Comparison of Dendritic Fields of Layer III Pyramidal

Comparison of Dendritic Fields of Layer III Pyramidal Neurons in Striate and Extrastriate Visual Areas of the Marmoset: a Lucifer Yellow Intracellular Injection Study

Introduction

The visual cortex of primates comprises areas which are distinct in terms of their connections, architecture, functional response properties and topographic representations of the visual field (for reviews see Felleman and Van Essen, 1991; Kaas and Krubitzer, 1991; Rosa, 1996). One of the response properties that varies dramatically amongst different visual areas is the size of neuronal receptive fields. In general, it has been observed that receptive field size progressively increases from caudal to rostral visual areas (Gattass et al, 1985). For example, for a given eccentricity, receptive fields of neurons in the first visual area (VI) are smaller than those in the second visual area (V2), and these are in turn smaller than the receptive fields of dorsomedial area (DM) neurons (Rosa and Schmid, 1995). The larger receptive fields observed in the rostral areas mean that neurons in these areas are able to integrate signals originating from larger portions of the retina than those in VI; this has been deemed necessary to facilitate the processing of image parameters such as colour (Zeki, 1983; Schein and Desimone, 1990), motion (Saito et al, 1986; Tanaka et al, 1986) and object shape (Desimone and Schein, 1987). One possibility is that there is a similar neuronal circuit in all visual areas, and that the larger integration is simply a function of the cortical magnification factor: given similar dendritic fields, a

Vision, Touch and Hearing Research Centre, Department of Physiology and Pharmacology, The University of Queensland, Queensland 4072, Australia

neuron located in an area with a small visual representation will be able to sample a larger fraction of the visuotopic map than a neuron in an area with a larger representation (Fig. \A, B). However, a neuron with a larger dendritic field superimposed on a small visual representation will be able to integrate even more inputs (Fig. 1Q. Only one study has specifically addressed this possibility (Lund et al, 1993). Using Golgi-stained sections, these authors demonstrated that the basal dendritic field areas of layer in pyramidal neurons in the macaque were larger in V4 than in VI or V2. In the present study we used intracellular injection of Lucifer Yellow to compare the basal dendritic field area of layer in pyramidal neurons in four visual areas of the marmoset cortex. Our results reveal an increase in average basal dendritic field area of layer III pyramidal neurons as one proceeds rostrally from VI to V2, and to DL (dorsolateral area; Allman and Kaas 1974) and FST (fundus of superior temporal area; Desimone and Ungerleider, 1986; Krubitzer and Kaas, 1990). Furthermore, we extend the observations of Lund and collaborators (1993) by demonstrating a 2.3-fold difference in mean basal dendritic field area of layer in pyramidal cells between different visual areas.

Materials and Methods Two adult male marmoset monkeys (CaUtthrixjacchus) were used in this study. As part of a related project, each animal was used in a single electrophysiological recording session aimed at defining the boundaries of extrastriate visual areas. At the end of the recording session one of the hemispheres (contralateral to the recordings) was used for the analysis of neuronal dendritic morphology, which is the subject of the present study. For details regarding the protocol of anaesthesia and visual stimulation during the electrophysiological recording session, see Rosa and Schmid (1995). Upon completion of electrophysiological recording the animal received a lethal dose of sodium pentobarbitone and was transcardially perfused with 0.9% saline followed by 4% paraformaldehyde in 0.1 mol/1 phosphate buffer (pH 7.2). In order to produce cortical slices parallel to the cortical layers, the dorsolateral cortex caudal to the sylvian sulcus was carefully dissected, flattened between two glass slides (e.g. Tootell etal, 1985) and left overnight in 4% paraformaldehyde in 0.1 mol/1 phosphate buffer at 4°C. Alternate 250 and 50 |iin sections were cut on a Vibratome, the 50 um sections being stained for cytochrome oxidase to reveal boundaries between visual areas and different cortical layers. The 250 um slices containing the base of layer HI were incubated for 10 min in 10"' mol/1 4,6-diamidino-2-phenylindole (DAPI) (Sigma D 9542, Sigma, St Louis, MO). This procedure enabled the visualization of neuronal somata under UV (341-343 nm) excitation. The slices were then mounted in a perspex injection chamber containing 0.1 mol/1 phosphate-buffered saline and placed on a fixed stage Zeiss microscope equipped with a Leitz micromanipulator. Neurons were injected with micropipettes (250-300 MO)filledwith 8% Lucifer Yellow (Sigma L 0259) in 0.05 mol/1 Tris buffer. Injection was by hyperpolarizing continuous current, up to 100 nA for -5-10 s. Neurons were injected in a semiregular array in closely spaced caudo-rostral rows crossing from VI to the tip of the sylvian sulcus, over an area of cortex -100 mm2.

Cerebral Cortex Nov/Dec 1996;6:8O7-813; 1047-3211/96/S4.00

Downloaded from cercor.oxfordjournals.org by guest on July 6, 2011

Basal dendritic field areas of layer III pyramidal neurons were compared between the first (V1), second (V2), dorsolateral (DL) and fundus of the superior temporal (FST) areas in marmoset monkey visual cortex. These areas correspond to early stages of visual processing (VI, V2) and to areas specialized for the analysis of shape (DL) and motion (FST). Neurons in fixed tangential cortical slices (250 urn) were injected with Lucifer Yellow and immunohistochemically processed for a diaminobenzidine reaction product Dendritic field areas were calculated for layer III pyramidal cells whose complete basal projection was judged to be within the section (n = 189). Borders between different visual areas were established based on cytochrome oxidase immunohistochemistry and myelin patterns in the experimental hemisphere, and electrophysiological recordings in the contralateral hemisphere. Pyramidal neurons in VI had a mean basal dendritic field area of 1.84 x 104 um 2 (SEM = 2.04 x 103 urn2; n = 21). Layer III pyramidal cells in V2 had a mean basal dendritic field 1.26 times larger (mean = 2.32 x 10 4 ± 1.78 x 103 um2; a = 42) than that of VI neurons. The mean dendritic field area of layer III pyramidal cells in DL (n = 76) was 1.5 times larger than that in VI (mean = 2.75 x 10 4 ± 1.59 x 10 3 |im 2 ), and that in FST (n = 50) was 2.3 times larger (mean = 4.26 x 10* ± 2.79 x 103 um2). Our results show that there is a correlation between tangential dendritic field area of basal dendrites of layer III pyramidal neurons and modality of visual processing. The increase in basal dendritic field area of layer III pyramidal cells may allow more extensive sampling of inputs as required by higher-order processing of visual information.

Guy N. Elston, Marcello G.P. Rosa and Michael B. Calford

10-

2040*

2.5

5*

10"

20°

5"

10"

20*

-Arooooooooooo

*-o-o-o-o-o-o-o

10"

20*

Figure 1. Relationship between visuotopic maps and recipient layer III pyramidal neurons. Left column: topographic maps of hypothetical visual areas, where the representation of the fovea is indicated by the star, the vertical meridian by squares, the horizontal meridian by circles and the isoeccentricity lines by dashed lines. The dendritic field of a single neuron is also shown. Right column: schematic view of the central 20° of the visual field. The portion of the visual field that is sampled by the neuron is indicated by the grey circle. A neuron with a small dendritic field located in a large visuotopic representation (top) samples a relatively small portion of topographic inputs. The same neuron located in a visuotopically compressed visual area (middle) samples inputs from a larger fraction of the topographic array. A neuron with a large dendritic field located in a similarly compressed visuotopic map (bottom) samples inputs from an even larger fraction of the topographic array.

Determination of the injection period was by visual cue, i.e. when the terminal dendrites fluoresced brightly the cell was deemed to be properly filled, and application of current was discontinued (Buhl etal,\ 989; Buhl and Dann, 1991). This technique allowed the experimenter to relate characteristics of DAPMabelled fluorescent somata with neuronal morphology whilst injecting. Previous intracellular injection studies in the rat (Elston et al, 1995) and the macaque (unpublished observations), as well as the present data, revealed that layer 111 pyramidal neurons had relatively large, nearly spherical DAPI-labelled cell bodies and brightly fluorescent nucleoli. Spiny stellate (multipolar) cells had smaller, spherical DAPI-labelled somata and bright nucleoli. Other cell types were easily distinguished by their characteristic DAP1 labelling, e.g. neurogliaform cells, which have small, irregular or crescent-shaped somata at certain planes of focus. While DAPMabelled neurons with spherical cell bodies were filled pseudo-randomly on a grid pattern, the cell injection technique may bias against small-bodied neurons. The presence of Lucifer Yellow in neurons was revealed using an antibody (raised in rabbits by Dr David Pow), at a concentration of

808 Layer III Pyramidal Cells in Marmoset Visual Cortex

• Elston et al.

1:400 000 in stock solution [2% bovine serum albumin (Sigma A3425), 1% Triton X-100 (BDH 30632) and 5% sucrose in 0.1 mol/1 phosphate buffer). Immunohistochemistry was carried out as described previously (Pow and Clark, 1990). The primary antibody was detected by a species-specific biotinylated secondary antibody (Amersham RPN 1004; 1:200 in stock solution for 2 h). The tissue was then processed in biotin-horseradish peroxidase complex (Amersham RPN1051; 1:200 in 0.1 mol/1 phosphate buffer); labelling was revealed using 3,3'-diaminobenzidine (DAB; Sigma D 8001; 1:200 in 0.1 mol/1 phosphate buffer) as a chromagen. In order to be included for analysis, neurons had to satisfy the following three conditions. (0 They had to be located at the base of layer m. The laminar location of cells in the cortex was determined on the basis of multiple, converging criteria, as follows: an appropriate cortical slice was selected based on the relationship between the thickness of different layers (established in coronal sections) and the number of tangential sections between the pia mater and the white matter. By focusing at different depths within this slice, the granular layer was identified by the closely spaced, small

Downloaded from cercor.oxfordjournals.org by guest on July 6, 2011

5'

~k