Dorsal midline thalamus, pretectum and responses to diffuse light in the rat

Physiology & Behavior, Vol. 26, pp. 873-886. Pergamon Press and Brain Research Publ., 1981. Printed in the U.S.A.

Dorsal Midline Thalamus, Pretectum and Responses to Diffuse Light in the Rat I R. M. C O O P E R , 2 J. A. B A T T I S T E L L A A N D A. M. R A T H

Department of Psychology, University of Calgary, Calgary, Alberta T2N 1N4, Canada Received 7 N o v e m b e r 1980 COOPER, R. M., J. A. B A T r l S T E L L A AND A. M. RATH. Dorsal midline thalamus, pretectum and responses to diffuse light in the rat. PHYSIOL. BEHAV. 26(5) ,873-886, 1981.--The present study examined dorsal midline thalamic and pretectal lesion effects on visual task performance, and tested current sensory interpretations of these effects. The f'wst three experiments supported a previous claim that the pretectal area plays a role in diffuse fight perception. Experiments 1 and 2 showed that dorsal midline thalamic lesions disrupt choice Of the correct illuminated goal box over the incorrect, dark goal box in a "simultaneous" two-choice discrimination task, but only when the rats wear light diffusing translucent eye masks. Experiment 3 showed that the deficit following a unilateral lesion is greatest when the animals are forced to use the diffusely occluded contralateral eye, which in the rat projects strongly to the damaged hemisphere. Despite this seemingly substantial support for an impairment in diffuse light perception, certain observations made during these experiments led to formulation of an alternative, nonvisual description of the lesion deficit. According to this new hypothesis, thalamic lesions undermine the abifity to inlaibit position responding, which in the diffuse light condition is critical for successful discrimination performance in such paradigms. Experiments 4--6 validated this new interpretation. First, Experiment 4 substantiated the literature indicating that thalamic lesions impair the animal's capacity to inhibit prepotent responses. Next, Experiments 5 and 6 showed that in a "successive" version of the two-choice task and in a task involving conditioned suppression of lever presslng to a light stimulus---two paradigms which probably make identical response demands in both the diffuse light and normal visual states---the lesion deficit registered is the same under both visual conditions. The study was judged to have undermined the view that dorsal midline and pretectal structures contribute directly to visual perception in the intact mammal.

Pretectum

Thalamus

Vision

DESTRUCTION of visual cortex (Brodmann's areas 17, 18 and 19) severely impairs the transoperative retention of form, pattern and other visual discriminations [12]. Diffuse light discriminations are an exception in this respect. Rats trained to perform two-choice, light-dark discrimination tasks while wearing light-diffusing translucent eye masks suffer little subsequent to visual cortex removal in contrast to their unoccluded counterparts permitted a spatially defined retinal image preoperatively [1,6]. One implication that can be drawn from these findings is that diffuse light intensity discriminations are subcortically mediated. In fact some evidence [4] suggests that posterior dorsal midline thalamus, which includes the pretectal region, is involved; if this region is subsequently destroyed, the animals wearing the diffusing masks, rather than the unoccluded subjects, are most impaired in their transoperative retention of the light-dark discrimination. The impairment resulting from such subcortical lesions is the principal concern of the present study. In particular, we wished to determine whether the deficit associated with the occluded state reflects an impairment in diffuse light sensitivity [4], or a disruption of certain nonvisual processes contributing to successful task performance. The experi-

ments outlined below deal principally with this issue, but in our concluding discussion, we briefly return to the major issue of the division of labor between the cortical and subcortical sectors of the visual system. EXPERIMENT 1

Blochert et al. [4] found that rats limited to the use of diffuse light cues, but not their visually unrestricted counterparts, were impaired on a "simultaneous" light-dark discrimination task by lesions involving posterior midline thalamus. Our first objective was to determine whether the differential effect could be demonstrated in individual rats trained under both visual conditions.

Method Four 300 g male black-hooded rats served as subjects. Under sodium pentobarbital anesthesia, Teflon posts were implanted in the skull, to anchor the light-diffusing and control headmasks [2,19]. The rats were then reduced to 80% of their ad lib weight and in stages adapted to an automated two-choice light-dark "shuttle discrimination" apparatus (Fig. 1). This apparatus is similar conceptually and dimen-

1This study was supported by a Canadian NSERC Grant APA-135. We are indebted to R. Ferrier, B. Sookochoff and G. Thurlow for their suggestions for revising the text. 2Send reprint requests to Prof. R. M. Cooper, Psychology Department, The University of Calgary, Calgary, Alberta, T2N IN4, Canada.

Copyright © 1981 Brain R e s e a r c h Publications Inc.--0031-9384/81/050873-14502.00/0

874

COOPER, BATTISTELLA AND RATH

ll i'

I-- 50 "'~o

~

!~!ll,,ll ' I t| I I

S_

f'l i~1

3o -~' I ~1 HI! II¢ ',l!1 ~:

I--

7

~-~ ~ - ~ v .

-I

/'4

~[

'!,' ,I II!

40-

o

o Unoccluded • Occluded

PRE OP POST-OP 20 , , t I l l I ; I

^/

-t

;\ /V

i i

l l l l l l

l | l l | l l l l | l l l l

II

l ] l l | | l

1|II

51 TRIAL SESSIONS 50-"%

A/"

i

FIG. I. Automated, two-choice shuttle discrimination apparatus.

,

9

50%.~

J

o°°:40N°21 ]~ <[ I.-

sionally to the "run-around" discrimination apparatus used earlier [4,5]. In the first pretrainlng stage, a rat was penned in one of the four goal boxes of the apparatus (Fig. 1) and trained to lever press for 45 mg food pellets. Two of the detachable goal boxes were then butted together, forming an alley in which the rat had to learn to run back and forth, alternately pressing the levers at each end for pellets. Preliminary training ended with two, 51-reinforcement sessions in the completely assembled discrimination box. During these sessions, the rat had to locate the uncued goal box in which a lever press delivered a pellet. This "correct" goal box alternated from one end of the apparatus to the other after a pellet was received. At the correct end the rat could sometimes obtain the pellet from the right and sometimes from the left goal box. By the end of this pretraining phase, a rat took approximately 30 min to complete a session. The rats were by this time accustomed to wearing headgear, introduced during the initial lever-pressing stage of pretraining, and to alternating between the two ends of the apparatus after receipt of a pellet. During discrimination training, alternation between ends of the apparatus was still required, but the correct goal box was cued by illumination of a white Lucite panel located at the goal box entrance (Fig. 1). The apparatus was housed in a dark room. The 51-trial training session was begun by placing the subject in the tunnel connecting the two ends of the apparatus and pointing it toward the end with one o f the cue panels illuminated. A lever press in the cued goal box was recorded on a "correct" counter, provided the rat had not fin-st pressed the lever in the uncued goal box at the same end of the apparatus. The trial ended once the rat pressed the lever in the lighted goal box; the light then went off, and one of the goal boxes at the other end of the apparatus became correct. Right and left goal boxes were programmed in a pseudorandom fashion to be correct equally often during a session, with the provision that one side was never correct more than three times consecutively. Cue panel luminosity was increased by about 1 log step in the diffuse light condition to compensate for the reduction in intensity produced by the white plastic masks. Two rats began the task wearing the diffusing masks and two wearing the visually-unrestrictiag control masks. After reaching criterion (43/51 correct on two consecutive sessions), the animals were switched to the other visual condition until they again reached criterion. A session-to-session alternation of headgear then ensued in overtraining sessions,

20

I'~ll

I

30

PRE-I OP I POST -OP

PRE~ 0P POST.OP

,,I,,,,,

~ I

1

•

~ i , Jl

OcctwJed l

i

I | I I I I I I I I I [ | i

i/

o Oe0ccl~led

• Occluded

iI

t

i

i]I ~,., 20

I I

51 TRIAL SESSIONS

51 TRIAL SESSIONS

5o \ [ u.. 4 0

|

1

/ ..._.../'~.1

\it

/

~ " ~"T -I~ , , 1 , , , , , ,IIII ,,~

II

l l l l l l l l l l l l l l l l l l l l l l

l l l l l l l l l l

51 TRIAL SESSIONS

FIG. 2. Unoccluded and occluded performance of Experiment ! subjects.

to ensure excellent light-dark task performance under both visual conditions. Under ether anesthesia, the rats were then placed in a Kriag stereotaxic instrument and cathodal electrolytic lesions were made by placing the bared tip of an insulated copper wire in the pretectal area and passing a 2 mA current for 60 sec. The wire tip was located 3.5 mm posterior to bregma, 1.5 mm lateral to the midline and 4.8 mm ventral to the skull surface. Following a 2 week recovery period, the animals were returned to the discrimination setting and m alternating sessions, wore either the diffusing or the control masks. After several sessions, some animals were tested solely in the diffuse light condition (Fig. 2), once it became clear they could perform well in the unoccluded state. At the conclusion of testing, the animals were perfused with formal-saline, and the brains were embedded in celloidin, sectioned, and then stained with thionin for histological examination.

Results and Discussion Postoperatively, all four animals demonstrated poorer performance in the diffuse light condition (Figl 2). The differential deficit was particularly striking for subjects 1 and 4: they took only one session to exhibit criterion performance in the unoccluded state, but 20 or more sessions when occluded. The difference was also clear-cut for subject 2,

MIDLINE THALAMUS AND VISION

875

FIG. 3. Left to right, lesion reconstructions for subjects 1, 2, 3 and 4 (Experiment 1). Diagrams after K6nig and Klippel [15].

although less pronounced. Subject 3's performance was somewhat anomalous since it produced poor scores in the fh-st three unoccluded sessions, but then displayed a sharp recovery in that state, while continuing to perform poorly when occluded. Experiment 1 demonstrates that a differential effect on performance can be observed in the same animal, dependent upon visual occlusion, and corroborates the earlier finding of Blochert et al. [4], who tested separate groups of occluded and unoccluded rats. Blochert et al. [4] proposed that the deficit reflected impaired diffuse light perception. However, one feature of the present experiment which undermines this visual interpretation is the nature of the induced lesions (Fig. 3). In subjects 1, 3 and 4, the destroyed area extended rostrally in the thalamus, well beyond the region receiving optic tract terminations. Animals 1, 3 and 4 also displayed the greatest postoperative difficulties, in contrast to subject 2, which sustained only a restricted lesion in posterior visual thalamus. Damage to anterior nonvisual thalamic structures could have contributed to the impairment, indicating that the deficit may not be visual in nature, a possibility pursued in Experiment 2. EXPERIMENT 2

The second experiment was methodologically similar to Experiment 1. Rats were trained preoperatively to perform the light-dark task under both visual states, then half of the rats underwent surgery aimed at producing lesions in anterior thalamus. The electrode placements were made 3 mm

anterior to the locus used in Experiment 1. The remaining half of the subjects received the usual posterior placements aimed at the more visual thalamic region. Experiment 2 was conducted over a two-year period, with three sets of the two experimental groups. These replications reflected our attempts to achieve similar-sized and appropriately placed lesions in both experimental groups. The lesion damage tended to flow more rostrally than caudally. Consequently, the animals given posterior placements tended to receive larger lesions. Since all lesions were made by passing 2 mA for 60 sec, the differences in lesion size between groups were presumably due to variations in resistance of the tissues comprising the dorsal thalamic brain region. The behavioral results are summarized in Table 1 as the number of 50-trial sessions needed to regain criterion (two consecutive sessions of 43/50 trials correct) in each of the two visual states or to the termination of testing. The lesions sustained in the three sets of subjects are depicted in Figs. 4, 5 and 6. Behavioral results from the first set of animals tested (Table l) indicated minimal and substantial impairments, respectively, for the anterior and posterior electrode placement groups in the diffuse light state. Figure 4 suggests, however, that the lesions of the anterior group were smaller than intended, confounding the interpretation of the behavioral data in terms of lesion locus. Unlike the first group tested, the anterior groups of the next two sets of animals did show clear-cut impairments in the diffuse light state (Table 1). Examination of lesions in subjects 104 and 1l0 of the second set and subject 28 of the

~°

II

i

..............."

ifT!

"'

"

,.-

" '~

' ~

i,!

f~

P.,,

8'

~.~ ~

D

•

-.....

i~i ~ "

~"

~,,~

,

g~J

:-:~

~.

~

,:: .

::

~i

.,---,-.

7i7i: ~

,

/

,

Z

> .g .j

e~ © ©

MIDLINE THALAMUS AND VISION

877

TABLE 1 NUMBER OF 50 TRIAL SESSIONS REQUIRED POSTOPERATIVELY TO REACH CRITERION IN THE UNOCCLUDED AND OCCLUDED STATES OF SUBJECTS WITH ANTERIOR (A) AND POSTERIOR (P) ELECTRODE PLACEMENTS (EXPERIMENT 2)

Rat

Electrode Placement

Unoccluded Sessions

Occluded Sessions

Set 1 52 53 58

A A A

0 0 1

2 3 1

54 55 56

P P P

1 2 0

30t 34t 20

Set 2 102 104 110

A A A

2 3 1

14 12 9

103 107 119

P P P

2 1 8

19 34t 19

Set 3 20

A

7

7

21 27 28

A A A

0 1 1

30 9 24

22 23 25 30

P P P P

1 2 3 1

31"~ 1~ 7 31t

tAnimals not trained to criterion.

third set, in particular, suggested that a substantial behavioral impairment could be induced by isolated rostral thalamic damage. We shall not report in detail additional experiments carried out with animals receiving small localized posterior thalamic and midbrain lesions. The impairment in the diffuse light state was always relatively slight, and we never observed the striking deficits seen in so many animals with large thalamic lesions. (The correlation between small lesions and a lack of disruption in discrimination performance can also be determined from examination of the coronal diagrams and behavioral data that have been presented in detail.) Experiment 2 was judged to have supported the idea that the deficit revealed in the diffuse light state was not a visuosensory one. The impairment in animals with rostral thalamic lesions, the relatively slight effect in rats with smaller yet well-localized lesions in posterior thalamus, and the correlation between lesion size and deficit, all suggest a nonvisual handicap. However, the evidence remains inconclusive: anterior thalamus does include the nucleus medialis dorsalis which projects to the frontal eye fields, and given the complex intercormectivity of midline thalamus, the functional delimitation of regions based on the lesion approach is hazardous. In Experiment 3, we switched from the lesion-

site tactic and adopted an experimental strategy which is widely believed to yield a decisive assessment of the braindamaged animal's visual status. EXPERIMENT 3

In the rat, most retinal ganglion fibres cross to the contralateral hemisphere. If thalamic damage produces a visual deficit, a unilateral lesion should therefore impair diffuselight performance more with vision restricted to the contralateral eye than to the ipsilateral eye. Experiment 3 tested this proposition, in a further attempt to define the nature of the handicap manifested in the diffuse light condition. As in the previous experiments, the animals were trained to perform the fight-dark task in both of the visual states. Near the end of preoperative training and thereafter, they wore new masks with one translucent and one opaque eyepiece so that each eye was employed separately in the diffuse-light condition over alternating sessions. The five subjects then underwent surgery, receiving a unilateral posterior thalamic lesion and, after recovery, were returned to the task for additional monocular diffuse light training sessions with the translucent and opaque covers again switched between right and left eyes over sessions. The behavioral results of Experiment 3 are summarized in Fig. 7 and the lesions are depicted in Fig. 8. Figure 7 shows, with the exception of subject 103, whose lesion was small, that performance was poorer when the rats were required to use the eye contralateral to the lesions. An input deficit might seem to have been shown: ff the animal performs well with one eye and not the other, how could the deficit not be visual in nature? However, an additional observation in Experiment 3 suggested a markedly different interpretation. Shortly after surgery, the animals had been placed on the floor and their behavior observed. Each animal was noted to move toward the side on which its lesion had been placed, with, for example, a right-sided lesion associated with locomotion to the right. This effect at first appeared to be only temporary, and was not obvious several days later. Nevertheless, this ipsiversive progression tendency was observed still later, during postoperative testing. Figure 7 reveals that on the initial postoperative sessions, whether using the ipsilateral or the contralateral eye, most animals performed well below criterion. Correlated with this poor performance were the animal's perseverative choices of the goal box corresponding to the side of the lesion, as shown by the numbers near the first postoperative points plotted in Fig. 7. They indicate the number of times out of 50 trials the animals chose the goal box, appropriately or not, on the lesion side. Animal 101, for example, chose the right-side box on 47 out of the 50 trials in the first session using the right-eye and 45 out of 50 on the first session using the left eye. The ipsiversive progression tendency provides grounds for a motor interpretation of the difference in monocular performance. For example, a rat with a right-hemisphere lesion and in the grip of a right going tendency might perform better in the apparatus using its right eye, since that eye surveys the right visual field which coincides with the animal's direction of travel. The view supplied by the contralateral 0eft) eye could be entirely normal, but since the animal does not move toward its left, that visual input will be less effective in governing movement. Consequently, the imbalance in reaction tendencies, biasing travel toward one alley or another consistently, at least makes Experiment 3 a

878

COOPER, B A T T I S T E L L A A N D RATH

FIG. 6. Above, left to right, lesion reconstructions for third set of subjects 20, 21, 27 and 28 (anterior placement) and below, 22, 23, 25 and 30 (posterior placement) of Experiment 2.

MIDLINE THALAMUS AND VISION

879

;

"~

,

_

~

.s ~

s

Y',,

'

,,,,,

,,,,

,= ,,,,,

5 0 TRIAL SESSIONS

°fg ft o,, z

-if)

40

8

, , . =

=, =E

,3o i.

:

:

=ILl 36 /

t~

/

-

50 TRIAL SESSIONS ,

z

Q

[3 RIGHT EYE

11:40

8

o L E F T EYE

e,,, DAMAGE TO

. 30

CON'rRALATERAL HEMISPHERE

:z:

I.-

i

20 ....

'

50 TRIAL SESSIONS FIG. 7. Monocular diffuse light performance of Experiment 3 subjects.

less-than-critical test of the effects of thalamic lesions on diffuse light perception. EXPERIMENT 4

We noted while conducting Experiment 2 that the thalamic lesions might be producing some nonvisual handicap. The weU-trained normal rat performs the light-dark task in an efficient manner, consistently choosing the illuminated goal box and after receiving a food pellet, quickly exiting that goal box to seek out the newly illuminated one at the far end of the apparatus. Postoperatively, however, some animals clearly registered an increase in total session lever presses, even in the unoccluded state, when choosing the correct goal box consistently. This observation is consistent with the more rigorous work of Delacour and his collaborators [9,10]. They studied the effects of midline dorsal thalamic lesions on lever press-

ing and found responding to be elevated under intermittent schedules of reinforcement. In related experiments, Vanderwolf [24], found that thalamic lesions accentuated the rat's tendency to remain immobile when fearful. This work suggests that such lesions counter the animals' inhibition of certain prepotent responses. To check the applicability of the Delacour and Vanderwolf work to the present study, subjects 103, 107 and 119 from Experiment 2, and three normal animals which had the same discrimination training, were tested on a differential reinforcement of low rate lever pressing schedule. The animals were required to delay a lever press for 20 sec to obtain a food pellet. A lever press during the 20 sec interval reset a timer, which required a complete new 20 sec delay before a press would deliver a pellet. Operant conditioning unit details and procedures were similar to those reported previously [7]. The performance of these subjects is summarized in the top part of Fig. 9, over 10 daily 1-hr test sessions. The lower part summarizes a later test of seven animals with scattered thalamic lesions (lesions not illustrated) and three normal control subjects. These data suggest that thalamic lesions do dispose animals to lever press inappropriately. Corroboration was obtained by comparing variable interval 30 sec lever pressing in still another group of 15 rats with scattered thalamic and midbraln lesions (not illustrated) and a group of 10 normal control animals. The brain-damaged animals' tendency to press at higher rates is shown in Fig. 10. This evidence, that midline thalamic lesions nonvisually handicap subjects in inhibiting prepotent responses, led to formulation of the hypothesis outlined in Experiment 3: that an impairment of motor control, and not a visuosensory deficit, underlies poor performance by thalamic subjects in the diffuse light condition. The unilateral lesions were proposed as having promoted ipsiversive progression in the apparatus, with the tendency more effectively governed when the visual field allowed the animal coincided with the direction of travel. The bilaterally thalamic rat was next examined in relation to a non-sensory interpretation of observed task deficits. EXPERIMENT 5

In the present discrimination apparatus, a subject seems logically compelled to adopt a "go, no-go" strategy to master the discrimination. Wearing the light-diffusing occluders, the subject cannot locate and run to the illuminated goal box, using retinal image cues. Rather, it must move to one side of ths choice chamber to determine if light is present on that side. If it is, the animal proceeds in that direction; if not, it must check this approach sequence, and move instead toward the opposite side. Thus, any movement bias toward one side of the box in the diffuse light condition should be more difficult for the brain-damaged animals to override (since they inhibit prepotent responses poorly), and should result in more lesion-group errors. This prediction in part follows Klfiver's emphasis on the need for such a response strategy in animals lacking spatial vision [14], and is partly derived from the observation that even normal rats with unrestricted vision seem to at first adopt a go, no-go strategy in performing the two-choice task. At the start of training, the animals exhibit only a "position habit", running consistently either to the right or the left goal box. They next reach the go, no-go phase where they begin to enter the favored goal box but, in the absence of the cue

880

COOPER, BATTISTELLA AND RATH

t '( \, t/, ~i ./). ;

,f

FIG. 8. Left to right, lesion reconstructions of subjects 101, 103, 105, 107 and 109 of Experiment 3.

light, stop, and run to the other, correct side. With further practice, the animals run directly to the illuminated goal box. According to our hypothesis, the occluded animal can never reach this third phase, but can only perform a go, no-go discrim&aation, based on the presence or absence of light when approaching the favored side of the apparatus. A variety o f experiments could be conducted to test this line of reasoning, including photographing the animal's movements, but as a first step, we adopted a less complicated technique. Subjects were trained to perform a typical "successive" discrimination (light in two alleys, go right; no light, go left) in the shuttle discrimination box. We assumed that in this new two-choice task, even the unoccluded animal would adopt and maintain the type of go, no-go strategy we believed to underlie the earlier, "one-lighted-alley" task in the diffuse light condition. Position responding is characteristic of the initial learning stages for any discrimination in the two-cboice apparatus. In addition, we saw little reason for the animal to relinquish its position habit to the preferred side even with prolonged training in the new task, since the subject could not "target" the correct goal box on the basis o f a contilcaous luminous cue: the correct box was no longer luminously distinct, and no increase in efficiency would result from abandoning a go, no-go response strategy involving examination of the same alley on each trail. If these inferences were correct, the differential effect should disappear in the new two-choice task, and the subjects with lesions would perform equally poorly in both occluded and unoccluded visual conditions. Except for the change to a successive, "two light versus

no light" task, Experiment 5 was procedurally identical to Experiments 1 and 2. Table 2 summarizes the postoperative performances of the four subjects and the lesions are shown in Fig. 11. There is no indication of a differential deficit, and the one subject (No. 36) which showed little effect in either visual condition, had a small lesion. The results of Experiment 5 supported our prediction that the differential deficit would disappear in the successive two-choice task. Even if our reasoning was faulty, given the equivalence of the deficit in the two visual conditions, an impairment in diffuse light perception had to be ruled out as an explanation of lesion-group inferiority in Experiments 1 and 2. EXPERIMENT6 While Experiment 5 demonstrated that the lesion-induced deficit does not appear solely in the diffuse light condition. the lesions might still increase the difficulty of discriminations which emphasize light intensity as a critical cue. The literature [23] indicates that rats given the type of task in Experiment 5, and faced with discriminanda differing in both form and intensity, reveal in generalization tests that intensity governs their discrimination behavior. As pointed out in Experiment 5, the positions of light stimuli in the two-light, no-light task are not spatially linked to the correct and incorrect goal boxes, as they are in the single-light version of the task. Consequently, the occluders might he needed in the single-light task to force the use of intensity rather than spatial cues, but become unnecessary in the two-light task, in

MIDLINE THALAMUS AND VISION

0.5

881

r

TABLE 2

Normal • Thalamic

0.4

NUMBER OF 50 TRIAL SESSIONS REQUIRED POSTOPERATIVELY TO REACH CRITERION IN THE UNOCCLUDED AND OCCLUDED STATES IN THE LIGHT-RIGHT, DARK-LEFT TASK (EXPERIMENT 5)

0.3 0.2 0

-

/

0.1 L

>.. ¢,.) Z

~

~

IJJ

~4~~'

",~

~m

~,

I

I

O.0

I

I

I

I

I

I

I

I

A Normal • Thalamic

0.4

t

--

A--

_

/

0.2

''zx/

_

A.a.AmA~A

0.1 0.0

I

I

J I

I

(~;m°'4

ONE-HOUR

I

I

I

I

SESSIONS

FIG. 9. Differential reinforcement of low rate schedule of performance of two sets of thalamic and normal rats (Experiment 4). Higher ratios mean better performance (see text).

IJ.I

50-

O A ....

O CONTROL GROUP A SUBCORTICAL GROUP ....A

rr 40IJJ

n tJ) IJJ tn 3 0 Ct) ILl r¢ n

z

Occluded Sessions

35 36 37 38

19 1 21t 10

18 2 21t 11

ONE-HOUR SESSIONS

0.3

D z

Unoccluded Sessions

tAnimals not trained to criterion.

dip .~-- e,,,,m

_w 0.5 (..) IJ. h

w

Subject

/XS S"

S

,..A S j,

o

20-

IJJ :S 0

I

I

I

I 2 3 I - V I 'IS-SEC -I t ~

I

I

4 5 VT 3 0 - S E C

I

6 I

DAYS O F T R A I N I N G ON A V I S C H E D U L E FIG. 10. Variable interval performance of thalamic and normal rats (Experiment 4).

which only a go, no-go strategy can be used in response to the light cue. From this perspective, the deficit always appears when the animal must respond to intensity cues-- and a lesion-induced impairment in intensity perception remains plausible. Thus, while the two visual conditions might lead to different response strategies, they might also lead to the use of different visual cues. Until these problems could be overcome, an adequate test of the thalamic animals was impossible. To circumvent the response problem we abandoned the two-choice apparatus and employed an operant-conditioning unit, with conditioned suppression of lever pressing as the behavioral indicator of the animals' vision. We found no reason to expect this simple withholding response to differ fundamentally between the occluded and unoccluded conditions. In addition, pilot experiments suggested that thalamic animals could withhold lever pressing well enough that their difficulty in inhibiting prepotent responses would not interfere with the anticipated experiment. The question of what visual cues the subjects used in the two visual conditions was answered by a successful demonstration of transoperative sparing and loss, respectively, in occluded and unoccluded animals, of a conditioned suppression response to a light stimulus following cortical lesions. This demonstration of a differential effect from cortical lesions ensured that the two visual states in fact led to different cue usage, and that the visual states per se, and not different responses engendered by those states, were responsible for previous observations of differential sparing after visual decortication. According to the present analysis, our previous demonstrations of habit sparing in occluded animals transoperatively were confounded, since they had always involved the use of the single light two-choice task, in which the responses of the occluded and unoccluded subjects would have differed. Finally, then, a satisfactory test of thalamic lesion effects could be conducted. Thirty normal rats were fu'st trained in operant conditioning units to lever press for 45 mg food pellets on a 30 sec variable interval schedule of reinforcement until response rates had stabilized. Posts were then implanted in the subjects' skulls, to anchor headgear, and the animals were returned to the units for additional lever press training and for adaptation to the light-diffusing occluders and control masks. Half of the subjects were assigned to the diffuse light state and half to the unoccluded condition. Next, the unconditioned responses to the 2 min light

882

COOPER, BATTISTELLA AND RATH

FIG. 11. Left to right, lesion reconstructions for subjects 35, 36, 37 and 38 (Experiment 5).

stimulus was examined by presenting the light on two occasions during each of two daily 1-hr lever pressing sessions. The Wight stimulus was provided by a jewelled 24 V lamp located 9.5 cm above the lever and 16.5 cm above the grid floor. The lamps were operated at a slightly-reduced voltage for the unoccluded subjects, to equate light intensity with that received by occluded subjects, whose diffusing masks do not transmit all of the light striking them. Conditioned suppression of lever pressing began by terminating the 2-min light stimulus with a 1 mA footshock lasting 0.5 sec. Preoperative training ended after the animals had received 14 light-shock pairings administered at a rate of two per day. The numerical index of conditioning was the "suppression ratio" of total presses during the 2-min light presentation divided by the total presses made during the 2 rain of darkness immediately preceding light onset. Thus a score of 1.00 indicated no suppression of lever pressing to the light, and a score of 0.00, asymptotic conditioned suppression. The subjects which had been trained wearing occluders were then divided into three subgroups, one of which underwent thalamic lesions, another posterior cortex removal, and a third, no surgery. The unoccluded subjects were subdivided and treated in the same manner. Following an 18-day recovery period, the animals were returned to the conditioning units, and after four leverpressing sessions, were tested for retention of the suppression response to the light stimulus. No shock accompanied the light on these extinction sessions. Figure 12 summarizes the course of the experiment for the six groups of animals.

Unfortunately, one each of the occluded and unoccluded animals receiving cortical lesions, and one of the unoccluded thalamic animals died. The mean curves for these groups shown in Fig. 12 are therefore based on only four animals each, while the other curves represent groups of five. As Fig. 12 illustrates, the results were unambiguous. The two unoperated groups, as well as the occluded cortical group, exhibited considerable suppression to the light during the postoperative extinction stage of the experiment. Retention was much poorer for the unoccluded cortical group as well as for the unoccluded and occluded animals with thalamic lesions. Of particular importance is the lack of any suggestion of a difference in retention between the two thalamic groups. This was reaffirmed by conducting a further, reacquisition phase of the experiment (Fig. 13), and once again, there was no indication that the diffuse-light subjects are more impaired by thalamic lesions. The thalamic and conical lesions are depicted in Figs. 14 and 15 respectively. The cortical lesions were deep, uncovering the underlying hippocampos and were accompanied by prominent lateral geniculate retrograde degeneration (not shown; see Bland and Cooper [3]). With no apparent basis for different response tendencies to have developed in the two thalamic groups, but with vision either restricted or not restricted to d ~ s e ~ t cues, both groups showed an equivalent performance deficit transoperatively. Experiment 6 flatly contradicts the claim that the pretectal area plays a special role in the perception of diffuse light or of light intensity. Rather, it appears that the

M I D L I N E T H A L A M U S A N D VISION

_

883

CONTROL GROUP • Occluded 0 Unoccluded

NR(3

CORT,CALGROUP

1.00

./

• Occluded

(~

[] Unoccluded

I~

SUSCORT,CALGROUP , occ,uded

it

~1 11

|I|

0.80

v.----

,~,

& Unoccluded INTACT - LESIONED . . . .

..Ja~/

/k ~

C O R T I C A L GROUP [] Unoccluded

I

ii A'

-

,

S U B C O R T I C A L GROUP & Occluded _ A Unoccluded

tt • t

It

,, •

t I

0

-

LESIONED ....

0.60

0

g

x

It I ,,os,-o,,.

-

tt

0.40

It

I

ix _

I',/\ f °°°F,

0.20 , , , , , , , , LIGHT ONLY

'~

3 5 ACQUISITION

7

t.~.

',',,.,,"

Yl

,,,-

, , ~ ,

-

_

I t

I EXTINCTION

"-

,.

s O - []

[3-.C]I

0.00 FIG. 12. Preoperative and postoperative conditioned suppression of occluded and unoccluded rats receiving cortical and subcortical lesions (Experiment 6). Lower ratios mean stronger suppression to the light CS (see text).

lesions undermine the ability to inhibit movement and this ability is stressed only in the occluded condition, in the single-light version o f the discrimination task.

I 2

I

I

I

4

I 6

REACQUlSITION

FIG. 13. Postoperative reacquisition of conditioned suppression by the occluded and unoccluded subcortical and unoccluded cortical animals (Experiment 6). Lower ratios mean stronger suppression to the light CS.

GENERAL DISCUSSION

The exact nature of the thalamic deficit still awaits specification but it seems to be most prominent when the animals must inhibit a prepotent response. This impairment is characteristic of damage to limbic system structures [7] and one wonders ff the hippocampal damage, frequently a feature of the subcortical lesions, contributed to the observed effects. It seems unlikely that hippocampal damage was wholly responsible, however, since in a number of cases, deficits were observed in subjects with minimal hippocampal damage. Moreover, we did not fmd that small hippocampal lesions produced the differential effect noted in Experiment 1. Nevertheless, we suspect that larger hippocampal lesions would duplicate all the subcortical effects reported in this paper. In this connection, Kimble [13] reported finding that hippocampal rats had difficulty in learning a successive but

not a simultaneous black-white Y-maze discrimination. According to the present analysis, Kimble's rats had difficulty with the successive task because it places a premium on the ability to check movement. F r o m our allusions to the limbic character of the observed deficits and our predictions that the deficits could also be obtained from damage to structures lying far from subcortical visual nuclei, it might appear that this study has no relevance to the problem of subcortical participation in perception. That conclusion we believe, cannot be justified. In an earlier project [5] we examined the claim that superior colliculus lesions disrupt visual perception. We found the claim to rest on nonvisual impairments which are related to those observed in the present study. Indeed, we question whether

0 (%

0

N

¢3

5-

~"

~ . . . .

~

~.~ ~ ~ ~ ~ ~ j l

<

.~

///

~

>

> ~.. Z

[...

> ,-..] ,-.]

© © "u

M I D L I N E T H A L A M U S AND VISION

885

FIG. 15. Dorsal view of brains showing posterior cortical lesions, Nos. 1-4 unoccluded subjects, Nos. 5--8occluded subjects. In all cases the lesions penetrate deeply, uncovering the hippocampus. Midline cortex remains in some subjects. The damage to the anterior part of the left hemisphere of No. 2 occurred when the skull was removed. The dimples lying anterior in the right hemispheres of Nos. 5 and 7 were produced by the screws anchoring the Teflon posts to the skull (Experiment 6).

any claims of subcortical visual perception in the normal animal can withstand critical analysis, and below we attempt to point out some of the hurdles making resolution of the problem difficult. The present study demonstrates the transoperative sparing of a diffuse light habit following visual cortex lesions. The phenomenon is apparently robust, now having been observed in quite dissimilar discrimination paradigms. This phenomenon in fact led us to search for the subcortical structures involved in the mediation of diffuse light discriminations. The present analysis provides no indication of pretectal area involvement. Perhaps some other structure such as the ventral part of the lateral geniculate body is important [17] but given some of the difficulties in analysis which this study points out, as well as the other problems mentioned below, a very cautious approach to the question seems imperative. It seems clear that dorsal midline thalamic and collicular damage affects many responses in an animal's repertoire. We, like others [20], have observed enlarged pupils and ocular deviations in animals with thalamic lesions. Tasks requiring acute vision or accurate retinal image placement could prove more difficult in the presence of such abnormalities. But the motor system effects go beyond receptor adjustments. In unpublished work R. Ferrier in our laboratory found that rats with colliculus lesions placed on an electrified grid, would not jump in order to escape or avoid electric footshocks, substantiating Lashley's earlier observation [16]. From the present study, as well as from the one involving the colliculus [5], it is also evident that even the animal's running movements are impaired. Such output complications of subcortical lesions can lead to faulty interpretation, as it has in some of our earlier work and as, we suspect, it has in many other studies. Nor does the placement of unilateral lesions provide a ready circumvention of the difficulty in ruling out nonvisual impairments. In Experiment 3 we argued that the better performance with the ipsilateral eye was consistent with the ipsiversive progression tendency created by the unilateral lesion. Certainly the appearance of ipsiversive progression, as well as the failure of the Experiment 3 discrimination find-

ings to correspond to those in Experiments 5 and 6, point to hazards in interpreting the behavior of unilateral preparations. Ipsiversive progression tendencies also occur in the animal suffering a unilateral coUiculus lesion, and it has been shown that these too can lead to faulty interpretation of the sensory status of the animal [5]. Many instances of so-called "sensory neglect" could just as well be described as "motor neglect" in studies of animals with unilateral lesions. A by-now extensive literature indicates that the decorticate animal has much greater visual capacity than was once supposed. Even the human hemidecorticate is said to possess coarse spatial discrimination in the affected visual hemifields [21]. Such work strongly suggests that at least in the decorticate condition subcortical structures can support complex stimulus registration. This need not imply, however, that these subcortical structures play any direct role in perception in the intact organism. Satinoff [22] has recently revived the Jacksonian stance of a hierarchially-organized neuraxis, including the idea that higher levels of control have evolved and inhibit lower levels. If so, perhaps only in the absence of cortical participation do subcortical structures play a direct role in the perceptual process. The problem of cortex-subcortex dynamics is just one more hurdle in the path of a clear answer to the question of subcortical participation in visual perception. It is a common finding that visually decorticate animals easily learn intensity discriminations. They can even demonstrate normal absolute and difference thresholds [8,18]. Moreover, as the present work has again demonstrated, diffuse light discriminations which might be regarded as " p u r e " intensity discriminations, are affected little by visual decortication. From such findings, we [4] and others [17] have proposed that the intensity dimension of vision is exclusively a product of subcortical structures in the normal animal. This proposal, however, may be challenged. There is work [18] which suggests that visually decorticate animals generalize improperly from one intensity to another. Ferrier [11] has also noted that even under diffuse light conditions decorticate rats are not as capable as normal rats of detecting an increase in the intensity of an illuminated lamp. Hence cortex may play a much greater role in intensity

886

COOPER, BATTISTELLA

p e r c e p t i o n t h a n is s o m e t i m e s s u p p o s e d , c o u n t e r i n g t h e ittea t h a t i n t e n s i t y v i s i o n is s u b c o r t i c a l l y m e d i a t e d . F i n a l l y , the m o s t c o n s i s t e n t f e a t u r e o f t h e s u b c o r t i c a t lesion l i t e r a t u r e is its i n c o n s i s t e n c y . C o n c e i v a b l y the p r e s e n t

AND RATH

s t u d y m a k e s a c o n t r i b u t i o n b y p o i n t i n g out, if only in o u t l i n e , s o m e o f t h e m e t h o d o l o g i c a l a n d c o n c e p t u a l pitfalls w h i c h h a v e c o n t r i b u t e d to the i n c o n s i s t e n c y in this r e s e a r c h area.

REFERENCES 1. Bauer, J. H. and R. M. Cooper. Effects of posterior cortical lesions on performance of a brightness discrimination task. J. comp. physiol. Psychol. 58: 84-92, 1964. 2. Birch, M. P., R. J. Ferrier and R. M. Cooper. Reversal set formation in the visually decorticate rat. J. comp. physiol. Psychol. 92- 1050-1061, 1978. 3. Bland, B. H. and R. M. Cooper. Experience and vision of the posterior neodecorticate rat. Physiol. Behav. 5:211-214, 1970. 4. Blochert, P. K., R. J. Ferrier and R. M. Cooper. Effects of pretectal lesions on rats wearing light-diffusing occluders, Brain Res. 104: 121-128, 1976. 5. Cooper, R. M., B. H. Bland, L. A. Gillespie and R. H. Whittaker. Unilateral posterior cortical and unilateral collicular lesions and visually guided behavior in the rat. J. comp. physiol. Psychol. 72: 286--295, 1970. 6. Cooper, R. M., K. P. Blochert, L. A. Gillespie and L. G. Miller. Translucent occluders and lesions of posterior neocortex in the rat. Physiol. Behav. 8: 693--697, 1972. 7. Cooper, R. M., R. J. Ferrier and M. P. Birch. Operant responding in rats with posterior midline cortical lesions. Brain Res. 125: 356-361, 1977. 8. Cooper, R. M., I. Freeman and J. P. J. Pinel. Absolute threshold of vision in the rat after removal of striate cortex. J. comp. physiol. Psychol. 64: 36-39, 1967. 9. Delacour, J. Effects of medial thalamic lesions in the rat: a review and an interpretation. Neuropsychologia 9: 157-174, 1971. 10. Dantzer, R. and J. Delacour. Modification of a phenomenon of conditioned suppression by a thalamic lesion. Physiol. Behav. 8: 997-1003, 1972. I I. Ferrier, R. J. Functions of the Cortical Visual System. Unpublished Ph.D. thesis, University of Calgary, 1979.

12. Horei, J. A., L. A. Bettinger, G. J. Royce and D. R. Meyer. Role of neocortex in the learning and relearning of two visual habits by the rat. J. comp. physiol. Psychol. 61: 66--78, 1966. 13. Kimble, D. P. The effects of bilateral hippocampal lesions in rats. J. comp. physiol. Psychol. 56: 273-283, 1963. 14. Kliiver, H. Functional significance of the geniculostriate system. Symp. Soc. exp. Biol. 1: 253-299, 1942. 15. Krnig, J. F. R. and R. A. Klippel. The Rat Brain. Baltimore: Williams and Wilkins. 1963. 16. Lashley, K. S. The mechanism of vision. XII. Nervous structures concerned in the acquisition and retention of habits based on reactions to light. Comp. Psychol. Monogr. 11: 43-79, 1935. 17. Legg, C. R. and A. Cowey. Effects of subcortical lesions on visual intensity discriminations in rat. Physiol. Behav. 19: 635646, 1977. 18. Levere, T. E. and J. Mills. Residual differential brightness thresholds following removal of visual cortex in rats. Physiol. Psychol. 5: 490--496, 1977. 19. Miller, L. G. and R. M. Cooper. Translucent occtuders and the role of visual cortex in pattern vision. Brain Res. 79- 45-59, 1974. 20. Pasik, P., T. Pasik and M. B. Bender. The pretectal syndrome in monkeys. I. Disturbance of gaze and body posture. Brain 92: 521-534, 1969. 21. Perenin, M. T. Visual function within the hemianopic field following early cerebral hemidecortication in man. II. Pattern discrimination. Neuropsychologia 16: 697-708, 1978. 22. Satinoff, E. Neural organization and evolution of thermal regulation in animals. Science 201: 16--22, I978. 23. Sutherland, N. J. and V. Hoigate. Two cue discrimination learning in rats. J. comp. physiol. Psychol. 61: 198-207, 1966. 24. Vanderwolf, C. H. Effects of medial thalamic damage on initiation of movement and learning. Psychon. Sci. 17: 23-25, 1969.