The superior colliculus and visual neglect in rat and hamster. II. possible mechanisms

Brain Research Reviews, 8 (1984) 143- 153 Elsevier

BRR 90015

143

The Superior Colliculus and Visual Neglect in Rat and Hamster. II. Possible Mechanisms

PAUL DEAN and PETER REDGRAVE

Department of Psychology, University of Sheffield, Sheffield SIOZTN (U.K.)

(Accepted July lOth, 1984)

Key words: superior colliculus - retinal ganglion cells - geniculostriate - rat - hamster

CONTENTS

1. Introduction .............................................................................................................................................

2. Unique retinal input to superior colliculus ........................................................................................................ 2.1. Size of collicular and geniculostriate projections .......................................................................................... 2.2. Nature of collicular and geniculostriate projections ......................................................................................

2.2.1. Visual field representation ............................................................................................................ 2.2.2. Retinal ganglion-cell properties in the two projections ..........................................................................

2.3. Unique retinal input to superior colliculus: conclusions .................................................................................

3. Superior colliculus affects geniculostriate function .............................................................................................. 3.1. Superior colliculus as output station for striate cortex .................................................................................... 3.2. Superior colliculus and ‘gating’ of striate function ........................................................................................

4. Mechanisms of neglect: conclusions ................................................................................................................

5. Summary .................................................................................................................................................

Acknowledgements ........................................................................................................................................

References ...................................................................................................................................................

143

144 144 145 146 146 148

148 148 149

150

151

151

151

1. INTRODUCTION

How can the visual neglect produced by collicular

damage in rats and hamsters be explained in terms of

the anatomy and physiology of their visual systems?

One widely-accepted proposal is illustrated diagra-

matically in Fig. 1 (cf. fig. 39 in ref. 62). This di- . . . . . . .-...

agram emlloclies two important assumpnons. rirsr, Fig. 1. Diagram of possible substrate of response-specific col- that the bulk of visual information passes from the

retina to the geniculostriate system, with only a sub-

set travelling to the superior colliculus. Secondly, the

superior colliculus is specifically responsible for visu-

ally triggered orienting movements, whereas striate

cortex (in conjunction with other areas of visual cor-

licular neglect, as discussed in section 1. The superior colliculus receives a subset of the information that travels to the geniculostriate system, but is the only visual structure with effective connections to the systems that produce orienting movements. Hence removal of the superior colliculus will impair visually- elicited orienting. LGd, dorsal lateral geniculate nucleus of the thalamus. (Retinal projections to other structures, such as pretectal nuclei, omitted. See section 1).

Correspondence: P. Dean, Department of Psychology, University of Sheffield, Sheffield SlO 2TN, U.K.

0165-0173/84/$03.00 0 1984 Elsevier Science Publishers B.V.

144

tex) controls most other responses to visual stimula-

tion.

The plausibility of this model depends on collicular

neglect being a response-dependent deficit. Howev-

er, recent behavioural evidence indicates that the

neglect is not confined to orienting movements, but is

in some instances a failure to detect or attend to cer-

tain kinds of visual stimulusg. The model illustrated in

Fig. 1 therefore requires substantial revision to ac-

commodate this new evidence.

Possible revisions are additionally constrained by

the observation that in rats and hamsters lesions of

striate cortex do not produce striking visual neg-

lect20.37.~.61. This would suggest the following possi-

bilities.

(i) Contrary to the traditional view that all impor-

tant visual information passes to striate cortex, some

visual information travels only to the superior collicu-

lus.

(ii) In rats and hamsters an intact superior collicu-

lus is essential for striate cortex to function complete-

ly normally. Destruction of the superior colliculus

therefore affects both major visual pathways.

These are not exclusive alternatives, and indeed

the proposal that collicular neglect has more than one

behavioural component9 would indicate that both

may be correct. Thus, one of the components, name-

ly a deficit in the detection of certain kinds of visual

stimulus, favours suggestion (i): whereas another be-

havioural component, a disorder of attention in

which other kinds of visual stimulus are detected only

under particular circumstances, is more consistent

with suggestion (ii).

Anatomical and electrophysiological evidence rel-

evant to each of the suggestions is considered in sec-

tions 2 and 3. In these sections only occasional refer-

ence is made to retinal projections to areas other

than the superior colliculus or dorsal lateral genicu-

late nucleus (dLGN); for example to the ventral lat-

eral geniculate nucleus, the pretectal nuclei, and the

accessory optic nuclei. One reason for the omission is

that in rats and hamsters, as in other animals, visual

capability appears to be extremely limited after de-

struction of both the superior colliculus and the corti-

cal target of the dLGN (refs. in ref. 7, 8). Combined

with electrophysiological evidence (refs. in ref. S),

this observation suggests that although the functions

of these other structures are not fully understood,

they are unlikely to play any major role in the per-

formance of the sorts of visual task that are the sub-

ject of this review. One point of terminology - the

pathway from dorsal lateral geniculate nucleus of the

thalamus to cerebral cortex is referred to as the ge-

niculostriate pathway. The term is used because cur-

rent evidence suggests that in rat and hamster either

the dLGN projects solely to striate cortex, or any

projection to extrastriate cortex is very

slight g , 3 6 10,23,49,59,65,

2. UNIQUE RETINAL INPUT TO SUPERIOR COLLICU-

LUS

There are two different kinds of evidence to sug-

gest that in rats and hamsters the retinal projection to

the superior colliculus probably carries visual infor-

mation that does not reach the dLGN. One is evi-

dence relating to size of retinal projections, which in-

dicates that the projection to superior colliculus may

be larger than that to the dLGN (section 2.1). The

second kind of evidence deals more specifically with

the nature of the information carried by the two path-

ways (section 2.2).

2.1. Size of collicular and geniculostriate projections

Linden and Perry33 found that after injection of the

retrograde tracer horseradish peroxidase into the su-

perior colliculus of rats, up to about 47% of cells in

the corresponding area of the ganglion cell layer in

the contralateral retina were labelled. Because in

rats about 50% of the cells in this layer are displaced

amacrine cells which do not send an axon into the op-

tic nerveso, it appears that over 90% of ganglion cells

project to the superior colliculus from the regions of

retina investigated (which excluded the most tempo-

ral region, corresponding to the central part of the

field). The authors give reasons for thinking that a

similar proportion of retinal ganglion cells may pro-

ject to the superior colliculus in hamsters. Prelimi-

nary results suggested that after injections of horse-

radish peroxidase into rat dLGN, about 50% of true

retinal ganglion cells were labelled.

These findings are very much at odds with the as-

sumption that in all mammals the bulk of retinal out-

put travels to the dLGN, with a (small) subset of it di-

verging to the superior colliculus (cf. Fig. 1). Howev-

145

er, it has recently been shown that almost all retinal

ganglion cells project to the superior colliculus in the

rabbit? and there are a number of other findings in

hamster and in rat that are consistent with Linden

and Perry’s results rather than the traditional view.

(1) It has been observed that ‘the greater number

of fibres of the optic tract pass over the outer surface

of the lateral geniculate nucleus to reach the optic

thalamus and colliculi’ (ref. 29, p. 345). Also, there

are about 120,000 fibres in the rat’s optic nerve, but

only about 40,000 cells in the dorsal lateral geniculate

nucleus74, and the volume of the superficial layers of

the superior colliculus (i.e. those layers that receive

the bulk of the retinal input) is about twice that of the

dLGN3s.

(2) Some evidence suggests that the conduction ve-

locities of optic nerve fibres in rat and hamster have a

trimodal distribution. It has been established that the

superior colliculus receives input from all 3 groups, in

both rati7J9,67@ and hamsters6,57.

(3) Toga and Collins73 indirectly measured the dis-

tribution of radioactivity within the brains of albino

rats, following injection of radioactively labelled

amino acid into the eye. They found that the superior

colliculus received 64-76% of the radioactivity

found in the brain contralateral to the injection (in al-

bino rats the visual pathways are almost entirely

crossed). By comparison, the dorsal lateral genicu-

late nucleus received ll-17% (and the pretectal nu-

clei 6-14%). Toga and Collins also measured visu-

ally-produced changes in neural metabolism using

[i4C]2-deoxyglucose. Glucose utilization in response

to light flashes was greater in the superior colliculus

than in dLGN or visual cortex, an effect more pro-

nounced with higher frequencies of stimulation. The

authors interpret these results as suggesting that the

retina projects much more heavily to the superior

colliculus than to other structures, and that this dif-

ference is reflected in the metabolic changes that oc-

cur in response to certain kinds of visual stimulation.

(4) Sefton67 recorded from 34 single cells in rat

dLGN and found that all would respond to stimula-

tion of the superior colliculus. Moreover, the latency

of the response to collicular stimulation correlated

well with the latency of the response to stimulation of

the optic nerve. For this reason and others Sefton

concluded that the fibres innervating the dLGN all

bifurcated to innervate the superior colliculus.

(5) Olavarria and Torrealba44 found that the re-

ceptive fields of cells in lateral prestriate cortex (area

18a of Krieg”) were only slightly affected by ablation

of striate cortex (area 17), and suggested that a tecto-

cortical path was probably responsible for the spar-

ing.

All these results taken together do not completely

rule out the possibility that the dorsal lateral genicu-

late nucleus receives a small input not shared with the

superior colliculus (perhaps from temporal retina):

however, they are all consistent with the possibility

that 40-50% of retinal ganglion cells in the rat pro-

ject to the superior colliculus but not to the dorsal lat-

eral geniculate nucleus. Whether these cells project

to the superior colliculus alone is still unclear. Large

lesions of the superior colliculus in adult rats do not

cause retrograde degeneration of retinal ganglion

cellssiJ3, whereas section of the optic nerve doe+.

Either all retinotectal axons have collaterals, in

which case it is puzzling that the superior colliculus

receives about twice as much label from an injected

retina as all other structures put together, or retino-

tectal axons possess some unknown property that

renders them vulnerable to damage relatively near

the cell body, but not to damage near the axon termi-

nal. At present it is not known which of these alterna-

tives is correct. However, both behavioural and elec-

trophysiological evidence (refs. in e.g. ref. 7) suggest

that if retinotectal axons do bifurcate to reach struc-

tures such as the pretectal nuclei, the information

that reaches those nuclei cannot be used on its own to

solve tests of visual neglect (see section 1).

To summarize, there is some evidence to suggest

that in rats little if any visual information fails to

reach the superior colliculus, and moreover the ret-

inotectal pathway carries information that the genic-

ulostriate pathway does not. Whether the superior

colliculus receives projections that no other retinal

target receives is not known for certain.

2.2. Nature of collicular and geniculostriate projections

The evidence above does not indicate what sort of

information the superior colliculus but not the dorsal

lateral geniculate nucleus receives from the retina.

The evidence considered now relates first to the rep-

resentation of different parts of the visual field in the

146

two structures (section 2.2.1), then secondly to the

properties of the retinal ganglion cells that make up

those representations (section 2.2.2).

2.2.1. Visual field representation

The experiments of Toga and Collins73 also pro-

vided information about visual field representation in

the retinotectal projection. Radioactive label after

eye-injection was not found uniformly across the su-

perficial layers of the superior colliculus: rather,

label was particularly dense in the medial portion of

the superior colliculus, where the upper part of the

visual field is representedi4@. A similar pattern was

revealed in the metabolic response to light flashes,

except that in this case the area of greatest response

extended further over the posterior part of the supe-

rior colliculus, where the far temporal part of the vis-

ual field is mapped.

These results raise the possibility that the informa-

tion received by the superior colliculus that is not

shared by the geniculostriate pathway concerns the

periphery of the visual field. Electrophysiological

studies of superior colliculus and striate cortex in rats

and hamsters provide some support for this idea.

There seems to be general agreement that in the su-

perior colliculus neither magnification factor nor re-

ceptive field size change greatly from central field to

periphery (refs. 5, 14, 68,72: though see ref. 12). In

contrast, there is a clear expansion of the central field

(corresponding to area centralis, which is nasal and

ventral to the projection of the optic disk in the visual

field) in dLGN and striate cortexiJ9Js,71. Thus, Fu-

kuda et al.18 found that about two-thirds of all genicu-

late units sampled had receptive-field centres within

25” of the centre of the area centralis, and Lennie and

Perry31 report a heavy preponderance of receptive

fields in the inferior nasal visual field (cf. ref. 55).

This difference in mapping could occur either be-

cause the central field is more fully represented in the

cortex, or the peripheral field more fully represented

in the superior colliculus, or some combination of

these. It appears that in the central parts of the visual

field, receptive fields of single units are similar in size

in both superior colliculus and striate cortex (about

5-15”: refs. 5, 12, 24, 71, 72, 76). However, in the

periphery of the field cortical cells may have larger

receptive fields than collicular cells, at least in ham-

sters71%72. Indeed in hamsters the extreme upper and

lateral portions of the visual field appear not to be

represented at all in striate cortex7i. This may also be

true for rats, in that recordings from dLGN have re-

vealed very few cells with receptive fields in the up-

per far peripheryss: however, the issue is still to be re-

solved, as other workers have claimed (ref. 39,

p. 230) that ‘The extent and contour of the visual

field represented in LGd is in close agreement with

that found in the superior colliculus’ (by Siminoff et

al.68).

Thus, to summarize, there is both anatomical and

electrophysiological evidence to suggest that the pe-

riphery of the visual field may be in some sense more

fully represented in the superior colliculus of rats and

hamsters, than in the geniculostriate system. Howev-

er, the precise nature of the difference is unclear. In

this context it would be useful to have a study in

which the peripheral field representation in striate

cortex (with superior colliculus removed) was specif-

ically compared with that in the superior colliculus

(with striate cortex removed).

2.2.2. Retinal ganglion-cell properties in the two pro-

jections

The evidence reviewed so far indicates that the su-

perior colliculus receives a heavier retinal projection

than the dorsal lateral geniculate nucleus, particular-

ly for peripheral regions of the visual field. In this

section we ask whether the retinal ganglion cells that

project to the superior colliculus but not to the dorsal

lateral geniculate nucleus have different morpholog-

ical or electrophysiological properties from other ret-

inal ganglion cells. Three studies suggest that they

are ceils with slowly conducting axons and small so-

mas.

(1) Schober and Gruschka65 measured the soma

sizes of retinal ganglion cells that were labelled with

horseradish peroxidase after injection of the enzyme

into either superior colliculus or dorsal lateral genic-

mate. A higher proportion of small retinal ganglion

cells (soma diameter less than about 9 pm) were la-

belled after injections into the superior colliculus

than into the dLGN.

(2) Suminoto et al.69 recorded the responses of sin-

gle cells in either the superior colliculus or the dorsal

lateral geniculate nucleus to stimulation of optic

nerve and optic chiasma. From the response latencies

to stimulation at these two sites, they were able to

147

calculate the conduction velocity of the fibres that produced the response in each cell. The superior colliculus received a higher proportion than the dLGN nucleus of slowly conducting cells (less than about 4 m/s). The retinal projection to the dLGN could con- tain a lower proportion of slowly conducting cells with small somas than the projection to the superior colliculus for one of two main reasons. Either (a) most cells project to dLGN, but only the small cells also project to the superior colliculus, or (b) most cells project to the superior colliculus, but only the larger cells also project to the dLGN. Linden and Perry’s32 data support the second alternative, as does, indirectly, a comparison of Sumitomo et al.‘s data with those obtained from measuring fibre diameters in rat optic nerve**. Ogden and Miller43 found that for rhesus monkey optic nerve, the external diameter of the fibre was related to conduction velocity by the equation

Velocity (m/s) = 3.2 x Diameter (inpm)

at least up to diameters of 5 pm. Applying this equation to the data of Sumitomo et al.69 suggests that the distribution of fibre diameters that supply the superi- or colliculus (calculating from velocities, 50% of input fibres < 1.22pm) comes much closer to the optic nerve distribution obtained by Hughes (50% c 1.08 pm) than does the distribution of dLGN fibres (50% < 1.92 pm). The discrepancy between collicular and optic nerve distributions may be the result of com- pression of the optic nerve fibres during sectioning**.

(3) Direct support for the view that it is small cells that project to the superior colliculus but not to the dorsal lateral geniculate nucleus is provided by a recent study of Ni and DreherQ, who were able to visu- alize retinal ganglion cell morphology after injections of horseradish peroxidase into either target area. On morphological grounds they divided retinal ganglion cells into 5 classes, 4 of which projected to both the dorsal lateral geniculate nucleus and the superior colliculus. In the remaining class (IIIa) however, many cells were labelled after injection of horseradish peroxidase into superior colliculus, but few after injection into dLGN. Class IIIA cells had small somas (mean 11.6 pm) and fine axons, as well as fine long primary dendrites with very few branches.

What sort of information is conveyed by these

slowly conducting retinal ganglion cells that project to the superior colliculus but not to the dorsal lateral geniculate nucleus? Unfortunately, there appear to have been few studies of the electrophysiological properties of retinal ganglion cells in the rat or hamster, and none relating those properties to conduction velocity. An indirect method is to examine the receptive field properties of cells in the superior colliculus that receive slowly conducting retinal affer- entsi6Js. The majority of these cells (87%) have been classified as either ON-OFF (responded phasically to both onset and offset of light spot), ON phasic (responded phasically to light onset alone) and movement sensitive (responded only to moving light spots). They typically preferred slow movement (C 30 deg/s) , and small stimuli (C 10’) by virtue of a suppressive surround which by itself did not give responses. ON-OFF and movement sensitive cells comprise just over 50% of all cell types recorded in the visual layers of the rat superior colliculus compared with about 20% in the dorsal lateral geniculate nucleusiQ0. The proportion of ON phasic cells is about the same in the two structures (about 32%); however, in dLGN but not superior colliculus the majority of these cells have concentric off surrounds*iJ4Ji, and respond to fast movementls.

Somewhat similar differences between the properties of cells in striate cortex and superior colliculus of hamsters have been described by Tiao and Blake- morerl,? (a) collicular cells frequently preferred small objects, by virtue of their strong suppressive surrounds. In contrast, cortical cells usually responded best to long moving contours; (b) most collicular neurones preferred slower stimulus velocities than did cortical neurones, although there were a few collicular cells that probably preferred faster movement than that needed by cortical cells; and (c) there was an overall preference for light objects on a dark ground in the cortex, and for dark on light in the colliculus.

These indirect observations therefore suggest that in rats and hamsters the small, slowly-conducting cells that project to the superior colliculus but not to dLGN give phasic responses to the appearance, disappearance and slow movement of small stimuli, and their receptive fields lack concentric surrounds. These properties are consistent with what is known about retinotectal and retinogeniculate projections

148

in cats, the animal that has been most extensively in-

vestigated (e.g. refs. 25,75). In cats, retinal ganglion

cells have been classified into 3 groups (for review

see e.g. ref. 30) on the basis of a number of prop-

erties. Put very briefly: (1) Y cells, with large cell

bodies and fast conducting axons, project both to

dLGN and superior colliculus; (2) X cells, with medi-

um-sized cell bodies and moderately fast conducting

axons, project to the dLGN but hardly at all to the su-

perior colliculus. There is currently disagreement

about the existence of a separate X class in rat~18.21,31,42.48.50,77~ and (3) W cells, with small cell

bodies and slowly’conducting axons, more than half

of which project to the superior colliculus but not to

the dLGN. Electrophysiologically, W-cells appear to

be a rather heterogeneous group, but some of them

respond phasically to the appearance, disappearance

and slow movement of small stimuli, and their recep-

tive fields lack concentric surrounds.

It therefore appears that the retinal ganglion cells

in rat that project to superior colliculus but not dLGN

correspond to a subclass of W-cell in cat (possibly, it

has been suggested, to subclass W2;17). The electro-

physiological properties and dendritic morphology of

W-cells have led to speculations that they function as

primitive ‘feature-detectors’ for local change: that is,

they perform a highly non-linear transformation

upon their inputs to generate information concerning

the existence, whereabouts and crude characteristics

of localized luminance changes relative to back-

ground (e.g. ref. 27,34,53,54).

2.3. Unique retinal input to superior colliculus: con-

clusions

The evidence reviewed here indicates that, con-

trary to traditional assumptions, in rats and probably

hamsters the superior colliculus receives a heavier

retinal projection than the dorsal lateral geniculate

nucleus. In fact it appears that in at least some areas

of the retina very nearly all retinal ganglion cells pro-

ject to the superior colliculus. In contrast, the genicu-

lostriate projection does not include one class of W-

like retinal ganglion cell, and appears to have an im-

poverished representation of the upper and temporal

periphery. It is not at present clear whether this im-

poverished representation arises because the class of

retinal ganglion cell that projects only to the superior

colliculus is relatively more abundant in the periph-

ery, or whether in the periphery additional classes of

ganglion fail to project to the dLGN. Nor is it known

whether there are any retinal ganglion cells, in any

part of the retina, that project to dLGN but not supe-

rior colliculus.

These uncertainties, combined with lack of de-

tailed electrophysiological data, make it impossible

to characterize precisely the kind of information that

the superior colliculus receives but the dLGN does

not. The best guess at present seems to be that the su-

perior colliculus may get unique information about

small, transient stimuli (local ‘events’?) throughout

the visual field, and perhaps about a wider range of

stimuli in the upper and temporal periphery.

3. SUPERIOR COLLICULUS AFFECTS GENICULO-

STRIATE FUNCTION

An intact superior colliculus might be necessary

for the geniculostriate system to function normally. If

so, collicular damage would disturb both major visual

pathways, and affect the processing of visual infor-

mation that reached the dLGN (but not necessarily

the superior colliculus itself). As a potential substrate

for collicular neglect this is therefore complementary

to the one discussed in the previous section.

There are two obvious arrangements whereby ge-

niculostriate function could be partly dependent on

the superior colliculus: (1) the superior colliculus is it-

self an important output station for striate cortex

(Section 3.1); and (2) the superior colliculus controls

some of the information that reaches striate cortex,

or leaves it for other areas of the brain (section 3.2).

3.1. Superior colliculus as output station for striate

cortex

Anatomical evidence indicates that in rat and ham-

ster, as in other animals, there are numerous path-

ways by which the geniculiostriate system could use

the superior colliculus as an output station. For ex-

ample, there are direct projections from both striate

cortex and extrastriate visual areas to the superior

colliculus (e.g. refs. 32, 44), as well as indirect con-

nections via the basal ganglia (e.g. ref. 56).

Direct attempts to evaluate the importance of

these pathways in collicular neglect have compared

the effects of lesions restricted to the superficial lay-

149

ers of the superior colliculus, with lesions that also in-

vade intermediate and deep layers. The rationale for

this procedure is that most cortical afferents arrive at

the intermediate and deep layers, whereas retinal in-

put is restricted to the superficial layers. For exam-

ple, Mort et al.4 and Finlay et al.11, found that

orienting to sunflower seeds by hamsters was com-

pletely abolished only if the deep layers of the super-

ior colliculus were destroyed or undercut. If damage

were restricted to the superficial layers, there was

near total loss of orienting to upper and peripheral

visual fields, but some preserved orienting in lower

and anterior fields. The suggestion was made that fi-

bres from visual cortex to the deep layers of the supe-

rior colliculus might be responsible for the spared

function seen after lesions restricted to the superficial

layers. Related observations that neglect is observed

only if deep collicular layers are damaged have been

made in ratss’, tree shrews4 and gerbil+.

The suggestion that neglect in central areas of the

field requires destruction of corticotectal fibres is

consistent with what is known about the organization

of the retinotectal and geniculostriate pathways (sec-

tion 2), and with the finding of Toga and Collins’s

that electrical stimulation of striate cortex in rat in-

creased metabolic rate much more in that region of

the superior colliculus dealing with the central part of

the visual field than in the peripheral representation.

Taken with Toga and Collins’ other findings (sec-

tion 2.2.1), this result suggests that the superior colli-

culus receives visual information from peripheral

field primarily from the retina, whereas information

from central areas arrives from both retina and visual

cortex.

However, evidence concerning the effects of su-

perficial-layer lesions is at present insufficient to per-

mit any definite inferences about the role of cortico-

tectal projections in collicular neglect.

(a) In some experiments4JrJ6 the lesions of the su-

perficial layers that failed to produce total neglect

were made in neonatal animals. It is clear that in

hamsters and rats damage to the superior colliculus at

this age gives rise to abnormal retinal projections to

surviving structures (e.g. refs. 26, 52, 62, 64). The

role played by these projections in alleviating neglect

is not known.

(b) In cases where superficial lesions have been

made in adult animals, there are other problems of

interpretation. In Mort et al.‘s experimenta, the su-

perior colliculus was damaged by undercuts that were

horizontal in the mediolateral direction. Such cuts

spare deep layers if they are made too superficially.

But because in hamsters the surface of the superior

colliculus slopes steeply downwards from medial to

lateral, an inadequate cut is likely also to spare lateral

portions of the superficial layers. These lateral por-

tions correspond to the inferior part of the visual

field, so that Mort et al.% incomplete neglect may re-

flect at least in part incomplete damage to the super-

ficial layers. There also appears to have been some

sparing in the experiment of Midgley and Teess’: in

the example illustrated (ref. 37, fig. 8) lateral, medi-

al and caudal parts of the superficial layers were not

destroyed. Moreover, although the authors say (p.

326) that lesions restricted to the more superficial

laminae of the superior colliculus did not impair

orienting, their data do not seem entirely to support

this conclusion, in that the extent to which licking was

suppressed by a novel stimulus was similar in animals

with superficial and large lesions in all but two of the

15 examples given.

Thus, it is not clear from current data whether total

destruction of the superficial layers in adult rats and

hamsters does produce a less complete visual neglect

than lesions which in addition invade the deeper lay-

ers. Moreover, even if it were found that superficial

lesions did produce milder effects, there would still

remain problems of interpretation, because although

the bulk of the retinotectal projection terminates in

the superficial layers, there is a detectable retinal in-

put to the deep layers of the superior colliculus in

ratls. The behavioural role of direct or indirect corti-

cal projections to deep layers of superior colliculus

has yet to be properly evaluated. Consequently, it is

at present impossible to say to what extent collicular

neglect arises because the superior colliculus is an

output station for the geniculostriate system.

3.2. Superior colliculus and ‘gating’ of striate function

Damage to the superior colliculus might interfere

with geniculostriate function because it turned off a

tonic or a phasic signal from the superior colliculus

necessary for certain kinds of information to pass to

or from striate cortex - that is, the superior collicu-

lus might act as a ‘gate’ on striate signals.

150

There are a number of collicular projections that

might plausibly carry such a signal. Examples are the

projections to: (i) the dorsal lateral geniculate nucle-

11~34~47.49 (it is also possible that the functions of the

dorsal lateral geniculate nucleus are altered by dam-

age to the superior colliculus, because at least some

retinogeniculate fibres are collaterals of retinotectal

fibres); (ii) extrastriate visual cortex via the lateral

posterior nucleus of the thalamus (e.g. refs. 2, 41,

49, 70); (iii) mesencephalic reticular formationis346;

and (iv) midline thalamic nuclei*.

Unfortunately, little is known concerning the

properties of these pathways. Olavarria and Torreal-

baa found that the receptive fields of cells in lateral

prestriate cortex (area 18a) were relatively slightly

affected by destruction of striate cortex, and they

suggested that the pathway from superior colliculus

to prestriate cortex, via the lateral posterior nucleus,

might be responsible for the spared function. Behav-

ioural evidence is consistent with this suggestion7JJ2.

However, to what extent this pathway is necessary

for normal prestriate function is unknown. There is

some evidences that the superior colliculus is in-

volved in the activation of the cortical EEG to brief

light flashes, but whether this mechanism involves

any gating of striate output is unclear. Finally, Midg-

ley and Tee@ have found that the some of the behav-

ioural symptoms of collicular destruction can be abol-

ished by administration of amphetamine, but the re-

lationship of this finding to putative gating systems

has yet to be established (see also ref. 9).

In summary, the evidence at present is insufficient

either to establish or to exclude gating of striate func-

tion as an important mechanism of collicular neglect

in rats and hamsters.

4. MECHANISMS OF NEGLECT: CONCLUSIONS

It was argued in section 1 that what is perhaps the

most widely-accepted mechanism underlying collic-

ular neglect (illustrated in Fig. 1) required revision in

the light of the findings: (a) the neglect is stimulus-

dependent and not confined to orienting responses;

and (b) it is not found after destruction of striate cor-

tex. Two possible revisions were proposed.

The first was that some visual information reaches

the superior colliculus but not the dLGN. The evi-

dence reviewed in section 2 indicates that this is al-

+?

->?

Fig. 2. Revision of Fig. 1 to illustrate multiple substrates for collicular neglect in rats and hamsters (section 4). (i) The superior colliculus receives a massive retinal projection, which con- veys to it visual information not received by the geniculostriate system. Thus, removal of the superior colliculus will affect detection of certain kinds of visual stimulus. (ii) The superior colliculus may be necessary for the geniculostriate system to function properly, for example by controlling striate output in extrastriate cortex, using the projection from superior colliculus to visual cortex via the thalamus. Removal of the superior colliculus may therefore affect processing of stimuli under particular conditions of cortical activation, producing an ‘attentional’ deficit. (iii) The precise connections of superior colliculus and visual cortex with response systems are unclear. It is thus possible that removal of the superior colliculus produces a response- specific impairment, although not necessarily of the orienting response.

most certainly the case (Fig. 2). One implication of

this conclusion is that after removal of the superior

colliculus there will be some visual information that

the animal is no longer able to process, i.e. there will

be some kinds of visual stimulation that the animal

can no longer detect. Unfortunately, current evi-

dence permits only a rough description of the infor-

mation that reaches the superior colliculus alone: it

appears to be related to localized, transient stimuli,

and to the periphery of the visual field. This is inter-

esting in the light of the behavioural evidence con-

cerning collicular neglects, which suggests that after

collicular damage rats and hamsters have difficulty

detecting small moving stimuli, an impairment that

may be more pronounced in the periphery. Thus, it is

possible that one behavioural component of collicular

neglect, namely an apparent failure to detect certain

classes of visual stimulus, arises because only the su-

perior colliculus (in rat and hamster) receives retinal

information concerning those kinds of stimulus. This

suggestion lends itself readily to experimental test,

with both electrophysiological and psychophysical

techniques.

The second revision proposed in section 1 was that

the superior colliculus might be necessary for normal

1.51

functioning of the geniculostriate system (Fig. 2).

Anatomical techniques have demonstrated numer-

ous pathways that could mediate such a dependence

of geniculostriate function upon the superior colliculus, but at present detailed information about the kind

of information that these pathways carry, and their

role in collicular neglect, is not available. One indi-

rect reason for suspecting their importance is the be-

havioural evidence9 that indicates collicular neglect

might have an attentional component. It appears that

collicular animals respond normally to certain stimuli

(possibly large stimuli in the central regions of the

field) only in particular circumstances (e.g. when the

animal is very alert). If this is the case, then informa-

tion about such stimuli must reach visual areas be-

sides the superior colliculus, but clearly those areas

are unable to function completely normally when the

colliculus is damaged. It is thus plausible to suppose

that dependence of the geniculostriate system upon

the superior colliculus underlies the attentional com-

ponent of collicular neglect, while recognizing the ex-

tent of the experimental work required to specify the

nature of that dependence in adequate detail.

Finally, Fig. 2 differs from Fig. 1 in leaving un-

specified the connections of superior colliculus and of

visual cortex with response mechanisms. It is argued

in the next paper that the superior colliculus in rats

and hamsters probably has widespread connections

with a variety of response mechanisms in addition to

those concerned with orienting movements. Howev-

er, it is unclear whether visual cortex also has effec-

tive (even if undirect) connections with the same

mechanisms. Thus, Fig. 2 leaves open the possibili-

ties that removal of the superior colliculus could pro-

duce a response-specific impairment; and that the re-

sponse concerned need not be orienting movements

(cf. Midgley and Tees finding concerning the effects

of collicular lesions on rearings’).

neglect, whereas damage to striate cortex does not,

require explanation in terms of the anatomy and

physiology of visual pathways in these animals. One

proposal is that neglect is produced because the supe-

rior colliculus is the only visual structure directly con-

cerned with the production of orienting movements.

However, recent behavioural evidence indicates that

this mechanism on its own is insufficient: collicular

neglect is not confined to orienting movements, but is

in part an inability to register particular kinds of visu-

al stimulus. Two additional mechanisms are consid-

ered.

(1) The superior colliculus receives visual informa-

tion that the geniculostriate pathway does not. A va-

riety of anatomical and electrophysiological evidence

suggests that this is the case. Although the exact na-

ture of the information is unclear, there is some sug-

gestion that the superior colliculus has a stronger rep-

resentation of the peripheral field than the geniculo-

striate pathway, and may be more concerned with

small transient stimuli throughout the visual field.

(2) An intact superior colliculus is essential for nor-

mal functioning of the geniculostriate system. Ana-

tomical evidence indicates that there are pathways

whereby: (a) visual cortex could use the superior col-

liculus as an output station; and (b) the superior colli-

culus could control signals entering or leaving the

geniculostriate system, although the nature of the in-

formation carried by the pathways is not yet under-

stood.

The precise contribution of these two additional

mechanisms to collicular neglect remains to be deter-

mined. However, it may be conjectured that the first

would underly the deficit in stimulus detection that

appears to be one component of collicular neglect,

whereas the second mechanism might underly an at-

tentional component.

ACKNOWLEDGEMENTS

5. SUMMARY

The observations that removal of the superior col-

liculus in rats and hamsters produces a striking visual

REFERENCES

1 Adams, A. D. and Forrester, J. M., The projections of the rat’s visual field on the cerebral cortex, Quart. J. exp. Phys- iol., 53 (1968) 327-336.

This work was supported by S.E.R.C. Grants

B/24707 and C/46963.

2 Antonetty, C. M., A Study of the Connections of the Superi- or Colliculus, Doctoral Thesis, University of London, 1973.

3 Bigl, V. and Schober, W., Cholinergic transmission in subcortical and cortical visual centers of rats: no evidence for the involvement of primary optic system, Exp. Bruin Res.,

152

27 (1977) 211-219. 4 Casagrande, V. A. and Diamond, I. T., Ablation study of

the superior colliculus in the tree shrew (Tupaia glis), 1. camp. Neural., 156 (1974) 207-238.

5 Chalupa, L. M. and Rhoades, R. W., Responses of visual, somatosensory, and auditory neurones in the golden hamster’s superior colliculus, J. Physiol. (Lend.), 270 (1977) 595-626.

6 Coleman, J. and Clerici, W. J., Extrastriate projections from thalamus to posterior occipito-temporal cortex in rat, Brain Res, 194 (1980) 205-209.

7 Dean, P., Grating detection and visual acuity after lesions of striate cortex in hooded rats, Exp. Brain Res., 43 (1978) 145-153.

8 Dean, P., Visual pathways and acuity in hooded rats, Be- hav. Brain Res, 3 (1981) 239-271.

9 Dean, P. and Redgrave, P., The superior colliculus and visual neglect in rats and hamsters. I. Behavioural evidence, Brain Res. Rev., 8 (1984) 000-000.

10 Dursteler, M. R., Blakemore, C. and Garey, L. J., Projec- tions to the visual cortex in the golden hamster, J. camp Neural., 183 (1979) 185-204.

11 Finlay, B. L., Marder, K. and Cordon, D., Acquisition of visuomotor behavior after neonatal tectal lesions in the hamster: the role of visual experience, J. camp. physiol. F’sychol., 94 (1980) 506-518.

12 Finlay, B. L., Schneps, S. E., Wilson, K. G. and Schneider, G. E., Topography of visual and somatosensory projections to the superior colliculus of the golden hamster, Brain Res., 142 (1978) 223-235.

13 Foreman, N., Spatial and Attentional Functions of the Mid- brain Visual System, Doctoral Thesis, University of Not- tingham, 1980.

14 Forrester, J. M. and Lal, S. L., The projection of the rat’s visual field upon the superior colhculus, J. Physiol. (Lond.), 189 (1967) 25P-26P.

15 Frankfurter, A., Beckstead, R. M. and Harting, J. K., Ret- inal projections to the deep layers of the superior colliculus in the rat and cat, Anat. Rec., 202 (1982) 60A.

16 Fukuda. Y. and Iwama, K., Visual receptive-field prop-

17

18

19

20

21

erties of single cells in the rat superior colliculus, Jap. J. Physiol., 28 (1978) 385-400. Fukuda, Y., Sumitomo, I. and Sugitani, M., Mode of retinal projections to the three visual centers of the rat (dorsal and ventral lateral geniculate nuclei and superior colliculus). In Y. Katsuki, R. Norgren and M. Sato (Eds.), Brain Mechanisms of Sensation, John Wiley, New York, 1981, pp. 91-104. Fukuda, Y., Sumitomo, I., Sugitani, M. and Iwama, K., Receptive-field properties of cells in the dorsal part of the albino rat’s lateral geniculate nucleus, Jap. J. Physiol., 29 (1979) 283-307. Fukuda, Y., Suzuki, D. A. and Iwama, K., Characteristics of optic nerve innervation in the rat superior colliculus as revealed by field potential analysis, Jap. J. Physiol., 28 (1978) 347-365. Goodale, M. A. and Murison, R. C. C.. The effects of lesions of the superior colliculus on locomotor orientation and the orienting reflex in the rat, Brain Res., 88 (1975) 243-261. Hale, P. T., Sefton, A. J. and Dreher, B., A correlation of receptive field properties with conduction velocity of cells in the rat’s retino-geniculo-cortical pathway. Exp. Brain Res., 35 (1979) 425-442.

22

23

24

25

26

27

28

29

Hughes, A., The pigmented-rat optic nerve: fibre count and fibre diameter spectrum, J. camp. Neural., 176 (1977) 263-268. Hughes, H. C., Anatomical and neurobehavioral investigation concerning the thalamo-cortical organization of the rat’s visual system, J. camp. Neurol., 175 (1977) 311-336. Humphrey, N. K., Responses to visual stimuli of units in the superior colliculus of rats and monkeys, Exp. Neural., 20 (1968) 312-340. Ilhng, R.-B. and Wassle, H., The retinal projection to the thalamus in the cat: a quantitative investigation and a comparison with the retinotectal pathway, J. camp. Neural.,

202 (1981) 265-285. Ingle, D. J., (1982) Organization of visuomotor behaviors in vertebrates. In D. J. Ingle, M. A. Goodale and R. J. W. Mansfield (Eds.), Analysis of Visual Behavior, M.I.T. Press, Cambridge, MA, 1982, pp. 67-109. Koch, C., Poggio, T. and Torre, V., Retinal ganglion cells. A functional interpretation of dendritic morphology, Phil. Trans. B, 298 (1982) 227-264. Krieg, W. J. S., Connections of the cerebral cortex. I. The albino rat. A. Topography of the cortical areas, J. camp. Neurol., 84 (1946) 221-275. Lashley, K. S., The mechanism of vision. VII. The projection of the retina upon the primary optic centers in the rat, J. camp. Neurol., 59 (1934) 341-371.

30 Lennie, P., Parallel visual pathways: a review, Vision Res., 20 (1980) 561-594.

31 Lennie, P. and Perry, V. H., Spatial contrast sensitivity of cells in the lateral geniculate nucleus of the rat, J. Physiol. (Lond.), 315 (1981) 69-79.

32 Lent, R., The organization of subcortical projections of the hamster’s visual cortex, 1. camp. Neural., 206 (1982) 227-242.

33 Linden, P. and Perry, V. H., Massive retinotectal projection in rats, Brain Res., 272 (1983) 145-149

34 MacKay-Sim, A., Sefton, A. J. and Martin, P. R., Subcor- tical projections to lateral geniculate and thalamic reticular nuclei in the hooded rat, J. camp. Neural., 213 (1983) 24-35.

35 Marr, D. C., Vision, W. H. Freeman, San Francisco, CA,

36

37

38

39

1982. Masterton, B., Skeen, L. C. and RoBards, M. J., Origins of anthropoid intelligence. II. Pulvinar-extrastriate system and visual reversal learning, Brain Behav. Evol., 10 (1974) 322-353. Midgley, G. C. and Tees, R. C.. Orienting behaviour by rats with visual cortical and subcortical lesions, Exp. Brain Res., 41(1981) 316-328. Midgley, G. C. and Tees, R. C., Reinstatement of orienting behavior in rats with superior colliculus lesions by n-amphetamine, Sot. Neurosci. Abstr., 8 (1982) 894. Montero, V. M., Brugge. J. F. and Beitel, R. E., Relation of the visual beld to the lateral gemculate body of the albino rat. J. Neurophvsiol.. 31 (1968) 221-236.

40 Mort, E., Cairns, S., He&h, H. and Finlay, B., The role of the superior colliculus in visually guided locomotion and visual orienting in the hamster, Physiol. Psychol., 8 (1980) 20-28.

41 Nagel, S. M. and Butter. C. M., Tectal and cortical afferents to nucleus lateralis posterior in the rat, Sot. Neurosci. Abstr., 5 (1979) 799.

42 Ni, S. Y. and Dreher, B., Morphology of rat retinal ganglion cells projecting to the thalamus and midbrain, Proc. aus-

43

44

45

46

47

tral. Physiol. Pharmacol. Sot., 12 (1981) 97P. Ogden, T. E. and Miller, R. F., Studies of the optic nerve of the Rhesus monkey: nerve fibre spectrum and physiological propetties, Vision Rex, 6 (1966) 485-506. Olavarria, J. and Torrealba, F., The effect of acute lesions of the striate cortex on the retinotopic organization of the lateral peristriate cortex in the rat, Brain Res., 151 (1978) 386-391. Olavarria, J. and Van Sluyters, C., The projection from striate and extrastriate cortical areas to superior colliculus in the rat, Brain Res., 242 (1982) 332-336. Otani, K., Kimura, M. and Yamada, J., The tectofugal projections in the rat and cat, Proc. 10th int. Congr. Anat., 1975, p. 183. Pasquier, D. A. and Villar, M. J., Subcortical projections to the lateral geniculate body in the rat, Exp. Brain Res., 48 (1982) 409-419.

48 Perry, V. H., The ganglion cell layer of the retina of the rat: a Golgi study, Proc. roy. Sot., 204, B (1979) 363-375.

49 Perry, V. H., A tectocortical visual pathway in the rat, Neu- rosci., 5 (1980) 915-927.

50 Perry, V. H., Evidence for an amacrine cell system in the ganglion cell layer of the rat retina, Neurosci., 6 (1981) 931-934.

51 Perry, V. H. and Cowey, A., Changes in the retino-fugal pathways following cortical and tectal lesions in neonatal and adult rats, Exp. Brain Res., 35 (1979) 97-108.

52 Perry, V. H. and Cowey, A., A sensitive period for ganglion cell degeneration and the formation of aberrant retino- fugal connections following tectal lesions in rats, Neurosci- ence, 7 (1982) 583-594.

53 Pinter, R. B., Product term non-linear lateral inhibition en- hances visual selectivity for small objects or edges, J. theor. Biol., 100 (1983) 525-531.

54 Pinter, R. B. and Harris, L. R., Temporal and spatial response characteristics of the cat superior colliculus, Brain Res., 207 (1981) 73-94.

55 Reese, B. E. and Jeffery, G., Crossed and uncrossed visual topography in dorsal lateral geniculate nucleus of the pigmented rat, J. Neurophysiol., 49 (1983) 877-885.

56 Rhoades, R. W. and Chalupa, L. M., Conduction velocity distribution of the retinocollicular pathway in the golden hamster, Brain Res., 159 (1978) 396-401.

57 Rhoades, R. W. and Chalupa, L. M., Conduction velocity distribution of the retinal input to the hamster’s superior colliculus and a correlation with receptive field characteristics, J. camp. Neural., 184 (1979) 243-264.

58 Rhoades, R. W., Kuo, D. C., Polcer, J. D., Fish, S. E. and Voneida, T. J., Indirect visual cortical input to the deep layers of the hamster’s superior colliculus via the basal ganglia, J. camp. Neural., 208 (1982) 239-254.

59 Ribak, C. E. and Peters, A., An autoradiographic study of the projections from the lateral geniculate body of the rat, Brain Res., 92 (1975) 341-368.

60 Schneider, G. E., Contrasting visuomotor functions of tec- turn and cortex in the golden hamster, Psychol. Forsch., 31

153

(1967) 52-62. 61 Schneider, G. E., Two visual systems, Science, 163 (1969)

895-902. 62 Schneider, G. E., Mechanisms of functional recovery fol-

lowing lesions of visual cortex or superior colliculus in neo- nate and adult hamsters, Brain Behav. Evol., 3 (1970) 295-323.

63 Schneider, G. E., Two visuomotor systems in the Syrian hamster. In D. J. Ingle and J. M. Sprague (Eds.), Sensori- motor Functions of Midbrain Tectum, Neurosci. Res. Progr. Bull., 13 (1975) 255-257.

64 Schneider, G. E., Is it really better to have your brain lesion early? A revision of the ‘Kennard Principle’, Neuropsycho- logia, 17 (1979) 557-583.

65 Schober, W. and Gruschka, H., Zur Projektion der einzel- nen Ganglienzellklassen der Retina der Albinoratte. Eine Studie mit Meerrettich-Peroxidase, Z. Mikrosk.-anat. Forsch. Leipzig, 92 (1978) 283-297.

66 Schober, W. and Winkelmann, E., Die geniculo-kortikale Projecktion bei Albinoratten, J. Hirnforsch., 18 (1977) l-20.

67 Sefton, A. J., The innervation of the lateral geniculate nucleus and anterior colliculus in the rat, Vision Res., 8 (1968) 867-881.

68 Siminoff, R., Schwassmann, H. 0. and Kruger, L., An electrophysiological study of the visual projection to the superior colliculus of the rat, J. camp. Neural., 127 (1966) 435-444.

69 Sumitomo, I., Ide, K., Iwama, K. and Arikuni, T., Conduc- tion velocity of optic nerve fibers innervating lateral geniculate body and superior colliculus in the rat, Exp. Neural., 25 (1969) 378-392.

70 Takahashi, T., Nuclear organization of the rat’s lateral thalamus, Sot. Neurosci. A&r., 8 (1982) 643.

71 Tiao, Y.-C. and Blakemore, C., Functional organization in the visual cortex of the golden hamster, J. camp. Neurol., 168 (1976) 459-482.

72 Tiao, Y.-C. and Blakemore, C., Functional organization in the superior colliculus of the golden hamster, J. camp. Neu- rol., 168 (1976) 483-504.

73 Toga, A. W. and Collins, R. C., Metabolic response of optic centers to visual stimuli in the albino rat: anatomical and physiological considerations, J. camp. Neural., 199 (1981) 443-464.

74 Vaney, D. I., Peichl, L., Wassle, H. and Illing, R.-B., Al- most all ganglion cells in the rabbit retina project to the superior colliculus, Brain Res, 212 (1981) 447-453.

75 Wassle, H. and Illing, R.-B., The retinal projection to the superior colliculus in the cat: a quantitative study with HRP, J. camp. Neural., 190 (1980) 333-356.

76 Wiesenfeld, Z. and Kornel, E. E., Receptive fields of single cells in the visual cortex of the hooded rat, Brain Res., 94 (1975) 401-412.

77 Winkelmann, E. and Brauer, K., Some aspects of neuronal organization in rat’s visual structures, Vision Res., 21 (1981) 1681-1683.

Documents

The superior colliculus and visual neglect in rat and hamster. II. possible mechanisms