002X 1932 Y, s30”+“00 Prrp‘mlon Pmr plc
CONTRIBUTION OF STRIATE CORTEX AND THE SUPERIOR COLLICULUS TO VISUAL FUNCTION IN AREA MT, THE SUPERIOR TEMPORAL POLYSENSORY AREA AND INFERIOR TEMPORAL CORTEX CHARLES G. GROSS Princeton University, Princeton. N.J. 08540, U.S.A.
studied the visual responses of single neurons in three extra-striate visual areas of the macaque following lesions of striate cortex, lesions of the tecto-pulvinar system or both. After striate lesions, there was (a) considerable specific activity remaining in area MT including direction selectivity, (b) only non-specific activity in the superior temporal polysensory area (STP). and (c) no visual responsiveness at all in inferior temporal cortex (IT). In animals with striate lesions. interruption of the tecto-pulvinar pathway eliminated the residual visual activity in MT and STP that survived the striate lesions. Interruption of the tecto-pulvinar pathway alone had little or no effect on visual evoked activity in any of the three areas. These results are related to the relative dependence of visual responsiveness in MT, STP and IT on striate cortex and the superior colliculus, to differences between the dorsal and ventral cortical processing streams. and to neural mechanisms underlying blind sight. Abstract-We
INTRODUCTION IN THE last 20 years there has been a virtual explosion in the number of cortical visual areas identified in primates and a concomitant increase in their known interconnections. A recent summary listed over 30 such areas with over 250 pathways among them [32]. As a result, it sometimes seems that a major pastime of extra-striate visual cortex mavens is the diagramming of these connections, diagrams that often resemble the circuit of a deranged computer, a Tokyo subway map or a plate of spaghetti and square meatballs. Common to virtually all these connectional schemata is that they are confined to the cortical mantle and begin at striate cortex. That is, there is the implicit assumption that all primate cortical visual areas are ultimately dependent on and only on striate cortex for their visual input. Yet, in addition to their input from striate cortex, the extra-striate visual areas receive projections from the pulvinar which, in turn, receives projections from the superior colliculus [8,9,20, 49, 761. They also receive a sparse connection directly from the dorsal lateral geniculate which, in turn, receives a tectal as well as retinal input [lo, 23,34,38, 1 IO]. These tecto-fugal routes to visual cortex presumably make some functional contributions to extra-striate visual cortex: they are unlikely to serve merely decorative or structural purposes. In order to assess the relative contributions of striate cortex and the tecto-pulvinar to extra-striate function in the macaque monkey, we compared the effects of damage to each on the visual properties of neurons in three extra-striate visual areas, uiz. area MT (or V5), the superior temporal polysensory area (or STP) and inferior temporal cortex (IT; cytoarchitectonic area TE).
49X
(‘. of MT cells not only atill respond to visual stimuli. but. in addition. she\\
STKIATL
(‘0Klt.X
A\11
SI PtKIOK
(‘OLLI(‘bLLS
1% VISl’\L
501
tUN(‘TION
40
VM
HM LESION ZONE 80
10 lj
0
I
I
I
0
360
180
90
DIRECTION
270
360
(deg.)
ilA&d
gohil&lkg
ibdiu 270
160
Responsesfrom a single
MT neuron chnractewcd as gi~inf a “strong” response. Pcristimulus histqrams for each direction of stimulu motion arc sho\cn in a semicircle around the bottom of the plots of rcaponscs vs stimulus direction. The rcsponsc plotted was the average spike rate in a tlmc window corrcspondinf to the width of the receptlvc field along it5 narrowest axis. The rcccptive Geld relative to the lesion zone ia shown in the schematic to the right. The dashed line indutcs the rate of spontaneous actiwty. The direction ofstimulus motion has been normalixd to 180 for the prefcrrcd dlrcction. The \tlmulus was a 0.5 x I5 bar of light mowng at 8 xc. Each histogram is baaed on 10 trials [7X]. Fig.
?.
the directional selectivity characteristic of MT cells in normal monkeys. This residual responsiveness is eliminated by superior colliculus lesions. The colliculus could provide visual information to MT either by way of the pulvinar or the lateral geniculate body and extra-striate cortex. However, neither tectal 137. 821 nor geniculate cells show much sensitivity to the direction of stimulus movement and the littlc directionality found in the lateral and inferior pulvinar is eliminated by striate lesions 161. Thus. the directional selectivity that survives striate lesions MT neurons must bc generated rlc JIOI‘Oin MT. WLRTZ and his collcagues 157, 581 came to a similar conclusion on the basis of their analysis of the rcceptivc field properties of MT cells. The second implication concerns blindsight, a term used to describe the surprising amount of visually guided behavior that survives damage to striate cortex in humans and, by analogy, the similar behavior that survives striate lesions in monkeys. These residual abilities include the ability to detect and IocaliLe visual stimuli and to discriminate the speed and direction of stimulus movement 131, 50, 62, 66. 68, 69. 74, 75,90, 107, 109, I I I]. Because of the spatial nature of thcsc residual abilities, the apparent dependence of the residual vision on the superior colliculus [6 2, 901, and because of the unconscious nature of the preserved vision in humans, it has usually been assumed that blindsight reflects subcortical processing [19, 62, 90. 1071. However. there is a close similarity between the properties of MT units remaining after striate lesions and one type of visually-guided behavior remaining after striate lesions.
C‘. G. GKOSS
6
-
0
90
270
180
DIRECTION
360
(deg.)
r~iz.the ability to detect the direction and speed ofmovement. Thus. MT may bc responsible for the ability of monkeys and humans to detect direction of stimulus movcmcnt within a scotoma produced by ;I striate Icsion. That some cortical mechanism, as opposed to a midbrain one, must be responsible for this function is indicated by two sets of findings. The first is that there is little 1371 or no 1821 directional sensitivity among tectal neurons in primates. The second is that hcmisphercctomi/ed patients cannot dctcct direction ol movement in their blind field (presumably because they have no contralateral MT). but thoy can detect target location and the presence of movcmcnt (presumably with their midbrain mechanisms) 168, 751. Further support for an extra-striate but cortical mechanism fol analysis of direction of movement is that the lasting delicit in makinp visually guided saccades in monkeys after near total unilateral lesions of the cerebral cortex, including MT. i\ grcatcr than after either unilateral or bilntcral striate Icsions [WI. (Rccovcry of the ability to make visual saccades after bilateral striate Icsions is actually markedly greater than after unilateral ones [X6, I I I],) Similarly. optokinetic nystagmus i\ more dcvastcd h! hcmidecortication in monkeys than by occipital lohcctomy [9X]. The paradoxical aspect of thcsc results is that whereas tectal Icxions have virtually no cffcct on the activity of MT cells in normal animals. they have a dc\,astatinp clrcct in animals u ith striate Icsions. This phenomenon. in which the superior colliculus 5ccms to bccomc important only in the absence ofstriate cortex, is not unique to arc;1 MT, as will bc discussed bCl(~W. SIIPERIOR
I EMPORAL
POI.YSENS(_)RY
ARl
CONTRIBUTION OF STRIATE CORTEX AND THE SUPERIOR COLLICULUS TO VISUAL FUNCTION IN AREA MT, THE SUPERIOR TEMPORAL POLYSENSORY AREA AND INFERIOR TEMPORAL CORTEX CHARLES G. GROSS Princeton University, Princeton. N.J. 08540, U.S.A.
studied the visual responses of single neurons in three extra-striate visual areas of the macaque following lesions of striate cortex, lesions of the tecto-pulvinar system or both. After striate lesions, there was (a) considerable specific activity remaining in area MT including direction selectivity, (b) only non-specific activity in the superior temporal polysensory area (STP). and (c) no visual responsiveness at all in inferior temporal cortex (IT). In animals with striate lesions. interruption of the tecto-pulvinar pathway eliminated the residual visual activity in MT and STP that survived the striate lesions. Interruption of the tecto-pulvinar pathway alone had little or no effect on visual evoked activity in any of the three areas. These results are related to the relative dependence of visual responsiveness in MT, STP and IT on striate cortex and the superior colliculus, to differences between the dorsal and ventral cortical processing streams. and to neural mechanisms underlying blind sight. Abstract-We
INTRODUCTION IN THE last 20 years there has been a virtual explosion in the number of cortical visual areas identified in primates and a concomitant increase in their known interconnections. A recent summary listed over 30 such areas with over 250 pathways among them [32]. As a result, it sometimes seems that a major pastime of extra-striate visual cortex mavens is the diagramming of these connections, diagrams that often resemble the circuit of a deranged computer, a Tokyo subway map or a plate of spaghetti and square meatballs. Common to virtually all these connectional schemata is that they are confined to the cortical mantle and begin at striate cortex. That is, there is the implicit assumption that all primate cortical visual areas are ultimately dependent on and only on striate cortex for their visual input. Yet, in addition to their input from striate cortex, the extra-striate visual areas receive projections from the pulvinar which, in turn, receives projections from the superior colliculus [8,9,20, 49, 761. They also receive a sparse connection directly from the dorsal lateral geniculate which, in turn, receives a tectal as well as retinal input [lo, 23,34,38, 1 IO]. These tecto-fugal routes to visual cortex presumably make some functional contributions to extra-striate visual cortex: they are unlikely to serve merely decorative or structural purposes. In order to assess the relative contributions of striate cortex and the tecto-pulvinar to extra-striate function in the macaque monkey, we compared the effects of damage to each on the visual properties of neurons in three extra-striate visual areas, uiz. area MT (or V5), the superior temporal polysensory area (or STP) and inferior temporal cortex (IT; cytoarchitectonic area TE).
49X
(‘. of MT cells not only atill respond to visual stimuli. but. in addition. she\\
STKIATL
(‘0Klt.X
A\11
SI PtKIOK
(‘OLLI(‘bLLS
1% VISl’\L
501
tUN(‘TION
40
VM
HM LESION ZONE 80
10 lj
0
I
I
I
0
360
180
90
DIRECTION
270
360
(deg.)
ilA&d
gohil&lkg
ibdiu 270
160
Responsesfrom a single
MT neuron chnractewcd as gi~inf a “strong” response. Pcristimulus histqrams for each direction of stimulu motion arc sho\cn in a semicircle around the bottom of the plots of rcaponscs vs stimulus direction. The rcsponsc plotted was the average spike rate in a tlmc window corrcspondinf to the width of the receptlvc field along it5 narrowest axis. The rcccptive Geld relative to the lesion zone ia shown in the schematic to the right. The dashed line indutcs the rate of spontaneous actiwty. The direction ofstimulus motion has been normalixd to 180 for the prefcrrcd dlrcction. The \tlmulus was a 0.5 x I5 bar of light mowng at 8 xc. Each histogram is baaed on 10 trials [7X]. Fig.
?.
the directional selectivity characteristic of MT cells in normal monkeys. This residual responsiveness is eliminated by superior colliculus lesions. The colliculus could provide visual information to MT either by way of the pulvinar or the lateral geniculate body and extra-striate cortex. However, neither tectal 137. 821 nor geniculate cells show much sensitivity to the direction of stimulus movement and the littlc directionality found in the lateral and inferior pulvinar is eliminated by striate lesions 161. Thus. the directional selectivity that survives striate lesions MT neurons must bc generated rlc JIOI‘Oin MT. WLRTZ and his collcagues 157, 581 came to a similar conclusion on the basis of their analysis of the rcceptivc field properties of MT cells. The second implication concerns blindsight, a term used to describe the surprising amount of visually guided behavior that survives damage to striate cortex in humans and, by analogy, the similar behavior that survives striate lesions in monkeys. These residual abilities include the ability to detect and IocaliLe visual stimuli and to discriminate the speed and direction of stimulus movement 131, 50, 62, 66. 68, 69. 74, 75,90, 107, 109, I I I]. Because of the spatial nature of thcsc residual abilities, the apparent dependence of the residual vision on the superior colliculus [6 2, 901, and because of the unconscious nature of the preserved vision in humans, it has usually been assumed that blindsight reflects subcortical processing [19, 62, 90. 1071. However. there is a close similarity between the properties of MT units remaining after striate lesions and one type of visually-guided behavior remaining after striate lesions.
C‘. G. GKOSS
6
-
0
90
270
180
DIRECTION
360
(deg.)
r~iz.the ability to detect the direction and speed ofmovement. Thus. MT may bc responsible for the ability of monkeys and humans to detect direction of stimulus movcmcnt within a scotoma produced by ;I striate Icsion. That some cortical mechanism, as opposed to a midbrain one, must be responsible for this function is indicated by two sets of findings. The first is that there is little 1371 or no 1821 directional sensitivity among tectal neurons in primates. The second is that hcmisphercctomi/ed patients cannot dctcct direction ol movement in their blind field (presumably because they have no contralateral MT). but thoy can detect target location and the presence of movcmcnt (presumably with their midbrain mechanisms) 168, 751. Further support for an extra-striate but cortical mechanism fol analysis of direction of movement is that the lasting delicit in makinp visually guided saccades in monkeys after near total unilateral lesions of the cerebral cortex, including MT. i\ grcatcr than after either unilateral or bilntcral striate Icsions [WI. (Rccovcry of the ability to make visual saccades after bilateral striate Icsions is actually markedly greater than after unilateral ones [X6, I I I],) Similarly. optokinetic nystagmus i\ more dcvastcd h! hcmidecortication in monkeys than by occipital lohcctomy [9X]. The paradoxical aspect of thcsc results is that whereas tectal Icxions have virtually no cffcct on the activity of MT cells in normal animals. they have a dc\,astatinp clrcct in animals u ith striate Icsions. This phenomenon. in which the superior colliculus 5ccms to bccomc important only in the absence ofstriate cortex, is not unique to arc;1 MT, as will bc discussed bCl(~W. SIIPERIOR
I EMPORAL
POI.YSENS(_)RY
ARl
