THE JOURNAL OF COMPARATIVE NEUROLOGY 278:157-180 (1988)

Organization of Primary Visual Cortex (Area 17) in the Ferret

MARGARET I. LAW, KATHLEEN R. ZAHS, AND MICHAEL P. STRYKER Department of Physiology, University of California, San Francisco,

California 94143-0444

ABSTRACT Anatomical and electrophysiological mapping techniques were used to

determine topographic organization and arrangement of ocular dominance columns in the primary visual cortex of ferrets. From its border with area 18 on the posterior lateral gyrus, area 17 extends around the caudal pole of the hemisphere and over the splenial gyrus to the caudal bank ofthe splenial sulcus. The visuotopic map is oriented with the isoazimuth lines approxi­mately parallel to the long axis of the posterior lateral gyrus and the isoelevation lines approximately perpendicular to the isoazimuths. Central azimuths are represented on the posterior lateral gyrus and peripheral azimuths are represented on the splenial gyrus; the inferior visual field maps medially and the superior visual field maps laterally. As in other species, the representation of the central visual field is expanded.

The ferret has a considerable degree of binocular vision. Receptive fields driven through the ipsilateral eye extended more than 20° into the contra­lateral visual field. Within the region of area 17 corresponding to the binoc­ular portion of the visual field, tritiated proline injected into one eye transneuronally labelled an ipsilateral projection as a series of patchy bands roughly complementary to gaps in the labelled contralateral projection. Physiological ocular dominance columns were evident as well in that neu­rons and groups of neurons recorded in this region showed clustered ocular dominance preferences. Most single neurons studied were binocularly driven.

Key words: ocular dominance columns, topography, visual field map

This study describes the anatomical and physiological organization of the primary visual cortex (area 17) of the normal adult ferret (Mustela putorius furo). Interest in the ferret and mink visual systems previously has concentrated on the retinogeniculate projection because of the availabil­ity of a catalog of coat color mutations that correlate with altered routing of ganglion cell axons to the dorsal lateral geniculate nucleus (dLGN) (Guillery, '71; Sanderson et aI., '74; Oberdorfer et aI., '77) and visual cortex (Huang and Guillery, '85). More recently, the ferret has appeared to be a potentially valuable subject for the study of early visual development because of the relative immaturity of its vi­sual system at birth (Linden et aI., '81; Guillery et aI., '85; McConnell, '85; reviewed in Jackson and Hickey, '85). Knowledge of the normal, adult organization is prerequi­site to further investigations into geniculostriate and corti­cal development.

In the present study we have located and mapped area 17 in the ferret and have compared its anatomical and phys­iological organization to that of area 17 in another carni-

© 1988 ALAN R. LISS, INC.

vore, the cat. Our results disclose the size, position, and topography of area 17 in the ferret and the segregation of the two eyes' inputs to layer IV into a system of ocular dominance columns. Some of these findings have appeared in abstract form (Law and Stryker, '83).

MATERIALS AND METHODS Physiological recordings

Animals and surgery. Adult ferrets (700-900 g) obtained from Marshall Farms (North Rose, NY) were initially an­esthetized with 0.8 ml/kg i.m. of a mixture ofketamine (50 mg/mn and acepromazine (0.5 mg/mn and were given atro­pine sulfate (0.2 mg/kg) to prevent mucus accumulation in

Accepted June 13, 1988. M.l. Law's present address is Dept. of Biology. The Explorato­

rium, 3601 Lyon St., San Francisco, CA 94123. Address reprint requests to Dr. M.P. Stryker, Dept. of Physiol­

ogy, University of California, San Francisco, CA 94143-0444.

158

the trachea during surgery. After intravenous and tracheal cannulation, anesthesia was maintained at surgical levels by using sodium thiopental (10-30 mg/kg, i.v.). The animals were placed in a modified stereotaxic holder on a feedback­controlled heating pad that maintained body temperature at 37.5°C, and a heart rate monitor was connected. The scalp was then retracted, and the portions of the skull and dura overlying area 17 were removed. In general, area 17 was made accessible by a large craniotomy extending from Horsley-Clarke anterior-posterior -2 mm to the caudal pole ofthe hemisphere (approximately Horsley-Clarke -8 mm), and from Horsley-Clarke lateral-medial 2 mm to the lateral margin of the skull (approximately Horsley-Clarke 11 mm). A skull screw was then placed over the hemisphere contra­lateral to that studied electrophysiologically, and the elec­troencephalogram (eeg) was monitored throughout the remainder of the experiment. Intravenous thiopental so­dium was administered at the first sign of desynchroniza­tion ofthe eeg.

Following surgery, a photograph of the exposed brain surface, to be used for noting the sites of electrode penetra­tions with respect to blood vessel landmarks, was taken and printed at 20-35 x. Pressure points and wound mar­gins were infiltrated with lidocaine, and neuromuscular blocade was induced with pancuronium bromide (0.1 mgl hour, i.v.). Thereafter the animals were ventilated with a mixture of nitrous oxide and oxygen (3:1) at a rate and stroke volume that maintained peak expired CO2 between 3.8 and 4.4% and peak inspiratory pressure less than lkPa. Anesthesia was supplemented by injecting thiopental (ap­proximately 2 mg/kg-hour, i.v.) as necessary. Pupils were then dilated with 2% atropine sulfate and the nictitating membranes were retracted with at 10% solution of phenyl­ephrine hydrochloride before contact lenses ofthe appropri­ate refractive power were fitted.

Thpogmphic organization. The topographic organization of area 17 was examined by making extracellular record­ings from single neurons and small clusters of neurons and relating their receptive-field positions to the locations of the corresponding recording sites within the posterior lat­eral gyrus.

Recordings were made with electropolished tungsten elec­trodes insulated to impedances of 700 KO to 3 MO with Stoner-Mudge lacquer (Hubel, '57). Visual responses ofneu­rons at each recording site were studied by projecting small rectangles of light from a hand-held lamp onto a tangent screen placed 570 mm from the eyes either directly in front of the animal or at an angle of 60° to the midline, Electro­lytic lesions (4-6 p.A cathodal for 4-6 seconds) marked the locations of many of the electrode penetrations, thus allow­ing us to determine the locations of the unmarked penetra­tions by interpolation. In addition, at the end of many recording sessions, injections of the dye fast green were made into the cortex to serve as additional markers.

Two different types of microelectrode penetrations were used. In one approach, electrode penetrations were made either normal to the surface of the brain, spaced at 300-500-p.m intervals on a square grid, or on a vertical axis along rostrocaudal rows spaced at I-mm intervals, with penetrations spaced at 500-p.m or I-mm intervals within a row. The entry points of the penetrations were recorded on photographs of the cortical surface with blood vessels as reference points. The electrodes were advanced vertically in 100-p.m steps until layer IV was encountered. Afferents from the lateral geniculate nucleus contributed to the multi-

M.1. LAW ET AL.

unit activity recorded within layer IV and could be recog­nized by their brisk responses to non-oriented, rapidly moved stimuli. Receptive fields were always determined through the dominant eye at each recording site, and were usually determined through each eye. After recording in layer IV, the electrode was advanced and receptive fields were often plotted again within layer VI.

The electrode was then advanced through the white mat­ter until the ventral tier of gray matter was recognized by a change in background activity. Within the ventral tier, receptive fields were plotted at recording sites in layer VI and again in layer IV. Because of the folding of the cortex, vertical electrode penetrations are not normal to the lami­nae in the ventral tier of gray matter. Some of the more caudal penetrations travelled almost parallel to the lami­nae within the caudal bank of the splenial sulcus, and recording sites were spaced at 100-p.m intervals here.

In the second approach, microelectrode penetrations were made more nearly tangential to the cortical surface at an­gles of 30-60° from the vertical. Such long penetrations allowed us to examine the shift of receptive field locations with small (50-200 p.m) changes in electrode position. In two cases, electrodes were angled 45-60° in a parasagittal plane, with the recordings made at 100-p.m intervals as the electrode advanced rostrally. In another case, a 30° angle in the coronal plane was used to allow access to the very most lateral portions of the posterior lateral gyrus.

Boundary with area 18. The boundary between areas 17 and 18 was determined physiologically by making rostro­caudal rows of vertical microelectrode penetrations spaced at 200-p.m intervals. The 17/18 border was defined by changes in receptive field size and neuronal response prop­erties and by the reversal of the progression of receptive field positions (Hubel and Wiesel, '62, '65; Tusa et aI., '79). Electrolytic lesions made at the physiologically defined bor­der in each row of penetrations guided our search for the anatomical features that might be used to define the bound­aries of area 17.

Visual response properties. The receptive field proper­ties of ocular dominance, preferred orientation, preferred direction of motion, and preferred sign of contrast were noted at most recording sites. The relative strength of each eye's multiunit response was judged qualitatively to fall into one of seven categories analogous to those defined by Hubel and Wiesel ('62) for single units.

In ten animals, higher-impedance electrodes were used to study single neurons in area 17. Findings from these ani­mals on ocular dominance are included in this report. Quantitative findings on orientation selectivity and other receptive-field properties will be presented elsewhere.

A scalar index was used to describe the bias toward one eye or the other in each animal's ocular dominance distri­bution (Reiter et aI., '86). This contralateral bias index (CBI) was calculated according to the following formula:

CBI = 100 x [(1-7) + (2/3)x(2-6) + (l/3)x(3-5) + nJ I (2 xn),

where bold numbers (1-7) equal the number of units in each ocular dominance group, and n equals the total number of visually responsive units. This index takes a value of 100 for exclusive dominance of the contralateral eye, a value of 50 for equal dominance of the two eyes, and a value of 0 for exclusive ipsilateral dominance. This index reflects the

FERRET VISUAL CORTEX

"weight" of the ocular dominance distribution toward one eye or the other.

Histology. At the completion of the recording session, animals were given lethal doses of pentobarbital (50-80 mg/kg, i.v.) and perfused intracardially with 0.1 M phos­phate-buffered saline followed by 4% paraformaldehyde in 0.1 phosphate buffer (pH 7.2-7.4). The head was postfixed by immersion in fixative solution for several days, after which the brain was blocked and sectioned at 30 or 40 tLm in a plane containing the electrode tracks. Cresyl violet stains were used to locate electrode tracks and marking lesions. Transneuronal autoradiographic tracing ofthe gen­iculocortical projection (see below) was carried out in two of the animals studied physiologically, and a variety of other stains (see below) were used to examine the 17/18 border region.

Camera lucida drawings and photomicrographs were made of sections containing electrode tracks. Recording sites were assigned by referring to the Horsley-Clarke co­ordinates ofthe electrode penetrations, microdrive readings at the lesion sites, and the locations of the dye marks. For the purpose of map reconstruction, recording sites were projected to cortical layer IV along the radial columns of cells evident in Nissl-stained sections.

The receptive field position at each recording site was noted on the photographs and drawings. The locations of these receptive fields were related to the fixation point for each eye and expressed as two angles, azimuth and eleva­tion, on a spherical polar coordinate system (Bishop et aI., '62). The fixation-point azimuth was determined by using the method of Sanderson and Sherman ('71). The fixation­point elevation was placed 3.5 ° below the projection of the optic disc in order to make the cortex maps consistent with the maps of the lateral geniculate nucleus (LGN) (Zahs and Stryker, '85). The uncertainty in assigning the true fixa­tion-point elevation will be discussed below.

Map reconstruction: graphically unfolding the cortex

From coronal sections. In one animal studied physiolog­ically we used a two-dimensional reconstruction technique (following Van Essen and Maunsell, '80) to obtain an over­view of the visual map. To make this flattened map from serial 40-tLm coronal sections, we began at the most caudal section through layer IV and traced every tenth section at the level of layer IV, aligning each section relative to its neighbor. To maintain this alignment accurately, we split the more rostral sections at the lateral and medial edges and straightened and aligned the ventral and dorsal halves separately. After completing the reconstruction, the azi­muth and elevation of each receptive field center were noted at the position on the reconstruction of the projection of each recording site onto layer IV. Isoazimuth and isoele­vation lines were then drawn over the flattened cortical reconstruction.

From parasagittal sections. In two animals prepared in parasagittal section, drawings of reconstructed electrode penetrations were used to create unfolded representations of area 17 and neighboring visual areas after building a physical three-dimensional model. In these reconstructions, data from four to ten serial 40-tLm sections were collapsed onto a drawing of a single section. Sections were selected for drawing at approximate I-mm intervals. Drawings were mounted on cardboard and assembled in accurate three­dimensional register. Strips of foil were then laid along the

159

border between layers IV and V, and the positions of the recording sites were projected radially onto these strips.

For all but the most medial and most lateral sections, the width of the strip of foil was the average distance between the section and its two neighbors of the series. Since, for both animals, the mapped region extended approximately to the medial border of area 17, the width of the most medial strip was chosen to be half the distance between the two most medial sections. For one animal (F81), electrode pene­trations had been placed quite far laterally, and recording sites were found in the sharply curved region of cortex, where the parasagittal plane is nearly tangential to the plane of layer IV. A tracing was made of the most lateral section of the series from F81, and this tracing was opened along an anteroposterior cut hinged at the caudal pole. In the second animal, sections containing recording sites did not extend as far laterally, and the most lateral section was represented as a strip half the width of the distance be­tween it and its nearest neighbor of the series.

The strips were then laid flat and aligned so the caudal poles of the sections were in the correct relative anterior­posterior positions. Isoazimuth and isoelevation lines were drawn by connecting recording sites with receptive fields having the same azimuths or elevations, respectively. In drawing these lines, interpolations were sometimes made. (For example, a 10° isoelevation line might be drawn be­tween recording sites with receptive field elevations of 9 and 11 0.)

Anatomy Anatomical features characterizing area 17. The bound­

aries of area 17 were delineated anatomically by using three techniques. Cresyl violet staining allowed visualiza­tion of the laminar structure of the cortex through which the 17/18 border could be discerned. Fiber staining carried out by using the Weil method was used to discriminate area 17 from 18 on the basis of differences in the caliber of myelinated axons innervating specific laminae. Finally, transneuronal autoradiography (Wiesel et aI., '74) was car­ried out in five ferrets to trace the projection from the LGN to the visual cortex. Ten to 12 days before perfusing these ferrets, we injected 2 mCi 3H-proline (specific activity 40 CilmM, New England Nuclear NET-323) dissolved in 7-25 ml sterile saline into the vitreous humor of the right eye by using a 33-gauge needle. Frozen 30-tLm sections from these animals were taken in the horizontal, parasagittal, or cor­onal plane, coated with Kodak NTB-2 emulsion, and devel­oped 3-8 weeks later. Autoradiographic labelling of the visual cortex was sufficiently intense to be visible at low magnification in darkfield but not brightfield optics.

Flattened preparations. In two animals that were not used for microelectrode studies, the posterior neocortex was physically flattened before sectioning by using the method described by Olavarria and Montero ('84) so that large regions of area 17 could be seen in a single section. These animals were first perfused with phosphate-buffered saline followed by brief perfusion with 2% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4). The posterior half to two­thirds of the two cortical lobes were quickly removed and the dura was peeled off. The pia covering the deeper sulci was cut and some internal white matter was removed. Two cuts were made at the posterior lateral and medial corners to facilitate the unfolding of the ventral aspect of the pos­terior lateral gyrus. The cortices were placed between sili­con-coated glass slides separated by 1-2-mm spacers and

160

were hardened in fixative in the flattened state for 2-7 days. The cortices were then embedded in a gelatin-albu­min mixture before sectioning at 40 /lm and autoradio­graphic processing.

Area and volume measurements. The total surface area of area 17 was calculated by using two different methods. In the two animals for which complete unfolded maps were prepared from series of parasagittal sections, the area en­compassed by the physiologically defined area 17 was a simply measured with a Summagraphics bitpad.

A second measure of the total area of area 17 in the plane of layer IV was calculated from measurements of the vol­ume of layer IV divided by its average thickness. The vol­ume of layer IV within area 17 was measured from a complete series of parasagittal autoradiographic sections taken at intervals of 300 /lm in two animals in which one eye had been injected with 2 mCi 3R-proline 1-2 weeks prior to perfusion. In such sections, layer IV stands out as a densely labelled band over the entire extent of area 17 (for example, see Fig. 5). The outline of each band on the side contralateral to the eye that had been injected, includ­ing the lightly labelled gaps that correspond to the unla­belled ipsilateral eye's ocular dominance patches, was traced in the camera lucida at a magnification of 15 X. The areas of the outlined bands were measured with a Summa­graphics bitpad position digitizer connected to the PDPlll 23 computer for planimetry. The sum of these areas, multi­plied by the spacing between sections, was taken to be the volume of layer IV within area 17. The surface area of area 17 in the plane of layer IV was then calculated by dividing this volume by the average thickness of layer IV measured in sections that intersected its plane at right angles.

This method of calculating surface area assumes a con­stant thickness of layer IV. This assumption has been veri­fied in other species (Van Essen and Maunsell, '80) and is confirmed by our repeated measurements in the ferret. It has the virtue that it accounts properly for the portion of layer IV that is cut in oblique or tangential section without the need for making an estimate of the surface orientation in the plane normal to that in which the sections were cut.

The area estimated by the second technique included both contra- and ipsilateral eye patches. The fraction of layer IV devoted to the ipsilateral eye was calculated from volume measurements of the densely labelled patches in the series of sections from the hemisphere ipsilateral to the injected eye. This fraction was expressed as a percentage ofthe total surface area, measured from the outer hemisphere. For comparison with our earlier findings in the LGN (Zahs and Stryker, '85), we have also expressed the ipsilateral area as a fraction of that of the contralateral area. The contralat­eral area was calculated by subtracting the ipsilateral area from the total area.

The relative areas innervated by the ipsilateral and con­tralateral eyes within the binocular portion of area 17 were measured from the flattened cortex illustrated in Figure 15. The projection of the ipsilateral eye delineated the re­gion in which the two sides were to be compared. Labelled and unlabelled areas within this region of area 17 in the two hemispheres were then measured. This reconstruction was also used to examine the period of ocular dominance patches. Comparable portions ofthe binocular region in the two hemispheres were outlined, and the periods were mea­sured in the direction normal to the elongate course of the patches at random, approximately I-mm, intervals within this region.

M.1. LAW ET AL.

Magnification factors. Areal magnification factors were derived from measurements on the two flattened maps pre­pared from parasagittal sections. Each area of cortex de­fined by a pair of isoazimuth lines and a pair of isoelevation lines was measured; areas bounded by isoazimuth or isoe­levation lines drawn from long interpolations between re­cording sites were excluded from these calculations. This area value was then divided by the area of visual field represented in that area of cortex to give the areal magni­fication factor in units of square millimeters of cortex per square degree of visual field.

RESULTS Location of area 17

Area 17 occupies about 80 mm2 on the most caudal por­tion of the neocortex of the ferret. This position, revealed by physiological mapping studies and a variety of anatomi­cal procedures, is illustrated in Figure 1. Area 17 wraps around the crown of the posterior lateral gyrus at the pos­terior pole of the hemisphere, traverses the occipitotem­poral sulcus (Lockard, '85) and splenial gyrus, and terminates on the caudal bank of the splenial sulcus. The occipitotemporal sulcus is a shallow sulcus that takes the same position with respect to area 17 as the suprasplenial sulcus in the cat (Tusa et aI., '78). This sulcus is not always apparent in parasagittal sections through the cortex, and its rostral extent varies greatly among animals. The sple­nial gyrus forms the tentorial surface of area 17.

The orientation of the visuotopic map in area 17 is also shown in Figure 1. The representation ofthe vertical merid­ian forms the dorsal rostral border of area 17 and is seen as a line that roughly follows the contour of the posterior lateral gyrus. The exact rostrocaudal position of this line varies among animals and may extend dorsally onto the lateral gyrus. Inferior elevations are represented dorsome­dially along this line and superior elevations map ventro­laterally. Isoazimuth lines are oriented nearly parallel to the long axis of the posterior lateral gyrus, and isoelevation lines run roughly perpendicular to the isoazimuth lines.

Visual responses The responses to visual stimuli of neurons in area 17 of

the ferret were remarkably similar to those described in the cat (Rubel and Wiesel, '62). Most single neurons in the cellular layers of the cortex were orientation-selective and responded optimally to moving stimuli. Receptive fields had sharp and easily defined borders and were typically a few to a few tens of degrees square in area.

Units recorded simultaneously had similar orientation selectivities and largely overlapping receptive fields. The receptive field of a multiple-unit cluster was similar in area to that of a single unit, and the receptive-field position was

Fig. 1. The location of area 17 on the surface of the ferret brain. The brainstem at the level of the medulla and the cerebellum have been re­moved. A: Lateral view showing the visual field projection onto area 17. Solid lines indicate isoazimuths and dashed lines indicate isoelevations. Labelling of sulci and gyri after Lockard ('85). B: Caudolateral view of area 17. C: Caudomedial view of area 17. D: Dorsal view of area 17. E: Relation of the surface appearance of area 17 to its appearance in section. The left hemisphere was cut in the parasagittal plane at the level shown in by the heavy arrowheads in B, the cortex lateral to the cut was removed, and the cut surface was outlined. Compare this figure to section 4 of Figure 5. The occipitotemporal and splenial sulci have been outlined on the left hemi­spheres in parts Band C. A, anterior; D, dorsal; L, lateral; P, posterior.

laleral sulcus

o

p -

suprasylvian sulcus

posterior rhinal lissure __ -'

pseudosylvian sulcuS

A

splenial gyrus

occipitotemporal sulcus

posterior 'a;,,,,, gyru.

B

A

L _ L

D

\6\

coronal sulcus

cruciale sulcus

piriform lobe

c

v

E

presylvian sulcus

olfactory bulb

anterior rhinallissure

poSlerior lateral gyrus

splenial sulcus occipitotemporal sulcus

posteflor lateral gyrus

occlpitoremporal sulcus

splenial gyrus

splemal sulcus

Figure 1

162

as easily defined. AJong electrode penetrations that ran oblique to the cortical surface, the preferred orientations of successively encountered units changed gradually and pro­gressively. In the middle cortical layers, from which the mapping data reported here were principally obtained, most multiple-unit responses were dominated by one eye, imply­ing that the eye preferences of simultaneously recorded units were similar.

Topographic organization of area 17

A single representation of the contralateral visual fie ld was found withi n area 17 of the fetTet. Physiological map­ping techniques were used to record from si ngle and multi-

\\\\\'\.

M.l. LAW J!;T AL.

pie units within much of area 17 in eight ferrets. The centers of the receptive fields plotted in four of these map­ping experiments are drawn in Figure 2 to illustrate the ext-ent of the visual field represented in area 17. Receptive fields of units in area 17 dominated by the contralateral eye were fou nd with azimuths as great DS 135° and eleva· tions ranbri ng from at least _ 40 0 to +70 degrees. Ipsilat· erally domina ted units (indicated by open circles in Fig. 2) were found only in the central portion of the visual field. The ferret's binocular visual fi eld was found to extend to at least 38° azimuth in the superior visual field and to at least 20° azi muth in the inferior visual field . This estimate of the extent of the binocular visual fi eld is consistent with

4 r////I \ .•• ///////.1 .,.

// /7/ f.t LLjj~LI 0- .'"'

Fig. 2. Monocular Md binocular segmenl8 olth" vil llBl field ~presented nl!!p(!ClIvely. Rooeptive fieldl ilIU8I.Tated all! from mapping experiments in in .~a 17. CIOIIed and open II}'mbois indicate re<:eptiv" fieldB 01 Wlil8 re- four aRimal1 (including Fal and 1"91 01 Fip. 4 and n Azimuth and \lle\'8-!fPOnding to stimulation through the contralateral and ip6ilateral eye&, tion anl indi~ted in de&n!ft.

FERRET VISUAL CORTEX

that found in mapping studies of the LGN (Zahs and Stry­ker, '85).

Reconstruction of map from coronal sections. In initial mapping experiments, rows of electrode penetrations were made in the coronal plane, and the brain was sectioned in that plane as well. Within each section, the more medial penetration sites represented more inferior portions of the visual field. For the dorsal surface at least, the more caudal sections at a given lateromedial position represented more peripheral portions of the visual field. However, for the major portion of area 17, lying beyond the dorsal surface, it was difficult to appreciate the organization of the visuotopic map from such sections.

In order to gain a sense of the overall organization and continuity of the visuotopic map, it is desirable to "unfold" the cortex to view it as a flat sheet. We did not find it possible to reconstruct electrode penetrations in a cortex that had been physically unfolded. However, it was possible to graphically unfold the cortex after locating recording sites in sections. The visuotopic map over a portion of area 17 from which recordings were obtained in one ferret is reconstructed in Figure 3 by using Van Essens and Maun­sell's ('80) technique as described in Materials and Meth­ods. The positions of marking lesions (made during the recording experiment and shown in Fig. 3A by filled circles) were placed on this reconstructed surface and the locations of remaining recording sites were inferred from these land­marks. Figure 3B shows isoazimuth and isoelevation lines drawn on the basis of receptive-field positions at these re­cording sites. On the dorsal surface (shown in the top half of Fig. 3B) more peripheral fields were located progres­sively more caudally and inferior fields progressively more medially. On the ventral (tentorial) surface (shown in the bottom half of Fig. 3B) lateromedial organization was simi­lar while receptive fields continued to progress peripherally as the surface continued rostrally and ventrally.

Reconstruction of map from sagittal sections. The major portion of area 17 was found in the previous experiments to lie on the tentorial surface of the splenial gyrus. Because this surface is inclined only slightly from the coronal plane, coronal sections of area 17 cut it very obliquely. Sagittal sections, on the other hand, were nearly normal to the surface of area 17 over most of its extent (except for the relatively small amount of area 17 lying at the lateral and medial edges of the hemisphere). For this reason, our most extensive analysis was conducted on material sectioned in the sagittal plane.

Receptive fields were determined at 143 sites in 54 elec­trode penetrations in one animal (F81) and at 134 sites in 41 penetrations in a second animal (F91). Electrode pene­trations were then reconstructed from serial parasagittal sections stained by a Nissl method. Figure 4 shows the centers of the receptive fields of the multiunit clusters recorded in F81. Because the location of area 17 is more easily seen in autoradiographically labelled sections than in Nissl-stained sections, the mapping data from F81 were superimposed on a series of labelled sections from another animal. This animal had received an injection of 3H-proline into one eye a week prior to a terminal recording experi­ment in which the boundary between areas 17 and 18 was determined electrophysiologically. The locations corre­sponding to the recording sites from the more extensive mapping experiment in Figure 4 were placed in these sec­tions, illustrated in Figure 5, by using the Horsley-Clarke coordinates of each site. The position of the 17/18 border

163

and the receptive-field elevations confirmed, respectively, the correct anterior-posterior and lateral-medial place­ments of the mapping data onto the sections.

Within parasagittal sections like those shown in Figure 5, area 17 may be traced caudally from its boundary with area 18 around the caudal pole and then rostroventrally along the splenial gyrus onto the caudal bank of the splen­ial sulcus. Receptive-field locations of units encountered along such a route follow an arc from central and inferior to peripheral and superior in the contralateral visual hemi­field. The elevations of receptive fields increase and the arcs become steeper in more lateral sections.

These parasagittal sections were also graphically un­folded (as described in Materials and Methods) in order to convey a sense of the overall organization of the map in area 17. The unfolded map constructed from F81 is shown in Figure 6. Features of the visuotopic map in area 17 are apparent in this reconstruction. Isoazimuth lines run roughly parallel to each other and are crossed at approxi­mately right angles by isoelevation lines. In this animal, the region of area 17 representing the central 10-15° of azimuth occupies approximately 2 mm on the dorsolateral surface of the posterior lateral gyrus. The most peripheral visual field (azimuths as large as 120°) maps to the crown of the splenial gyrus.

Although the pattern illustrated in Figures 4-6 from animal F81 above was the most common, in some animals the precise position of area 17 on the cortical hemisphere differed by more than 2 mm from this case. Such variability in position is significant when one considers that the entire extent of area 17 (along the representation of the horizontal meridian) is no more than 10 mm.

This variability is illustrated by data from another case, F91. The receptive-field centers of the sites recorded in this animal are shown in Figure 7. Figure 8 shows the corre­sponding unfolded map. This figure illustrates the less fre­quently observed, but not rare, case in which area 17 is shifted back so that only the central 5 ° of azimuth are represented on the dorsal surface ofthe cortex, and the very far periphery (azimuths larger than 85°) comes to be repre­sented on the caudal bank of the splenial sulcus. In other cases (not illustrated), less common than those above, area 17 appeared to be shifted as much as 1.5 mm rostral to the position occupied in the case illustrated in Figures 4-6.

Size of area 17 The size of area 17 was measured in two independent

ways as described in Materials and Methods. Mter label­ling the geniculocortical projections by transneuronal au­toradiography, the volume of layer IV was measured from a complete series of sections and was then divided by the average thickness of layer IV measured in sections normal to the cortical surface to yield an estimate of surface area in the plane of layer IV. In one 860-g female analyzed by this method, area 17 occupied 76.7 mm2

• In another 890-g female, area 17 occupied 87.2 mm2.

In two other ferrets, the region of area 17 studied electro­physiologically was measured on the flattened maps illus­trated in Figures 6 and 8. For these measurements, area 17 was bounded dorsorostrally by the representation of the vertical meridian (00 isoazimuth line), medially by the most medial section containing recording sites, and laterally by the most lateral section containing recording sites. The rostroventral boundary was somewhat arbitrarily drawn as a line skirting the most rostroventral recording sites in

164

A

B

OOflSAL SURFACE

VEN TRA~ SURFACE

AOSTAA~

AOS1'111A~

Figure 3

0.--:;40' S·

o ISOAZIMUTH

........

1\1 . 1. LAW F.'l' AL.

FERRET VISUAL CORTEX

+50

+40

+30

+20

+10

0

-10

-20

-30

-40

-50

-60

-70

Fig. 4. The receptive field centers of 143 multiunit clusters recorded in area 17 and neighboring visual areas in ferret F8I. Locations marked with circles are receptive fields of sites recorded in area 17; locations marked with squares are receptive fields of sites recorded in area 18; locations marked with triangles are fields recorded in other visual area(s). Numbers and letters within the symbols refer to recording sites shown in Figures 5

165

+60

+50

+40

+30

+20

+10

0

-10

-20

-30

-40

-50

-60

-70

and 6. The receptive field at the site marked 8g (8" azimuth, +40" elevation) is designated as being from "other visual areas" in Figures 4 and 6 on the basis of the response characteristics of the units recorded at the site. This site appears to be well within area 17 on the section illustrated, an artifact of superimposing data on this particular representative section.

------------------------ area 17, except for the region between rows 2 and 6 in Fig. 3. Flattened representation of area 17 in one ferret showing the

organization of the visual map. A: Serial coronal sections were recon­structed by using a two-dimensional mapping technique (modified from Van Essen and Maunsell, '80). Sections were split for reconstruction at the positions indicated by the arrow in the inset and the dorsal and ventral halves were aligned separately. Electrolytic lesions and penetration sites (filled circles) confirmed the alignment and allowed the estimation of the locations of many other penetration sites (open circles). B: Isoelevation (dashed) and isoazimuth (solid) lines describe the organization of the map over the flattened cortical surface. More peripheral fields are located pro­gressively more caudally and wrap around the tentorial surface of the gyrus. The inferior portion of the visual world maps medially. D, dorsal; V, ventral; L, lateral; M, medial; WM, white matter.

Figure 6, where the 120 0 isoazimuth was used as the bound­ary_ The areas so measured were 54 mm2 in F81 and 52 mm2 in F91. The mapping experiment in F81 did include slightly more of the far peripheral visual field than did the experiment in F91. These areas represent an underestimate of the size of area 17, because they include only the regions studied electrophysiologically_ In the case of F81, it was possible to estimate the area not included in the electro­physiological map from the extent of transneuronal labeL In this case, the total area of area 17 was found to be 65 mm2 _

166 M.1. L.4.W ET AJ~ .

,

~'igure 5

fI<~ lmET VISUAL COHTEX 167

~'igure 5 continued

168

filii:. 5. M.o.ppinll dutll from the ~xp"rimenl IIlu~I",IA,d in ~'ilr"'" 4 have ~"""I"'rimpclilled '''' lI .... ri..sol'pa,.,.,... .. .;1.t>I1 """Iio"" through r .... rt>ttOl"lu. s..,.1.ionf; hllVIl .... n lt1l.1WlJt:uronally Jabel1lod IIftcr an il\ienioo or Irililtled pmline inl6 the conl raiMh,ral ~ and lire shc)\oll In dIIrldicld. ~I" rv of 1111,,, Ii i& _II a~ !l l/loolIl'd band in I~ lJCction~. ~'lIr ....u::h $IlCI.ion. locations corresponding to the rerording 8it~ from I .... ITI!iI"f! uten9h.., "",ppin/:" <!:t~rh"cnl of FJI(l.U'C 4 I1re .ndiC1l.lcd by the lI'lIm.1. Roowding

'''"~.' • _ _ QuoII _ ,,,.Ira'

• Ie a

M.1. LAW ET AJ~ .

~;te!I "'C,." in lI.ye. IV, lett,,~ mark the top (>flh" pnu..'<'lion lin .. (.'.m',,;n,ng the 1'\.'(onI;ng ~ite. The recI!Jlt,,-e-field jIClf!ilion for cmch llill! i/llOho .... n in F'il.'W"1! 4 and ill 00.., .. .,.. .. 1..,<1 by a rererence numOOr for the K'CI.ion tnumbe. in " 'hite Circle) (lOiItaI:"ptt'<l ,,',th thl: letter marl""JI 0." """""'ing sile. Number IM:n"ath t .... ref" ... ",,,,, numbe.- ill th .. H()I'>IIey-Ch.rke luteroJ·mooial coonIinaw "r thlt Aeclion. Am,>WII point to tho tuudal polo .nd the edge af lhll ~plenjlllllUlcu~, WhL .... ;.ppi ieabie.

'" :'"~ 1;) .. ee / w '"

6

'"

'mm

.... opIenIl' -lUICY.

Fig. 6. Flflttened repreoJ<'n tlltion of flr1!lI 17 find neighboring vUoualllreaa C'OlI<Ilructed r""n. the mllppinll data dillph.yed on the led.ions of' Figure 5. CirciN indicate rewrdinlf l ito!!! in area 17; iquan'f indicate ,ites in aNa 18; and tri"nll:l~ iIIh<:n>.' !lites in moro I'OIJtl'1li vialill area(l~ Th_ symbol!! cor'rolpOnd to the illl" in Figure:; gnd the mces.li,·e field!! in Figu ..... 4. Thin oolid line.! 1lN: .-.:imulhll; ~hed linN IU'fI iBoele ... tion& Hetovy Ii""" ~~.~ ,I. ..... . . " I 1 • ..... ... ~ _'~~'~I_.I ....

FERRET VISUAL CORTEX 169

+70

+60 +60

+50 A +50

+40 +40

+30 +30

+20 +20

+10 +10

0 0

-10

-20

-30

-40

-50

-60

Fig. 7. Receptive-field centers of 134 multiunit clusters recorded in area 17 and neighboring visual areas of F91. Conventions are the same as in Figure 4. Receptive field 9c was omitted to reduce crowding; it lay between 9b and 9d.

Magnification factor

An expanded representation of the central visual field is seen in the flattened maps of Figures 6 and 8. In Figure 6, for example, the central 30 0 of azimuth maps to an area of cortex (30 mm2) r-eater than that representing the periph­eral 90 0 (24 mm ).

From the flattened maps, a greater area of cortex also appears to be devoted to the inferior visual field (34 mm2

)

than to the superior visual field (20 mm2), but this conclu­sion depends on the assignment of the fixation-point eleva­tion. The assignment of elevations in the cortical maps was made to be consistent with previously published maps of the LGN (Zahs and Stryker, '85). This assignment was on the basis of the distribution of geniculate receptive-field

sizes as a function of elevation, a function that changed little between -30 and +20 0 of elevation. Inspection ofthe flattened cortical maps reveals that the elevations near _100 have the most expanded representation, suggesting that the true horizontal meridian may be closer to the elevation labelled _100 in the maps than to that labelled 0 0

• If the line labelled - 10 a represents the projection ofthe horizontal meridian more accurately than the line labelled 0 0

, then the representation of the superior visual field (31 mm2) is somewhat larger than that of the inferior visual field (23 mm2 ). This issue will be considered in the Discussion.

The expanded representation of the central visual field may be expressed quantitatively as the magnification fac­tor (Daniel and Whitteridge, '61). Areal magnification fac-

170 M.1. LAW ET AL.

medial

rostral ..... 1----- caudal ----I ... rostral

caudal pole

5 \ 10

edge of splenial sulcus

o / -35

~ (j) () -24 .!!' '-::J U)

co U) -20 .' '-0

"

~ M ~ ~ ~

~ ~~~

caudal pole

rostral ..... t__--- caudal ----ool .... rostral

lateral

-5

@ @

(j) ()

.!!' '-::J U)

~ c: (j) 0-

0

+10

...... .. @/ @

'+15

11J~ edge of

splenial sulcus

£1

1mm

Fig. 8. Flattened representation of area 17 and neighboring visual areas constructed from mapping data from F91. Conventions the same as for Figure 6. Receptive fields are shown in Figure 7.

tors were calculated as described in Materials and Methods. Because ofthe spacing ofthe electrode penetrations during the recording sessions, as well as the expansion of the central visual field which occurs in the visuotopic map in area 17, smaller areas of the visual field were represented in areas of cortex measured in the central representation than in the peripheral representation. Areas of cortex de­voted to 5 ° x 5 ° areas ofthe visual field were measured for the region of area 17 representing 0-20° of azimuth and - 20° to + 25 ° of elevation. For the region representing 20-50° azimuth and - 20° to + 10° elevation, areas of cortex were measured representing areas of the visual field not exceeding 50°2. For azimuths greater than 50° or eleva­tions less than -20°, the cortical representations of visual field areas not exceeding 100°2 were measured.

Figure 9A shows a scatter plot and best-fitting smooth curve of the relation between magnification factor and ec­centricity, defined as the square root of the sum of the squares of azimuth and elevation. In calculating eccentric­ity, 10° was added to the values of elevation shown on the maps, since the data presented here suggest that the _10° isoelevation line is closer to the true representation of the horizontal meridian than is the 0° line. Magnification falls off quite gradually by this analysis, in which azimuthal and elevational changes are combined.

Magnification also changes gradually as a function of azimuth, as shown in Figure 9B. Values on the abscissa indicate the center of the visual field areas represented. Magnification remains fairly constant across the central 25 ° of azimuth and then declines steadily. This trend, ob-

FERRET VISUAL CORTEX

0 0

X

5

4

3

2

0

A

0

. . . .

171

.. . . . .. . . 20 40 60 80

Oi Eccentricity (deg) OJ "0

&4 B (fJ

E 3 11

E

552

0 t5 ro

LL c 0 0 0 10 20 30 40 50 60 70 80 '@ u Azimuth (deg) ~

~4 C ro 15 ::;:;:

3

2

o -40 -30 -20 -10 0 10 20 30 40 Elevation (deg)

Fig. 9. Areal magnification factors in area 17 of the ferret. A: Areal magnification vs. eccentricity. Pooled data from F81 and F91. Smooth curve is the second-order polynomial that best fits these points. A single datum (eccentricity = 3.5, magnification factor = 7.48) has been omitted from this figure to save space, but this point was included when fitting the curve. B: Areal magnification vs. azimuth. Mean and standard error of the mean

served when data are pooled across elevations as shown, is also present when magnification factor vs. azimuth is plot­ted for a series of elevations (not shown).

Figure 9C shows magnification factor plotted as a func­tion of elevation. Magnification is greatest at -12.5 0 and decreases sharply above and below this elevation. When magnification factor vs. elevation is plotted separately for each of a series of azimuths placed at 50 intervals, a series of curves with peaks at -12.5 0 is obtained (data not shown).

The 17/18 border The border between areas 17 and 18 was delineated both

physiologically and anatomically. Physiology. Microelectrode recordings made in two fer­

rets to define the rostral border of area 17 on the posterior lateral gyrus revealed a reversal in the visual-field repre­sentation similar to that across the 17/18 border in other species. Figure 10 shows three examples of this transition. At successive electrode positions from rostral to caudal along row A, for example, fields 1-3 moved from peripheral to more central portions of the visual field while receptive fields recorded in the more caudal penetrations 3-6 re­versed the direction of progression and moved back toward the periphery of the contralateral visual field. We take this

(SEM) at each azimuth were calculated from the pooled data from F81 and F91 at all elevations. Number above each error bar is the number of areas that contributed to the mean. C: Areal magnification vs. elevation. Mean and SEM at each elevation were calculated from the pooled data from F81 and F91 at all azimuths.

progression of receptive-field positions to be evidence that the more rostral penetrations were within area 18 and the more caudal ones within area 17. Note also that, as in other species, receptive fields in area 18 were considerably larger than those in area 17. Possibly as a result of the large receptive fields, neurons and neuron clusters within area 18 responded much better to extremely rapid motion than did neurons in area 17. These changes in receptive-field size and character always occurred within 500 JLm of the point at which the progression in receptive-field position reversed (rows Band C of Fig. 10). Electrolytic marking lesions were made at the points of transition between areas 17 and 18 (marked with asterisks in the inset). These lesions allowed us to identify the physiologically defined border in histolog­ical sections (see Fig. 12D below).

Normal anatomy. In the cat four features visible in the normal anatomy of the visual cortex are useful in locating the border between areas 17 and 18 (Otsuka and Hassler, '62); 1) area 17 contains fewer large pyramidal cells in layer III; 2) at the border between 17 and 18, a cluster of large pyramidal cells is often found; 3) layer IV is nearly twice as thick within area 17 as within area 18; and 4) coarse bun­dles of myelinated fibers are more prominent within area 18. The Nissl preparation shown in Figure 12D illustrates

172 1\1.1. L.AW ET AJ~ .

10·

A 6 2

1

1

0 20·

2

7

-10·

6

GJ CD -20·

0 3

4

-30·

f ig. 10. Eleetrophyaiological detennination of the 17/18 border. Ph0to­graphic ineet on right IIhow:a tlw ,urf_ of an adult ferret. brain in the N'gion mapped. lnaet at IO"'er left. Bho .... doraal view of left. hemisphere (compan! to fil:. 10); bos indicate& area shown in phocogmphic inset: !ICa1e bar .. 2 mm. Entry poinl4 of 19 of 63 pe1\eU"ationa made in thiB region are indicat.ed in three 1'0 ...... : A. B, and C. Numbenl in each row indicate penetra-

the third of these feotu res clearly, and the first two may be seen with some study. Figure 11 shows a Wei! myelin prop­aration that illustrates the fourth feature. All of these features tha t distinguish area 17 from area IS in the cat appear to be useful as well in the ferret.

Bxperimentai anatomy, 'I'ransneuronal autoradio­graphic labelling of the geniculocortical projection was also

6

lion aites that ron-espond to numbered ~ptive liellb to the left.. Receptive roehl. in row B are llIown hatclled. Aateriak.s on photograph mark lbe 17n8 border as defined by the reve .... ] of dired.ion of ~ion of reoeptive fi81d poeitiona. Note lal""g<!r size of!ll"1!a I S Tf!OI!ptive fielcla (e.!:. , A~and A2L C, caudal; L, hI,teral: M. medial; R, rostral: LS, IIIt.eralaukua; 55, Wpra&yi · "ian IUlcua.

helpful in defining the 17118 border. Figure 12A shows a low·magnification dark-field photomicrograph ofa parasag­ittal section through the region of the 17118 border. Area 17 in the ferret receives a projection from the dLGN that is densely labelled with the trans neuronal method; the projec­tion to area 18 is labelled very much more lightly and possibly not throughout its fu ll extent. The vertical white

I<' I<; I{HET VI SUAL CORn:X

,--~~.

~·ig. 12. The 17/18 border. A: Low.power darkfield photomicrograph of a parasagitlal eeclion through the vi$ual cortex contralaoorl1\ to an eye which bad been inj.ecl.ed with 3H.proline. Note deTl8e labelling (light) of layer IV within area 17. Region betw..,.,n ..ertkal ... ·hite li~ enlarged in B. 8 : Enlargement of . ulOr1l<liograph in A Ihowing the 17118 border region. Electrolytic lesion (white arrow) marks 17118 border defined physiologically. Note eJ:pansion of labelling into upper and lower layen between Il.IIteriska. extending less than I mm into area 17 and Ie8/! than 2 mm into area 18. C: Micrograph shown in A photographiclIlly overexposed to demonstrate the

173

Fig. 11. Weil alain of pal1lSllgiual section through ferret visual corteJ:. Dotled line shows pial surface; area 17 to righl of white arrow; lU"ea 18 to len. As in cat. coanoe bundl .... of myelinal.ed fibers are mOTe prominent in area 18. Donal is up: caudal i. to right; acale bar equal, 200 "m.

" ~ .. :

.... '. '

upansion of labelling into upper and lower layers near the 17/ 18 border and extension of weak labelling of layer rv into llN!a 18. I): Nilllli alain of aerial section a<ljacent 10 eection Ilhown in A- C. Electrolytic lesion (black arrow) marks phYBiologically.tJe.fined border bet ... ·een llf""tQ 17 and 18. Ni. l _in ",veals histological tnlnsition between the two areas. n.e lin,,", on t M! photomicrograph delineate the ,ill; cortical layers. demonstrating greeter thickn". of layer IV in area 17 (right 'ide). Scale bara equal I mm. A, anterior; D. donal: P. posterior; V, ventral.

174

bars in Figure 12A delineat.e the region shown in higher magnification in Fil,,'ure 128. The adjacent section, contain­ing an electrolytic lesion (alTow) made at the physiologi­cally determined border, is shown in a Nissl preparation in Figure 12D. This lesion confirms that the physiological 171 18 border is whel'e the dense label ends.

Close exami nation of the autoradiographs revealed the unexpected finding that the transneuronal label was dis· tributed more widely within the upper and lower layers in the 1~2 mm near the 17/18 border than within either area 1701' 18. Figure 12C is a print of the same photomicrograph shown in 12A, exposed for a longer time in order to make visible the increased density of grains in layers II, ill, V, and VI near the 17/18 border. This is more clearly seen in the high-magnification darkfield photomicrograph of Fig­ure 12B. Labelling considerably above background was ob· served in all cortical layers within tbe region shown between the two asterisks in 12B.

Oculal' dominance columns in area 17 Auatomy. lntraocular injections of 3H_proline followed

by transneuronal autoradiography revealed a pattern of ocular dominance patches-within area 17. Transneuronally labelJed geniculocortical terminations were found to fill layer IV of the contralater al hemisphere of area 17 contin­uously in more peripheral representations corresponding to the monocular segment of the visual field and with a patchy distribution in the representation of the binocular segment of the visual field. In t he ipsilateral hemisphere, the label·

M.I. LAW ET AI..

ling was confined to patches within the postel'Olat.eral pole roughly complementary to patches seen in the other hemi­sphere. F igure 13 shows the transneuronallabelling of the two hemispheres in coronal sections taken about 1.8 mm from the caudal pole. Note the patchy labelling of layer IV and the particular clarity of the patches on the ipsilateral side, in wh.ich faint patchy labelling is also seen i,n lay· er VI.

The overall nature of the. geniculocortical termination pattern is best seen in the reconstructions in Figure 14 and 15. Figure 14 is an expanded reconstruction of serial hori· zontal sections. The right eye of the animal from which these sections were made was injected with 3H_pl'Oline 11 days prior to perfusion. In this figure drawings of sections from the cortex contralateral to the injl.>cted eye are at'rayed above those from the ipsilateral hemisphere such that the caudal portions of each reconstruction abut. allQwing the complementary pattern of t he patches to become more evi· dent. The widths of the ipsilateral eye patches varied be­tween 200 and 1,500 /tm when viewed in the horizonal plane.

Figure 15 shows the labelled regions from a ferret in which the visual cortex was physically flattened (see Mate­rials and Methods) before sectioning parallel to the plane of the flattened surface. Tracings of serial sections from this brain were superimposed to make this reconstruction, which makes visible the r ather striped pattern of ocular domi­nance patches. The period of the ocular dominance patches measured in t his brain was 820 /tm ± 25 /tm (SEMl.

Fig. 13. (kular dominance patches in coronal section through area 17 following transneuronlll transport of 2 n,Ci of ~H.pro1ine injected into Tight eye. Left (contralaUlral) hemisphere shown on left; right (i psilateral) hen,isphere shown on right). Dorsal is up. medial is toward the middle; scale bar I!<juals I mm.

F t<~ I{ nE'I' VISUAl , COUTEX 175

RIGHT VISUAL CORTE)(

Figure 14

Fig. 14. Expandlod reconst ruction of ocular dominance pateheo; in hori. zontal I!eCtion. Hemillphere oontralaterlll t.o lin eye that had been injected with 3n ·proline it Bilown above ipsilateral hemisphere. Ar(!u of dense autoradiographk label are drawn in blacl< while lightly labelled areu appcllor lIlippled. Note complementary rultUJ"e oIlabelling pattern in the two

hemispheres: patchell appear on ipsilateral side where holfl$ occur in the otherwUse uniform labelling of the oon trlliatel"ll.l hemisphere. Note alllO that ocular dominance paleh.,. occur only in the portion of area 17 repr.· .. imting cent ral ... i.wal fields. R, rostral; C, caudal; M, medial: L.latual.

'" '\,

IPSILATERAL

, C ... ~~J

T~-·8.

• L

1MM ~~~~ CONTRALATERAL i"0

L. Figure IS

Fig. IS. Pattern of lI.ulOl"adiographic labelling in layer IV of physically flattened yiaual cortex from a fe<Tet that had receh-ed intraocular injectiOD of:lJi'proline. Tracings T«Onstrocted from .erial !JectiOrul near ly tangt!ntial to flattened layer IV. The most pO!!teril}l" pOrti(lns of each hemillphere ",-ere only partially flftlumoo, and therefore the fidelity of the rt!COnstruction is lest aoo;urote in Lhelle relli\l""_ The nlC\Irultruction thus appea .. in four ~ and the allgnment of caeh of the four pam il ahown by indiyidual axeL

Axe. with solid line. indicate the orientation of the central. Ilatumed port ion of the hemiflPheres, while the uee: with broken IiI\e8 ahow the orientation of the partially flattened, po8I.erior portiOIlll. Label' ... in Figure 14. Note the nearly ronlinuous, I"OIIlIU.udally elongated ocular dominance pntchell. Note all!O lha~ ip.!lilllteral patche. are nlllTOwer ,h I} contralateral palchcft.

176

In two transneuronally labelled brains studied in sagittal section, the ipsilateral eye patches occupied 15 and 14% of the total volume oflayer IV in area 17, while the contralat­eral eye patches occupied 90% of the total volume. The ipsilateral projection thus represents approximately 17% of the volume ofthe contralateral projection.

Even within the binocular segment, transneuronally la­belled geniculocortical afferents serving the contralateral eye occupied a greater portion of area 17 than did those serving the ipsilateral eye: the labelled contralateral eye bands in Figures 13 and 15 tended to be wider than the gaps separating these bands, and wider as well than the labelled bands in the ipsilateral hemisphere. The volume of layer IV occupied by the afferents serving each eye within the most binocular segment (region extending halfway from the 17/18 border to the beginning of the monocular seg­ment) was measured on a series of autoradiographically labelled coronal sections. Approximately 77% of the most binocular region of the contralateral hemisphere was occu­pied by label while only 49% of the corresponding region was labelled ipsilaterally. The sum of these two figures, 126%, would indicate that there is some degree of overlap between afferents serving the two eyes. ThE) absolute mag­nitude of this overlap is difficult to determine from such autoradiographic material because of the somewhat arbi­trary definition of "borders" of the labelled patches, when, in fact, grain density declines gradually rather than abruptly (see LeVay et aI., '78). The relative values of 77 and 49% coverage by the contralateral and ipsilateral gen­iculocortical afferents are, however, probably genuine, as they were drawn by using the same criterion. These values would suggest that even within the binocular segments there should be a substantially greater influence of the contralateral than the ipsilateral eye. Physiological find­ings below are in accord with this suggestion.

Physiology. The relative efficacy of the two eyes was assessed physiologically within the binocular portion ofthe ferret visual cortex. Single units in area 17 with receptive fields within 10° of the vertical meridian were usually driven binocularly. Figure 16 shows an ocular dominance histogram compiled by using the seven-point scale of Hubel and Wiesel ('62) from 220 single units (all receptive field centers between + 20 and -30° elevation and < 15° azi­muth) recorded in ten ferrets. The overall greater domi­nance of the contralateral eye is evident from this histogram: about twice as many units favored the contra­lateral eye as favored the ipsilateral eye (contralateral bias index = 66, see Materials and Methods).

In multiunit recordings from layer IV, which constituted most of the mapping data, eye preference was much more marked, and responses were purely monocularly driven at 68% of the binocular segment recording sites (visual field coverage illustrated in Fig. 2). The multiunit data revealed that the two eyes are much more nearly equal in their influence within 10° of the vertical meridian (contralateral bias index = 59) than in more peripheral parts of the binocular representation (contralateral bias index = 78). More than four times as many recording sites were domi­nated by the contralateral eye as by the ipsilateral eye in peripheral portions of the binocular representation while the corresponding figure was less than twice for the central 10° of azimuth. A clustering of neurons according to eye preference was also evident physiologically. Recording sites within a penetration normal to the surface tended very strongly to have the same eye preference. Neighboring pen­etrations also tended to be similar in eye preference.

M.1. LAW ET AL.

DISCUSSION Topographic organization

Electrophysiological mapping techniques were used to determine the topographic organization of primary visual cortex in ferrets. From its border with area 18 on the pos­terior lateral gyrus, area 17 extends around the caudal pole of the hemisphere and over the splenial gyrus to the caudal bank of the splenial sulcus. The visuotopic map is oriented with the isoazimuth lines approximately parallel to the long axis of the posterior lateral gyrus and the isoelevation lines approximately perpendicular to the isoazimuths. Cen­tral azimuths are represented on the posterior lateral gyrus and peripheral azimuths are represented on the splenial gyrus; the inferior visual field maps medially and the su­perior visual field maps laterally.

How much of area 17 was mapped in this study? An estimated 83% of area 17 was mapped electrophysiologi­cally in each of the two animals from which maps were constructed from parasagittal section. This value is based on a comparison of the mapped areas to the area of area 17 defined by transneuronal labelling. The unmapped area includes the ventromedial splenial gyrus and the ventral portion of the caudal bank of the splenial sulcus. It can be inferred from Figures 6 and 8 that this region of area 17 contains the representation of the superior peripheral vi­sual field. The extent ofthe visual field represented in area 17 is best illustrated in Figure 2. Receptive fields of units in area 17 have been found with azimuths as great as 135° and elevations ranging from at least -40 to +70°.

Use of maps. These mapping experiments were under­taken in part to provide a guide for future studies of ferret area 17. Figures 4-6 should aid the neurophysiologist in relating locations in the visual field to sites in the cortex.

en --1

u:: 40

() 30 u.. o 20 f-

a] 10 () 0:: 0 W a..

57 46

39 31

20

1 2 3 4 5

contralateral equal <

6 7

ipsilateral >

OCULAR DOMINANCE Fig. 16. Ocular dominance histogram compiled from recordings of 220

single units in ten ferrets. Histogram plots percentage of units in each of seven ocular dominance groups defined by Hubel and Wiesel ('62): group 1 (seven) for units driven exclusively through contralateral (ipsilateral) eye' group 2 (six) for units strongly dominated by the contralateral (ipsilateral) eye; group 3 (five) for units weakly dominated by the contralateral (ipsilat· eral) eye; and group four for units driven nearly equally through the two eyes. Number of units in each bar of histogram is indicated above the bar.

FERRET VISUAL CORTEX

In order to locate a site in the cortex representing a known location in the visual field, the physiologist should refer to Figure 4 to find the reference number of the receptive field closest to the desired visual field location. The reference number can then be used to find the corresponding cortical site in the parasagittal sections of Figure 5 or in the flat­tened representation of area 17 in Figure 6. Similarly, to find the receptive field of a recording site more rapidly when the Horsley-Clarke coordinates of that site are known, the physiologist should first refer to Figure 5 to get the reference number of the appropriate receptive field, next look at the flattened representation in Figure 6 to find the approximate azimuth and elevation of that receptive field, and finally refer to Figure 4 for a more precise location of the receptive field. When it is necessary to adjust an elec­trode position, the following guidelines may be helpful. Caudal movement of the electrode results in a peripheral­superior (the main effect of the electrode movement on the receptive-field location being given first) shift in the recep­tive fields of sites recorded on the dorsal surface of area 17 and in a central-inferior shift in the receptive fields of sites on the tentorial surface. Lateral movement of the electrode results in a superior-central shift in the receptive-field lo­cation for sites recorded on either surface.

Comparison with the cat The cat and ferret have in common the general plan of

the carnivore visual system and geniculate lamination (Sanderson, '74). The two species have similar tapetal reti­nas, and a similar classification of retinal ganglion cells into a, {3, and 'Y types has been made on anatomical grounds (Vitek et aI., '85). The two species differ in that the ferret has much smaller, more laterally placed eyes, a much-less­pronounced area central is, and a stronger visual streak than does the cat (Vitek et aI., '85). Receptive fields in the ferret are also substantially larger than in the cat, as pre­dicted from the angular subtense of retinal ganglion cell dendritic fields, and the ferret has a smaller fraction of {3 cells and apparently lacks an ipsilaterally directed a-cell projection (Vitek et aI., '85; Zahs and Stryker, '85). Several features of ferret visual cortex are similar to the cat and differ from those of the monkey, including the laminar patterns of staining for cytochrome oxidase and acetylcho­linesterase and the extent of callosal connections (Rock­land, '85). The segregation of ON- and OFF-center cells in ferret LGN and cortex also differs from cat, although it is similar to the mink (Sanderson, '74; LeVay and McConnell, '82; Stryker and Zahs, '83; LeVay et aI., '87; Zahs and Stryker, '88).

Topographic organization. The topography of ferret area 17 may be compared to that of the cat, a much more widely studied carnivore. Figure 17 schematically compares the position and organization of the cat and ferret visual cor­tices. The organization of area 17 in the cat shown in Figure 17 A is redrawn from Tusa et aI. ('78). In both figures iso­elevation lines appear as dotted lines, while isoazimuth lines are solid. In the cat, the fixation point (the intersection between the vertical and horizontal meridians) occurs at the caudal-medial aspect of the dorsal surface of the hemi­sphere at the junction of the lateral and posterior lateral gyri (asterisk). In the ferret, the representation of the fixa­tion point occurs more laterally and ventrally, approxi­mately two-thirds of the way down the dorsolateral surface of the posterior lateral gyrus (asterisk). Viewed from its caudal aspect, the cat's brain could be made to resemble the ferret's brain by spreading the caudal poles upward and

177

A CAT

FERRET

B

Fig. 17. Comparison of the visual field representation within area 17 of the cat and of the ferret. A: Area 17 in the cat (redrawn from Tusa et aI., '78) resides more anteriorly and medially than area 17 in the ferret, shown in B. For each species, the representation of the fixation point is indicated by an asterisk.

outward from a hinge at the dorsal midline. The tentorial surface of area 17 in the ferret thus plays the role of the medial bank of area 17 in the cat.

The ferret's visual field extends farther along the horizon­tal meridian than does the cat's, as would be expected from the ferret's more laterally placed eyes. Recording sites were found in ferret area 17 with receptive fields having azi­muths as great as 135°; the cat's visual field extends to the 90° azimuth.

Magnification. There is an expanded representation of the central visual field within area 17 of ferrets. Such central magnification has been described in the cat as well (Tusa et aI., '78). However, cortical magnification factor changes much more gradually as a function of azimuth in ferrets than in cats, probably reflecting the ferret's visual streak (Vitek et aI., '85; Henderson, '85) vs. the cat's well­developed area centralis (Stone, '78).

In cats, cortical magnification factor appears to be deter­mined by the gradient of retinal ganglion cell density (Tusa et aI., '78). However, central magnification in macaque striate cortex appears to be greater than that which would be predicted on the basis of retinal ganglion cell densities or LGN volumes (Malpeli and Baker, '75). This study pro-

178

vides insufficient data to estimate the precise extent to which magnification factors in ferret cortex are determined by retinal ganglion cell densities. However, magnification factor in cortex does change in parallel with ganglion cell density in retina: cortical magnification factor changes more rapidly as a function of elevation than as a function of azimuth, just as retinal ganglion cell density changes more rapidly along the inferior-superior axis than along the na­sotemporal axis (Vitek et aI., '85). The similarity between the fraction of the LGN devoted to the ipsilateral eye (12, 16, and 19% in three cases studied by Zahs and Stryker, '85) and the fraction of cortical layer IV occupied by affer­ents serving that eye (14 and 15% in the two present cases) suggests that magnification changes little between LGN and cortex.

Monocular and binocular segments. The binocular seg­ment of the visual field is smaller in ferret than in cat, in accordance with the more lateral placement of the ferret's eyes. The binocular field in the cat encompasses up to 45 ° on each side ofthe vertical meridian (Sanderson, '71). Based on the sample of receptive fields illustrated in Figure 2, the ferret's binocular visual field extends to at least 38° azi­muth in the superior visual field and to at least 20° azi­muth in the inferior visual field. This estimate ofthe extent of the binocular visual field is consistent with that found in mapping studies ofthe LGN (Zahs and Stryker, '85).

The binocular segment occupies about half of the visual cortex in the ferret, leaving a large and accessible monocu­lar segment. By contrast, in the cat, only a rather inacces­sible 13% of area 17 is devoted to the monocular segment of the visual field (Tusa et aI., '78). Thus, the ferret may provide an excellent model system for investigations of binocular interactions and future studies that will compare the organizations of the monocular and binocular representations.

Ocular dominance columns. The present study provides anatomical and physiological evidence that binocular re­gions of area 17 are arranged into a pattern of ocular dominance columns similar to those observed in the cat (Shatz et aI., '77). These columns are more or less continu­ous and tend to be elongated in the rostrocaudal direction­roughly perpendicular to the border with area 18 over much of the dorsal surface of the posterior lateral gyrus. These columns differ from those in the cat in the smaller extent of the cortex in which they occur and in the greater degree of imbalance between the sizes of the two eyes' patches. In the cat, Shatz and Stryker (,78) measured the contralateral patches to be 33% larger than the ipsilateral patches while the present results show a 57% difference in size in the ferret. Rockland ('84) has used transneuronal transport of wheat germ agglutinin conjugated to horseradish peroxi­dase to reveal ocular dominance patches in the ferret, and McConnell and LeVay ('84) have evidence for them in a closely related species, the mink.

Relation to intrinsic cortical connections Rockland (,84) has found that local projections labelled

retrogradely by injections within area 17 of the ferret ap­pear as multiple patches in coronal section. Reconstructions of these patches reveal rostrocaudally elongated bands on the dorsal surface of the posterior lateral gyrus (Rockland, '85), reminiscent of earlier findings on tree shrew and pri­mate (Rockland et aI., '82; Rockland and Lund, '83). The similar course of ocular dominance bands in present results is consistent with the possibility that these patches might

M.1. LAW ET AL.

be related to the ocular dominance system in the ferret, although the present findings do not, of course, exclude other possibilities. Patches that may be similar in area 18 of the cat appear to be related to orientation rather than to ocular dominance columns (Matsubara et aI., '85), while in the tree shrew, no strict relation between such patches and deoxyglucose-Iabelled orientation columns was evident <Rockland et aI., '82). Further work will be required on the organization of orientation columns in the ferret to resolve this point.

Elevation of the fixation point Consideration of cortical magnification factors suggests

that the true fixation point elevation is below that assigned in Figures 2-8. In this study, the fixation point was chosen to make the cortical maps consistent with maps of the LGN previously published (Zahs and Stryker, '85). When con­structing the geniculate maps, the fixation point elevation was placed 3.5 ° below the elevation of the projection of the optic disc on the basis of the elevation of the smallest receptive fields found in two geniculate mapping exper­iments. There was, however, little change in the sizes of geniculate receptive fields between elevations 34° below to 16° above the optic disc projection, leaving the precise ele­vation of the fixation point relatively uncertain.

We failed in our attempts to locate the fixation point more precisely by relating the location of area centralis to the locations of retinal ganglion cells with known receptive fields. In two animals, we attempted to label retinal gan­glion cells corresponding to known regions of the visual field by making horseradish peroxidase injections within the LGN at sites with physiologically defined receptive fields. When the retinas of these animals were prepared in flat mount, we found that both the widespread extent of retrograde ganglion-cell labelling and the imprecision in locating area centralis on the basis of its higher cell density resulted in a broad range of estimates of location of the fixation point. Since these experiments failed to define the location of the fixation point more precisely than had our geniculate recordings, we do not illustrate them here.

Because of the presence of a visual streak, it is, however, reasonable to assume that magnification factor in ferret area 17 is greatest along the representation of the horizon­tal meridian. If this assumption is valid, the real fixation point elevation is approximately 12° below that assigned in the figures, or 15 ° below the elevation of the optic disc projection. This location is well within the range of eleva­tions provided by a consideration of geniculate receptive field sizes and the results of the retinal labelling exper­iments. This estimate is also consistent with that expected from relating the retinal whole mount data of Vitek et aI. ('85) to the extents ofthe visual field observed in our studies of the cortex and LGN (Zahs and Stryker, '85).

Neighboring visual areas During these experiments, electrode penetrations were

made into visually responsive areas rostral to area 17. Areas were distinguished by differences in the most effective stim­uli for eliciting responses and by reversals in the visuotopic map when penetrations crossed from one area to another. Two visual areas were encountered rostral to area 17 on the dorsal surface of the posterior lateral and lateral gyri; these areas are presumed to be areas 18 and 19. Visual responses were also recorded on the caudal bank of the splenial sulcus rostral to area 17. This area may be analo-

FERRET VISUAL CORTEX

gous to area 20b in the cat, described by Tusa et aI. ('78) as being located in the posterior portion of the splenial sulcus abutting the representation of the upper visual field in area 17. Alternatively, it may be analogous to the splenial visual area of Kalia and Whitteridge ('73).

Border with area 18 Direct geniculocortical input, as revealed in the present

study by transneuronally transported label following an eye injection of 3H -proline, goes primarily to area 17 in the ferret, although a very much weaker projection also can be observed in layer IV of at least part of area 18. The border region between areas 17 and 18 can be visualized in the ferret by using the same criteria applied to the cat (O'Leary, '41; Otsuka and Hassler, '62; LeVay and Gilbert, '76): that is, bundles of coarse myelinated fibers appear to enter area 18; layer IV in area 18 is half as thick as layer IV in area 17; larger pyramidal cells are often seen at the border; and area 17 contains fewer large pyramidal cells, especially in layer V, than area 18. Thus, the border between areas 17 and 18 can be identified in normal histological sections. Rockland ('85) has shown that an acetylcholinesterase stain also reveals the border between areas 17 and 18 with par­ticular clarity; the low levels in layer IV of area 17 stand out.

The 17/18 border in the ferret has one feature not promi­nent in many other species. Transneuronal autoradiogra­phy of the visual cortex following an injection into one eye reveals weak labelling above background of layers II, III, V, and VI, as well as IV, extending less than 1 mm into area 17 and about 2 mm into area 18 (Fig. 6). Such a pattern of labelling appears not to be present in mouse (Drager, '74), a gray squirrel (Weber et aI., '77), cat (Shatz et aI., '77; Shatz and Stryker, '78), bush baby (Hubel et aI., '76), owl monkey (Kaas et aI., '76), squirrel monkey (Hubel et aI., '76), or rhesus monkey (Hubel and Wiesel, '77), although if it were present but fainter than in the ferret, it might well have failed to show up on the published autoradiographs. Figure 2 of Hubel (,75) shows a tree shrew with a slight suggestion of a similar pattern of labelling. However, it is not possible to determine from the literature whether this is a general finding in the tree shrew. The significance of this finding is unknown, but a differential distribution of geniculocortical fibers to the border region between areas 17 and 18 in the ferret and the tree shrew may reflect the presence of some other input unique to this border region.

The transneuronally labelled geniculocortical projection to layer IV of area 18, as compared to the label in area 17, appears from the present results to be much less strong in the ferret than it is in the cat (Shatz et aI., '77). The gray squirrel (Weber et aI., '77) is the only species we have found in the literature in which a patern of labelling similar in this respect to that in the ferret appears.

Unresolved questions The present experiments focused on topographic organi­

zation and on ocular dominance columns. It will also be of interest to determine whether the receptive-field classes that are recognized in primary visual cortex in the cat and in other binocular mammals are present in ferrets. Exper­iments on this question and on the organization of orienta­tion columns in the ferret are in progress (Waitzman and Stryker, in preparation).

A particularly interesting question regarding the colum­nar organization of ferret visual cortex concerns the ar-

179

rangement of its ON- and OFF-center geniculate inputs. ON- and OFF-center responses are segregated into sublam­inae within the LGN (Stryker and Zahs, '83). The projection of these sublaminae onto the cortex takes the form of patches reminiscent of ocular dominance patches in layer IV (Zahs and Stryker, '88, in the ferret; LeVay et aI., '87, in the mink). We hope to understand the development and functional significance of such an organization.

The lack of an ipsilaterally directed a-cell projection in the ferret raises questions about parallel X and Y path­ways, their development, and effects of deprivation (Vitek et aI., '85). Area 18 in the cat is thought to receive exclu­sively Y-type geniculate input originating from the retinal a-cells (Sherman, '85). This is difficult to reconcile with the present finding that area 18 in the ferret has properties similar to those of the corresponding area in the cat, includ­ing cells driven by the ipsilateral eye. An answer to this difficulty for the cortex is that there appear to be cells with Y-like morphology and physiological properties in layer Al ofthe LGN, leaving open the question ofthe retinal inputs to these geniculate cells (Esguerra et aI., '87).

ACKNOWLEDGMENTS This work was supported by grants R01-EY-02874 and

K04-EY-00213 (to M.P.S.) and 5-F32-EY-05603 (to M.LL.) from the National Institutes of Health, by National Science Foundation and University of California Graduate Oppor­tunity Fellowships (to K.R.Z.), and by grants from the McKnight and March of Dimes Birth Defects Foundations. We are grateful to W. Jenkins for assistance with the Weil staining and to K.S. Rockland for comments on an earlier version ofthis manuscript.

LITERATURE CITED Bishop, P.O., W. Kozak, and G.J. Vakkur (1962) Some quantitative aspects

of the cat's eye: Axis and plane of references, visual field co-ordinates and optics. J. Physiol. (Lond.) 163:466-502.

Daniel, P.M., and kD. Whitteridge (1961) The representation of the visual field on the cerebral cortex in monkeys. J. Physiol. (Lond.) 159:203-221.

Drager, U.C. (1974) Autoradiography of tritiated proline and fucose trans­ported transneuronally from the eye to the visual cortex in pigmented and albino mice. Brain Res. 82:284-292.

Esguerra, M., A.W. Roe, and M. Sur (1987) Morphology of identified ferret LGN neurons characterized in vivo and in vitro. Soc. Neurosci. Abstr. 13:1434.

Guillery, R.W. (1971) An abnormal retinogeniculate projection in the albino ferret. Brain Res. 33:482-485.

Guillery, R.W., M. Ombrellaro, and A.L. LaMantia (1985) The organization of the lateral geniculate nucleus and of the geniculocortical pathway that develops without retinal afferents. Brain Res. 352:221-233.

Henderson, Z. (1985) Distribution of ganglion cells in the retina of adult pigmented ferret. Brain Res. 358:221-228.

Huang, K., and R.W. Guillery (1985) A demonstration of two distinct geni­culocortical projection patterns in albino ferrets. Brain Res. 352:213-220.

Hubel, D.H. (1957) Tungsten microelectrode for recording from single units. Science 125:549-550.

Hubel, D.H. (1975) An autoradiographic study of the retino-cortical projec­tions in the tree shrew (Tupaia glis). Brain Res. 96:41-51.

Hubel, D.H., and T.N. Wiesel (1962) Receptive fields, binocular interaction, and functional architecture in the cat's visual cortex. J. Physiol. (Lond.) 160:106-154.

Hubel, D.H., and T.N. Wiesel (1965) Receptive fields and functional archi­tecture in two non-striate visual areas (18 and 19) of the cat. J. Neuro­physiol. 28:229-289.

180

Hubel, D.H., and T.N. Wiesel (1977) Functional architecture of macaque monkey visual cortex. Proc. R Soc. Lond. [BioI.] 198:1-59.

Hubel, D.H., T.N. Wiesel, and S. LeVay (1976) Functional architecture of area 17 in normal and monocularly deprived Macaque monkeys. Cold Spring Harbor Symp. Quant. BioI. 40:581-587.

Jackson, C.A., and T.L. Hickey (1985) Use of ferrets in studies of the visual system. Lab. Anim. Sci. 35:211-215.

Kaas, J.H., C. Lin, and V.A. Casagrande (1976) The relay of ipsilateral and contralateral retinal input from the lateral geniculate nucleus to the striate cortex in owl monkey: A transneuronal study. Brain Res 106:371-378.

Kalia, M., and D. Whitteridge (1973) The visual areas in the splenial sulcus of the cat. J. Physiol. (Lond.) 232:275-283.

Law, M.l., and M.P. Stryker (1983) The projection of the visual world onto area 17 ofthe ferret. Invest. Ophthalmol. [Suppl.] 24:227.

LeVay, S., and C.D. Gilbert (1976) Laminar patterns of geniculocortical projection in the cat. Brain Res. 113:1-19.

LeVay, S., and S.K. McConnell (1982) ON and OFF layers in the lateral geniculate nucleus of the mink. Nature 300:350-351.

LeVay, S., S.K McConnell, and M.B. Luskin (1987) Functional organization of primary visual cortex in the mink (Mustela vision), and a comparison with the cat. J. Compo Neurol. 257:422-441.

LeVay, S., M.P. Stryker, and C.J. Shatz (1978) Ocular dominance columns and their development in layer IV of the cat's visual cortex: A quanti· tative study. J. Compo Neurol. 179:223-224.

Linden, D.C., RW. Guillery, and J. Cucchiaro (1981) The dorsal lateral geniculate nucleus of the normal ferret and its postnatal development. J. Compo Neurol. 203:189-211.

Lockard, B.l. (1985) The forebrain of the ferret. Lab. Anim. Sci. 35:216-228. Malpeli, J.G., and F.H. Baker (1975) The representation of the visual field

in the lateral geniculate nucleus of Macaca mulatta. Matsubara. J., M. Cynader, N.V. Swindale, and M.P. Stryker (1985) Intrin·

sic projections within visual cortex: Evidence for orientation·specific local connections. Proc. Natl. Acad. Sci. USA 82:935-939.

McConnell, S.K (1985) Migration and differentiation of cerebral cortical neurons after transplantation into the brains of ferrets. Science 229: 1268-1271.

McConnell, S.K., and S. LeVay (1984) Segregation of on· and off·center afferents in mink visual cortex. Proc. Natl. Acad. Sci. USA 81:1590-1593.

O'Leary, J.L. (1941) Structure of area striata of the cat. J. Compo Neurol. 75:131-161.

Oberdorfer, M.D., RW. Guillery, and H.M. Murphy (1977) Geniculocortical pathways in the mink that have abnormal retinogeniculate pathways. Anat. Rec. 187:669-670.

Olavarria, J., and V.M. Montero (1984) Relation of callosal and striate· extrastriate cortical connections in the rat: Morphological definition of extrastriate visual areas. Exp. Brain Res. 54:240-252.

Otsuka, R, and R Hassler (1962) Ueber aufbau und gliederung der cortica· len sehsphaere bei der katze. Arch. Psychiatr. Nervenkr. 203:212-234.

Reiter, H.O., D.M. Waitzman, and M.P. Stryker (1986) Cortical activity blockade prevents ocular dominance plasticity in the kitten visual cor­tex. Exp. Brain Res. 65:182-188.

Rockland, KS. (1984) Anatomical organization of area 17 in the ferret. Invest. Ophthalmol. [Suppl.] 25:211.

M.1. LAW ET AL.

Rockland, KS. (1985) Anatomical organization of primary visual cortex (area 17) in the ferret. J. Compo Neurol. 241:225-236.

Rockland, KS., and J.S. Lund (1983) Intrinsic laminar lattice connections in primate visual cortex. J. Compo Neurol. 216:303-318.

Rockland, KS., J.S. Lund, and A.L. Humphrey (1982) Anatomical banding of intrinsic connections in striate cortex of tree shrew (Tupaia glis). J. Compo Neurol. 209:41-58.

Sanderson, KJ. (1971) The projection of the visual field to the lateral geniculate and medial interlaminar nuclei in the cat. J. Compo Neurol. 143:101-117.

Sanderson, KJ. (1974) Lamination of the dorsal lateral geniculate nucleus in carnivores of the weasel (Mustelidae), raccoon (Procyonidae), and fox (Canidae) families. J. Compo Neurol. 153:239-266.

Sanderson, KJ., RW. Guillery, and RM. Shackelford (1974) Congenitally abnormal visual pathways in mink {Mustela vision) with reduced reti­nal pigment. J. Compo Neurol. 154:225-248.

Sanderson, KJ., and S.M. Sherman (1971) Nasotemporal overlap in the visual field projected to the lateral geniculate nucleus in the cat. J. Neurophys. 34:453-466.

Shatz, C.J., S.H. Lindstrom, and T.N. Wiesel (1977) The distribution of afferents representing the right and left eyes in the cat's visual cortex. Brain Res. 131:103-116.

Shatz, C.J., and M.P. Stryker (1978) Ocular dominance in layer IV of the cat's visual cortex and the effects of monocular deprivation. J. Physiol. (Lond.) 281:267-283.

Sherman, S.M. (1985) Functional organization of the W-, X-, and Y-cell pathways in the cat: A review and hypothesis. Prog. Psychobiol. Psy­chol. 2:233-313.

Stone, J. (1978) The number and distribution of ganglion cells in the cat's retina. J. Compo Neurol. 180:753-772.

Stryker, M.P., and KR Zahs (1983) ON and OFF sublaminae in the lateral geniculate nucleus of the ferret. J. Neurosci. 3:1943-1951.

Tusa, RJ., L.A. Palmer, and A.C. Rosenquist (1978) The retinotopic organi­zation of area 17 (striate cortex) in the cat. J. Compo N eurol. 177:213-236.

Tusa, RJ., A.C. Rosenquist, and L.A. Palmer (1979) Retinotopic organiza­tion of areas 18 and 19 in the cat. J. Compo Neurol. 185:657-678.

Van Essen, D.C., and J.H.R Maunsell (1980) Two-dimensional maps of the cerebral cortex. J. Compo Neurol. 191:255-281.

Vitek, D.J., J.D. Schall, and A.G. Leventhal (1985) Morphology, central projections and dendritic field orientation of retinal ganglion cells in the ferret. J. Compo Neurol. 241:1-11.

Weber, J.T., V.A. Casagrande, and J.K Harting (1977) Transneuronal trans­port of [3H]proline within the visual system of the grey squirrel. Brain Res. 129:346-352.

Wiesel, T.N., D.H. Hubel, and D.M.K Lam (1974) Autoradiographic dem­onstration of ocular-dominance columns in the monkey striate cortex by means of transneural transport. Brain Res. 79:273-279.

Zahs, K.R, and M.P. Stryker (1982) The projection of the visual field onto the lateral geniculate nucleus of the ferret. Soc. Neurosci. Abstr. 8:207.

Zahs, KR, and M.P. Stryker (1985) The projection of the visual field onto the lateral geniculate nucleus of the ferret. J. Compo Neurol. 241:210-224.

Zahs, KR, and M.P. Stryker (1988) Segregation of ON and OFF afferents to ferret visual cortex. J. Neurophysiol. 59:1410-1429.