MATURATIONAL INFLUENCES ON PERSEVERATION OF AVOIDANCE AND REVERSAL LEARNING AFTER SELECTED BRAIN DAMAGE IN RATS

Transcription

1 ACTA NEUROBIOL. EXP. 1988, 48: MATURATIONAL INFLUENCES ON PERSEVERATION OF AVOIDANCE AND REVERSAL LEARNING AFTER SELECTED BRAIN DAMAGE IN RATS James F. BRENNAN, Carolyn A. COHEN, and Peter A. BERTUCCI Department of Psychology, University of Massachusetts at Boston Boston, MA 02125, USA Key words: avoidance, frontal cortex, caudate, hippocampus, ontogeny, perseveration, rat Abstract. In Experiment I, 18 weanling and 18 adult male rats received bilateral electrolytic lesions of the dorsomedial or ventrolateral prefrontal cortex, or the hippocampus, while 12 additional pups and adults served as nonoperated controls (n = Glgroup). Subjects were observed for perseverative responding in acquisition of a step-up avoidance task, followed by reversal training and extinction. Thirty days later, subjects were retrained and tested in the same manner. During initial training, the numbers of trials to criteria, errors, and latencies on the last 5 trials all indicated significant effects from age, primarily, and surgery, secondarily. After 30 days, surgical effects assumed a more dominant role, with hippocampal and ventrolateral damage producing the greatest extent of response perseveration. Experiment 11 replicated the essential procedural sequence as Experiment I, but included damage from combined lesions. 24 weanlings and 24 adults received bilateral lesions to the dorsomedial prefrontal cortex, the dorsomedial prefrontal cortex plus caudate nucleus, or sham lesions. Subjects from this experiment were trained after 7 days or 60 days recovery (n = G/group). The acquisition results indicated that all independent variables were significant, but only the age effect attained significance in the extinction data. Both experiments point to the prafound influence of age at the time of damage in accounting for recovery of avoidance behavior.

2 INTRODUCTION Developmental comparisons of motor behavior under conditions of aversive motivation typically indicate that weanlings show deficits relative to older rats in the acquisition and retention of passive avoidance tasks (20) and active avoidance (see 5, 7, 15, 17, 19). The avoidance deficit has been interpreted as a result of the functional immaturity of underlying response inhibitory systems. Weanlings may acquire conditioned fear as rapidly as adults, but be unable to inhibit the motor response that places them in the presence of fear eliciting stimuli. Likewise, there is evidence that weanlings process information within sensory modalities at different rates from adults (14). Accordingly, weanlings show a persistence of prepotent responses that may be viewed as stereotypic. The integrative behavior that is exhibited by mature organisms presumably requires a greater neural capacity than earlier developed behavioral responses. However, when levels of arousal are increased due to stress, the neural capacity may become overloaded and earlier behavior patterns may appear (21). Thus, perseveration seems to be coupled with behavioral arousal and becomes controlled by the maturation of the nervous system. Studies of selective damage to the prefrontal cortex indicate that rats become hyperreactive to noxious unconditioned stimuli (US), a pattern exaggerated in younger subjects (1). In an early review, Brutkowski (2) suggested that the orbital subdivision of the prefrontal cortex is implicated in perseverative responding. Brutkowski (2) noted that the descending pathways from the prefrontal cortex consist of a dual fiber system to the limbic and subcortical structures. The hippocampal formation receives fibers through the fasciculus cinguli from the medial prefrontal cortex. The loss of inhibition in prefrontal rats may reflect the relation of the prefrontal cortex to the limbic system (see 12, 13, 18). On the basis of similarities of the behavioral effects of hippocampal and prefrontal lesions in different species, Nonneman, Voight and Kolb (16) suggested that these structures are functionally interrelated. Lesions in both the prefrontal cortex and the hippocampus produce deficits in passive avoidance, discrimination reversal and extinction. However, adult hippocampal rats are able to learn 2-way active avoidance faster than normals or rats with neocortical lesions (6, 22), while differences between hippocampal rats and normals in l-way active avoidance acquisition is less clear (8). The deficits in passive avoidance acquisition suggest that the hippocampus is involved in the cessation of responses

3 rather than in their initiation. Moreover,.the enhancement of 2-way active avoidance behavior appears to result from increased levels of locomotion after hippocampal damage. EXPERIMENT I Experiment I focused on age related differences in perseverative stereotypy defined as the difficulty in altering a response pattern to a given stimulus. Dissociated lesion effects were investigated through comparisons among various groups of young and adult rats, which included controls, bilateral lesions in two sites of the prefrontal cortex, and bilateral hippocampal lesions. Some data (e.g., 10) indicate that recovery of function in species typical behavior in pups after cortical lesions is not as substantial as once believed. Since young rats show inhibitory deficits, lesions to these areas at a critical age may result in a disruption in the development of mechanisms that are involved with the inhibition of responses. Method Subjects. Male pigmented rats were bred in the laboratory of the University of Massachusetts at Boston, and the entire colony was maintained on a 12 h light/dark cycle. Of the 48 naive rats used in this experiment, 24 were day old weanlings and 24 were day old adults at the time of surgery. All pups were housed with the mother until weaning at 25 days of age. To preclude detection of the operated pups, all of the pups of the entire litter were washed with alcohol before replacing the operated subjects with the mother prior to weaning. All adults and pups after weaning were housed alone. Food and water were available ad lib. except for a 24-h food deprivation period prior to surgery in the adult subjects. Apparatus. A three compartment clear acrylic chamber was used throughout the experiment. The floor of the center compartment was slightly lower than the side compartments. The total dimensions of the center compartment measured 20 cm on each side and 35.5 cm high, from the floor, for the adults. Due to the addition of a false floor used t.o accommodate the smaller size of the pups, the height of the center compartment was reduced to 33.5 cm. Each side compartment measured 20- X 20- X 30-cm and could be entered through guillotine doors, 9- X 15-cm, located on each side of the center compartment. The floor of the entire apparatus consisted of steel rods spaced 1.5 cm apart, through which scrambled shock (BRS Lehigh Valley) from a matched impedance source (4) could be delivered. There are psychophysical data showing

4 that this type of shock is equally aversive across ages at moderate intensities (3). Latency to the correct response was recorded by a Lafayette Instruments timer to 0.01 s. The only light source was 15 W bulb suspended 55 cm above the center of the chamber. Procedure. Surgery. Eighteen rats at each age level were randomly assigned to one of three conditions: Bilateral lesions of the dorsomedial (DM) prefrontal cortex, the ventrolateral (VL) prefrontal cortex, or the hippocampus (H). Subjects were anesthetized with S.C. injection of Nembutal (50 mglkg). For surgery, subjects were mounted in a stereotaxic frame manufactured by Kopf Instruments, and trephine holes were drilled using an Emesco Dental Drill. A radio frequency lesion generator (Radionics, Model RFG-4) activated a temperature sensing electrode with 1 mm exposure. Electrolytic lesions were made by maintaining a constant electrode tip temperature of 70' C for 60 s. The coordinates used for the three lesion procedures were as follows: DM lesions were made 2 mm bilateral to the midline and 2.3 and 4.0 mm anterior to bregma in pups- and adults, respectively, and the electrode was lowered 1.0 and 1.5 mm from the dura, respectively in the pups and adults. VL lesion coordinates were 4.3 and 5.0 rnm bilateral and 2.6 and 3.0 mm anterior to bregma, and the electrode was lowered 3.4 and 6.0 mm for pups and adults, respectively. The hippocampal lesions were made 1.5 and 3.0 mm bilateral and 1.5 and 3.0 mm posterior to bregma, and the electrode was lowered 1.5 and 3.0 rnm from the dura in the pups and adults respectively. The adults were allowed to recover in the home cage for 7 days, and the pup were returned to their mother and litter mates after awakening, and also recovered for 1 week. After surgery, the subjects were placed in an Armstrong X4 Infant Incubator (Ohio Medical Products, Model 500). The remaining 12 rats at each age level served as nonerated controls. Behavior. Each subject was randomly assigned to a step-up active avoidance task of go right or go left, and training began after 5 min of free exploration of the chamber. All training procedures at the initial stage of the experiment took place on one day. On a given trial, the subject was placed in the center compartment, oriented toward the rear of the compartment, and exposed to the nondifferentiated conditioned stimulus (CS) that consisted of raising the guillotine doors on both sides of the center compartment. If the subject did not emit the correct response choice within 5 s, a 1 ma footshock US was delivered until the ccrrect response terminated it. F01low;ng the first shock trial, subjects remained fairly still upon subsequent placement hto the center compartment, insuring that the appropriate orientation of the correct res-

5 ponse was consistent for subsequent trials. Trial latency was recorded, and entry into the wrong compartment was scored as an error and did not affect the shock contingency. Between trials the subjects were placed in a holding cage for approximately 30 s. Acquisition training terminated after 5 consecutive nonshock trials, and on the next trial, reversal training began in which subjects were required to switch orientation to the opposite side compartment. The same criterion was used for reversal learning as during acquisition training. Following reversal training, extinction testing began on the next trial by omitting any possibility of shock presentation, and this procedure continued until the subject remained in the center compartment for a total of 10 s on 5 consecutive trials. Extinction trials were terminated if the subject did not respond within a total of 15 s. Thirty days later, all subjects were retrained in the same manner on Day 1 until they reached the same acquisition, reversal and extinction criteria. Eight subjects did not meet the initial acquisition criterion of 5 consecutive avoidance responses within 75 training trials and were replaced with naive subjects. These animals included 1 nonoperated adult, 1 adult and 1 pup given DM lesions, 1 adult and 2 pups with VL damage, and 2 H damaged adults. In addition, 1 H lesioned adult failed to meet the 30 day reacquisition criterion and was excluded from the analysis of that stage of the experiment. Histology. Following 30 day extinction, the surgical subjects were deeply anesthetized and intracardially perfused with 10 /o formalin. The brains were removed and stored.in 10 /o formalin for several days, after which they were blocked, frozen, sectioned at 100 pm and stained with cresyl violet. Reconstructions of the extent of lesion damage followed the patterns of Konig and Klippel (11). Results Histology. Summaries of the histological examination of the tissue samples are depicted in Fig. 1, which shows the extent of the smallest and largest of the respective lesions for each age level. It should be noted that the young subjects were at least 50 days of age at the time of perfusicn, and age related differences in lesion extent were not pronounced. DM damage in both pups and adults was mostly confined to the neocortex, corpus callosum and caudate. One adult subject also sustained damage to the cingulum. VL lesions produced damage to the orbital frontal cortex, external capsule, as well as to the caudate in one adult

6 and one pup. The H damage in both age groups was seen in the dorsomedial prefrontal cortex, corpus callosum, caudate and dorsal hippocampus. One of the younger subjects also received damage to the fimbria. Fig. 1. Composite reconstructions of smallest and largest lesion extent following dorsamedial prefrontal damage (top), ventrolateral prefrontal damage (middle), and hippocampal damage (bottom) for the young (Y) and adult (A) subjects of Experiment I. Behav?or. Trials to criterion. The mean numbers of trials to criteria and associated standard errors for the acquisition, reversal, and extinction phases during the 7- and the 30-day training stages of the experiment are displayed in Table I. To examine each training phase initially, three 2 X 4 ANOVAS of the 7 day training means revealed that Age was a significant factor in acquisition (F [1/40] = 8.82, P < 0.05), reversal learning (F [1/40] =6.09, P < 0.01), and extinction (F [1/40] = 7.80, P < 0.05). Significant Duncan comparisons (all Ps < 0.05) showed that adult acquisition required more trials than the pups after DM and H damage. For reversal learning, post-hoc comparisons showed that the adults needed more trials to switch the response choice in the VL group. Additionally, the pups had more rapid extinction than the adults of the H damaged group, while the remaining groups had comparable extinction rates.

7 Mean number of trials (i) with associated standard errors (SE) for acquisition (Acq), reversal (Rev), and extinction (Ext) during the 7- and 30-day stages for control (C), dorsomedial (DM), ventrolateral (VL), and hippocampal (H) subjects at adult (A) and young (Y) age levels Surgical Treatment C DM VL H 7 Rev Ext A f9.26 Y 16.50f k f 2.80 A 19.17k f k f4.77 Y k f * Rev Ext A 7.674~ Y 9.83 f ~ f5.71 A k k k6.34 Y 24.83k k f5.41 A separate analysis of performance during the 30 day stage indicated that Surgery was a significant source of variation in acquisition trials (F [3/39] = 3.54, P < 0.05). Duncan comparisons for the acquisition phase showed that the H damaged subjects required significantly more trials than the DM groups as well as the VL subjects (P' s < 0.05). Surgery was also a significant factor for reversal training during this stage (F [3/39] = 4.99, P < 0.05). Post-hoe comparisons revealed that there was a difference in the number of trials between the control and VL groups, control and H subjects, as well as between VL and H subjects (P' s < 0.05). During both acquisition and reversal training at the 30 day stage, Age only approached acceptable significance (P < 0.10). These findings indicated that Surgery exerted a differential effect during the 30 day stage, with the H damaged subjects showing the great-

8 est deficit. No significant differences were found in the number of criterion trials among the groups during the 30 day extinction test. Combined analyses of the earlier and later stages of training for the acquisition phase found that fewer overall trials were required during 30 day reacquisition (F [1/40] = 13.36, P < 0.001). A similar analysis of reversal learning trials showed that the interaction of Time of Testing x Surgery was significant (F [3/40] = 3.32, P < 0.05). As evident in Table I, there was an overall decrease in the number of trials required across time, and during the 30 day stage, the differences between age groups appeared to be overridden by the effect of Surgery. That is, the surgical treatments exerted a similar long lasting effect in both the young and adult subjects, which becaqe salient as time passed. The combined analysis for extinction showed that the 30 day retraining stage required more extinction trials than initial training (F [3/40] = 15.5, P < 0.001). Errors. The mean numbers of errors for both acquisition and reversal during the 7- and 30-day stages may be seen in Table 11. A 2 X 4 ANOVA of initial acquisition revealed that Age was significant (F[1/40] = 8.62, P < 0.01), and Duncan tests confirmed that there were Mean numbers of errors (3 with associated standard errors (SE) for acquisition (Acq) and reversal (Rev) learning during the 7- and 30-day stages for control (C), dorsomedial (DM), ventrolateral (VL), and hippocampal (H) subjects in the adult (A) and young (Y) age groups Surgical Treatment 7 Acq Rev A kl.W 13.50k3.41 Y 4.17k k kO rt0.77 A k f k3.39 Y f k kO Acq Rev A 4.17rt k k4.44 Y 2.33,t k k k1.31 A 3.17k f k2.76 Y 3.33k k k k1.57

9 more errors in the adults of the hippocampal treatment compared to similarly treated pups (P < 0.05). The effect of Surgery during initial acquisition (F [3/40] = 4.17, P < 0.01) was clarified by Duncan comparisons indicating that the H damaged subjects committed more errors than the groups with VL damage (P < 0.05). There were no significant differences in errors among the groups during the 7-day reversal learning phase of the experiment. Although no significant effects were found in the number of errors made during 30-day reacquisition, the effect of Surgery did emerge as a source of variance in errors emitted during reversal learning (F [3/ 1401 = 7.77, P < 0.001). Duncan test found a difference between the errors of the control and H groups P < 0.05). As in the number of trials to criterion measure, the errors of the hippocampal damaged subjects were markedly high. Latencies. Since subjects acquired criterion performance levels for various stages of the experiment at different rates, trial latencies were analyzed by examining performance during the 5 trials just prior to the criterion trials of a given stage. A 5 X 2 X4 analysis indicated that only the effect of surgical treatment accounted for systematic variation in latencies on the last 5 trials before acquisition criterion trials (F [3/ /40] = 5.93, P < 0.002), and Duncan Tests indicated that latencies among the 4 surgical treatment groups were different from each other (Ps< O.Ol), except for the difference between the controls and the DM group which was not significant. A similar analysis of trial latencies during reacquisition 30 days later showed only a significant effect of surgical treatment (F [3/40] = 2.98, P < 0.05). The VL group was significantly faster and the H group significantly slower than the DM and groups, which did not differ from each other (Ps < 0.05). Since latencies did not change systematically over the last 5 acquisition trials before the criterion trials at either stage, the mean latency for these trials was calculated for each subject and the stages of acquisition were compared in a 2 X 2 X 4 ANOVA of the effects of time of training, age and surgical treatments. The results indicated that mean trial latencies varied with the surgical treatment (F [3/40] = 5.97, P < 0.002) and the time of training (F [1/40] = 9.34, P < 0.01). Figure 2 shows the joint relationship between these variables. It is clear that the greatest improvement from 7- to 30-day acquisition occurred in the VL group, primarily, and in the H group as well, while the DM and control groups did not show much change. Indeed, the latter group showed somewhat slower responding during 30 day reacquisition, although this comparison was not significant.

10 ffl n z 0 U W ffl 0 7- DAY ACQ 2'H1] 30-DAY REACQ - - SURGICAL TREATMENTS Fig. 2. Mean latencies of directional avoidance responding during the last 5 trials of acquisition training prior to criterion trials during the 7- and 30-day phases of Experiment I. The analysis of latencies on the trials immediately preceding the initial reversal criterion indicated only that the adults were slower overall than their younger counterparts (F /1/40] = 6.33, P < 0.02). In the 30 day reversal results, only the progressive effect of the 5 trials turned out to be marginally significant (P < 0.10). In the comparison between stages of reversal training, the analysis of the mean latencies of the last 5 precriterion trials indicated significant effects from Age (F [I/ 1401 = 6.95, P < 0.01) and time of reversal training (F [1/40] = 4.49, P < 0.05). Figure 3 depicts both of these main effects and suggests a si- V) n z U W V) Z DAY REVERSAL C3P 30-DAY REVERSAL ADULTS PUPS Fig. 3. Mean latencies of responses by each age group during the last 5 trials of reversal training prior to criterion trials during the 7- and 30-day phases of Experiment I. AGE

11 milar improvement in mean reversal latencies between 7- and 30-days, with the adults tending toward more improvement than the younger subjects. The extinction process was gradual in all subjects, reflected in the lengthening of response latencies across progressive trials. However, specific age or surgical effects were not dramatic. Analysis of the 7 day extinction latencies on the 5 trials just prior to the extinction criterion trials revealed that only the interaction of Age x Surgery approached significance (P < 0.10). A similar finding for the Age x Surgery interaction effect was obtained in the 30 day extinction latencies (P < 0.10), while the expected effect of progressive extinction trials attained acceptable significance (F [4/160] = 2.60, P < 0.05). In the comparison between 7- and 30-day extinction testing, analysis of the mean of the last 5 trials before the criterion trials indicated a significant effect of time of training (F [1/40] = 4.22, P < 0.05). In addition, the interactions of Age X Surgery (F [3/40] = 3.90, P < 0.02) and Time of Testing X Surgery (F 13/40] = 2.84, P < 0.05) attained significance. Figure 4 depicts these interactive relations. While the 7 day extinction test produced slightly longer overall latencies, it was the DM pups that showed the O /U)ULTS 7-DAY EXTINCTION DAY EXTINCTION SURGICAL TREATMENTS Fig. 4. Mean latencies of the last 5 exti)nction trials before criterion for each age group and surgical treatment during both training phases of Experiment I. longest latencies, while the hippocampal and VL groups of both ages had overall faster responding immediately after surgery. These differences seemed minimized by the 30 day test.

12 Discussion The measures of trials to criterion and trial latencies just prior to criterion during initial training tended to show that adults of the three surgical treatments were inferior to their younger counterparts, which was not found between the age levels of the control condition. This finding, in relation to hippocampal damage, conflicts with previous (6) claims that adult hippocampal damaged animals should acquire an active avoidance task faster than the other groups. However, upon closer examination of the task requirements, it may be that the ambiguity of the CS, consisting of the simultaneous raising of both compartment doors, may have been too difficult for the hippocampal damaged adults. Since the same cue was used &ring all three stages of the experiment, the correct response could not have been predicted. The degree of perseverative responding appears to vary directly with the difficulty of the discrimination (9). Thus, the task may have rendered the critical cues too ambiguous, producing a difficulty in the emergence of avoidance behavior and maintenance of escape performance. Moreover, the locomotor aspects of the task seemed to be closer to those of 1-way active avoidance, rather than 2-way avoidance requirements. Accordingly, the more profound hippocampal effects expected on the basis of results of 2-way avoidance may have not emerged due to the lessened locomotor element of the task requirement (8, 22). The comparison of the numbers of trials required to meet criterion between 7- and 30-day acquisition showed that there was a decrease in meeting the latter criterion. This finding would appear to support an overall facilitation from the initial acquisition stage and finds support in the latency measure as well. Indeed, this positive transfer effect was strongest in the VL adults that showed the fastest 30 day acquisition with fewest errors. This finding is.especially interesting because the 30 day acquisition task was actually a switch in orientation from the last task performed during earlier extinction which, in effect, produced another series of reversal tasks. The transfer effect from initial acquisition to reacquisition 30 days later facilitated the adult VL subjects, but was minimal in the subjects of both ages with partial hippocampal damage. Response persevertion should heve been most clearly measured during reversal training and extinction. Overall, the adults showed longer latencies during both reversal phases, but especially during 30 day reversal. The analysis of the number of trials needed for reversal training found that the adults with VL damage required more trials than their younger counterparts. That the adults showed a difficulty inhibiting the previously acquired response was in agreement with Brutkowski (2)

13 who argued that VL lesioned adult rats show a regression to perseverative behavioral patterns similar to younger subjects, which become manifested as a difficulty in altering response direction. In the number of errors during the initial reversal performance of the young and adult hippocampal damaged groups, the latter showed a greater degree of perseverative responding. However, this age difference did not hold for the 30 day reversal because both age levels were comparable in the persistence of errors, reflecting equivalent hippocampal damage. This finding would seem to indicate that the you~~g hippocampal rats did not show as much plasticity in subsequent brain development when lesioned at 18 days of age. The results tended to show a dissociation between the effects of DM and VL lesions. From the resslts there appeared to be two distinct response patterns to the task for the subjects with cortical lesions. The post-hoc comparisons tended to show that the VL subjects made less errors and required less trials, except for the initial reversal task. In contrast, the DM subjects did not show differences from normals, except for the persistence in response latencies during extinction in the DM pups. With the exception of initial reversal, the trials to criterion measures seem to support the conclusion that the VL groups showed little incidence of perseverative responding in this task. EXPERIMENT I1 In the first experiment, immediate training and testing of directional behavior were explained primarily from age specific differences known to occur in avoidance learning; surgical effects seemed to exaggerate age related effects in accounting for acquisition and retention. Conversely, 30-day retraining was affected by residual DM or hippocampal lesion effects, while age influences at the time of initial training and surgery were minimal. The purpose of Experimental I1 was to extend these findings by examining combined lesion effects implicating both dorsomedial prefrontal cortex and caudate nucleus and testing subjects over a longer recovery period. It has been suggested (23, 24) that overtraining may result in differential habituation of arousal in DM rats, possibly interfering with reacquisition and recall of directional learning. Another factor in explaining recovery from prefrontal cortical damage relates to functional shifts from damaged areas to others that might subserve the learning deficit (16). Accordingly, the comparison of recovery between the more severe lesion and the prefrontal damage only may be informative if removal of the caudate in the former results in persistent deficits over the long recovery period of 60 days.

14 Method Subjects. A total of 48 male hooded rats served as subjects in this experiment. At the time of surgery, 24 rats were day old weanlings and 24 were day old adult rats. All pups were housed initially with their mothers and litter mates, prior to weaning at 25 days of age. After surgery, all pup in a given litter were washed with alcohol prior to replacing the operated subjects, so that detection of the treated pups by the mother was obscured. After weaning, all pups, along with the adult subjects, were housed singly. Food and water were available ad libitum in the home cage throughout the experiment, except for a 24 h food deprivation schedule imposed on the adult subjects just prior to surgery. Appuratus. The same apparatuses for behavioral training and testing and for the surgical procedures as described in Experiment I were used in this experiment. Procedure. Surgery. Subjects at each age level were evenly assigned to one of three surgical treatments: Bilateral ablations of the dorsomedial prefrontal cortex (DM), bilateral destruction of the dorsomedial prefrontal cortex and the caudate nucleus (DM-I-C), or a sham operate control procedure (Sh). Twenty-four hours prior to surgical procedures, the adult subjects were deprived of food, and just prior to surgery, each subject was anaesthetized with an s.c. injection of sodium pentobarbital (Nembutal, 50 mg/kg). All lesions were made by maintaining a constant tip temperature of the electrode at 70 C for 60 s. The surface coordinates for the lesion procedures were as follows: DM lesions were made 2 mrn bilateral to the midline and 2.30 and 4.00 mm anterior to bregma, the surface skull reference point, in the pups and adults respectively. The electrode was lowered 1.00 and 1.50 mm from the dura respectively in the pups and adults. The DM+C coordinates were the same bilaterally and anterior to bregma as the DM treatment, but the electrode was lowered 1.50 and 2.00 mm for the pups and adults respectively. The sham operated subjects were treated the same as the DM-I-C groups, but the electrode was not activated. The adult subjects were allowed to recover in their home cages for either 7- or 60-days, while the pups were returned to their mother and litter mates upon awakening and allowed 5 days recovery prior to weaning. Training and testing. The same training procedure as described in Experiment I was followed in this experiment. After meeting the acquisition criterion, subjects were tested in extinction beginning on the next trial and continued until 5 consecutive trial latencies of 10 s each were recorded. There were no retraining or reversal learning stages, as in

15 Experiment I. Accordingly, the 2 X 2 X 3 design called for combinations of age, time of training/testing, and surgical treatment, respectively. A total of 6 pups and 2 adults died during surgery, and 4 pups, 3 after DM+C lesions and 1 after DM treatment, failed to meet the acquisition criterion of 5 consecutive nonshock trials. These subjects were replaced with naive animals. Histology. After extinction testing, subjects received overdoses of Nembutal and were intracardially perfused with 10 /a formalin, the brains removed and stored in 10 /o formalin, and then blocked, frozen, sectioned at 100 pm, and stained with cresyl violet. As with the results of Experiment I, composite reconstructions of damaged areas were made following the patterns of Konig and Klippel (11). Results and discussion Histology. Figure 5 shows summary reconstructions of the lesion extent obtained from subjects at both age levels in Experiment 11. It should be noted that at the time of sacrifice, the young subjects were Fig. 5. Composite reconstructions of the extent of the largest and smallest dorsomedial prefrontal (DM) Lesions and combined dorsmedial and caudate (DM+ C) lesions for the young (Y) and adult (A) rats of Experiment 11. at least 26 days of age, and examination of the extent of damage did not reveal any systematic observations specific to the age at the time of surgery. DM lesions at both age levels were largely confined to the neo-

16 cortex and corpus callosum. In DM+C subjects, damage was similar to the DM groups and included the dorsomedial portion of the caudate nucleus as well. Behavoior. Acquistion. The mean unmber of acqisition trals, along with respective standard errors, are indicated for each group in Table 111. Overall analysis of acquisition trials revealed a significant main Mean numbers of trials 6) and standard errors (SE) during acquisition in adult and pup groups with dorsomedial prefrontal lesions (DM), prefrontal and caudate combined or sham lesions (Sh) trained either 7- or 60-days after surgery 7-Days Adults PUPS 60-Days Adults Pups 7-Days Adults PUPS 60-Days Adults PUPS Trials 22.00& &3.34 Errors effect of surgical treatment (F [2/36] = 10.86, P < 0.001), with the sham subjects of both ages acquiring the task most rapidly and the DMS- C subjects least.quickly (all Ps < 0.01, Duncan Tests). A significant effect from the Time after surgery (F [1/36] = 4.20, P < 0.05) confirmed that subjscts trainde after 7 days required more trials (Z = 20.5) than those trained 60 days (5 = 17.5) after sugery. A significant interaction was found between Age and Time of Training (F [1/36] = 4.02, P < 0.053), with the adults trained after 7 days requiring significantly more trials and the aduljs trained after 60 days requiring the least. The young groups trained at 7- and 60-days differed from the overall adult number of trials, but did not differ from each other.

17 Figure 6 shows the mean number of errors, and associated standard errors, committed during the immediate and 60 day acquisition training stages. Analysis of mean errors in acquisition showed general agreement DM DM+C S h SURGICAL TREATMENTS Fig. 6. Mean numbers of errors cammitted during acquisition training in adults and pups from the surgical treatments trained either 7- or 60-days after surgery in Experiment 11. with the trials to criterion measure. Surgical treatment accounted for a significant amount of variability (F [1/36] = 10.21, P < 0.001), with the DM-kC groups committing more errors than subjects in the other surgical conditions, which were also significantly different (Ps < 0.01). Age approached acceptable significance (P < 0.10), suggesting that the young subjects tended to commit more errors overall. As Fig. 6 suggests, a significant interaction emerged between Age and Surgical Treatment (F [2/36] = 4.11, P < 0.05), and the DM+C groups appeared to account for this effect by the pups committing greater numbers of errors than their adult counterparts at both acquisition levels. This trend is also seen in Fig. 6 from the DM 60 day age groups. A significant interactior~ between Age and Time of Training was also found (F [I1361 = 4.50, P < 0.05), and again the same DM-kC adults and pups were at variance with the trends of their respective age levels and training conditions. As might be expected a significant 3-way interaction emerged for the effect of Age X Surgery X Time of Training (F[2/36] = 3.64, P < 0.05). The final measure of acquisition was trial latency. As in Expeximent I, the latencies of the last 5 trials prior to the acquisition criterion trials were analyzed for each subject in order to compare trial latencies among subjects that differed in the total number of trials to attain acquisition criterion. 2 - Acta Neurobiol. Exp. 5/88

18 An overall analysis of Age X Time of Training X Surgery X Trials indicated the presence of a main effect from Age (F[1/36] = 15.36, P <0.001), which was explained by the overall longer latencies in the pups. The main effect of the Time of Training was highly significant (F[1/36) = 33.03, P < ), reflecting overall longer latencies obtained 7 days after surgery than after 60 days recovery. There was an interaction of Age X Time of Training (F[1/36] = 9.98, P < 0.01), which is depicted in Fig. 7 and was largely the product of the pups tested 7 days 3.50, 0 7-DAY ACQUISITION EiY 60-DAY ACQUISITION 3.00 I ADULTS PUPS AGE Fig. 7. Overall mean latencies of each age group during acquisition training in Experiment 11. after surgery having overall longer latencies than either the adults tested then or both age groups tested 60 days after surgery. There was also a significant Age X Trials interaction (F[4/144] = 2.42, P < 0.05), DAY 60-DAY ADULTS A-A 7-DAY PUPS A-A 60-DAY PUPS LAST ~ V ACQUISITION E TRIALS Fig. 8. Mean latencies over the last 5 acquisition trials prior to criteribn for both age groups at the 7- and 60-day training stages of Experiment 11.

19 which was largely explained by both adult and 60-day pup groups having reached asymptotic acquisition performance, while the 7-day pups still showed improvement in latencies. This effect may be seen in Fig. 8 showing mean trial latencies for each age group at both acquisition phases, and these relationships reflect the significant interaction of Trials X Age X Time of Testing (F[8/144] = 3.86, P < 0.01). The interaction of Trials X Age X Surgery approached significance (P = 0.07) and indicated that any differential surgical effect appearing at the beginning of the trial block was largely obscured by the beginning of the criterion trials. These trial latency interactions support the error results in showing severe disruption in the pups trained after 7 days and after 60 days within the DM4-C surgical treatment. Extinction. The mean numbers of extinction trials and associated standard errors are presented in Table IV for ali groups. Only the main Mean numbers of trials (i) and related standard errors (SE) in extinction for groups receiving dorsomedial prefrontal lesions (DM), combined damage to the dorsomedial prefrontal cortex and the caudate nucleus (DM+C), or sham lesions (Sh) 7- prior to training DM DM+C SH 7-Days Adults 32.25f & f Pups 10.00& s & Days Adults 32.25~t s ~ effect of Age (F[1/36] = 6.5, P < 0.01) attained significance with the adults showing overall greater resistance to extinction (5 = 35.29) than the young subjects (2 = 20.79). For extinction, latencies on the last 5 trials prior to the criterion trials were collected for each subject and analyzed in a 2 X 2 X 3 X 5 ANOVA. Only the overall main effects of Age (F[1/36] = 11.20, P< 0.002) and progressive Trials (F [4/144] = 9.31, P < 0.01) were significant. Essentially, the age effect was the result of slower responding in extinction on the part of the pups. The results of Experiment I1 tended to support the findings of Experiment I in suggesting the overall more dramatic short term effect of surgical treatment in the pups than in the adults. As in Experiment I,

20 the DM lesions produced deficits in measures of trials to criterion and latencies, and this effect was enhanced after the more severe combined lesions. With the recovery period extended to 60 days, surgical effects were minimal, and subjects at both ages seemed to acquire the response at a rate comparable to the sham controls. Once the response was acquired extinction responding was comparable across ages, surgical treatments, and time of testing. GENERAL DISCUSSION The results of this study indicate that both age and surgical treatment differentially affected perseverative responding and recovery of directional avoidance performance, depending upon the time of testing. In general, the quality of the lesion (Experiment I) produced differential effects that were more pronounced at the younger ages. When extent and severity of damage were compared (Experiment 11), the younger subjects also showed greater disruption and loss over time under the combined condition than after prefrontal damage alone. Expectations concerning recovery of functions were largely met. By the retesting phase of both experiments, subjects recovered response levels. While not always to the level of control performance, surgical subjects nevertheless met criteria with only a few exceptions. Indeed, after the 60 day recovery period, there were few differences in performance, especially during extinction. These data are supportive of other findings, although the exact mechanism of recovery was not clear from the present findings. Interpretations suggesting a shift of function to other brain areas, particularly the caudate (24), are not inconsistknt with the present results. The performance of the pups in these experiments lends further support to the view offered by Kolb and Wishaw (10) that recovery of function in young subjects may not be as pronounced or predicted as once asserted. If greater plasticity in the development of the brain in the young subjects is manifested in aversively motivated tasks, then age should have accounted for group differences during the retraining phase, with the pups showing more successful responding that the adults in the surgical group, while the intact groups should have attained comparable results. Within the context of the present measures of directional avoidance learning, the expected recovery of function in the pups was not seen to the extent as the controls. Instead the latter stage of both experiments was characterized by the tendency to perseverate a previously acquired response, to the degree which was comparable to adults with similar damage.

21 Experiment I was part of an undergraduate honors thesis presented by the second author to the Department of Psychology, University of Massachusetts at Boston. Experiiment I1 was part of an undergraduate honors thesis presented by the third author to the Department of Psychology, University of Massachusetts at Boston. The authors wish to thank Ms. Mikala Brennan for assistance with the construction of the figures. REFERENCES 1. BRENNAN, J. F., POWELL, E. A., and VICEDOMINI, J. P Differential effects of dorsomedial prefrontal lesions on active and passive avoidance in young and adult rats. Acta Neurobiol. Exp. 37: BRUTKOWSKI, S Functions of the prefrontal cortex in animals. Physiol. Rev. 45: CAMPBELL, B. A Learning in infra-primate mammals. In H. W. Stevenson, E. H. Hess, and H. L. Rheingold (ed.), Early behavior: cmparative and developmental approaches. New York, John Wiley, p CAMPBELL, B. A. and TEGHTSOONIAN, R Electrical and behavioral effects of different types of shock stimuli on the rat. J. Comp. Physiol. Psychol. 51: COULTER, X., COLLIER, A,, and CAMPBELL, B. A. (1976). Long-term retention of Pavlovian fear conditioning by infant rats. J. Exp. Psychol.: Anim. Behav. Proc. 2: DOUGLAS, R. J The hippocampus and behavior. Psychol. Bull. 67: EGGER, G. J. and LIVESEY, P. J Age effects in the acquisition and retention of active and passive avoidance learning by rats. Develop. Psychobiol. 5: GRAY, J. A. and McNAUGHTON, N Comparison between behavioural effects of septa1 and hippocampal lesions: a review. Neurosci. Biobehav. Rev. 7: KIMBLE, D. P The effects of bilateral hippocampal lesions in rats. J. Comp. Physiol. Psychol. 56: KOLB, B. and WISHAW, I. Q Neonatal frontal lesions in the rat: Sparing of learned but not species-typical behavior in the presence of reduced brain weight and cortical thickness. J. Camp. Physiol. Psychol. 95: KONIG, J. F. R. and KLIPPEL, R. A The rat brain: a stereotaxic atlas of the forebrain and lower parts of the brain stem. William and Wilkias, Baltimore. 12. LEONARD, C The prefrontal cortex of the rat. I. Cortical projections qf the mediodorsal nucleus. 11. Efferent connections. Brain Res. 12: LEONARD, C The connections of the dorsomedial nuclei. Brain Behav. & Evol. 6: MARKIEWICZ, B., KUCHARSKI, Q., and SPEAR, N. E Ontogenic cmparison of memory for Pavlovian conditioned aversions to te~mperature, vibration, odor, or brightness. Dev. Psychobiol. 19: MILLER, R. R. and BERK, A. M Sources of infantile amnesia. In N. E. Spear and B. A. Campbell (ed.), Ontogeny of learning and memory. Erbaum, Hillsdale, NJ.

22 16. NONNEMAN, A. J., VOIGT, J., and KOLB, B Comparisons of behavioral -effects of hippocampal and prefrontal lesions in the rat. J. Comp. Physiol. Psy~hol. 87: RICCIO, D. C., ROHRBAUGH, M., and HODGES, L. A Developmental aspects of passive and active avoidance learning in rats. Dev. Psychobiol. 1: ROSVOLD, H. E The frontal lobe system: cortical-subcortical interrelationships. Acta Neurobiol. Exp. 32: SCHWEITZER, L. and GREEN, L Acquisition and extended retention of a conditioned taste aversion in preweanling rats. J. Camp. Physiol. Psychol. 96: SCHULENBERG, C. J., RICCIO, D. C., and STIKES, E. R Acquisition and retention of a passive avoidance response as a function of age in rats. J. Comp. Physiol. Psychol. 74: THELEN, E Rhythmical behavior in infancy: An ethological perspective. Dev. Psychol. 17: WERKA, T., GRADKOWSKA, M., and DANIELSSON, I Partial lesions of dorsal hippocampal afferents and GM1-ganglioside treatment effects on 2-way avoidance in rats. Acta Neurobiol. Exp. 48: WILCOTT, R. C Preoperative habituation and effects of prefrontal lesions on an autonomic response of the rat. Physiol. Psychol. 11: WILCOTT, R C Preoperative overtraining and effects of prefrontal lesions on delayed alternation in the rat. Physiol. Psychol. 14: Accepted 18 March 1988