Physiological segregation of geniculo-cortical afferents in the visual cortex of dark-reared cats

Brain Research, 362 (1986) 281-286

281

Elsevier BRE 11348

Physiological Segregation of Geniculo-Cortical Afferents in the Visual Cortex of Dark-Reared Cats N.V. SWINDALE and M.S CYNADER

Departments of Physiology and Psychology, Dalhouste Umversity, Hahfax, NS (Canada) (Accepted May 7th, 1985)

Key words" dark rearing - - visual cortex - - geniculate afferent - - conduction failure - - dark-reared cat - - ocular dominance

Transneuronai autoradlography in cats dark-reared from birth shows that gemculate terminals from left and right eyes are spread throughout layer IV in area 17 of the visual cortex In area 18 however, segregation into alternating left and right eye ocular dominance patches appears to be normal. We attempted to correlate this difference between the two areas with the ocular dominance distribution of physiologicalactivity of the geniculate afferents. To do this, mlcroelectrode recordings of multi-unit activity, believed to represent the activity of a number of genlculate terminal axons in the vicinity of the electrode tip, were made m layer IV of areas 17 and 18 in dark-reared cats. The ocular dominance distribution of this activity was predominantly monocular in area 18, as the anatomical results predicted, but surprisingly there were also many regions in area 17 where activity could only be elicited by stimulation of one eye. This has to be reconciled with the anatomical results showing that inputs from both eyes are present throughout layer IV in area 17 of dark-reared cats. Reasons for the discrepancy are discussed. INTRODUCTION In kittens less than two weeks old, the two sets of geniculate afferents from the left and right eyes are uniformly intermixed in layer IV of the visual cortex. During the following 2 - 3 weeks these inputs u n d e r g o a process of segregation that results in the formation of alternating left and right eye ocular d o m i n a n c e patches 13 and of physiologically defined ocular d o m inance columns s,9,25. If instead kittens are r e a r e d from birth in total darkness, ocular d o m i n a n c e patches, as shown by transneuronal autoradiography, fail to develop normally. In a r e a 17, afferents carrying information from left and right eyes are distrIbuted t h r o u g h o u t layer IV, with at most only slight periodic fluctuations in density 11,17,1s,22. In area 18 on the other hand (Fig. 1), the two eyes' afferents appear to segregate into a n o r m a l p a t t e r n of alternating left and right eye patches17.ls, 2a. W e studied the physiological consequences of the redistribution of geniculate afferents brought a b o u t by dark-rearing, by recording extracellularly with low impedance microelectrodes in layer IV, where

the geniculate afferents terminate. It p r o v e d easy to record vigorous and non-habituating responses to visual stimuli from neuronal elements in this layer. Since cortical neurons are relatively unresponsive and habituate rapidly in d a r k r e a r e d cats 4,I0 we believe that this visually e v o k e d activity resulted mainly from action potentials in geniculate axons. The evoked activity was always multi-unit, p r o b a b l y reflecting the presence of action potentials in a n u m b e r of geniculate axons close to the tip of the recording electrode. A simdar technique was used by McConnell and LeVay15 to record from geniculate axons in layer IV of the mink after silencing the actwity of cortical neurons with rejections of kainic acid Given the overlapping distribution of left and right eye inputs in area 17 of d a r k - r e a r e d animals, we expected to be able to evoke binocular responses throughout area 17 with few, if any, monocular responses. In area 18 on the other h a n d we expected the m a j o r i t y of responses to be monocular. W e report here that, on the contrary, regions where responses were m o n o c u l a r were equally c o m m o n in both areas.

Correspondence: N V Swindale, Department of Physiology, Tupper Budding, Dalhousie University, Hahfax, NS, Canada B3H 4J1 0006-8993/86/$03 50 © 1986 Elsevier Science Publishers B.V (Biomedical Division)

282

Fig. 1. Dark-field autoradiograph of a horizontal section through the visual cortex of a cat dark reared for one year, following Injection of trttmted prohne into the ~psflateral eye The arrow marks the boundary between areas 17 and 18. Note the nearly continuous d~stnbutlon of label in layer IV m area 17 and the patchy &stnbut~on m area 18. (Taken from data presented by Swlndale23)

MATERIALS AND METHODS Recordings were made from areas 17 and 18 m 6 dark-reared cats. These animals had been reared in total darkness from the time of natural eye opening for periods ranging from 3 months to 3 years. Preparation for recording was by conventional means3.4: anesthesia was reduced with i.v. sodium pentothal, a tracheotomy was performed, and anesthesia subsequently maintained by respiration under paralysis with 70% N20 and 30% 02. In some ammals this anesthesia was supplemented with doses of i.v. sodium pentothal: this helped to further depress the responsiveness of cortical neurons. Contact lenses with 4 mm artificial pupils were selected by retinoscopy to focus the eyes onto a tangent screen about 1 5 m m front of the animal. A large hole in the skull was made, exposing about 0.5 cm2 of visual cortex on the left- or right-hand side, from which the overlying dura was removed. Two methods for recording from layer IV were used. For area 18, and for area 17 where it is exposed at the posterior end of the marginal gyrus, multiple penetrations normal to the surface of the cortex were made. With this method, one ocular dominance measurement per penetration could be obtained A photograph, or a drawing of the blood vessel pattern on the cortical surface served as a reference for establishing the position of electrode penetrations

These were spaced at roughly 200-300/~m intervals over the surface of the cortex Layer IV was identified by the appearance of visually evoked responses at penetration depths between 800 and l l 0 0 # m . The other method, used exclusively for the area 17 measurements, involved making a smaller number (3-6 per animal) of long penetrations down the medial bank of the marginal gyrus. Visual responsiveness was tested at intervals of 100 ~m along the electrode track Ocular dominance was assessed by listening to responses evoked by flashing or moving small spots of hght across the multi-unit receptive field. Response strength in the dominant eye was measured on a scale from 1 to 4, 1 being a barely discernible response and 4 being a strong one. Ocular dominance was measured on Hubel and Wiesel's 7-point scale. One problem m measuring ocular dominance if overall response strength is weak is that monoculanty may be exaggerated. For example, a response might be categorized with an ocular dominance in group 2 if the response in the dominant eye is strong and a response m the non-dominant eye can just be detected. If overall sensitivity of the measurements declines, the response m the non-dominant eye may no longer be detectable and the recording would be classified as group 1 To avoid this problem, measurements showmg response strengths of 1 m the dominant eye were &scarded from the final data analysis. However, the

283 ocular dominance distribution of measurements with response strengths of 1 or 2 turned out to be no more monocular than that in categories 3 and 4. At the end of each experiment the animal was given a lethal dose of barbiturate and perfused with 4% paraformaldehyde. Unstained freshly sectioned tissue was examined under a dissecting microscope: this allowed drawings to be made of the electrode tracks and of the upper and lower boundaries of layer IV. To confirm the boundaries of layer IV, some sections were subsequently stained for cytochrome oxidase which serves as a marker for this layer 26. Ocular dominance histograms were compiled taking one measurement from each of the vertical penetratlons through layer IV. In the tangential penetrations in area 17 we had data from points spaced 100 pm apart. Because measurements of ocular dominance from adjacent points were obviously correlated, the number of measurements in each ocular dominance category was divided by two. This effectively weighted the measurements from tangential penetrations in a way that was comparable to those made from the vertical penetrations, which were spaced about twice this distance (i.e. 200-300pm) apart.

sponses could be obtained until depths of 500-800 /~m, at which point visually evoked activity became detectable. This activity was almost always multiunit, with individual units being impossible to resolve either aurally or by examining the oscilloscope trace. The responses differed from those of cortical neurons in being: (a) insensitive to the orientation or direction of motion of a bar stimulus; (b) in following repeated flashes without habituation up to several Hz; (c) in having circular receptive fields where on and off responses, if both were present, were invariably superimposed; and (d) in being most sensitive to a spot stimulus moved or flashed in the centre of the receptive field. Increasmg the size of the spot beyond the receptive field size produced a decrease in the multi-unit response. Responses with these properties contmued as the electrode was advanced, with no change m ocular dominance for a further 200-300 /~m, at which point we presume the electrode left layer IV and responses either vamshed, or became sigmflcantly weaker. S~milar behaviour was found in the penetrations down the medial bank of area 17, except that ocular dominance changed as the electrode was advanced, with the regions of high visual responsiveness invariably correlating with the histologically determined location of layer IV. An ocular dominance histogram derived from 170 measurements m area 18 in 5 dark-reared animals, made as described above, is shown in Fig. 2a. As would be expected on the basis of the results of transneuronal autoradiography, this shows a preponderance of monocular responses over binocular ones, and an overall dominance of input from the contralateral eye. A similar histogram based on 191 measurements of

RESULTS Neurons in the visual cortex of cats that have been reared tn the dark are known to be at best poorly and sluggishly responsive to visual stimuli and to habituate rapidly 1,2,4,5,10,14,20. Under the conditions of our experimentation, cortical neuronal responses were nearly always undetectable or hard to evoke. Typically as the electrode was advanced through the upper layers of the cortex in a vertical penetration, no re6050-

(a)

50-

40-

(b)

(c)

40-

30-

3020-

20-

10-

10-

O1area 218

3 4 5 d a r k reared

6

7

O-

I

2010-

1 1

2 area 17

3 4 5 dark reared

6

7

O-

1

2 3 area 17

4 5 normal

6

7

Fig 2 Ocular dominance histograms a genlculate responses in layer IV of area 18 in dark-reared cats. b. gemculate responses in layer IV of area 17 in dark-reared cats c: genlculate responses in layer IV of area 17 in two normal cats m which cortical responses were suppressed by a relatively large dose of barbiturate anaesthetic

284 ocular dominance in area 17 is shown in Fig. 2b. This also shows a high proportion of monocular points, 45% of all the measurements being either group 1 or 7. The slight overall predominance of input from the lpsilateral eye is unexpected, although its stanstlcal slgmficance is uncertain. In one long penetration through layer IV, a periodic alternation between left and right eye dominated responses was observed (Fig. 3) Recordings from tangential penetrations through layer IV in two normal cats (160 measurements), where the visual responsiveness of cortical neurons was depressed by use of a relatively high dose of barblturate, showed an ocular dominance distribution similar to that of the dark-reared animals (Fig 2c), with 44% of points being monocular DISCUSSION It seems unlikely that the activity we recorded in layer IV originated from cortical neurons. In cats dark-reared for several months or more a large proportion of cortical neurons is visually unresponsive2,4,sA4,20. A m o n g those that do respond however, a proportion is orientation selective, while there is a still larger proportion of units that are non-oriented with large circularly symmetric receptive fields 4.520. There are several reasons why such responses are unlikely to have contributed to our recordings. Firstly, the responses of cortical neurons in dark-reared cats are usually weak, slow in onset and habituate rapidly,

ocuI,q r domin i

,s""

"'"

i

ln~ i

4-

• o ~,.

V

• o •

D

ot '

,IV

•

~i

Fxg. 3 Drawing of a parasaglttal section through area 17 of a 3-year-old dark-reared cat, showing the boundaries of layer IV determined by cytochrome oxidase staining, and an electrode penetration made m the plane of the section. On the right is a graph showing a periodic alternation in the ocular dominance of responses recorded as the electrode passed through layer IV m the region indicated.

and could easily be swamped by a gemculate response. This is especnally so m the present case, because in assessing responses for ocular dominance we used stimuh - - rapidly flashing or moving small spots of light - - that are relatively ineffective in driving cortical units, but are optimal for geniculate neurons, which retain normal response properties during dark rearing 16. Secondly, neither orientation nor direction selectivity, characteristics of at least some cortical units in dark-reared cats, was observed m any of the responses on which our measurements of ocular dominance were based. Thirdly, if cortical responses were contributing to our measurements it should have been possible to evoke activity in all layers of the cortex. However, we were never able to evoke visually responsive multi-unit activity m the upper layers of the cortex, the single units that were occasnonally isolated invariably faded to respond to the stimuli that were used routinely to activate the layer IV multi-unit activity. The weak responses that could sometimes be evoked below layer IV are most likely to have been due to the presence of a relatively low density of geniculate terminals in layer VI ~2 or to the presence of axons en route to layer IV from white matter. Finally, the ocular dominance distribution obtained from conventional recordings from isolated single units in the visual cortex of dark reared cats is predominantly binocular 4.5, quite unlike the monocular distribution obtained here. One other factor affecting the interpretation of our results is the possibility that some component of the recorded responses might be postsynaptlc in origin (e.g. due to local subthreshold depolarization of dendrites) One would expect such responses to be much smaller in amplitude than those due to the presynaptic action potentials however Another pmce of evidence against this possibility comes from the results of Singer and Tretter20 who showed by measurements of latency of responses to electrical snmulatlon that dark-rearing has a primary effect on conduction across the geniculo-cortical synapse, rather than at later stages in cortical processing Assuming that our measurements of ocular dominance derive from the presence of action potentials m genlculate afferent axons m layer IV, how can our finding of a high proportion of monocular responses from area 17 of dark-reared cats be reconciled with the anatomical findings~L17,18,22 that both left and

285 right eye afferents from the geniculate are to be found throughout layer IV? One complication affecting the interpretation of autoradiographs such as that in Fig. 1, is the existence of spillover in the lateral geniculate nucleus. This occurs when radioactive label transported from one eye leaks into geniculate laminae innervated by the unlabelled eye. This results in a more uniform distribution of label in layer IV of the visual cortex than would otherwise be present13. The possibility that increased spillover might account for the more continuous pattern of labelling seen in area 17 of dark-reared cats seems unlikely however. Firstly, measurements of spillover (refs. 18, 23, Swindale, unpublished results) have not shown any difference between dark-reared and normal animals. Secondly, segregation of geniculate inputs appears to be normal in area 18 of the dark-reared cats, although increased spillover would be expected to degrade the labelling pattern equally in both area 17 and 18. It could be objected that spillover might be more severe in geniculate X-cells which innervate area 17, than in Y-cells, which are the predominant source of geniculate input to area 18 (ref. 6). However, X-cells are much more numerous than Y-cells, and it is likely that splllover measurements reflect mainly the behaviour of this cell type. While the autoradiographic results from area 17 of dark-reared animals do show periodic fluctuations in the density of inputs from the ipsilateral eye, these are small in amplitude, and not present in all regions of cortex12,2L It seems unlikely that they could account for more than small fluctuations in ocular dominance. Inputs from the contralateral eye are probably even more uniformly distributed than those from the ipsilateral eye, even allowing for the greater degree of spillover on the contralateral side 13 and it is even more difficult in this case to explain the existence of a large proportion of regions in area 17 where no input from the contralateral eye could be evoked at all, despite the presence of a vigorous response from the ipsilateral eye. Our results are also consistent with the

REFERENCES 1 Blakemore, C and Van Siuyters, R C, Innate and environmental factors in the development of the kitten visual cortex, J Phystol (London), 248 (1975) 663-716. 2 Buisseret, P. and Imbert, M, Vtsual cortical cells: their de-

unpublished observation of Thompson et al. 24 that ocular dominance patches are detectable in deoxyglucose autoradiographs from dark-reared cats stimulated through one eye. If the foregoing arguments are accepted, then we can suggest two likely explanations of our results. One is the possibility that the electrodes used recorded selectively from a sub-population of geniculate afferents which was segregated, while a remaining more numerous and unsegregated population remained unrecorded. The segregated population might have been the Y-cell afferents, which are segregated in area 18, and have larger axons, perhaps making them more amenable to extracellular recording. Another possible explanation of our results is conduction failure. If this were to occur in some regions of the geniculate terminal arborizations, they would be incapable of giving rise to physiologically detectable activity, though perhaps still capable of transporting anatomical labels. Conduction failure is believed to occur in the axonal arborizations of Ia afferents to spinal motoneurons in normal cats 7 and in other neuronal systems19,21. Given a tendency for conduction failure to occur, it seems quite probable that there might be a stage during normal segregation of geniculate afferents when parts of axons that are being removed, or will ultimately be removed, become electrically silent It is possible that this stage of electrical silence might precede the removal or retraction of the axon during normal development: dark-rearing may perhaps arrest the process of segregation at this intermediate stage.

ACKNOWLEDGEMENTS We thank L. Colpitts, technical assistance and on the manuscript. The Grants PG-29 (MRC) M.S.C.

J. Hickey and G. Troop for J. Matsubara for comments research was supported by and A9939 (NSERC) to

velopmental properties in normal and dark-reared kittens, J Physiol. (London), 255 (1976) 511-525 3 Cynader, M S. and Berman, N., Receptive-field organisaUon of monkey superior colliculus, J. Neurophystol., 35 (1972) 187-201 4 Cynader, M.S., Berman, N. and Hem, A , Recovery of

286 tunction in cat visual cortex following prolonged deprivation. Exp Bram Res. 25 (1976) 139-156 5 Fregnac. Y and Imbert. M . Early development of visual cortical cells m normal and dark reared kittens relationship between orientation selectivity and ocular dominance. J Phystol (London). 278 (1978) 27-44 6 Garey. L J and Blakemore. C . The effects of monocular deprivation on different neuronal classes m the lateral gemculate nucleus of the cat. Exp Bram Res. 259 (1977) 259-278 7 H e n n e m a n . E . Luscher. H - R . Mathls. J . Simultaneously active and mactwe synapses of single la fibres on cat spinal motorneurones. J Phystol (London). 352 (1984) 147-161 8 Hubel. D H and Wlesel. T N . Receptive fields, binocular interaction and functional architecture in the cat's visual cortex. J Phy, tol (London). 160 (1962) 106-154 9 Hubel. D H . Wlesel. T N and LeVay. S . Plasticity of ocular dominance columns in monkey strlate cortex. Phil Trans R Soc London Ser B . 278 (1977) 377-409 10 Imbert. M and Bmsseret. P . Receptive field characteristics and plastic properties of visual cortical cells in kittens reared with or without visual experience. Exp Bram Res. 22 (1975) 25-36 11 Kahl. R E . D e v e l o p m e n t of ocular dominance columns in cats reared with binocular deprivation or strabismus. Soc Neuroscl Abstr. 8 (1982) 4 12 LeVay. S and Gilbert. C D . Laminar patterns of gemculocortical projection in the cat. Brain Research. 113 (1976) 1-19 13 LeVay. S . Stryker. M P and Shatz. C J . Ocular dominance columns and their development in layer IV of the cat's visual cortex a quantitative study. J Comp Neurol. 179 (1978) 223-244 14 Leventha]. A G and Htrsch. H V B . Receptive field properties of different classes of neurons in visual cortex of normal and dark reared cats. J Neurophystol. 43 (1980) 1111-1132 15 McConnell. S K and LeVay. S . Segregation of on- and off-centre afferents in mink wsual cortex. Proc Natl Acad

Scl U S A . 81 (1984) 1590-1593 16 Mower. G D . Burchfield. J L a n d D u f f y . F H . T h e effects of dark rearing on the development and plasticity of the lateral gemculate nucleus. Dev Brain Res. 1 (1981) 418-424 17 Mower. G D . Christen. W G and Kaplan. C J . Absence of ocular dominance columns in binocularly deprwed cats. Invest Ophthalmol. S u p p l . 25 (1984) 214 18 Mower. G D . Caplan. C J . Christen. W G and Duffy. F H . Dark rearing prolongs physiological but not anatomical plasticity of the cat visual cortex. J Comp Neurol. 235 (1985) 448-466 19 Parnas. I . Differential block at high frequency of branches of a single axon innervating two muscles. J Neurophystol. 35 (1972) 903-914 20 Singer. W and Tretter. F . Receptive field properties and neuronal connectivity in stnate and parastrmte cortex of contour deprived cats. J Neurophystol. 39 (1976) 613-630 21 Smith. D O . Mechanisms of action potential propagation fadure at sites of axon branching in the crayfish. J Phystol (London). 301 (1980) 243-259 22 Swlndale. N V . Absence of ocular dominance patches in dark reared cats. Nature (London). 290 (1981) 332-333 23 Swmdale. N V . The effects of restricted visual experience on the development of ocular dominance patches in the cat. Soc Neuroscl Abstr . 8 (1982) 297 24 T h o m p s o n . I D . Kossut. M and Blakemore. C . Development of orientation columns in stnate cortex revealed by 2-deoxyglucose autoradiography. Nature (London). 301 (1983) 712-715 25 Wlesel. T N . Hubel. D H and L a m . D M K . Autoradlographic demonstration of ocular-dominance columns m the monkey stnate cortex by m e a n s of transneuronal transport. Bram Research. 79 (1974) 273-279 26 Wong-Raley. M . Changes in the visual system of monocularly sutured or enucleated cats demonstrable with cytochrome oxtdase histochemistry. Bram Research. 171 (1979) 11-28