1997) demonstrate that
there is no exact correspondence between morphological
characteristics of layer III pyramidal neurones and the
hierarchical levels of processing proposed on the basis of
anatomical tracing (Felleman and Van Essen, 1991). Pyramidal
cell morphology remains to be determined in many other visual
areas, both in the occipitotemporal and occipitoparietal
pathways. It will be interesting to determine the morphological
characteristics of pyramidal neurones in layer III in areas
associated with both pathways, such as the superior temporal
polysensory area (Bruce et al., 1981; Boussaoud et al., 1991),
and the frontal eye fields, to determine to what extent the trends
for larger and more densely spined dendritic trees persist.
Conclusions
We have demonstrated that basal dendritic field areas of layer III
pyramidal neurones differ significantly between visual areas of
the occipitotemporal pathway. Moreover, the serial increase in
basal dendritic field area of neurones in these areas is correlated
with the size of intrinsic axonal clusters. The increase in area,
complexity and spine density of the basal dendrites of layer III
pyramidal neurones in successive areas of the occipitotemporal
pathway allow neurones in the more rostral areas such as V4 and
TEO to integrate a larger number of inputs than those in caudal
areas V1 and V2.
Notes
We would like to thank Dr David Vaney for generously allowing us access
to his laboratory and Dr David Pow for providing his superb antibody to
Lucifer Yellow. We would also like to thank Dr Mike Calford for advice on
statistical analysis, Rowan Tweedale for assistance with the Sholl analysis
and Rowan Tweedale, and Professor Jack Pettigrew for comments on the
manuscript. Supported by project grants from the Australian National
Health and Medical Research Council.
Address correspondence and reprint requests to Guy Elston, Vision,
Touch and Hearing Research Centre, Department of Physiology and
Pharmacology, The University of Queensland, Queensland 4072,
Australia. Email: [email protected].
References
Abel PL, O’Brien BJ, Lia B, Olavarria JF (1997) Distribution of neurons
Cerebral Cortex Apr/May 1998, V 8 N 3 291
projecting to the superior colliculus correlates with thick cytochrome
oxidase stripes in macaque visual area V2. J Comp Neurol 377:
313–323.
Amir Y, Harel M, Malach R (1993) Cortical hierarchy ref lected in the
organization of intrinsic connections in macaque monkey visual
cortex. J Comp Neurol 334: 19–64.
Beck PD, Kaas JH (1994) Interhemispheric connections of neonatal owl
monkeys (Aotus trivirgatus) and galagos (Galago crassicaudatus).
Brain Res 651: 57–75.
Blasdel GG, Fitzpatrick D (1984) Physiological organization of layer IV in
macaque striate cortex. J Neurosci 4:880–895.
Blasdel GG, Lund JS, Fitzpatrick D (1985) Intrinsic connections of
macaque striate cortex: axonal projections of cells outside lamina 4C.
J Neurosci 5:3350–3369.
Bosking WH, Zhang Y, Schofield B, Fitzpatrick D (1997) Orientation
selectivity and the arrangement of horizontal connections in tree
shrew striate cortex. J Neurosci 17:2112–2127.
Boyd J , Matsubara J (1991) Intrinsic connections in cat visual cortex: a
combined anterograde and retrograde tracing study. Brain Res
560:207–215.
Boussaoud D, Desimone R, Ungerleider LG (1991) Visual topography of
area TEO in the macaque. J Comp Neurol 306:554–575.
Boussaoud D, Ungerleider LG, Desimone R (1990) Pathways for motion
analysis: Cortical connections of the medial superior temporal and
fundus of the superior temporal visual areas in the macaque. J Comp
Neurol 296:462–495.
Brodmann K (1909) Vergleichende Lokalisationslehre der Grosshirnrinde. J Psychol Neurol 4:177–226.
Bruce CJ, Desimone R, Gross CG (1981) Visual properties of neurons in a
polysensor y area in superior temporal sulcus of the macaque.
J Neurophysiol 46:369–384.
Buhl EH, Lübke J (1989) Intracellular lucifer yellow injection in fixed
brain slices combined with retrograde tracing, light and electron
microscopy. Neuroscience 28:3–16.
Buhl EH, Singer W (1989) The callosal projection in cat visual cortex as
revealed by a combination of retrograde tracing and intracellular
injection. Exp Brain Res 75:470–476.
Bullier J, Nowak LG (1995) Parallel versus serial processing: new vistas on
the distributed organization of the visual system. Curr Opin Neurobiol
5:497–503.
Bullier J, Girard P, Salin P-A (1994) The role of area 17 in the transfer of
information to extrastriate visual cortex. In: Cerebral cortex, Vol. 10:
Primary visual cortex in primates (Peters A, Rockland KS, eds), pp.
301–331. New York: Plenum Press.
Burkhalter A, Van Essen DC (1986) Processing of color, form and disparity
information in visual areas VP and V2 of ventral extrastriate cortex in
the macaque monkey. J Neurosci 6: 2327–2351.
Casagrande VA, Kaas JH (1994) The afferent, intrinsic and efferent
connections of primary visual cortex in primates. In: Cerebral cortex,
Vol. 10: Primary visual cortex in primates (Peters A, Rockland KS,
eds), pp. 201–259. New York: Plenum Press.
Colonnier ML (1964) The tangential organization of the cortex. J Anat
98:327–344.
Crook JM, Kisvarday ZF, Eysel UT (1996) GABA-induced inactivation of
functionally characterized sites in cat striate cortex (area 18) — effects
on direction selectivity. J Neurophysiol 75:2071–2088.
Crook JM, Kisvarday ZF, Eysel UT (1997) GABA-induced inactivation of
functionally characterized sites in cat striate cortex — effects on
orientation tuning and direction selectivity. Vis Neurosci 14:141–158.
Cusick CG, Gould HJI, Kaas JH (1984) Interhemispheric connections of
visual cortex of owl monkeys (Aotus trivirgatus), marmosets
(Callithrix jacchus), and galagos (Galago crassicaudatus). J Comp
Neurol 230:311–336.
Daniel PM, Whitteridge D (1961) The representation of the visual field on
the cerebral cortex in monkeys. J Physiol 159:203–221.
DeFelipe J, Fariñas I (1992) The pyramidal neuron of the cerebral cortex:
morphological and chemical characteristics of the synaptic inputs.
Prog Neurobiol 39:563–607.
Desimone R, Ungerleider LG (1989) Neural mechanisms of visual
processing in monkeys. In: Handbook of neuropsychology (Boller F,
Grafman J, eds), Vol. 2, pp. 267–299. New York: Elsevier.
DeYoe EA, Van Essen DC (1988) Concurrent processing streams in
monkey visual cortex. Trends Neurosci 11:219–226.
Diamond IT, Conley M, Itoh K, Fitzpatrick D (1985) Laminar organization
292 Supragranular Pyramidal Neurones in Macaque Ventral Stream • Elston and Rosa
of geniculocortical projections in Galago senegalensis and Aotus
trivirgatus. J Comp Neurol 242:584–610.
Diao YC , So KF (1991) Dendritic morphology of visual callosal neurons in
the golden hamster. Brain Behav Evol 37:1–9.
Distler C, Boussauod D, Desimone R , Ungerleider LG (1993) Cortical
connections of inferior temporal area TEO in macaque monkeys.
J Comp Neurol 334:125–150.
Dow BM, Snyder AZ, Vautin RG , Bauer R (1981) Magnification factor and
receptive field size in foveal striate cortex of the monkey. Exp Brain
Res 44:213–228.
Eccles JC (1984) The cerebral neocortex: a theory of its operation. In:
Cerebral cortex, Vol. 2: Functional properties of cortical cells (Jones
EG, Peters A, eds), pp. 1–32. New York: Plenum Press.
Einstein G (1988) Intracellular injection of lucifer yellow into cortical
neurons in lightly fixed sections and its application to human autopsy
material. J Neurosci Methods 26:95–103.
Elston GN, Rosa MGP (1997) The occipitoparietal pathway of the
macaque monkey: comparison of pyramidal cell morphology in layer
III of functionally related cortical visual areas. Cereb Cortex
7:432–452.
Elston GN, Rosa MGP, Calford MB (1996) Comparison of dendritic fields
of layer III pyramidal neurones in striate and extrastriate visual areas
of the marmoset: a lucifer yellow intracellular injection study. Cereb
Cortex 6:807–813.
Elston GN, Pow DV, Calford MB (1997) Neuronal composition and
morphology in layer IV of two vibrissal barrel subfields of rat cortex.
Cereb Cortex 7:422–431.
Felleman DJ , Van Essen DC (1987) Receptive field properties of neurons
in area V3 of macaque monkey extrastriate cortex. J Neurophysiol
57:889–920.
Felleman DJ, Van Essen DC (1991) Distributed hierarchical processing in
primate cerebral cortex. Cereb Cortex 1:1–47.
Fitzpatrick D, Itoh K, Diamond IT (1983) The laminar organization of the
lateral geniculate body and the striate cortex in the squirrel monkey
(Saimiri sciureus). J Neuroscience 3:673–702.
Fitzpatrick D, Lund JS, Blasdel GG (1985) Intrinsic connections of
macaque striate cortex. Afferent and efferent connections of lamina
4C. J Neurosci 5:3329–3349.
Florence SL, Kaas JH (1992) Ocular dominance columns in area 17 of Old
World macaque and talapoin monkeys: Complete reconstructions and
quantitative analyses. Vis Neurosci 8:449–462.
Fujita I, Fujita T (1996) Intrinsic connections in the macaque inferior
temporal cortex. J Comp Neurol 368:467–486.
Gattass R, Gross CG, Sandell JH (1981) Visual topography of V2 in the
macaque. J Comp Neurol 201:519–539.
Gattass R, Sousa APB, Gross CG (1988) Visuotopic organization and
extent of V3 and V4 of the macaque. J Neurosci 8:1831–1845.
Gegenfurtner KR, Hawken MJ (1996) Interactions of motion and color in
the visual pathways. Trends Neurosci 9:394–401.
Gilbert CD, Wiesel TN (1983) Clustered intrinsic connections in cat visual
cortex. J Neurosci 3:1116–1133.
Hassler R (1966) Comparative anatomy of the central visual system in dayand night-active primates. In: Evolution of the forebrain (Hassler R,
Stephan H, eds), pp. 419–434. Stuttgart: Thieme Verlag.
Hawken MJ, Parker AJ (1984) Contrast sensitivity and orientation
selectivity in lamina IV of the striate cortex of Old World monkeys.
Exp Brain Res 54:367–372.
Hendrickson A E, Wilson JR, Ogren MP (1978) The neuroanatomical
organization of pathways between the dorsal lateral geniculate
nucleus and visual cortex in Old World and New World primates.
J Comp Neurol 182:123–136.
Horton JC, Hocking DR (1996) Intrinsic variability of ocular dominance
column periodicity in normal macaque monkeys. J Neurosci
16:7228–7339.
Horton JC, Hubel DH (1981) Regular patchy distribution of cytochrome
oxidase staining in primary visual cortex of macaque monkey. Nature
292:762–764.
Hubel DH, Wiesel TN (1974) Uniformity of monkey striate cortex: a
parallel relationship between field size, scatter, and magnification
factor. J Comp Neurol 158:295–306.
Hübener M, Bolz J (1992) Relationship between dendritic morphology
and cytochrome oxidase compartments in monkey striate cortex.
J Comp Neurol 324:67–80.
Jones EG (1975) Varieties and distribution of non-pyramidal cells in the
somatic sensor y cortex of the squirrel monkey. J Comp Neurol
160:205–268.
Kaas JH (1987) Why does the brain have so many visual areas? J Cogn
Neurosci 1:121–135.
Kennedy H, Dehay C, Bullier J (1986) Organization of the callosal
connections of visual areas V1 and V2 in the macaque monkey. J Comp
Neurol 247:398–415.
Kim CBY, Pier LP, Spear PD (1997) Effects of ageing on numbers and sizes
of neurons in histochemically defined subregions of monkey striate
cortex. Anat Rec 247:119–128.
Kisvarday ZF, Eysel UT (1992) Cellular organization of reciprocal patchy
networks in layer III of cat visual cortex (area 17). Neuroscience
46:275–286.
Kisvarday ZF, Kim DS, Eysel UT, Bonhoeffer T (1994) Relationship
between lateral inhibitory connections and the topography of the
orientation map in the cat visual cortex. Eur J Neurosci 6:1619–1632.
Krubitzer LA, Kaas JH (1989) Cortical integration of parallel pathways in
the visual system of primates. Brain Res 478:161–165.
Lachica EA, Casagrande VA (1992) Direct W-like geniculate projections to
the cytochrome oxidase (CO) blobs in primate visual cortex. J Comp
Neurol 319:141–158.
Lachica EA, Beck PD, Casagrande VA (1992) Parallel pathways in
macaque monkey striate cortex: Anatomically defined columns in
layer III. Proc Natl Acad Sci USA 89:3566–3570.
Lachica EA, Beck PD, Casagrande VA (1993) Intrinsic connections of layer
III of striate cortex in squirrel monkey and bush baby: correlation with
patterns of cytochrome oxidase. J Comp Neurol 329:163–187.
LeVay S, Connoly M, Houde J, Van Essen DC (1985) The complete pattern
of ocular dominance stripes in the striate cortex and visual field of the
macaque monkey. J Neurosci 5:486–501.
Leventhal AG, Thompson KG, Liu D, Zhou Y, Ault SJ (1995) Concomitant
sensitivity to orientation, direction, and color of cells in layers 2, 3,
and 4 of monkey striate cortex. J Neurosci 15:1808–1818.
Levitt JB, Yoshioka T , Lund JS (1994) Intrinsic cortical connections in
macaque visual area V2: evidence for interactions between different
functional streams. J Comp Neurol 342:551–570.
Li L, Miller EK, Desimone R (1993) The representation of stimulus
familiarity in anterior inferotemporal cortex. J Neurophysiol
69:1918–1929.
Livingstone MS, Hubel DH (1984) Anatomy and physiology of a color
system in the primate visual cortex. J Neurosci 4:309–356.
Lorente De Nó R (1938) The cerebral cortex: architecture, intracortical
connections and motor projections. In: Physiology of the nervous
system (Foulton JF, ed), pp. 291–325. London: Oxford University
Press.
Lund JS (1973) Organization of neurons in the visual cortex, area 17, of
the monkey (Macaca mulatta). J Comp Neurol 147:455–496.
Lund JS (1984) Spiny stellate neurons. In: Cerebral cortex, Vol. 1: Cellular
components of the cerebral cortex (Peters A, Jones EG, eds), pp.
255–308. New York: Plenum Press.
Lund JS, Yoshioka T, Levitt JB (1993) Comparison of intrinsic connectivity
in different areas of macaque monkey cerebral cortex. Cereb Cortex
3:148–162.
Martinez-Garcia F, Gonzalez-Hernandez T, Martinez-Milan L (1994)
Pyramidal and non-pyramidal callosal cells in the striate cortex of the
adult rat. J Comp Neurol 350:439–451.
Matsubara JA, Chase R, Thejomayen M (1996) Comparative morphology
of three types of projection-identified pyramidal neurons in the
superficial layers of cat visual cortex. J Comp Neurol 366:93–108.
Maunsell JHR, Newsome WT (1987) Visual processing in monkey
extrastriate cortex. Annu Rev Neurosci 10:363–401.
Maunsell JHR, Van Essen DC (1983) Functional properties of neurons in
middle temporal visual area of the macaque monkey. I. Selectivity
for stimulus direction, speed and orientation. J Neurophysiol
49:1127–1147.
Merigan WM, Maunsell JHR (1993) How parallel are the primate visual
pathways? Annu Rev Neurosci 16:369–402.
Miyashita Y, Okuno H, Hasegawa I (1993a) Tuning and association —
neural memor y mechanisms of complex visual forms in monkey
temporal cortex. Biomed Res 14:89–94.
Miyashita Y, Date A, Okuno H (1993b) Configurational encoding of
complex visual forms by single neurons of monkey temporal cortex.
Neuropsychologia 31:1119–1131.
Moran J, Desimone R (1985) Selective attention gates visual processing in
the extrastriate cortex. Science 229:782–784.
Nakamura K, Kubota K (1995) Mnemonic firing of neurons in the monkey
temporal pole during a visual recognition memory task. J
Neurophysiol 74:162–178.
Nieuwenhuys R (1994) The neocortex: an overview of its evolutionary
development, structural organization and synaptology. Anat Embrol
190:307–337.
Rock land KS (1985) A reticular pattern of intrinsic connections in
primate area V2 (area 18). J Comp Neurol 235:467–478.
Rock land KS (1989) Bistratified distribution of terminal arbors of
individual axons projecting from area V1 to middle temporal area
(MT) in the macaque monkey. Vis Neurosci 3:155–170.
Rockland KS (1998) Elements of cortical architecture: hierarchy revisited.
In: Cerebral cortex, Vol. 12: Extrastriate cortex (Kaas JH, Rockland
KS, Peters A, eds). New York: Plenum Press (in press).
Rockland KS, Pandya DN (1979) Laminar origins and terminations of
cortical connections of the occipital lobe in the rhesus monkey. Brain
Res 179:3–20.
Rockland KS, Pandya DN (1981) Cortical connections of the occipital
lobe in the rhesus monkey: interconnections between areas 17, 18, 19
and the superior temporal sulcus. Brain Res 212:249–270.
Rock land KS, Saleem KS, Tanaka K (1994) Divergent feedback
connections from areas V4 and TEO in the macaque. Vis Neurosci
11:579–600.
Roe A, Ts’o DY (1995) Visual topography in primate V2: multiple
representation across functional stripes. J Neurosci 15:3689–3715.
Rosa MGP (1998) Visuotopic organization of primate extrastriate cortex.
In: Cerebral cortex, Vol. 12: Extrastriate cortex (Kaas JH, Rockland
KS, Peters A, eds). New York: Plenum Press (in press).
Rosa MGP, Gattass R, Fiorani M (1988) Complete pattern of ocular
dominance stripes in V1 of a New World monkey, Cebus apella. Exp
Brain Res 72:645–648.
Rosa MGP, Soares JGM, Fiorani M , Gattass R (1993) Cortical afferents of
visual area MT in the Cebus monkey: Possible homologies between
New and Old World monkeys. Vis Neurosci 10:827–855.
Sato H, Katsuyama N, Tamura H, Hata Y, Tsumoto T (1996) Mechanisms
underlying orientation selectivity of neurons in the primary visual
cortex of the macaque. J Physiol 494:757–771.
Schein SJ, Desimone R (1990) Spectral properties of V4 neurons in the
macaque. J Neurosci 10:3369–3389.
Shipp S, Zeki S (1989) The organization of connections between area V5
and V1 in macaque monkey visual cortex. Eur J Neurosci 1:309–332.
Sobotka S, Ringo JL (1993) Investigation of long term recognition and
association memory in unit responses from inferotemporal cortex.
Exp Brain Res 96:28–38.
Spatz WB, Vogt DM , Illing RB (1991) Delineation of the striate cortex,
and the striate–peristriate projections in the guinea pig. Exp Brain Res
84:495–504.
Szentagothai J (1975) The ‘module concept’ in cerebral cortex
architecture. Brain Res 95:475–496.
Thejomayen DM , Matsubara JA (1993) Confocal microscopic study of the
dendritic organization of patchy, intrinsic neurons in area 18 of the
cat. Cereb Cortex 3:442–453.
Tigges J, Tigges M, Anschel S, Cross NA, Letbetter WD, McBride RL (1981)
Areal and laminar distribution of neurons interconnecting the central
visual cortical areas 17, 18, 19, and MT in squirrel monkey (Saimiri).
J Comp Neurol 202:539–560.
Tootell RBH, Silverman MS, De Valois RL, Jacobs GH (1983) Functional
organization of the second cortical visual area of primates. Science
220:737–739.
Ts’o DY, Gilbert CC (1988) The organization of chromatic and spatial
interactions in the primate striate cortex. J Neurosci 8:1712–1727.
Ungerleider LG, Mishkin M (1982) Two cortical systems. In: Analysis of
visual behavior (Ingle DJ, Goodale MA, Mansfield RJW, eds), pp.
549–586. Cambridge, MA: MIT Press.
Van Essen DC, Newsome WT, Bixby JL (1982) The pattern of interhemispheric connections and its relationship to extrastriate visual
areas in the macaque monkey. J Neurosci 2:265–283.
Van Essen DC, Newsome WT, Maunsell JHR (1984) The visual field
representation in striate cortex of the macaque monkey: asymmetries,
anisotropies and individual variability. Vis Res 24:429–448.
Vidyasagar TR, Pei X , Volgushev M (1996) Multiple mechanisms
underlying the orientation selectivity of visual cortical neurones.
Trends Neurosci 19:272–277.
Vogt-Weisenhorn DM, Illing R-B , Spatz WB (1995) Morphology and
connections of neurons in area 17 projecting to the extrastriate areas
Cerebral Cortex Apr/May 1998, V 8 N 3 293
MT and 19DM and to the superior colliculus in the monkey Callithrix
jacchus. J Comp Neurol 362:233–255.
Voigt T, LeVay S, Stamnes M A (1988) Morphological and immunocytochemical observations on the visual callosal projections in the cat.
J Comp Neurol 272:450–460.
von Bonin G, Bailey P (1947) The neocortex of Macaca mulatta. Urbana,
IL: University of Illinois.
Webster MJ, Bachevalier J, Ungerleider LG (1994) Connections of inferior
temporal areas TEO and TE with parietal and frontal cortex in
macaque monkeys. Cereb Cortex 4:470–483.
294 Supragranular Pyramidal Neurones in Macaque Ventral Stream • Elston and Rosa
Weisskopf M , Innocenti GM (1991) Neurons with callosal projections in
visual areas of newborn kittens: an analysis of their dendritic
phenotype with respect to the fate of the callosal axon and of its
target. Exp Brain Res 86:151–158.
Wong-R iley M (1979) Changes in the visual system of monocularly
sutured or enucleated cats demonstrable with cytochrome oxidase
histochemistry. Brain Res 171:11–28.
Yoshioka T, Levitt JB, Lund JS (1992) Intrinsic lattice connections of
macaque monkey visual cortical area V4. J Neurosci 12:2785–2802.