Comparison of retention performance between young rats with fimbria-fornix lesions and aged rats in a 14-unit T-maze

Behavioural Brain Research, 35 (1989) 253-263 Elsevier

253

BBR 00970

Comparison of retention performance between young rats with fimbria-fornix lesions and aged rats in a 14-unit T-maze Hideki Kametani 1, Elaine L. Bresnahan ~,2, Mark E. Chachich 1,3, Edward L. Spangler 1 and Donald K. Ingram 1 IMolecular Physiology and Genetics Section, Laboratory of Celhdar and Molecular Biology, Gerontology Research Center, National hlstitute on Aging, Francis Scott Key Medical Center. Baltimore, MD 21224 (U.S,A.), 2Department of Psychology, Essex Community College, Baltimore, MD 21237 (U.S.A.) and 3Department of Psychology, Towson State University, Towson, MD 21204 (U.S.A.) (Received 26 October 1988) (Revised version received 27 March 1989) (Accepted 30 March 1989) Key words: Aging; Fimbria-fornix; Maze; Learning; Retention; Memory; Hippocampus; Rat

Young (3-months) and aged (24-months) male F-344 rats were pretrained in one-way active avoidance in a straight runway for 3 days. Then two 10-trial daily sessions were given in a 14-unit T-maze in which the response requirement was to negotiate each of 5 maze segments within 10 s to avoid footshoek. One day or one week after acquisition, bilateral electrolytic lesions were made in the fimbria-fornix of young rats (1-day lesion or 1-week lesion). Corresponding sham operations were made for remaining young rats (l-day sham or l-week sham). Aged animals did not receive any surgical treatment. One week after surgery, a 10-trial retention test was conducted to assess the lesion effects on retention and to manipulate the interval between acquisition and lesions. Aged animals were tested in the maze 1 week after acquisition. Results revealed that rats with fimbria-fornix lesions exhibited significant impairment compared to sham-operated groups on all retention performance measures including errors, runtime, number of shocks, duration of shock, and alternation errors. The number of errors and alternation errors of lesioned animals were still higher than those of sham-operated animals at the second half of the retention test, whereas other non-cognitive measures for lesioned animals recovered to control levels. The interval between acquisition training and lesions had no influence on retention performance. Although performance of aged rats during acquisition and retention trials was significantly worse than that of young controls and lesioned animals, a similar recovery pattern during retention testing was found for young rats with fimbria-fornix lesions and aged rats, i.e. both groups showed significant declines in non-cognitive measures with less decline in cognitive measures. These results suggest that the fimbria-fornix is partially involved in retention of 14-unit T-maze performance and that the age-related retention deficit observed in this task may be related to impaired transmission through this pathway.

INTRODUCTION T h e i n v o l v e m e n t o f the h i p p o c a m p u s and its r e l a t e d s t r u c t u r e s in l e a r n i n g a n d m e m o r y h a s b e e n w i d e l y d o c u m e n t e d in y o u n g r o d e n t models. H i p p o c a m p a l d a m a g e in y o u n g rats p r o d u c e s p r o f o u n d p e r f o r m a n c e i m p a i r m e n t in a variety o f

spatial t a s k s including the M o r r i s w a t e r m a z e 27 mad the r a d i a l a r m m a z e t8,2~ I n t e r r u p t i o n o f h i p p o c a m p a l afferents, e.g. the fimbria-fornix ( F F ) p a t h w a y , in rats also elicits m a r k e d deficits in t h e s e b e h a v i o r a l t a s k s ]s,31,32,39 N u m e r o u s theories o f k i p p o c a m p a l f u n c t i o n in learning a n d m e m o r y have been p r o p o s e d for r o d e n t models.

Correspondence to: D.K. Ingram, Molecular Physiology and Genetics Section, Gerontology Research Center, NIA, NIH, Francis Scott Key Medical Center, 4940 Eastern Avenue, Baltimore, MD 21224, U.S.A.

254 For example, the working memory/reference memory theory3~ the data-based memory theory21, the cognitive mapping theory29, the representational/dispositional memory theory38, and the interference theory4~ all implicate the involvement of hippocampal circuitry in specific aspects of memory processing. In a recent review, Barnes 3 distilled many of the apparently diverse views ofhippocampal involvement for memory processing in rodents into a more parsimonious perspective focusing on concepts of cue versus place learning. If the information to be acquired or retained is highly stimulusbound such that removal of a particular cue is sufficient to impair performance in the task, then hippocampal damage does not appear to interfere greatly with performance either during acquisition training or retention testing in such cue learning tasks. However, if the task requires informatioe processing about a spatial environment that involves efficient encoding of the relationships among diverse sensory features of the environment, then hippocampal damage prior to training clearly impairs acquisition in these place learning tasks. If hippocampal dmnage occurs after training in such tasks, then the results from numerous studies are variable with respect to whether impairment performance is observed. As a reason for these diverse results, Barnes 3 suggested that the hippocampus may have only transient involvement in storage/retrieval processes. Hippocampal lesions after specified intervals between training and retention testing may prove ineffective because the information resides in extrahippocmnpal storage sites. Results supporting this hypothesis have been supplied. For example, Flexner et al. 9 could induce retrograde amnesia in mice for retention performance in a shock-motivated Y-maze task when a protein synthesis inhibitor, puromycin, was injected into the hippocarnpus 3 days after training but not when injected 1 week after training. Aging is associated with impairment in memory processing 4. Compared to their young counterparts, aged rats exhibit performance deficits in spatial memory tasks such as the radial arm maze 17, the Morris water maze 1~ and the circular platform task ~. These age-related deficits are pos-

sibly related to dysfunction in hippocampal neural circuitry2'4'7'23. For example, reduction of perforated axospinous synapses 7 and a fast decay speed of long-term potentiation (LTP) recorded in the dentate gyrus 1 have been correlated with impaired performance in spatial memory tasks among aged rats. Rats and mice exhibit marked and reliable agerelated performance deficits in a 14-unit Tmaze 5,1z,~3,15'16'25. This maze has 14 consecutive position discriminations with a left or right turn at each choice-point. The age-related impairment in this task occurs with procedural variations such as the type of motivation used (food, water or footshock) and distribution of practice, and it has been observed in a variety of rodent genotypes and strains (for reviews, see IngramZh'~6). Although the age-related decline in 14-unit Tmaze performance has been well documented ~5,~6, less parmnetric research has been conducted on the retention aspect. In a foodmotivated version of the 14-unit T-maze, Goodrick ~2 observed that retention performance of aged Wistar rats did not appear different from that of their young counterparts if the acquisition level was equated. Knowlton et al. 22 used the terminology of Olton et al. 3~ to describe the 14-unit T-maze as a reference memory task because the response pattern to be learned is fixed across trials. According to the working/reference memory theory of hippocampal function, the hippocampus is not primarily involved in a trial-independent or reference type of memory processing 3~ In contrast to other spatial tasks in which extramaze visual cues are critical for efficient performance, negotiation of the 14-unit T-maze does not appear to require visual cues. Animals can learn this task under dark conditions, which indicates that vision is not necessary for accurate performance ~5. It has been suggested that accurate maze performance relies upon an internally-driven response algorithm rather than attention to external visual stimul? 5. Despite these distinctive features of the 14-unit T-maze from other spatial tasks, Bresnahan et al. 6 recently reported that young rats with bilateral F F lesions exhibited marked learning impairment in this maze. The impairment asso-

255

Apparatus

ciated with FF lesions appeared to parallel that previously observed in scopolamine-treated young rats 36 as well as that observed in aged rats 13'1s, and the results suggested involvement of the septohippocampal system and cholinergic dysfunction in the age-related learning deficit observed in this maze. The current experiment was conducted to explore further the involvement of the septohippocampal system as well as an age-related difference for retention performance in the 14-unit T-maze. The effects of bilateral FF lesions on retention performance in the maze were examined, and the effect of the time interval between acquisition and lesion was manipulated. To this end, FF lesions were made either one day or one week after acquisition training, and a retention test was conducted one week after surgery. According to the prediction based upon the Flexner et al. 9 results, retention might be impaired when the FF lesion occurred one day after maze acquisition but not one week after training. Additionally, the retention performance of aged rats was compared to that of young rats with either FF or sham lesions, which would establish (a) whether or not an age-related retention deficit exists in this avoidance paradigm, and (b)whether or not retention performance of FF-damaged young rats would be similar to that of aged rats.

The 14-unit T-maze and the straight runway used for one-way shock avoidance pretraining have been described in detail elsewhere6,36. In brief, the straight runway was made of transparent Plexiglas, 100 cm in length and 12 cm in width. The floor consisted of stainless steel bars placed diagonally to the path of the maze and connected to a Coulbourn Instruments (Model E13-08) shock unit to deliver footshock (0.8 mA). Start and goal boxes were identical black Plexiglas boxes and were interchangeable for start and goal positions. White noise (85 dB) was provided through a speaker located near the runway, The 14-unit T-maze was constructed of transparent Plexiglas, 2 m in width and 2 m in length, surrounded by plywood walls that were painted gray. Only the ceiling and overhead fluorescent lighting were visible from the floor of the maze. To prevent backtracking, the maze was divided into 5 segments by guillotine doors, and these doors were closed after the animal moved to the next segment. Movement of animals in the maze was recorded by photosensors wired to a microprocessor. These data were stored on computer PROMS for further off-line data analysis. To mask extraneous sounds, white noise (85 dB) was delivered from 4 speakers located at opposite corners under the maze.

MATERIALS AND METHODS

Procedures Straight runwaypretraining.A 20-rain period of

Subjects Forty-one male Fisher-344 rats, 3-months (n = 30) and 24-months (n = 11), were obtained from Harlan Sprague-Dawley, Inc., Indianapolis, IN. The animals were housed doubly in plastic cages in a vivarium maintained on a 12-h light/dark photocycle (lights on at 06.00 h). The room temperature was controlled at approximately 22 ~ Food (NIH-07 formula) and water were provided ad libitum. The rats were permitted about 2 weeks to acclimate to the vivarium prior to treatment. All animals were tested during the light cycle of the photo-period.

adaptation to the maze room preceded one-way active avoidance pretraining. After the adaptation period, a rat was placed in the start-box and was pushed into the runway. The animal had 10 s to avoid footshock by running into the goalbox at the opposite end of the runway. If an animal did not avoid shock within 10 s, 0.8mA scrambled footshock was delivered via the grid floor until the rat entered the goalbox or until 60 s elapsed which resulted in removal from the runway and placement in the goalbox. The intertrial interval was approximately 2 rain with 10 trials given on 3 consecutive days. The avoidance criterion was 8 successful avoidances in 10 consecutive trials on the third pretraining day. For those animals that met criterion, maze training began 24 h later.

256

14-unit T-maze training. The procedure for maze training has been described elsewhere 6,36. In brief, a rat was allowed 20 rain to adapt to the maze room. A 10-s shock avoidance contingency existed in each of 5 maze segments. If the rat could not avoid shock within 10 s in any one segment, 0.8mA scrambled footshock was delivered until the animal entered the next segment. The 10-s avoidance contingency was reset each time a door was closed behind the animal as it entered the next segment in the sequence. If the rat received shock for 600 s before reaching the goalbox, it was removed from the maze and placed into the goalbox. The intertrial interval was 2 min, and 10 trials were given daily for 2 consecutive days. Maze performance was assessed by several behavioral measures: number of errors, percent alternation errors, runtime, number of shocks, and shock duration, and each was averaged over blocks of 5 trials. As previously defined 36, the alternation error measure served to assess the tendency for each rat to use an alternation strategy for solving the maze. This measure is based on the ratio of errors committed in forward-going position discriminations of at least 3 turns to the frequency of possibilities to make alternation errors. Surgery. Young rats were assigned randomly to a 1-day FF lesion group (FF1, n = 11), a 1-week FF lesion group (FF7, n = 9), or to corresponding sham-operated control groups (CON1 or CON7, n = 5 for each group). Stereotaxic surgery was performed on the 1-day group about 24 h after the last acquisition trial, and on the 1-week group 7 days after the last acquisition trial. Each rat was anesthetized using ketamine with acepromazine (10 ml/1 ml) at a dose of 100 mg/kg i.m. The anterior coordinates 33 for FF lesions were 2.2 mm posterior to bregma, 2.6 mm lateral to midline, and 4.2 mm below the surface of the skull. The posterior coordinates were 3.0 mm posterior to bregma, 4.0 mm lateral to midline, and 4.6 mm below the surface of the skull. These coordinates were selected to damage FF connections to the hippocampus without completely transecting all connections. This approach would appear to be more relevant as a model of agerelated impairment compared to a complete F F

transection. For lesion animals, 4 holes were drilled in the skull at the specified coordinates, and the dura was cut beneath each hole. A 1.5-mA and 15-s duration direct current from a Grass Model D C - L M 5 lesion maker was passed through a lowered Teflon-insulated nichrome wire which was scraped approximately 0.5 mm at the tip. For sham animals, holes were drilled in the skull but no wire was lowered into FF. Aged animals (AGED) did not receive any surgical treatment. Retention test. One week after surgery, the retention test was given to all animals. Therefore, the intervals between acquisition and retention were 8 days for FF1 (n = i 1) and CON1 groups (n = 5) and 14 days for F F 7 (n = 9) and corresponding CON7 rats (n = 5). The retention interval for A G E D rats (n = 8) was 7 days. All animals again received straight runway training before the retention test. Criterion for successful completion &this training was 8 successful avoidances within a maximum 20 consecutive trials. Approximately 4-5 hr after straight runway training, a 10-trial session in the 14-unit T-maze began. The procedure for this retention session was identical to that for acquisition training. Histology. One week after all behavioral testing was completed, lesioned animals were given a lethal injection of sodium pentobarbital and perfused intracardially with 10~/o formalin. Brains were removed from the skull and stored in 10Z formalin for 10 days. Coronal sections of 30 #m cut through the lesion sites were stained with thionin Nissl stain and were examined for histological verification of lesion locations. Drawings of brain sections were made from a Bausch and Lomb microprojector and showed the site and extent of lesion damage.

RESULTS

Histological analyses. Damage of the FF was found in 10 of 11 animals in the FF1 group, and in 8 of 9 animals in the FF7 group. Performance data from the two animals with no F F damage were omitted f?om all subsequent analyses. In both FF1 and FF7 groups, the extent of lesion damage was variable. Some damage was found in

257 SMALL LESION

LARGE LESION

Fig. 1. Coronalsections from two representativeanimals,one with largeand anotherwith smalllesions(lesions:dottedareas). Sections were drawn from the Paxinos and Watson33 atlas. adjacent structures, including the dorsal and medial edge of the caudate, the dorsal and lateral edge of the thalamus, the lateral edge of the hippocampus, and the overlying cortex. Drawings of representative damage from relatively small and large lesions are shown for two different animals in Fig. 1. Straight runway pretraining. Performance in one-way active avoidance pretraining between combined young rats and aged rats was significantly different. Mean percent avoidance responses across the 3 daily sessions of 10 trials each was 88 ~ of all trials among the young animals, and 63?/0 for the AGED animals (t-test, P < 0.01). Thus, learning of the shock avoidance response was impaired in A G E D animals compared to their young counterparts. Because one of the aged animals did not meet the avoidance criterion (8 out of 10 correct responses in the third training day), it was omitted from training in the 14-unit T-maze. 14-unit T-maze acquisition. Three AGED ani-

reals did not negotiate the maze within 600 s on at least two consecutive trials; therefore, these animals were omitted from subsequent analyses. Fig, 2 disp!ays 14-unit T-maze performance for each group on each behavioral variable. All young groups quickly improved performance during the acquisition phase. Separate two-factor (group by block) repeated measures analyses of variance (ANOVAs) for young groups revealed significant main effects of blocks on all variables (all Because the main effect of groups was not significant on any measure, the 4 groups of young animals were combined into a single young group for comparisons to the AGED group. Two-factor (age by block) repeated measures ANOVAs revealed that AGED rats exhibited significantly poorer performance on all measures compared to young animals (all P's < 0.01). Additionally, significant interactions between blocks and age on all behavioral variables (P's < 0.01) indicated that AGED rats learned this maze more slowly than did their young counterparts (see Fig. 2).

258 1.s 0"c 0.~

~

0.6

~

Fr/t FF',

0.4 0,~ 0,~ 0.1 0

1

~

~

8 B

4 2 0

2

3

4

5

1

2

3

4

5

100 tZ

80 60 40 2D 0

6

ijjj 1

FF7

M

AGED

D.

I

2 3 4 5 6 ACQUISITION RETENTION BLOCKS OF FIVE TRIALS

2

3

4

5

6

201 18, ,-E. 16, 14,

160 140 120

I~

6

JJJJ

.B.

$

CON1

m CeN7

10,

~

sc 4C

~

2c o

jiJ, 1,

2 3 4 5 6 ACQUISITION RETENTION BLOCKS OF FIVE TRIALS

Fig. 2. Per trial means and standard errors for specificperformance measures ofyoung FF-lesioned (FF1 and FF7), young sham operated (CON1 and CON7), and aged control (AGED) F-344 male rats in the 14-unit T-maze. A: number of errors; B: alternation errors; C: runtime; D: number of shocks; E: shock duration. FF lesions were made 1 or 7 days after last acquisition session.

Straight runway retraining. Mean percent avoidance responses during straight runway training among young groups were 100% for CON1, 97~ for CON7, 9 5 ~ for FF1, and 9 9 ~ for FF7. A one-way ANOVA for young groups revealed no significant effect of groups (F3,24 = 0.67, P > 0.1). In contrast, mean percent avoidance responses in A G E D rats was 7 2 ~ , and was significantly lower than that of all young rats combined (t-test, P < 0.05). However, all A G E D rats met the criterion for successful completion of this training. 14-unit T-maze retention. As shown in Fig. 2, CON1 and CON7 groups exhibited equivalent performance on the retention test. Compared to controls, rats with F F lesions exhibited significantly impaired maze performance during the first block of retention testing. All performance measures except errors and alternation errors showed recovery to baseline levels during the second retention test block. The interval between

training and F F lesion did not affect these results, as both FF1 and F F 7 groups were equally impaired. The A G E D group also exhibited a retention performance deficit that paralleled that of FFlesioned animals but appeared worse because of the earlier impaired acquisition. For each performance variable, a repeated measures ANOVA for the two young sham groups (CON1 and CON7) and the two lesion groups (FF1 and FF7) revealed no significant main effects of the acquisition-surgery interval (all P's > 0.1). Therefore, for all subsequent analyses the two sham groups were combined into a single control group (CON), and the two lesion groups were combined into a single lesion group (FF). (1) Number of errors. Because of the large age difference in level of acquisition, retention performance was inferred from a comparison between the last trial of acquisition training and the first trial of the retention test for each group. Results of within-group comparisons indicated

259 that CON animals maintained perfect retention performance on this error measure; whereas, both F F animals and A G E D rats exhibited a higher number of errors on the first retention trial compared to the last acquisition trial (paired t-tests, P's < 0.01). A repeated measures two-factor (group by block) ANOVA for between-group (CON, FF and AGED) comparisons of retention errors across blocks revealed significant main effects of groups and blocks (F2.32 = 21.95 and FI,32 = 26.00, P's < 0.01, respectively). The interaction between groups and blocks was significant (F2.32 = 4.76, P < 0.05). Subsequent multiple comparisons by Tukey tests (1~< 0.05) revealed that FF rats made more errors than CON rats during the two retention blocks. Furthermore, the number of errors among A G E D animals was significantly higher than that of FF rats at both blocks of the retention test. (2) Alternation errors. On this measure, the within-group paired t-test comparisons of the last acquisition trial to the first retention trial again showed no retention effect for the CON group (t < 1.0). In contrast, alternation errors on the retention trial for both FF and A G E D rats were higher than those on the last acquisition trial (P's < 0.01 and 0.05, respectively). A repeated measures ANOVA revealed a significant main effect of both groups and blocks (F2.32 = 4.58, P < 0 . 0 5 ; F1.32 = 6.45, P < 0 . 0 1 , respectively); however, the interaction of groups and blocks was not significant. Post-hoc group comparisons using the Tukey test (P < 0.05) revealed that alternation errors in both the FF and AGED groups were significantly higher than those of the CON group at retention block 5, and that no difference existed between F F and A G E D groups. At retention block 6, F F animals still showed significantly higher alternation errors compared to CON animals, but they exhibited lower alternation errors than did the A G E D rats. (3) Runtime. FF and A G E D groups both required more time to negotiate the maze at the first retention trial than at the last acquisition trial (within-group paired t-tests, P's < 0.01 and 0.05, respectively). Performance on the first retention trial for CON animals was not statistically different from that of the last acquisition trial. A

repeated measures ANOVA revealed significant main effects of groups and blocks (F2,32 = 12,95 and F~,32 = 9.21, P's < 0.01 respectively), and the interaction between groups and blocks also was significant (F2,a2 = 9.21, P < 0.01). Tukey comparisons (P < 0.05) revealed that FF animals were impaired at the first retention block on this runtime measure but did recover to a level similar to the CON group on the second retention block; whereas AGED rats exhibited longer runtimes than those of young groups throughout the retention test. (4) Number of shocks. Within-group paired ttest comparisons revealed retention impairments in FF animals (P < 0.01) but not in AGED or CON rats. The repeated measures ANOVA indicated significant main effects of groups and blocks (F2,32= 16.06 and F1,32-- 41.45, P's<0.01, respectively), and the interaction between groups and blocks again was significant (F2.32 = 9.62, P < 0.01). Tukey tests (1~< 0.05) indicated a significantly greater number of shock episodes in the AGED group compared to FF and CON groups. Furthermore, FF animals again exhibited significant impairment during the initial but not the second retention block. (5) Shock duration. On this measure, withingroup paired t-test comparisons revealed no retention deficits in CON or AGED animals; only FF rats exhibited a deficit. However, the repeated measures ANOVA indicated significant main effects of groups and blocks (F2,32 = 7.64, P < 0.01; F2.32 = 10.58, P < 0.01, respectively) as well as a significant interaction between groups and blocks (F2.32 = 7.56, P < 0.01). Results of Tukey tests ( P < 0.05) indicated significantly higher shock duration at the first retention block in FF and Aged rats compared to that of CON animals. However, during the second retention block, both FF and AGED groups showed recovery similar to the level of CON animals. DISCUSSION In a previous study, Bresnahan et al. 6 reported that bilateral FF damage disrupted acquisition performance of young rats in the shock-motivated 14-unit T-maze. This deficit in maze performance

260 appeared similar to that previously observed in aged rats 1~.~5 and in scopolamine-treated young rats a6. Thus, involvement of the septohippocampal cholinergic system in the acquisition of this maze has been suggested. The present study, which focused on retention performance in the 14-unit T-maze, has shown that bilateral F F lesions produced modest but significant retention deficits. Savings of learning was evident as retention performance for these lesion animals was superior to performance during early acquisition trials. Although the number of errors and alternation errors of lesioned animals were still higher than those of sham-operated rats on the last block of the retention test, FF-lesioned animals did show rapid improvement on the non-cognitive behavioral measures - runtime, number of shocks, and shock duration. Therefore, the retention deficit induced by the FF lesion appeared relatively transitory, especially with these noncognitive measures. Secondly, the results of the retention study indicate that the interval between acquisition training and lesions had no influence on retention performance. Animals with F F lesions one-day after acquisition or one-week after acquisition exhibited similar performance on the retention test. Finally, performance of aged rats during retention trials was significantly worse than that of either young controls or lesioned animals; and as previously observed, aged animals exhibited poor acquisition performance compared to their young counterparts 15'~6. In contrast to the view that the hippocampus is involved primarily with working memory, Barnes 3 has suggested that the hippocampus and its related structures appear to be critical for acquisition of both working and reference memory tasks when the task heavily involves place learning. For example, lesions of the hippocampus or its related structures have been shown to disrupt acquisition of both working and reference memory components of spatial tasks a~,~4,34. However, in these tasks, once efficient performance is acquired, hippocampal damage appears to be less effective for inducing performance deficits. For example, DiMattia and Kesner s demonstrated that hippocampal lesions after acquisition of the Morris water maze produced only

small deficits in retention and that lesioned rats quickly recovered to the prelesion performance level. Also, J a r r a r d 19 reported temporary performance impairment of both working and reference memory tasks in rats with CA1 and CA3 hippocampal fields damaged by kainic acid. Taken together, these findings suggest that the hippocampus and its related structures are primarily involved in acquisition of new memories regarding the integration of information about spatial environments, regardless of a working or reference memory requirement, and that hippocampal structures are less involved in retention or retrieval of this information after acquisition. That is, other structures such as the neocortex might store and maintain memories in a more permanent fashion. In general, the current results are consistent with the above perspective. Although the FF lesions did produce significant impairment in error performance during the retention test, the level of impairment was modest (about the same level as that of blocks 2 and 3 during training), and lesioned animals did show rapid recovery at block 6. It was noted that recovery rate as measured by alternation errors and number of errors was different from recovery rate as measured by runtime, number of shocks, and shock duration. In lesioned rats, non-cognitive variables including runtime, number of shocks, and shock duration returned to levels similar to those of sham groups more quickly than did those observed for the cognitive measures, error and alternation errors. The tendency to commit alternation errors induced by F F lesions appears to be more resistant to recovery. While young rats manifested the strong alternation strategy for solving the maze during early training, this strategy was gradually abandoned for a more efficient one. The well-trained animal apparently does not rely upon learning a configuration of extramaze stimuli to orient itself in space. Instead, the animal must develop a response algorithm that codes the correct sequence of left-right turns in the maze 15,16. In contrast to CON rats which exhibited a low alternation level in the retention test, the fact that FF-damaged rats relied heavily upon an alternation strategy in the retention phase is likely due to the difficulty for

261 these animals to maintain an internal response algorithm acquired before the lesion. A similar behavioral phenomenon after F F lesions has been observed in the Morris water maze as a stereotypic circling swim 28. If the hippocampus is transitorily involved in acquisition and storage of new information as suggested above, the timing of hippocampal damage and acquisition should be a critical factor in the observed lesion-induced impairment. Flexner et al. 9 noted that a 1-week interval appeared critical for disrupting performance of mice during retention testing in a Y-maze avoidance task. This interval, however, did not prove important in the current study. The retention deficit was equivalent whether the lesion was made one day or one week after acquisition training. Using a Morris water maze task, Sutherland et al. 37 noted that about 8 weeks post-training was the critical period for producing retention deficits with hippocampal lesions. Rats with lesions made 12 weeks after training exhibited better retention performance compared to groups lesioned earlier. Thus, perhaps in the current experiment if the FF lesion had occurred at a later interval following training, then no disruption of retention performance would have been observed. Alternatively, if only one short acquisition session had been used, then the retention deficit might have been greater than that observed. Discrepancies among studies regarding the length of the interval for disruption by hippocampal damage may be related to the nature of the lesions, to the task or both, as well as to the amount of prelesion acquisition training. Thus, the transitory nature of hippocampal functioning involved in information processing remains unclear. In the present study, F F lesions may have disrupted some of the cholinergic medial septal afterents to the hippocampus. Previous pharmacological results did suggest that the central cholinergic system is involved in acquisition but not retention of the 14-unit T-maze. Spangler et al.35 reported that a muscarinic receptor antagonist, scopolamine (1.0 mg/kg, i.p.), administered prior to acquisition training of the 14-unit T-maze disrupted acquisition, whereas the same dose injected prior to retention testing did not impair performance.

Thus, it is possible that other neurotransmitter systems coursing through the F F 24 could be responsible for the transitory retention deficits observed in FF-lesioned rats in this experiment. Further experiments that focus on lesions of certain neurotransmitter systems involved in this pathway are recommended. Also additional analysis using acetylcholinesterase staining of cholinergic hippocarnpal input such as medial septal nucleus is suggested26. Regarding the age effect on 14-unit T-maze performance in the current experiment, the results were straightforward. Similar to many past results involving this paradigm 16, aged rats were markedly impaired in acquisition compared to their younger counterparts. When error performance between the last acquisition trial and the first retention trial was compared in the present study, aged rats exhibited a retention deficit one week after acquisition training. Young control rats showed perfect retention over both 1- and 2-week intervals. In a food-motivated version of this task, Goodrick 12 reported an increase in errors in both young (6 months) and aged (26 months) Wistar rats over a 45-day interval. However, this retention deficit appeared equivalent across agegroups if differences in original learning were considered. The discrepancy between our results and those of Goodrick's may be due to methodological factors, but the perfect retention performance evidenced in young control groups over this shorter interval (1-2 weeks) in the current experiment is a viable explanation for differing results. It should be noted, however, that the age-related retention deficit in the current experiment was not a marked one as the aged rats exhibited recovery between retention blocks 5 and 6 on all parameters except alternation errors. Whether or not aged rats would exhibit a retention deficit if they had achieved the same level of acquisition as that of young rats is questionable. It is our experience in pilot studies and that of Goodrick's in past studies ~2'13 that the level of maze mastery achieved by young rats can never be obtained by most aged rats, even after extensive training. Several parallels were evident when the retention performance of aged rats was compared to that of FF-lesioned rats. First, the retention

262 deficit was not manifested as a dramatic retrograde amnesia in either group. Second, both groups showed recovery during retention testing. Third, in terms of recovery, the cognitive measures of errors and alternation errors appeared more affected in both groups than were the non-cognitive performance variables. Thus, behavioral parallels between aging and FF damage are evident for both acquisition 6,16 and retention performance in this task. Whether or not these similarities stem from common neurobiological mechanisms remains uncertain. The role of severN neurotransmitter systems in transitory retention deficits observed in both aged and FFlesioned animals should be further investigated in this complex maze paradigm. ACKNOWLEDGEMENTS

The authors acknowledge the valuable contribution of Gunther Baartz, Richard Hiner, Maurice Zimmerman, and Raymond Bannar for construction of apparatus, William Yee for data analyses and computer-generated figures, and Nancy Muth for histological assistance.

8

9

10

11

12

13

14

15

16 REFERENCES 1 Barnes, C.A., Memory deficits associated with senescence: a tleurophysiologieal and behavioral study in the rat, J. Comp. Physiol. Psychol., 93 (1979) 79-104. 2 Barnes, C.A., Selectivity of neurological and mnemonic deficits in aged rats. In T.D. Petit and G.O. Ivy (Eds.), Neural plasticity: a ltfespan approach, Alan R. Liss, New York, 1987, pp. 235-264. 3 Barnes, C.A., Spatial learning and memory processes: the search for their neurobiological mechanisms in the rat, TINS, 11 (1988) 163-169. 4 Bartus, R.T., Dean, R.L., Beer, B. and Lippa, A.S., The eholinergic hypothesis of geriatric memory dysfunction, Science, 217 (1982) 408-412. 5 Berman, R.F., Goldman, H. and Altman, H.J. Agerelated changes in regional cerebral blood flow and behavior in Sprague-Dawley rats, NeurobioL Aging, 9 (1988) 691-696. 6 Bresnahan, E., Kametani, H., Spangler, E., Chaehieh, M., Wiser, P. and Ingram, D.K., Fimbria-fornix lesions in young rats impair acquisition in a 14-unit T-maze simi!ar to prior observed performance deficits in aged rats, Psychobiology, 16 (1988) 243-250. 7 De Toledo-Morrell, L., Geinisman, Y. and Morrell, F., Individual differences in hippocampal synapfic plasticity as a function of aging: behavioral, electrophysiological

17

18

19

20

21

22

23

and morphological evidence. In T.L. Petit and G.O. Ivy (Eds.), Neural plasticity : a lifespan approach, Alan R. Liss, New York, 1987, pp. 283-328. DiMattia, B.V. and Kesner, R.P., Posterior parietal cortex: a part of the cognitive map*. N'eurosci. Abs~r., 10 (1984) 136. Flexner, L.B., Church, A.C., Flexner, J.B. and Rainbow, T.C., The effect of the beta-adrenergic receptor antagonist, propranolol, on the cerebral spread of a memory trace in mice, Pharmacol. Biochem. Behav., 21 (1984) 633-639. Gage, F.H., Dunnett, S.B. and Bj~Srklund, A., Spatial learning and motor deficits in aged rats, Neurobiol. Aging, 5 (1984) 543-548. Gage, P.D., Performance of hippocampectomized rats in a reference/worklng memory task: effect of preoperative versus postoperative training, Physiol. Psychol., 13 (1985) 235-242. Goodrick, C.L., Learning, retention and extinction of a complex maze habit for mature young and senescent Wistar albino rats, J. Gerontol., 23 (1968) 298-304. Goodrick, C.L., Error goal-gradient of mature young and aged rats during training in a 14-unit spatial maze, Psycho1. Rep., 32 (1973) 359-362. Harrell, L.E., Barlow, T.S. and Parsons, D., Cholinergie neurons, learning, and recovery of function, Behav. Neurosci., 101 (1987) 644-652. Ingrain, D.K., Analysis of age-related impairments in learning and memory in rodent models, Ann. N.Y. Acad. Sci., 444 (I985) 312-331. Ingrain, D.K., Complex maze learning in rodents as a model of age-related memory impairment, Neurobiol. Aging, 9 (1988) 475-485. Ingrain, D.K., London, E.D. and Goodrick, C.L., Age and neurochemical correlates of radial maze performance in rats, Neurobiol. Aging, 2 (1981) 41-47. Jarrard, L.E., Selective hippocampal lesions: differential effects on performance by rats of a spatial task with preoperative versus postoperative training, J. Comp. Physiol. Psychol., 92 (1978) 1119-1127. Jarrard, L.E., Selective hippocampal lesions and behavior: effects of kainic acid lesions on place mad cue tasks, Behav. NeuroscL, 97 (1983) 873-889. Kosher, R.P. and DiMattia, B.V., Posterior parietal association cortex and hippocampus: equivalent of rmaemonic function in animals and humans. In L.R. Squire and N. Butters (Eds.), Neuropsychology of Memory, Guilford, New York, 1984, pp. 385-398. Kesner, R.P., Neurobiologieal views of memory. In J.L. Martinez and R.P. Kesner (Eds.), Learning and Memoly: A Biological View, Academic, Orlando, 1986, pp. 389-438. Knowlton, B.J., Wenk, G., Olton, D.S. and Coyle, 7.T., Basal forebrain lesions produce a dissociation of trialdependent and trial-independent memory performance, Brain Res., 345 (1985) 315-321. Landfleld, P.W., McGaugh, J.L. and Lynch, G., Impaired synaptio potentiation processes in the hippocampus of aged, memory-deficient rats, Brain Res., 150 (1978) 85-101.

263 24 MacLean, P.D., Fiber systems of the forebrain. In G. Paxinos (Ed.), The Rat Nervous System, VoL I, Academic, Orland, 1985, pp. 417-440. 25 Michel, M.E. and Klein, A.W., Performance differences in a complex maze between young and aged rats, Age, 1 (1978) 13-16. 26 Mitchell, S.J., Rawlins, J.A.., Steward, O. and Olton, D.S., Medial septal lesions disrupt theta rhythm and cholinergic staining in medial entorhinal cortex and produce impaired radial arm maze behavior in rats, J. Neurosci., 2 (1982) 292-302. 27 Morris, R.G.M., Garrard, P., Rawlins, J.N.P. and O'Keefe, J., Place navigation impaired in rats with hippoeampal lesions, Nature (Lond.), 297 (1982) 681-683. 28 Nilsson, O.G., Shapiro, M.L., Gage, F.H., Olton, D.S. and Bj0rklund, A., Spatial learning and memory following fimbria-fornix transection and grafting of fetal septal neurons to the hippocampus, Exp. Brain. Res., 67 (1987) 195-215. 29 O'Keefe, J. and Nadel, L., TheHippocampusasa Cognitive Map, Oxford University Press, Oxford, 1978. 30 Olton, D.S., Becker, J.J. and Handelmann, G.E., Hippocampus, space and memory, t~ehav. Brain Sci., 2 (1979) 313-322. 31 Olton, D.S. and Feuster, W.A., Hippocampal function acquired for spatial working memory, Exp. Brain Res., 41 (1981) 380-389. 32 Olton, D.S., Walker, J.A. and Gage, F.H., Hippocampal

connections and spatial discrimination, •rain Res., 139 (1978) 295-308, 33 Paxinos, G. and Watson, C., The Rat Brain in Stereotaxic Coordinates, Academic, New York, 1982. 34 Rasmussen, M., Barnes, C.A. and McNaughton, B.L., The relative role ofentorhinal prefrontal cortical systems in spatial and non-spatial reference learning, Neurosei. Abstr., 12 (1986) 749. 35 Spangler, E.L., Chachich, M.E. and Ingram, D.K., Scopolamine in rats impairs acquisition but not retention in a 14-unit T-maze, Pharrnaeol. Biochem. Behav., 30 (1988) 949-955. 36 Spangler, E.L., Rigby, P. and Ingrain, D.K., Scopolamine impairs learning performance of rats in a 14-unit T-maze, Pharmacol. Biochem. Behav., 25 (1986) 673-679. 37 Sutherland, R.L, Arnold, K.A. and Rodriquez, A.R., Anterograde and retrograde effects on place memory at~er limbic or diencephalic damage, Neurosci. Abstr., 13 (1987) 1066. 38 Thomas, G.J., Memory: time binding in organisms. In L.R. Squire and N. Butters (Eds.), Neuropsychology of Memory, Guilford, New York, 1984, pp. 374-384. 39 Walker, S.A. and Olton, D.S., Fimbria-fornix lesions impair spatial working-memory but not cognitive mapping, Behav. Neurosci., 98 (I984) 226-242. 40 Winocur, G., The hippocampus and cue-utilization, Physiol. PsychoL, 8 (1980) 280-288.