3 r11'\_ a»\'."); 8 It VII; . ~I Q a o. 1;! l I 4 ’ r I I nfli - .-. I .1 II ‘ . . v V.\ fl... .a H3"? ,Jv: ISL.“ ,1 {a .5 «f ’3‘. \ .76.. . LI B RA R Y ' Michigan State University ABSTRACT RETENTION AND LEARNING IN NORMAL AND BRAIN-LESIONED RHESUS MONKEYS AFTER.A SIX-YEAR INTERVAL By Bruce Ray Marshall Four experiments were conducted to assess retention and new visual discrimination learning by four normal and three operated (posterior association area lesions) rhesus monkeys remaining from a previous study by Raisler (1966). In the previous study, one lesioned male showed perfor- mance comparable to the controls while the two lesioned females showed little evidence of postOperative retention or new learning. In Experiment I, all subjects received training on five new three-dimensional pairs. Each pair was presented for a maximum of three days (40 trials per day). In addi- tion, those subjects who did not reach criterion on two or more of the five pairs also received more extended training on additional pairs. The results indicated that all lesioned subjects could learn three-dimensional dis- crimination problems if given extended training on indi- vidual pairs. Performance on pairs presented for a max- imum of three days of training, however, was comparable to performance on similar pairs in Raisler's (1966) study. EXperiment II involved a retention measure of Bruce Ray Marshall previously presented three-dimensional pairs. Each subject received two trials with each of 32 of a set of 6# pairs learned six years previously under four condi- tions of trial and problem sequences. Three of the four normal subjects showed significant retention scores (about 75%) while none of the lesioned subjects demon- strated any evidence of retention. The lack of retention by the lesioned male who had originally shown learning was attributed to a memory deficit. Some evidence was also found suggesting that conditions of trial and problem sequence during initial learning can affect retention. In EIperiment III, the effect of reversing part of the sequence of operations in the presentation of a trial was tested. The manipulation of raising the Opaque door of the WGTA following the placement of the Kluver tray near the test cage (rather than the reverse sequence) resulted in increased errors for two lesioned subjects but no differences for the normal controls. In Experiment IV, three planometric form and three planometric color problems were presented to determine the saliency of color and form cues to the lesioned sub- jects. The lesioned male who had demonstrated learning comparable to the normal controls met criterion on nearly all color problems and one form problem. The marked increase in errors for this subject, however, indicated that color cues alone were not sufficient to account for this subjects normal learning scores on three- Bruce Ray Marshall dimensional problems. Moreover, the other lesioned sub- jects showed no evidence of discrimination of either color or form. ft“) 4 .7!”- 0 Approved flffwfl £141,041 Robert L. Raisler Committee Chairman Lester M. Hyman John I. Johnson Committeemen Date flufi 1‘ A” / ¢7/ RETENTION AND LEARNING IN NORMAL AND BRAIN-LESIONED RHESUS MONKEYS AFTER A SIX-YEAR INTERVAL BY Bruce Ray Marshall A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1971 To Grandpa (671) 11 ACKNOWLEDGMENTS I first wish to thank the members of my committee. Dr. Lester M. Hyman's criticisms of my first pr0posal provided me with a better understanding of the assets and liabilities of various research designs. Dr. John I. Johnson's suggestions of the possibility of cueing led to what I consider to be the most interesting and as yet uninterpretable aSpect of this study, that is, the fac- tor reSponsible for the sudden learning in the latter part of EXperiment I. In addition, Dr. Johnson's later suggestions concerning the organization of this thesis provided me with a more intuitive understanding of the purpose and format of a research report. I am especially indebted to Dr. Robert L. Baisler, who as my thesis chairman and advisor, has devoted both a considerable portion of his time and the resources of his laboratory. He has provided continuous direction and constructive suggestions throughout this, as well as other, undertakings. This research would not have been possible without his contribution. In addition he has served as a model of how a scientist struggles with the problems inherent in an unexplored area. I am eSpecially thankful for this latter experience. 111 My family also deserves acknowledgment. My mother and brother have provided both affection and support throughout my years as a student. Without their faith, this endeavor might never have begun. And finally, my wife as both critic and typist, has substantially aided me in the preparation of this thesis. Without her talents and constant encouragement this endeavor might never have been completed. This investigation was supported by Biomedical Sciences Support Grants 5805FR07049-02 and -03 to Robert L. Raisler from the General Research Support Branch, Division of Research Resources, Bureau of Health. Professions Education and Manpower Training, N. I. H. iv TABLE OF CONTENTS LIST OF TABLES . . . . INTRODUCTION. . . . . METHODS . . . . . Subjects . . . . Apparatus . . Adaptation Procedure EXPERIMENT I: NEW LEARNING methOd O O 0 O O 0 Results 0 O O O O O EXPERIMENT II: RETENTION AFTER SIX methOd C O O O O 0 Results. . . . . . EXPERIMENT III: PRESENTATION METHOD NethOd O O O O O 0 Results. . . . . . EXPERIMENT IV: COLOR OR FORM CUES. methOd O O 0 O O 0 Results 0 O O O O 0 DISCUSSION . . . . Learning . . . . Retention . . . Presentation Method. Color-Form. . . Individual Differences LIST OF REFERENCES. . . LIST OF TABLES Table Page 1. Number of Problems Learned to Criterion and Total Number of Errors (Experiment I). 14 2. Six-Year Retention Scores of Pairs Pre- sented under Four Conditions of Trial and Problem Sequences. . . . . . . . 19 3. Trial Two Relearning Scores . . . . . 21 4. Number of Errors on Pairs with Two _ Methods of Presentation . . . . . . . 25 5. Total Errors on Planometric Color and Form Problems . . . . . . . . . . 29 vi INTRODUCTION The functional significance of the posterior associa- tion area in the rhesus monkey has been a major research question over the past three decades. Since Kluver and Bucy's (1938) report of "visual agnosia" following bilat- eral temporal ablation, numerous investigators (for a review, see Mishkin, 1966, or Wilson, 1968) have dealt extensively with this issue. Unfortunately the results of their efforts have in general only magnified the com- plexity of the problem. To date, no parsimonious explan- ation exists for the detrimental effect of the posterior association area lesion on visual discrimination behavior nor has the controversy been completely resolved concerning the importance of various subareas of the posterior associa- tion area and their possible interconnections with striate areas. Pribram (1954) has strongly implicated the infero- temporal area as prepotent for visual discrimination problems, particularly the two-dimensional patterns. The finding of a severe, but not permanent, pattern discrim- ination deficit following bilateral inferotemporal lesions has been consistently replicated both by Pribram (Pribram, Blehert, & Spinelli, 1966) and by others (Mishkin, 1966; 1 2 Wilson,Wilson, & Sunenshine, 1968). Pribram (1958) suggests that the inferotemporal cortex exerts efferent control over striate cortex via subcortical structures and also denies the possibility of any cortico-cortical connections from inferotemporal cortex to striate cortex via prestriate areas. Mishkin (1966), on the other hand, argues against the possibility of efferent control via subcortical nuclei, and proposes that the prestriate areas exhibit a high degree of equipotentiality as a cortical relay between inferotemporal and striate areas. He suggests that removals or crosshatching of prestriate areas have been incomplete possibly due to the danger of intrusion into geniculostriate radiations. Mishkin further proposes that recovery of pattern discrimination following bilateral inferotemporal lesions is contingent upon intact prestriate areas, and has presented evidence (Ettlinger, Iwai, Mishkin, & Rosvold, 1968) that the combination of inferotemporal and complete prestriate lesions performed in either order results in a permanent pattern discrimination deficit. The finding of a permanent pattern discrimination deficit, however, is not new. Harlow and his colleagues (Warren & Harlow, 1952; Raisler & Harlow, 1965) have previously reported such an effect following posterior association lesions although their subjects did not receive such extensive ablations of inferotemporal cortex. Returning to the question of the nature of the 3 observed deficit, it was stated that no parsimonious explanation has been presented. That is not to say that simple explanations have not been proposed but rather that one-factor explanations of attention or sampling deficits (Butter, 1968; Stamm & Knight, 1963), retention deficits (Weiskrantz, 196M; Wilson et al., 1968), or inability to withhold reSponding (Schilder, Pasik, & Pasik, 1968), have not been sufficient to encompass all of the reported results. Any theory of cortical func- tioning in the rhesus must be able to account for the results of a number of intriguing studies. First, the effects of inferotemporal or posterior association lesions seem to be Specific to certain aSpects of the visual discrimination task. It is a well established fact that visual field defects or deficiencies in eye-hand coordina- tion in picking up small objects, are not usually found following inferotemporal lesions (Weiskrantz & Cowey, 1963; Wilson & Mishkin, 1959). Yet these inferotemporal animals show gross deficiencies in discriminating pairs of objects or patterns in the formal testing situation. Second, the behavioral effects of the inferotemporal lesion appear to be Specific to the visual modality. Experiments using analogous training procedures in olfactory, tactile, and auditory modalities have not resulted in such large deficits following inferotemporal lesions (Brown, Rosvold, & Mishkin, 1963; Schilder et al., a 1968; Weiskrantz & Mishkin, 1958). Third, preoperative overtraining on a stimulus pair does not lead to the pronounced deficit following temporal ablation (Chow & Orbach, 1957; Orbach & Fantz, 1958). The effect of overtraining is Specific to the over- trained pair, i.e., non-overtrained pairs result in the characteristic deficit in the same subjects. Fourth, the degree of deficit and permanence of the deficit following inferotemporal or posterior association area lesions depends upon the type of discrimination problem. Warren and Harlow (1952) report, following posterior association lesions, almost complete recovery of the ability to discriminate three-dimensional object pairs (varying in color, form, and size). Although object discriminations were not impaired after a 14- month postoperative interval containing extensive dis- crimination training, no substantial evidence for learning was found for two-dimensional form or size discriminations. A later study by Raisler and Harlow (1965) suggests that recovery of three-dimensional discriminations may depend upon utilization of color cues. This finding is con- sistent with the above chance performance on color dis- criminations in Warren and Harlow's data as well as Warren's (1953) results showing the saliency of color cues in the intact rhesus. Although it is clear that extended postoperative training can lead to high levels of performance on 5 certain visual tasks, the question of whether this recovery can be spontaneous remains unanswered. Barlow (1939) found no evidence of spontaneous recovery in monkeys on patterned strings problems following unilateral occipital lobectomies. He interpreted the results of re- training as learning to counteract the effects of a hemianopsia. Chow (1952) found only minimal evidence of recovery without retraining. Of two monkeys that had been in darkness for three months postOperatively, one monkey showed improved post-darkness retention scores (in comparison with postoperative retention scores) on both preoperatively learned color and brightness dis- criminations. In addition two other monkeys had received three months of postOperative training on new discrim- inations (rather than darkness). They showed even greater (and more consistent) retention of the three preopera- tively learned discriminations following the additional training. Moreover, other studies have led to somewhat contradictory results. Stewart and Ades (1951) have pre- sented evidence of retention of an auditory discrimi- nation in a shock avoidance task where the superior temporal gyri were removed in two successive Operations separated by seven days or longer. Shorter interoperative intervals did not result in such high levels of retention. In a later study Meyer (1958) has attempted to clarify Stewart and Ades' finding. Meyer's results indicate that with an interoperative interval of 12 days, a brightness 6 discrimination was retained in rats who spent the inter- operative interval in their home cages but was not re- tained in rats who spent the interval in darkness. As Meyer has stated, his results demonstrate the impor- tance of postoperative experience and suggest that re- learning may not have to be specific to the formal test situation. The aforementioned studies of spontaneous reorgan- ization have in general dealt with animals lesioned during their adult life. The effects of lesions per- formed early in life suggest a rather different picture. Experiments on monkeys with lesions of frontal associa- tion cortex (Harlow, Akert, & Schiltz, 196“), motor cor- tex (Kennard, 1942), and posterior association cortex (Raisler & Harlow, 1965), all show less severe deficits from early lesions than from lesions performed on the adult monkey. Such findings are analogous to the recov- eries from aphasia in human patients following cerebral trauma at an early age (Penfield & Roberts, 1959). A later study by Raisler (1966), however, has re- sulted in contradictory evidence regarding the effects of early versus late lesions. This study, which dealt with posterior association area lesions including major portions of inferotemporal cortex, did not result in the expected decreased deficits in the early lesion group. Although the histological processing has been partially completed for the early lesion group, the older 7 lesioned subjects as well as the older controls have not yet been sacrificed. The present study is a followup assessment of the visual discrimination capabilities Of the remaining subjects. The assessment approach, involving a series of short tests (e.g., Milner, Corkin, & Teuber's (1968) periodic testing of H.M., a patient with a severe hippo- campal amnesia syndrome), was chosen to provide the most information concerning both retention over a six year interval and the nature and severity of the visual dis- crimination deficiencies of the remaining lesiOned sub- jects. More Specifically, this study attempted to answer ‘ four questions: (a) Can the lesioned animals which have demonstrated little evidence of postoperative learning meet criterion if given extended training on a single problem pair? (b) After a six-year interval, is there any evidence of retention of previously learned object discriminations and is this retention related to the lesion? (c) Can the sequence of Operations in the presen- tation of a trial affect performance? and finally (d) Is the magnitude of the deficit in the lesioned animals related to the type of differences between the positive and negative stimulus (1.6. color versus form)? METHODS Subjects The seven rhesus monkeys (Macaca mulatta) tested in all stages of this study served in a previous exper- iment by Raisler (1966). Four subjects (two males, two females) constituted the unOperated adolescent controls (Group 38N) of that study. The remaining subjects (one male, two females) were part of the adolescent Operated group (38T) that received extensive bilateral posterior association area lesions, probably including major portions of inferotemporal cortex. Although histological confirm- ation of the extent of the lesions has not yet been accomplished, there is some suspicion that Monkey 672 may have a partial visual field deficiency due to an unplanned intrusion into the geniculostriate radiations. Preoperatively all subjects received training using 16 three-dimensional (stereometric) and 16 two-dimen- sional (planometric) Object pairs. Postoperatively test- ing included new three-dimensional pair learning, and retention and relearning of the two- and three-dimensional pairs learned preoperatively. In addition, all subjects postoperatively received two months of training on 64 pairs of Objects under four different conditions of trial 8 and problem sequences. The results of the extensive postoperative testing indicated that two of the surviving subjects from the adolescent Operated group showed essentially no evidence of retention or new object pair learning while the third subject showed abilities approaching, but not equalling, the unOperated controls. Apparatus Testing in all stages was accomplished using a mod- ified Wisconsin General Test Apparatus (Harlow, 19h9) attached directly to the front of the test cage. The test cage was located in a semi-soundproof room and was illuminated during testing by a light from the WGTA itself. The test cage consisted of a frame of wood to which the sides, floor, and roof of woven stainless steel cable were attached. The inside dimensions of the cage were #8”w;x 30"h.x 2h"d, thus allowing considerable movement of even the largest male. The WGTA consisted of a box- like structure, Open at two ends, suSpended from the front of the test cage. The experimenter sat at one Open end and looked through the box to the other Open end at the front of the test cage. Between trials a vertical sliding door near the front of the cage could be lowered to occlude the subject's vision. During a trial, a Klfiver tray (described in Raisler & Harlow, 1965) was pushed along the floor of the box to the front of the test cage. A piece of white cloth, suspended from the experi- 10 menter's end of the WGTA allowed viewing of the sub- ject during a trial but prevented eye-tO-eye contact. Ample lighting was provided by a flush mounting ceiling light attached to the inside roof of the WGTA. Adaptation Procedure Adaptation prior to testing consisted of two stages. First each subject was placed in the test cage for a mini- mum of 15 minutes per day for a total of eight days. When necessary, additional training was given to those subjects who entered the tranSport cage reluctantly. When each subject had displaced a wooden block to obtain a food reward with latencies consistently under five seconds in 10 daily trials on two consecutive days, then all sub- jects received an additional 10 trials on the day immedi- ately preceding the onset of testing. EXPERIMENT I: NEW LEARNING The first stage of testing provided a rough esti- mate Of new object pair learning. Method A total of 9 stimulus pairs were randomly selected from a population of 75 new stereometric pairs con- sisting of multicolored, multisized, and multishaped junk objects mounted on white 3" x 3" masonite bases. These new pairs had been randomly paired subject to the restriction that the members of a pair differ in one color. All subjects received training with five new three-dimensional pairs at the rate of one pair per day, #0 trials per day, until the criterion of 32 cor- rect of 40 trials on a single day had been met or until three consecutive days of training on that pair had failed to produce criterial performance. If at the end of the fifth problem, a subject had not reached criterion on two or more of the five problem pairs, at least one additional pair was given until criterial performance was met or until at least 12 days of testing had failed to produce criterial performance. 11 12 In the first part of Experiment I (testing on five pairs for a maximum of 120 trials each) as well as in part of the extended training given to individual sub- jects, a trial was defined in the usual manner. That is, the Klfiver tray with the positive stimulus covering the baited foodwell and the negative stimulus covering the empty foodwell was pushed to the front of the test cage after or during the time the door was raised. Following a displacement of either stimulus by the subject, the tray was immediately pulled back and the door lowered. During the intertrial interval, the previous response was recorded and the preparation for the next trial was accomplished (baiting and positioning of the positive and negative stimuli). Positioning of the positive stimu- lus was varied according to a Gellermann (1933) series. During the latter part of the extended training, how- ever, certain procedural variations in the presentation of a trial were achieved. These variations included pushing the tray to the front of the cage prior to raising the door, the inclusion of an occasional cor- rection trial, and deliberate attempts to mask any audi- tory cues that occurred in the baiting process. These variations were presented primarily as a check for cueing. Results Number of problems learned and total errors for the first five new stereometric pairs are presented in I o . . 1L . . N . A J \ O a - . ~| J . I I. .I.‘ l 4. 1 .L . . I I I a a... . a . a . . . i - e. . A a I I! I . l . . , . _ L t, . . . . s . . I. . . e l.‘ I I v . .1 .. . c. o . 1 . u . .. \, J _ . u . cl. I us i l l I l e l. I ._ 1 II J I. ,4 a J \ n . O . . . I . . a . . . 13 Table 1. None of the normal subjects had any diffi- culty in making any of the discriminations whereas the lesioned subjects showed highly variable perfor- mance. The lesioned male (671) showed performance com- parable to the controls while the two lesioned females showed either chance performance throughout (672) or above chance but below criterial performance (670) on the third training day of four of the five problems. Since subjects 670 and 672 did not reach criterion on two or more of the first five problems, they received training on three and four additional pairs, reSpec- tively. For the first 10 days of testing on the first additional problem, Monkey 672 remained at chance. On days 11 and 12, however, above chance but below cri- terial performance began to appear. As a result an additional day of training was given. On this day condi- tions Of presentation were varied partially to elim- inate the possibility of cueing the subject through un- intended means. Surprisingly enough, on day 13 Monkey 672 made only 2 errors in 36 trials. One additional problem given with the varied presentation conditions was learned within two days. The third new problem for Monkey 672 was given with the standard method Of presentation, and again criterion was met, but only after six days of training. Monkey 670 not criterion in four days each on two additional pairs with the standard method of presentation. '\ 14 Table 1 Number of Problems Learned to Criterion and Total Number of Errors (Experiment I) Group Subject Problems Learned Total Errors 666 5 12 667 5 16 Normal 668 5 26 669 5 6 670 1 221 Temporal 671 5 10 672 0 304 15 This subject also received two pairs with the varied presentation conditions. Criterion on these last two pairs was met within two and three days respectively. Summarizing the results of the additional testing, it appears that given sufficient training on individual pairs, both subjects 670 and 672 can learn new stereo- metric problems, although their performance on the first three days may not suggest any signs of improvement. The possibility that method Of presentation of a trial can affect performance was suggested and provided the basis for Experiment III testing. EXPERIMENT II: RETENTION AFTER SIX YEARS The second experiment provided a retention measure of stereometric pairs that had been postoperatively learned six years previously. The first trial with each pair was considered as the pure retention measure, where- as a second trial performance was included to provide a crude measure of the learning set performance if, and only if, no retention was demonstrated. This stage of testing was of interest primarily because of the unusually long interval since the original learning. Method Stimuli consisted of a sample of 32 of the 6“ stereo- metric problem pairs (the first four pairs of each Of the eight sets of eight pairs) from Experiment II of the previous study (Raisler, 1966). In that study all subjects received training on 64 stereometric problem pairs under four conditions of trial and problem sequences. A prob- lem was defined as the presentation of a stimulus pair, i.e., a positive and a negative stimulus. During two months of testing, each subject twice received one week of training on each of the four conditions, subject to the restriction that both the first and the second months 16 17 of training each include all of the four conditions. During each week of testing (during any condition), each subject received five days Of training on 8 new problems. On any day of testing, for all conditions, each subject received 8 problems per day, 5 trials per problem, for a total of #0 trials per day. What varied in the four conditions was the sequence of the #0 daily trials with the 8 problems. In condition "1", each subject received 5 trials with problem 1, then 5 trials with prob- lem 2, then 5 trials with problem 3, etc., until 8 prob- lems had been presented for a total of 40 trials per day. For the remaining four days of that condition, the same. 8 problems were presented in the same sequence of trials. In condition "2", each subject received trial 1 with each of the first 2 problems, followed by trial 2 of each of the first 2 problems, etc., until each of the 2 problems had been presented for 5 trials. This sequence was fol- lowed by three more pairs of problems presented in the same sequence for a total of 8 problems and 40 trials. The remaining four days of condition "2" training each contained the same sequence of #0 trials with the same 8 problems. In condition "4", each subject received trial 1 with each Of the first 4 problems, followed by trial 2 of each of the first h problems, etc., until each Of the first 4 problems had been presented for 5 trials. Four more problems were presented in the same manner for a _ - , . 4‘ , i i. _ ; . - i. . . - - .t ‘ . I \ ., ‘ n I - s Q 18 total of 40 trials and 8 problems. As in the other con- ditions, the remaining four days of training contained the same sequence. In condition "8", the sequence of trials was presented in an analogous fashion, i.e., trial 1 of each of the 8 problems, followed by trial 2 of each of the 8 problems, etc. In the present study, during eight testing days, each subject received a sample of 32 of the 6# pairs from the four conditions varying in problem and trial sequences. Four problems, two trials per problem, were presented on each of eight days. To maximize retention the 32 problems were presented in the same order as each subject initially experienced them. The same method of presentation of a trial (i.e. without correction, etc.) was used as in the first part of Experiment I. Results The retention scores on trial one for both the lesioned and control subjects are presented in Table 2. Summing across conditions for each subject, it is evi- dent that three Of the normals had total correct scores greater than or equal to 23 (Ho: Néé 16, p <..01, one-tailed test, binomial distribution). Although the use of the binomial does allow treatment of the individ- ual subject, the overall °