- Email: [email protected]
The effect of excitotoxic hippocampal lesions on simple and conditional discrimination learning in the rat
Behavioural Brain Research 99 (1999) 103 – 113
Research report
The effect of excitotoxic hippocampal lesions on simple and conditional discrimination learning in the rat T.K. Murray a,*, R.M. Ridley b a
b
Astra Arcus AB, S-151 85 So¨derta¨lje, Sweden MRC Comparati6e Cognition Team, Department of Experimental Psychology, Downing Street, Cambridge, CB2 3EB, UK Received 13 March 1998; received in revised form 2 June 1998; accepted 2 June 1998
Abstract The effect of excitotoxic lesions of the hippocampus on acquisition and reversal of simple and conditional tasks was investigated using a Y-maze. Hippocampal-lesioned rats were severely impaired on acquisition and reversal of a conditional visuo-spatial task (where different pairs of visually distinctive choice arms indicated whether a left or right arm choice was correct on that trial) and were unable to acquire a visuo-visual conditional discrimination (where the appearance of the start arm indicated which of the visually distinctive choice arms was correct irrespective of their left/right position). They were not impaired on acquisition or reversal of a simple spatial left/right discrimination task (where all arms had the same visual appearance) nor on acquisition of a visual discrimination (where the correct, visually distinctive, choice arm varied in its left/right position). Hippocampal-lesioned rats were, however, impaired on reversal of this visual discrimination task and on acquisition and reversal of another visual discrimination task in which the visually distinctive choice arms were less different from each other than in the first version of this task. The degree of impairment in the lesioned rats was related to task difficulty for the sham-operated rats and was not specific to tasks requiring spatial choices, visual discrimination or conditional responding. The impairment on conditional tasks was greater than the impairment on those non-conditional tasks which happened to be matched for task difficulty for the sham-operated rats, suggesting that the conditional demand may target the function of the hippocampus rather closely. Statistically worse than chance performance by hippocampal-lesioned (and sham-operated) rats at the beginning of reversal testing, which was given 24 h after achieving criterion on acquisition of that task, indicated that hippocampal-lesioned rats simultaneously exhibited good memory but impaired learning for the type of information required for those tasks. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Learning; Hippocampus; Rat; Reversal; Spatial discrimination; Conditional discrimination; Y-maze
1. Introduction Rats with hippocampal lesions have frequently been found to be impaired on spatial tasks including those presented in a water maze [19], a radial arm maze [4,12] or a Y-maze [6]. A general role for the hippocampus in spatial learning has been suggested [5,11] although * Corresponding author. Lilly Research Centre, Erlwood Manor, Windlesham, Surrey GU20 6PH, UK.
hippocampal-lesioned rats may not be impaired on very simple spatial tasks [10,22]. Hippocampal-lesioned rats have also been found to be impaired on a variety of conditional [6], configural [17] or contextual [20] tasks where the correct choice of a discriminative stimulus depends on information contained in the context, or in some other stimulus, proximity or circumstance present at the time of response choice. This has lead to a view of hippocampal function in the domain of configural or relational memory [3,23] although several studies have
0166-4328/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved. PII S0166-4328(98)00077-1
104
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
failed to find impairment by rats with hippocampal lesions on apparently configural tasks [1,2]. In this paper, we use a Y-maze to assess the effect of hippocampal lesions made using the excitotoxin, Nmethyl-D-aspartate (NMDA) on (i) simple left/right spatial discrimination tasks, where all arms had the same visual appearance; (ii) visual discrimination tasks, where the correct, visually distinctive, choice arm varied in its left/right position; (iii) a conditional visuo-spatial discrimination task, where different pairs of visually distinctive choice arms indicated whether a left or right arm choice was correct on that trial; and (iv) a conditional visuo-visual discrimination task, where the appearance of the start arm indicated which of the visually distinctive choice arms was correct irrespective of their left/right position. Comparison of performance on these tasks will indicate the extent to which the hippocampus is involved in spatial and/or conditional cognitive processing.
2. Methods
2.1. Animals Male Lister Hooded rats (Harlan UK, Bicester, UK), weighing 250–300 g at the start of the experiment, were used. They were housed in groups of five rats per cage and maintained on a 12-h light/12-h dark cycle (lights on 7:00–19:00 h). All testing was carried out during the light phase of the cycle. The rats were maintained at 90% of their free-feeding body weight and fed at the end of each test session. Water was available ad libitum.
2.2. Lesions Rats were anaesthetized with a gaseous anaesthesic consisting of Halothane® (Rhoˆne Me´rieux, Essex, UK) /nitrous oxide/oxygen (induction: 1.0 l/min N2O, 0.5 l/min O2, 4–8% Halothane; maintenance: 1.0 l/min N2O, 0.5 l/min O2, 1 – 1.5% Halothane, adjusted according to each rat’s respiration rate and reflex response). Lesions were made using 0.5 ml 0.12 M N-methyl-D-aspartate (NMDA) injected stereotaxically into three sites per hemisphere. NMDA was purchased from Sigma (Dorset, UK) dissolved in phosphate-buffered saline with the aid of 2 M sodium hydroxide and the pH titrated to 6.5–7.5 by addition of 2 M hydrochloric acid. Stereotaxic co-ordinates were AP: −3.14, L:9 1.6, V: − 2.8; AP: −4.16, L: 94.0, V: − 3.2; AP: − 4.8, L: 9 3.0, V: −2.6, from bregma according to the atlas of Paxinos and Watson [13]. Each injection was made over a period of 5 min, followed by a 5 min equilibration time, with the needle remaining in place. Shamoperated rats received identical surgery to the NMDA
group but without the injection of the neurotoxin. Diazepam (10 mg/kg, Roche Products, UK) was administered (i.p.) to all rats (including sham-operated) to prevent seizures following surgery. Seven days post surgery rats began training on the Y-maze tasks.
2.3. Apparatus and pre-training Rats were tested in a Y-maze, which was made of grey perspex and had three arms of equal size, 60 cm long, 11.5 cm wide and 25 cm high, positioned 31 cm above the floor. The start arm had a guillotine door behind which the rats were held until the start of each trial, which commenced when the guillotine door was raised. A food cup was located at the end of each choice arm. Rats were pre-trained for 4 days so they would become familiar with the maze. On day 1 of pre-training, pairs of rats were placed for 10 min in the start arm of the maze, with food pellets (45 mg, Noyes sucrose pellets, Sandown Scientific, UK) scattered in all arms. On day 2, rats were placed singly in the maze for 10 min; again, pellets were scattered in the arms. On days 3 and 4 the rats were placed in the maze singly for 5 min; on day 3, pellets were placed at the end of both choice arms, and on day 4, pellets were placed only in the food cups.
2.4. Experimental design On all tasks, correct responses were rewarded with four food pellets placed in a food cup at the end of the correct arm. If a rat made an incorrect response, it was allowed to go to the empty food well at the end of the incorrect arm and was removed after 5 s. This was recorded as an error. Maximum trial length was 3 min. If the rat did not make a choice within this time or did not eat all four reward pellets within the time allowed, it was removed from the maze. This was regarded as a non-trial and was not recorded as an error. Rats were run successively in groups of five such that the intertrial interval (ITI) for each rat was approximately 3 min. The rats were placed in their home cage between trials. Rats were trained for ten trials per day, to a criterion of nine correct responses in ten consecutive completed trials on all tasks. The behavioural studies described here were performed over many months and it was necessary to use three sets of rats for this purpose. Thus not all rats were tested on all behavioural tasks: Set 1 (sham n=8, lesion n= 9) were tested on the spatial and visual (version II) discrimination tasks; Set 2 (sham n=8, lesion n= 9) were tested on the visual (version I) discrimination and the visuo-spatial conditional tasks; Set 3 (sham n=9, lesion n=9) were tested on the visuo-visual conditional task only.
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
105
Table 1 Trials to criterion ( 9 S.E.M.) Task
Sham operated
Hippocampal lesioned
Unmatched t-test
Spatial discrimination Spatial reversal Visual discrimination version I Visual reversal version I Visual discrimination version II Visual reversal version II Visuospatial conditional Visuospatial reversal Visuovisual conditional Visuovisual reversal
29.1 97.3 34.4 94.5 42.3 95.1 54.6 99.2 50.1 910.1 80.7 911.1 53.4 98.2 62.6 96.2 79.5 91.4 91.2 917.1
35.7 9 5.0 34.3 9 4.7 51.9 98.2 78.1 910.1 83.3 9 12.4 156.2 9 20.7 122.4 917.4 145.0 98.5 300 fail Not done
n.s. n.s. n.s. PB0.05 PB0.05 PB0.001 PB0.005 PB0.001 Not applicable Not applicable
2.5. The spatial discrimination task Rats were trained on acquisition and then on reversal of a spatial discrimination task where all arms were grey. Half of the rats in each group were allocated left arm positive and the remaining rats were allocated right arm positive for acquisition. At 24 h after completion of acquisition, these conditions were reversed and the rats were trained on this reversal to criterion.
2.6. The 6isual discrimination task 6ersion I In this task, discriminative stimuli consisted of black or white, perspex, ‘whole arm’ inserts, which covered the walls and floor of the whole of the choice arms. The start arm was coloured grey. Each insert measured 59 cm long, 11 cm wide and 25 cm high. The left/right position of the black and white inserts varied according to a pseudorandom schedule. Half of the rats in each group were allocated black arm positive and the remaining rats were allocated white arm positive for acquisition. Twenty-four hours after completion of acquisition, these conditions were reversed and the rats were trained on this reversal to criterion.
2.7. The 6isual discrimination task 6ersion II Acquisition and reversal of another visual discrimination task was also performed as above but using black or white ‘floor’ inserts rather than the ‘whole arm’ inserts used for the visual discrimination task version I. It was supposed that this might make the task more difficult by providing less salient discriminative information.
2.8. The 6isuo-spatial conditional task This task used the same discriminative stimuli that were used for the visual discrimination task version 1, i.e. black or white ‘whole arm’ inserts. On each trial, two
identical inserts were used but the inserts were swapped between right and left arms, according to a pseudorandom schedule, to prevent the rats from learning to choose one of the inserts on the basis of distinguishing blemishes. Half of each group of rats was rewarded for choosing the left arm when both inserts were white, and the right arm when the inserts were black, while the remaining rats was tested according to the opposite reward contingency. Twenty-four hours after completion of acquisition, these conditions were reversed and the rats were trained on this reversal to criterion.
2.9. The 6isuo-6isual conditional task In the visuo-visual conditional task, ‘whole arm’ inserts made of black, white, striped or grey perspex were used and testing followed the same procedure as the visuo-spatial task. Each arm, including the start arm, contained a different insert and each rat was assigned a rule such as ‘if start arm grey, choose the black (not the white) choice arm; if start arm striped, choose the white (not the black) choice arm’. The start arm was grey or striped according to a pseudorandom schedule and the left/right position of the black/white choice arms was determined according to another, superimposed, pseudorandom schedule. Twenty-four hours after completion of acquisition, these conditions were reversed and those rats which had attained criterion on acquisition were trained on this reversal to criterion.
2.10. Histology On completion of behavioural testing, rats were perfused transcardially with saline followed by 10% formaldehyde in 5% sucrose. Brains were removed and placed in 5% formaldehyde in 30% sucrose at 4°C for 2 days. Coronal 20 mm (1 per 100 mm) sections were cut through the hippocampus using a cryostat and stained with Cresyl Violet.
106
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
Fig. 1. Learning curves for spatial discrimination learning in the Y-maze. =sham-operated rats (n =8); = hippocampal-lesioned rats (n= 9). Ordinate shows mean number of errors per block of five trials ( 9S.E.M.). Solid horizontal line: chance performance (2.5 errors/five trials); dotted horizontal lines: limits of chance at 5% level when all rats are performing. A. Acquisition. B. Reversal.
3. Results Trials to criterion ( 9 S.E. of the mean) for each group for each task are shown in Table 1. Groups were compared on each task using unmatched t-tests. Learning curves were plotted for each task (see Figs. 1–5). The limits of chance for each block of trials, when all animals in each group were performing the task, were calculated and are shown as dotted lines in the figures. Thus the mean chance number of errors performed in a block of five trials is 2.5, error scores in excess of 3.2 errors per block of five trials (calculated using the
binomial test for eight to nine animals per group) indicate that the group is performing at statistically worse than chance; error scores which are less than 1.8 errors per block of five trials indicate that the group is performing at statistically better than chance at that point in the learning curve.
3.1. The spatial discrimination task There was no difference in the number of trials to criterion on acquisition of the spatial discrimination task (sham-operated: 29.19 7.3 trials; hippocampal-le-
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
107
Fig. 2. Learning curves for the visual discrimination task version I in the Y-maze. = sham-operated rats (n = 8); =hippocampal-lesioned rats (n =9). Ordinate shows mean number of errors per block of five trials ( 9 S.E.M.). Solid horizontal line: chance performance (2.5 errors/five trials); dotted horizontal lines: limits of chance at 5% level when all rats are performing. A. Acquisition. B. Reversal.
sioned: 35.795.0 trials, see Fig. 1A). On the reversal of this task, both groups began by performing at just worse than chance and rapidly learnt the reversal at the same rate (Fig. 1B). The number of trials to criterion did not differ (sham-operated: 34.49 4.5 trials; hippocampal-lesioned: 34.394.7 trials).
3.2. The 6isual discrimination task 6ersion I There was no difference in the number of trials to criterion on acquisition of this visual discrimination
task (sham-operated: 42.39 5.1 trials; hippocampal-lesioned: 51.99 8.2 trials, see Fig. 2A). On the reversal of this task, both groups began by performing at worse than chance although hippocampal-lesioned rats learnt this reversal more slowly than sham-operated rats (sham-operated: 54.69 9.2 trials; hippocampal-lesioned: 78.19 10.1 trials, (unmatched t-test, t=2.31, PB 0.05, see Fig. 2B).
3.3. The 6isual discrimination task 6ersion II Hippocampal-lesioned rats required more trials than
108
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
Fig. 3. Learning curves for the visual discrimination task version II in the Y-maze. = sham-operated rats (n = 8); = hippocampal-lesioned rats (n =9). Ordinate shows mean number of errors per block of five trials ( 9S.E.M.). Solid horizontal line: chance performance (2.5 errors/five trials); dotted horizontal lines: limits of chance at 5% level when all rats are performing. A. Acquisition. B. Reversal.
sham-operated rats to reach criterion on acquisition of version II of the visual discrimination task (sham-operated: 50.1910.1 trials; hippocampallesioned: 83.3912.4 trials, unmatched t-test, t = 2.26, P B0.05, see Fig. 3A). On the reversal of this task, both groups began by performing at worse than chance and returned to chance performance at the same rate; hippocampal-lesioned rats then spent a substantial proportion of the total number of trials performing at chance before approaching criterion at an almost normal rate (sham-operated: 80.79 11.1 trials;
hippocampal-lesioned: 156.29 20.7 trials, unmatched t-test, t= 4.12, P B 0.001, Fig. 3B).
3.4. The 6isuo-spatial conditional task Hippocampal-lesioned rats were severely impaired on acquisition of the visuo-spatial task (sham-operated: 53.49 8.2 trials; hippocampal-lesioned: 122.49 17.4 trials; unmatched t-test, t= 3.76, PB0.005). Both groups started performance at chance but, while the sham-operated group showed a steady improvement over trials,
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
109
Fig. 4. Learning curves for visuo-spatial conditional discrimination learning in the Y-maze. = sham-operated rats (n =8); =hippocampal-lesioned rats (9). Ordinate shows mean number of errors per block of five trials ( 9 S.E.M.). Solid horizontal line: chance performance (2.5 errors/five trials); dotted horizontal lines: limits of chance at 5% level when all rats are performing. 4. Acquisition. 4. Reversal.
reaching criterion after 8 days of training, the hippocampal-lesioned group continued to perform at chance for 15 days before better than chance performance was consistently achieved (Fig. 4A). Hippocampal-lesioned rats were also severely impaired on reversal learning of the visuo-spatial task (sham-lesioned: 62.69 6.2 trials; hippocampal-lesioned: 145.098.5 trials; unmatched t-test, t = 4.09, PB 0.001). Fig. 4B shows that both sham-operated and hippocampal-lesioned rats started reversal learning at worse than chance performance. Sham-lesioned rats then learnt at a steady, consistent rate whereas the
hippocampal-lesioned rats learnt at a consistent, but slower, rate.
3.5. The 6isuo-6isual conditional task Only one of the hippocampal-lesioned rats was able to acquire the visuo-visual task within 350 trials whereas the sham-lesioned group learnt this task in 79.59 14.0 trials. Fig. 5 shows that the hippocampal-lesioned group were largely unable to improve above chance performance. Reversal testing on this task could not be done using the hippocampal-lesioned rats since
110
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
Fig. 5. Learning curves for visuo-visual conditional discrimination learning in the Y-maze. = sham-operated rats (n =9); = hippocampal-lesioned rats (n=9). Ordinate shows mean number of errors per block of five trials ( 9 S.E.M.). Solid horizontal line: chance performance (2.5 errors/five trials); dotted horizontal lines: limits of chance at 5% level when all rats are performing. Data are for acquisition; reversal could not be undertaken because hippocampal-lesioned rats were unable to learn the task.
they had not completed acquisition. Sham-lesioned rats learnt this reversal in 91.2917.1 trials.
3.6. Histology The extent of the hippocampal lesion is illustrated in Fig. 6 and represents the minimum and maximum size of the lesion. All animals received bilateral damage to the hippocampus, primarily to the CA1 region. In the dorsal hippocampus, complete pyramidal cell loss in the CA1 region was seen with some damage in the CA2 and CA3 regions occurring in approximately half of the animals. In the ventral hippocampus, part of the CA1 field was spared and only minimal damage to the CA3 region was seen in a small number of rats. Damage occurred in the dentate gyrus in approximately one third of the animals. Very little damage was observed outside the hippocampal formation, occurring only in the region of the needle tracts in some animals. There were no systematic differences between the lesions in the three groups of lesioned rats.
4. Discussion These experiments demonstrate that hippocampal lesions produce a severe learning impairment on conditional discrimination tasks and an impairment on some forms of visual discrimination learning. The circumscribed nature of the impairment is illuminated by the equivalent worse than chance performance at the beginning of the reversal tasks shown by both
hippocampal-lesioned and sham-operated rats. This indicates that both groups possessed equivalent retention of information acquired 24 h previously and were equally competent to make non-random choices based on that retention. This excludes the possibility that the impairment resulted from motoric, perceptual or motivational difficulties in the lesioned rats. Furthermore, this worse than chance performance excludes the possibility that impairment could result from motoric or spatial perseveration; persistent choice of one arm on either of the conditional tasks would have resulted in chance (rather than worse than chance) performance since the rewarded arm in these tasks was random with respect to left/right position. Impairment on reversal learning could be described as ‘cognitive perseveration’ or ‘interference’ if the rats were continuing to make cognitive choices based on previously acquired information even when currently available information was suggesting that a new choice would be appropriate. This explanation is unlikely, however, because of the similarity in the magnitude of the impairment on visual discrimination version I reversal (where interference would be expected) compared to visual discrimination version II acquisition (where interference would not be expected) and where task difficulty for the sham-operated rats was well matched. For visual discrimination version II reversal, the performance of both groups was very similar at the beginning of reversal (where interference would be expected to be greatest); the impairment in the lesioned rats consisted mainly of chance performance in the middle portion of testing. Similarly, for reversal of the visuo-spatial task,
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
111
Fig. 6. Schematic representation of coronal sections from rat brain (based on Paxinos and Watson [13]) illustrating the histology results. The solid area illustrates the damage in the rat with the smallest lesion; the solid + hatched areas illustrate the damage in the rat with the largest lesion.
the hippocampal-lesioned rats returned to chance performance only slightly more slowly than the sham-operated rats; the greater part of the hippocampal-lesioned rats’ impairment accrued from errors in the middle and last third of the learning curve. Lastly, the impairment on reversal was highly comparable to the impairment on acquisition of the visuo-spatial task, where these tasks were well matched for difficulty in the sham-operated rats. All these comparisons suggest that the impairment exhibited by the hippocampal-lesioned rats did not consist of excessive sensitivity to interference. The difference between visual discrimination tasks version I and version II which led to the difference in the effect of the hippocampal lesion on these tasks is not clear. It might have been expected that version II would have been more difficult than version I because, in the former, the visually distinctive part of the choice arms was confined to just the floor rather than the floor and walls, but, in fact, the learning scores for version I and version II did not differ significantly for the shamoperated rats. Previous studies have found that
hippocampal lesions usually do not impair acquisition of simple visual discriminations [18,21,24] although impairment has been demonstrated on acquisition under certain circumstances [25] and is frequently seen on reversal of visual discrimination tasks [18,21]. One possibility is that unlesioned rats are able to direct their attention to the relevant part of the choice arm without needing to learn to do so (such that the two versions of the visual discrimination task are of equal difficulty) but that hippocampal-lesioned rats do not do this automatically and therefore have to learn to do so (thereby producing long periods of chance performance). Examination of the learning curves (Figs. 2 and 3) indicates that there was no difference in performance between the two groups on acquisition of version I, whereas on acquisition of version II the hippocampal-lesioned rats performed at chance for 8 days (the total time taken by the sham-operated rats to reach criterion) before approaching criterion. On reversal of version I, the hippocampal-lesioned rats showed a learning rate which was slower than for the sham-
112
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
operated rats but which was consistent across the whole learning curve, whereas on reversal of version II, the hippocampal-lesioned rats ‘unlearned’ the original discrimination at the same rate as the sham-operated rats but then performed at chance for a considerable period of time before approaching criterion. This could occur if, on reversal of version II, the hippocampal-lesioned rats not only ‘unlearned’ the association of one floor colour with reward, but also ‘unlearned’ the relevance of the floor rather than the wall colour to solving the task. Hippocampal-lesioned rats would then have to relearn to orient to this aspect of the choice arms (producing the persistent chance performance in the middle portion of the learning curve) before reversal learning could be achieved. In contrast, on reversal of version I, learning would occur at a steady rate, since attention to any part of the choice arm would provide the necessary information to learn the task. The extent to which the aspect of cognitive dysfunction in hippocampal lesioned rats which is demonstrated by the visual discrimination task version II is also responsible for the impairment on conditional tasks is not clear. Hippocampal-lesioned rats were unimpaired at learning a non-conditional spatial response equivalent to the spatial response required for the visuo-spatial conditional task, and were also unimpaired at learning the non-conditional visual discrimination task version I, which used the same discriminative stimuli in the choice arms as the visuo-visual conditional task. This indicates that the lesioned rats’ difficulty does not lie in some motoric or perceptual impairment. Hippocampal-lesioned rats were generally more impaired on those tasks which the sham-operated rats found most difficult (see Table 1), suggesting that the hippocampus is recruited to solve certain types of difficult task. The results cannot, however, be explained entirely in terms of non-specific task difficulty because, when pairs of tasks were compared which were well matched for task difficulty in the sham-operated rats, hippocampal-lesioned rats were more impaired on conditional than non-conditional tasks. Thus lesioned rats were more impaired on visuo-visual conditional acquisition than on visual discrimination version II reversal, and on visuo-spatial conditional acquisition than on visual discrimination version I reversal, even though learning scores for the relevant sham-operated groups were almost identical for the two tasks in each comparison. In a formally equivalent test situation using the Wisconsin General Test Apparatus, monkeys with dysfunction of the hippocampus induced by excitotoxic lesion, fornix transection, or neurotoxic lesion of the cholinergic input to the hippocampus, have been found to be specifically impaired on acquisition of the visuospatial and visuo-visual conditional tasks [15]. The
hippocampal impairment is critically dependent on the order of presentation of the two types of trial in conditional tasks [14] suggesting that the hippocampus is involved in the cognitive processing of conditionality rather in learning mechanisms per se. We have previously demonstrated that the NMDAreceptor antagonist, dizocilpine, also produces a severe impairment of conditional learning in rats [8] and a milder effect on simple discrimination tasks [9] which resemble the effects of hippocampal-lesions. Robinson et al. [16] have also found similarities between the effects of dizocilpine treatment and hippocampal lesions on learning in rats. Since the hippocampus contains the highest levels of NMDA receptors in the brain [7] it would seem likely that it is glutamatergic activity (NMDA is a receptor for glutamate) in the hippocampus (rather than glutamatergic activity elsewhere) which is necessary for acquisition of tasks requiring conditional analysis. In summary, we have demonstrated that hippocampal lesions produce an impairment on conditional learning in rats, and on some aspects of visual discrimination learning, but have no effect on simple spatial learning. These results are largely in agreement with the view that the hippocampus is important for learning tasks which require a flexible response to stimuli depending on context or relations with other stimuli [3], although this may be only one aspect of hippocampal function.
References [1] Davidson TL, McKernan MG, Jarrard LE. Hippocampal lesions do not impair negative patterning: a challenge to configural association learning. Behav Neurosci 1993;107:227 – 34. [2] Deacon RMJ, Rawlins JNP. Effects of aspiration lesions of hippocampus or overlying neocortex on concurrent and configural object discrimination in rats. Behav Brain Res 1996;77:165 – 74. [3] Eichenbaum H, Otto T, Cohen NJ. Two functional components of the hippocampal memory system. Behav Brain Sci 1994;17:449 – 518. [4] Jarrard LE. Selective hippocampal lesions and behavior: implications for current research and theorizing. In: Isaacson RL, Pribram KH, editors. The Hippocampus, 4. New York: Plenum Press, 1986. [5] Jarrard LE. On the role of the hippocampus in learning and memory in the rat. Behav Neural Biol 1993;60:9 – 26. [6] Kimble DP. The effect of bilateral hippocampal lesions in rats. J Comp Physiol Psychol 1963;56:273 – 83. [7] Monaghan DT, Monaghan CW. Distribution of NMDA-sensitive, L-3[H]glutamate binding sites in rat brain as determined by quantitiative antoradiography. J Neurosci 1985;5:2909 –19. [8] Murray TK, Ridley RM. The effect of dizocilpine (MK801) on conditional discrimination learning in the rat. Behav Pharmacol 1997;8:383 – 8. [9] Murray TK, Ridley RM, Snape MF, Cross AJ. The effect of dizocilpine (MK-801) on spatial and visual discrimination tasks in the rat. Behav Pharmacol 1995;6:540 – 9.
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113 [10] Oliveira MGM, Bueno OFA, Pomarico AC, Gugliano EB. Strategies used by hippocampal- and caudate-putamen-lesioned rats in a learning task. Neurobiol Learn Mem 1997;68:32 – 41. [11] Olton DS, Becker JT, Handelmann GE. Hippocampus, space and memory. Behav Brain Sci 1979;2:313–22. [12] Olton DS, Papas BC. Spatial memory and hippocampal function. Neuropsychologia 1979;17:669–82. [13] Paxinos G, Watson C. The rat brain. Stereotaxic Coordinates. Australia: Academic Press, 1986. [14] Ridley RM, Baker HF. Evidence for a specific information processing deficit in monkeys with lesions of the septohippocampal system. Cortex 1997;33:167–76. [15] Ridley RM, Baker HF, Harder JA, Pearson C. Effects of lesions of different parts of the septo-hippocampal system in primates on learning and retention of information acquired before or after surgery. Brain Res Bull 1996;40:21–32. [16] Robinson GS, Crooks GB, Shinkman PG, Gallagher M. Behavioural effects of MK-801 mimic deficits associated with hippocampal damage. Psychobiology 1989;17:156–64. [17] Rudy JW, Sutherland RJ. The hippocampal formation is necessary for rats to learn and remember configural discriminations. Behav Brain Res 1989;34:97–109.
.
113
[18] Samuels I. Hippocampal lesions in the rat: effects on spatial and visual habits. Physiol Behav 1972;8:1093 – 8. [19] Schenk F, Morris RGM. Dissociation between components of spatial memory in rats after recovery from the effects of retrohippocampal lesions. Exp Brain Res 1985;58:11 – 28. [20] Seldon NRW, Everitt BJ, Jarrard LE, Robbins TW. Complementary roles for the amygdala and hippocampus in aversive conditioning to explicit and contextual cues. Neuroscience 1991;42:335 – 50. [21] Silveira JM, Kimble DP. Brightness discrimination and reversal in hippocampally-lesioned rats. Physiol Behav 1968;3:625–30. [22] Stevens R. Effet of amount of training on reversal learning in hippocampectomized rats. Physiol Psychol 1973;1:377 –9. [23] Sutherland RJ, Rudy JW. Configural association theory: the role of the hippocampal formation in learning, memory, and amnesia. Psychobiology 1989;17:129 – 44. [24] Wible CG, Shiber JR, Olton DS. Hippocampus, fimbria-fornix, amygdala, and memory: object discrimination. Behav Neurosci 1992;106:751 – 61. [25] Winocur G. Effects of interference on discrimination learning and recall by rats with hippocampal lesions. Physiol Behav 1979;22:339 – 45.
Research report
The effect of excitotoxic hippocampal lesions on simple and conditional discrimination learning in the rat T.K. Murray a,*, R.M. Ridley b a
b
Astra Arcus AB, S-151 85 So¨derta¨lje, Sweden MRC Comparati6e Cognition Team, Department of Experimental Psychology, Downing Street, Cambridge, CB2 3EB, UK Received 13 March 1998; received in revised form 2 June 1998; accepted 2 June 1998
Abstract The effect of excitotoxic lesions of the hippocampus on acquisition and reversal of simple and conditional tasks was investigated using a Y-maze. Hippocampal-lesioned rats were severely impaired on acquisition and reversal of a conditional visuo-spatial task (where different pairs of visually distinctive choice arms indicated whether a left or right arm choice was correct on that trial) and were unable to acquire a visuo-visual conditional discrimination (where the appearance of the start arm indicated which of the visually distinctive choice arms was correct irrespective of their left/right position). They were not impaired on acquisition or reversal of a simple spatial left/right discrimination task (where all arms had the same visual appearance) nor on acquisition of a visual discrimination (where the correct, visually distinctive, choice arm varied in its left/right position). Hippocampal-lesioned rats were, however, impaired on reversal of this visual discrimination task and on acquisition and reversal of another visual discrimination task in which the visually distinctive choice arms were less different from each other than in the first version of this task. The degree of impairment in the lesioned rats was related to task difficulty for the sham-operated rats and was not specific to tasks requiring spatial choices, visual discrimination or conditional responding. The impairment on conditional tasks was greater than the impairment on those non-conditional tasks which happened to be matched for task difficulty for the sham-operated rats, suggesting that the conditional demand may target the function of the hippocampus rather closely. Statistically worse than chance performance by hippocampal-lesioned (and sham-operated) rats at the beginning of reversal testing, which was given 24 h after achieving criterion on acquisition of that task, indicated that hippocampal-lesioned rats simultaneously exhibited good memory but impaired learning for the type of information required for those tasks. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Learning; Hippocampus; Rat; Reversal; Spatial discrimination; Conditional discrimination; Y-maze
1. Introduction Rats with hippocampal lesions have frequently been found to be impaired on spatial tasks including those presented in a water maze [19], a radial arm maze [4,12] or a Y-maze [6]. A general role for the hippocampus in spatial learning has been suggested [5,11] although * Corresponding author. Lilly Research Centre, Erlwood Manor, Windlesham, Surrey GU20 6PH, UK.
hippocampal-lesioned rats may not be impaired on very simple spatial tasks [10,22]. Hippocampal-lesioned rats have also been found to be impaired on a variety of conditional [6], configural [17] or contextual [20] tasks where the correct choice of a discriminative stimulus depends on information contained in the context, or in some other stimulus, proximity or circumstance present at the time of response choice. This has lead to a view of hippocampal function in the domain of configural or relational memory [3,23] although several studies have
0166-4328/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved. PII S0166-4328(98)00077-1
104
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
failed to find impairment by rats with hippocampal lesions on apparently configural tasks [1,2]. In this paper, we use a Y-maze to assess the effect of hippocampal lesions made using the excitotoxin, Nmethyl-D-aspartate (NMDA) on (i) simple left/right spatial discrimination tasks, where all arms had the same visual appearance; (ii) visual discrimination tasks, where the correct, visually distinctive, choice arm varied in its left/right position; (iii) a conditional visuo-spatial discrimination task, where different pairs of visually distinctive choice arms indicated whether a left or right arm choice was correct on that trial; and (iv) a conditional visuo-visual discrimination task, where the appearance of the start arm indicated which of the visually distinctive choice arms was correct irrespective of their left/right position. Comparison of performance on these tasks will indicate the extent to which the hippocampus is involved in spatial and/or conditional cognitive processing.
2. Methods
2.1. Animals Male Lister Hooded rats (Harlan UK, Bicester, UK), weighing 250–300 g at the start of the experiment, were used. They were housed in groups of five rats per cage and maintained on a 12-h light/12-h dark cycle (lights on 7:00–19:00 h). All testing was carried out during the light phase of the cycle. The rats were maintained at 90% of their free-feeding body weight and fed at the end of each test session. Water was available ad libitum.
2.2. Lesions Rats were anaesthetized with a gaseous anaesthesic consisting of Halothane® (Rhoˆne Me´rieux, Essex, UK) /nitrous oxide/oxygen (induction: 1.0 l/min N2O, 0.5 l/min O2, 4–8% Halothane; maintenance: 1.0 l/min N2O, 0.5 l/min O2, 1 – 1.5% Halothane, adjusted according to each rat’s respiration rate and reflex response). Lesions were made using 0.5 ml 0.12 M N-methyl-D-aspartate (NMDA) injected stereotaxically into three sites per hemisphere. NMDA was purchased from Sigma (Dorset, UK) dissolved in phosphate-buffered saline with the aid of 2 M sodium hydroxide and the pH titrated to 6.5–7.5 by addition of 2 M hydrochloric acid. Stereotaxic co-ordinates were AP: −3.14, L:9 1.6, V: − 2.8; AP: −4.16, L: 94.0, V: − 3.2; AP: − 4.8, L: 9 3.0, V: −2.6, from bregma according to the atlas of Paxinos and Watson [13]. Each injection was made over a period of 5 min, followed by a 5 min equilibration time, with the needle remaining in place. Shamoperated rats received identical surgery to the NMDA
group but without the injection of the neurotoxin. Diazepam (10 mg/kg, Roche Products, UK) was administered (i.p.) to all rats (including sham-operated) to prevent seizures following surgery. Seven days post surgery rats began training on the Y-maze tasks.
2.3. Apparatus and pre-training Rats were tested in a Y-maze, which was made of grey perspex and had three arms of equal size, 60 cm long, 11.5 cm wide and 25 cm high, positioned 31 cm above the floor. The start arm had a guillotine door behind which the rats were held until the start of each trial, which commenced when the guillotine door was raised. A food cup was located at the end of each choice arm. Rats were pre-trained for 4 days so they would become familiar with the maze. On day 1 of pre-training, pairs of rats were placed for 10 min in the start arm of the maze, with food pellets (45 mg, Noyes sucrose pellets, Sandown Scientific, UK) scattered in all arms. On day 2, rats were placed singly in the maze for 10 min; again, pellets were scattered in the arms. On days 3 and 4 the rats were placed in the maze singly for 5 min; on day 3, pellets were placed at the end of both choice arms, and on day 4, pellets were placed only in the food cups.
2.4. Experimental design On all tasks, correct responses were rewarded with four food pellets placed in a food cup at the end of the correct arm. If a rat made an incorrect response, it was allowed to go to the empty food well at the end of the incorrect arm and was removed after 5 s. This was recorded as an error. Maximum trial length was 3 min. If the rat did not make a choice within this time or did not eat all four reward pellets within the time allowed, it was removed from the maze. This was regarded as a non-trial and was not recorded as an error. Rats were run successively in groups of five such that the intertrial interval (ITI) for each rat was approximately 3 min. The rats were placed in their home cage between trials. Rats were trained for ten trials per day, to a criterion of nine correct responses in ten consecutive completed trials on all tasks. The behavioural studies described here were performed over many months and it was necessary to use three sets of rats for this purpose. Thus not all rats were tested on all behavioural tasks: Set 1 (sham n=8, lesion n= 9) were tested on the spatial and visual (version II) discrimination tasks; Set 2 (sham n=8, lesion n= 9) were tested on the visual (version I) discrimination and the visuo-spatial conditional tasks; Set 3 (sham n=9, lesion n=9) were tested on the visuo-visual conditional task only.
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
105
Table 1 Trials to criterion ( 9 S.E.M.) Task
Sham operated
Hippocampal lesioned
Unmatched t-test
Spatial discrimination Spatial reversal Visual discrimination version I Visual reversal version I Visual discrimination version II Visual reversal version II Visuospatial conditional Visuospatial reversal Visuovisual conditional Visuovisual reversal
29.1 97.3 34.4 94.5 42.3 95.1 54.6 99.2 50.1 910.1 80.7 911.1 53.4 98.2 62.6 96.2 79.5 91.4 91.2 917.1
35.7 9 5.0 34.3 9 4.7 51.9 98.2 78.1 910.1 83.3 9 12.4 156.2 9 20.7 122.4 917.4 145.0 98.5 300 fail Not done
n.s. n.s. n.s. PB0.05 PB0.05 PB0.001 PB0.005 PB0.001 Not applicable Not applicable
2.5. The spatial discrimination task Rats were trained on acquisition and then on reversal of a spatial discrimination task where all arms were grey. Half of the rats in each group were allocated left arm positive and the remaining rats were allocated right arm positive for acquisition. At 24 h after completion of acquisition, these conditions were reversed and the rats were trained on this reversal to criterion.
2.6. The 6isual discrimination task 6ersion I In this task, discriminative stimuli consisted of black or white, perspex, ‘whole arm’ inserts, which covered the walls and floor of the whole of the choice arms. The start arm was coloured grey. Each insert measured 59 cm long, 11 cm wide and 25 cm high. The left/right position of the black and white inserts varied according to a pseudorandom schedule. Half of the rats in each group were allocated black arm positive and the remaining rats were allocated white arm positive for acquisition. Twenty-four hours after completion of acquisition, these conditions were reversed and the rats were trained on this reversal to criterion.
2.7. The 6isual discrimination task 6ersion II Acquisition and reversal of another visual discrimination task was also performed as above but using black or white ‘floor’ inserts rather than the ‘whole arm’ inserts used for the visual discrimination task version I. It was supposed that this might make the task more difficult by providing less salient discriminative information.
2.8. The 6isuo-spatial conditional task This task used the same discriminative stimuli that were used for the visual discrimination task version 1, i.e. black or white ‘whole arm’ inserts. On each trial, two
identical inserts were used but the inserts were swapped between right and left arms, according to a pseudorandom schedule, to prevent the rats from learning to choose one of the inserts on the basis of distinguishing blemishes. Half of each group of rats was rewarded for choosing the left arm when both inserts were white, and the right arm when the inserts were black, while the remaining rats was tested according to the opposite reward contingency. Twenty-four hours after completion of acquisition, these conditions were reversed and the rats were trained on this reversal to criterion.
2.9. The 6isuo-6isual conditional task In the visuo-visual conditional task, ‘whole arm’ inserts made of black, white, striped or grey perspex were used and testing followed the same procedure as the visuo-spatial task. Each arm, including the start arm, contained a different insert and each rat was assigned a rule such as ‘if start arm grey, choose the black (not the white) choice arm; if start arm striped, choose the white (not the black) choice arm’. The start arm was grey or striped according to a pseudorandom schedule and the left/right position of the black/white choice arms was determined according to another, superimposed, pseudorandom schedule. Twenty-four hours after completion of acquisition, these conditions were reversed and those rats which had attained criterion on acquisition were trained on this reversal to criterion.
2.10. Histology On completion of behavioural testing, rats were perfused transcardially with saline followed by 10% formaldehyde in 5% sucrose. Brains were removed and placed in 5% formaldehyde in 30% sucrose at 4°C for 2 days. Coronal 20 mm (1 per 100 mm) sections were cut through the hippocampus using a cryostat and stained with Cresyl Violet.
106
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
Fig. 1. Learning curves for spatial discrimination learning in the Y-maze. =sham-operated rats (n =8); = hippocampal-lesioned rats (n= 9). Ordinate shows mean number of errors per block of five trials ( 9S.E.M.). Solid horizontal line: chance performance (2.5 errors/five trials); dotted horizontal lines: limits of chance at 5% level when all rats are performing. A. Acquisition. B. Reversal.
3. Results Trials to criterion ( 9 S.E. of the mean) for each group for each task are shown in Table 1. Groups were compared on each task using unmatched t-tests. Learning curves were plotted for each task (see Figs. 1–5). The limits of chance for each block of trials, when all animals in each group were performing the task, were calculated and are shown as dotted lines in the figures. Thus the mean chance number of errors performed in a block of five trials is 2.5, error scores in excess of 3.2 errors per block of five trials (calculated using the
binomial test for eight to nine animals per group) indicate that the group is performing at statistically worse than chance; error scores which are less than 1.8 errors per block of five trials indicate that the group is performing at statistically better than chance at that point in the learning curve.
3.1. The spatial discrimination task There was no difference in the number of trials to criterion on acquisition of the spatial discrimination task (sham-operated: 29.19 7.3 trials; hippocampal-le-
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
107
Fig. 2. Learning curves for the visual discrimination task version I in the Y-maze. = sham-operated rats (n = 8); =hippocampal-lesioned rats (n =9). Ordinate shows mean number of errors per block of five trials ( 9 S.E.M.). Solid horizontal line: chance performance (2.5 errors/five trials); dotted horizontal lines: limits of chance at 5% level when all rats are performing. A. Acquisition. B. Reversal.
sioned: 35.795.0 trials, see Fig. 1A). On the reversal of this task, both groups began by performing at just worse than chance and rapidly learnt the reversal at the same rate (Fig. 1B). The number of trials to criterion did not differ (sham-operated: 34.49 4.5 trials; hippocampal-lesioned: 34.394.7 trials).
3.2. The 6isual discrimination task 6ersion I There was no difference in the number of trials to criterion on acquisition of this visual discrimination
task (sham-operated: 42.39 5.1 trials; hippocampal-lesioned: 51.99 8.2 trials, see Fig. 2A). On the reversal of this task, both groups began by performing at worse than chance although hippocampal-lesioned rats learnt this reversal more slowly than sham-operated rats (sham-operated: 54.69 9.2 trials; hippocampal-lesioned: 78.19 10.1 trials, (unmatched t-test, t=2.31, PB 0.05, see Fig. 2B).
3.3. The 6isual discrimination task 6ersion II Hippocampal-lesioned rats required more trials than
108
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
Fig. 3. Learning curves for the visual discrimination task version II in the Y-maze. = sham-operated rats (n = 8); = hippocampal-lesioned rats (n =9). Ordinate shows mean number of errors per block of five trials ( 9S.E.M.). Solid horizontal line: chance performance (2.5 errors/five trials); dotted horizontal lines: limits of chance at 5% level when all rats are performing. A. Acquisition. B. Reversal.
sham-operated rats to reach criterion on acquisition of version II of the visual discrimination task (sham-operated: 50.1910.1 trials; hippocampallesioned: 83.3912.4 trials, unmatched t-test, t = 2.26, P B0.05, see Fig. 3A). On the reversal of this task, both groups began by performing at worse than chance and returned to chance performance at the same rate; hippocampal-lesioned rats then spent a substantial proportion of the total number of trials performing at chance before approaching criterion at an almost normal rate (sham-operated: 80.79 11.1 trials;
hippocampal-lesioned: 156.29 20.7 trials, unmatched t-test, t= 4.12, P B 0.001, Fig. 3B).
3.4. The 6isuo-spatial conditional task Hippocampal-lesioned rats were severely impaired on acquisition of the visuo-spatial task (sham-operated: 53.49 8.2 trials; hippocampal-lesioned: 122.49 17.4 trials; unmatched t-test, t= 3.76, PB0.005). Both groups started performance at chance but, while the sham-operated group showed a steady improvement over trials,
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
109
Fig. 4. Learning curves for visuo-spatial conditional discrimination learning in the Y-maze. = sham-operated rats (n =8); =hippocampal-lesioned rats (9). Ordinate shows mean number of errors per block of five trials ( 9 S.E.M.). Solid horizontal line: chance performance (2.5 errors/five trials); dotted horizontal lines: limits of chance at 5% level when all rats are performing. 4. Acquisition. 4. Reversal.
reaching criterion after 8 days of training, the hippocampal-lesioned group continued to perform at chance for 15 days before better than chance performance was consistently achieved (Fig. 4A). Hippocampal-lesioned rats were also severely impaired on reversal learning of the visuo-spatial task (sham-lesioned: 62.69 6.2 trials; hippocampal-lesioned: 145.098.5 trials; unmatched t-test, t = 4.09, PB 0.001). Fig. 4B shows that both sham-operated and hippocampal-lesioned rats started reversal learning at worse than chance performance. Sham-lesioned rats then learnt at a steady, consistent rate whereas the
hippocampal-lesioned rats learnt at a consistent, but slower, rate.
3.5. The 6isuo-6isual conditional task Only one of the hippocampal-lesioned rats was able to acquire the visuo-visual task within 350 trials whereas the sham-lesioned group learnt this task in 79.59 14.0 trials. Fig. 5 shows that the hippocampal-lesioned group were largely unable to improve above chance performance. Reversal testing on this task could not be done using the hippocampal-lesioned rats since
110
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
Fig. 5. Learning curves for visuo-visual conditional discrimination learning in the Y-maze. = sham-operated rats (n =9); = hippocampal-lesioned rats (n=9). Ordinate shows mean number of errors per block of five trials ( 9 S.E.M.). Solid horizontal line: chance performance (2.5 errors/five trials); dotted horizontal lines: limits of chance at 5% level when all rats are performing. Data are for acquisition; reversal could not be undertaken because hippocampal-lesioned rats were unable to learn the task.
they had not completed acquisition. Sham-lesioned rats learnt this reversal in 91.2917.1 trials.
3.6. Histology The extent of the hippocampal lesion is illustrated in Fig. 6 and represents the minimum and maximum size of the lesion. All animals received bilateral damage to the hippocampus, primarily to the CA1 region. In the dorsal hippocampus, complete pyramidal cell loss in the CA1 region was seen with some damage in the CA2 and CA3 regions occurring in approximately half of the animals. In the ventral hippocampus, part of the CA1 field was spared and only minimal damage to the CA3 region was seen in a small number of rats. Damage occurred in the dentate gyrus in approximately one third of the animals. Very little damage was observed outside the hippocampal formation, occurring only in the region of the needle tracts in some animals. There were no systematic differences between the lesions in the three groups of lesioned rats.
4. Discussion These experiments demonstrate that hippocampal lesions produce a severe learning impairment on conditional discrimination tasks and an impairment on some forms of visual discrimination learning. The circumscribed nature of the impairment is illuminated by the equivalent worse than chance performance at the beginning of the reversal tasks shown by both
hippocampal-lesioned and sham-operated rats. This indicates that both groups possessed equivalent retention of information acquired 24 h previously and were equally competent to make non-random choices based on that retention. This excludes the possibility that the impairment resulted from motoric, perceptual or motivational difficulties in the lesioned rats. Furthermore, this worse than chance performance excludes the possibility that impairment could result from motoric or spatial perseveration; persistent choice of one arm on either of the conditional tasks would have resulted in chance (rather than worse than chance) performance since the rewarded arm in these tasks was random with respect to left/right position. Impairment on reversal learning could be described as ‘cognitive perseveration’ or ‘interference’ if the rats were continuing to make cognitive choices based on previously acquired information even when currently available information was suggesting that a new choice would be appropriate. This explanation is unlikely, however, because of the similarity in the magnitude of the impairment on visual discrimination version I reversal (where interference would be expected) compared to visual discrimination version II acquisition (where interference would not be expected) and where task difficulty for the sham-operated rats was well matched. For visual discrimination version II reversal, the performance of both groups was very similar at the beginning of reversal (where interference would be expected to be greatest); the impairment in the lesioned rats consisted mainly of chance performance in the middle portion of testing. Similarly, for reversal of the visuo-spatial task,
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
111
Fig. 6. Schematic representation of coronal sections from rat brain (based on Paxinos and Watson [13]) illustrating the histology results. The solid area illustrates the damage in the rat with the smallest lesion; the solid + hatched areas illustrate the damage in the rat with the largest lesion.
the hippocampal-lesioned rats returned to chance performance only slightly more slowly than the sham-operated rats; the greater part of the hippocampal-lesioned rats’ impairment accrued from errors in the middle and last third of the learning curve. Lastly, the impairment on reversal was highly comparable to the impairment on acquisition of the visuo-spatial task, where these tasks were well matched for difficulty in the sham-operated rats. All these comparisons suggest that the impairment exhibited by the hippocampal-lesioned rats did not consist of excessive sensitivity to interference. The difference between visual discrimination tasks version I and version II which led to the difference in the effect of the hippocampal lesion on these tasks is not clear. It might have been expected that version II would have been more difficult than version I because, in the former, the visually distinctive part of the choice arms was confined to just the floor rather than the floor and walls, but, in fact, the learning scores for version I and version II did not differ significantly for the shamoperated rats. Previous studies have found that
hippocampal lesions usually do not impair acquisition of simple visual discriminations [18,21,24] although impairment has been demonstrated on acquisition under certain circumstances [25] and is frequently seen on reversal of visual discrimination tasks [18,21]. One possibility is that unlesioned rats are able to direct their attention to the relevant part of the choice arm without needing to learn to do so (such that the two versions of the visual discrimination task are of equal difficulty) but that hippocampal-lesioned rats do not do this automatically and therefore have to learn to do so (thereby producing long periods of chance performance). Examination of the learning curves (Figs. 2 and 3) indicates that there was no difference in performance between the two groups on acquisition of version I, whereas on acquisition of version II the hippocampal-lesioned rats performed at chance for 8 days (the total time taken by the sham-operated rats to reach criterion) before approaching criterion. On reversal of version I, the hippocampal-lesioned rats showed a learning rate which was slower than for the sham-
112
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113
operated rats but which was consistent across the whole learning curve, whereas on reversal of version II, the hippocampal-lesioned rats ‘unlearned’ the original discrimination at the same rate as the sham-operated rats but then performed at chance for a considerable period of time before approaching criterion. This could occur if, on reversal of version II, the hippocampal-lesioned rats not only ‘unlearned’ the association of one floor colour with reward, but also ‘unlearned’ the relevance of the floor rather than the wall colour to solving the task. Hippocampal-lesioned rats would then have to relearn to orient to this aspect of the choice arms (producing the persistent chance performance in the middle portion of the learning curve) before reversal learning could be achieved. In contrast, on reversal of version I, learning would occur at a steady rate, since attention to any part of the choice arm would provide the necessary information to learn the task. The extent to which the aspect of cognitive dysfunction in hippocampal lesioned rats which is demonstrated by the visual discrimination task version II is also responsible for the impairment on conditional tasks is not clear. Hippocampal-lesioned rats were unimpaired at learning a non-conditional spatial response equivalent to the spatial response required for the visuo-spatial conditional task, and were also unimpaired at learning the non-conditional visual discrimination task version I, which used the same discriminative stimuli in the choice arms as the visuo-visual conditional task. This indicates that the lesioned rats’ difficulty does not lie in some motoric or perceptual impairment. Hippocampal-lesioned rats were generally more impaired on those tasks which the sham-operated rats found most difficult (see Table 1), suggesting that the hippocampus is recruited to solve certain types of difficult task. The results cannot, however, be explained entirely in terms of non-specific task difficulty because, when pairs of tasks were compared which were well matched for task difficulty in the sham-operated rats, hippocampal-lesioned rats were more impaired on conditional than non-conditional tasks. Thus lesioned rats were more impaired on visuo-visual conditional acquisition than on visual discrimination version II reversal, and on visuo-spatial conditional acquisition than on visual discrimination version I reversal, even though learning scores for the relevant sham-operated groups were almost identical for the two tasks in each comparison. In a formally equivalent test situation using the Wisconsin General Test Apparatus, monkeys with dysfunction of the hippocampus induced by excitotoxic lesion, fornix transection, or neurotoxic lesion of the cholinergic input to the hippocampus, have been found to be specifically impaired on acquisition of the visuospatial and visuo-visual conditional tasks [15]. The
hippocampal impairment is critically dependent on the order of presentation of the two types of trial in conditional tasks [14] suggesting that the hippocampus is involved in the cognitive processing of conditionality rather in learning mechanisms per se. We have previously demonstrated that the NMDAreceptor antagonist, dizocilpine, also produces a severe impairment of conditional learning in rats [8] and a milder effect on simple discrimination tasks [9] which resemble the effects of hippocampal-lesions. Robinson et al. [16] have also found similarities between the effects of dizocilpine treatment and hippocampal lesions on learning in rats. Since the hippocampus contains the highest levels of NMDA receptors in the brain [7] it would seem likely that it is glutamatergic activity (NMDA is a receptor for glutamate) in the hippocampus (rather than glutamatergic activity elsewhere) which is necessary for acquisition of tasks requiring conditional analysis. In summary, we have demonstrated that hippocampal lesions produce an impairment on conditional learning in rats, and on some aspects of visual discrimination learning, but have no effect on simple spatial learning. These results are largely in agreement with the view that the hippocampus is important for learning tasks which require a flexible response to stimuli depending on context or relations with other stimuli [3], although this may be only one aspect of hippocampal function.
References [1] Davidson TL, McKernan MG, Jarrard LE. Hippocampal lesions do not impair negative patterning: a challenge to configural association learning. Behav Neurosci 1993;107:227 – 34. [2] Deacon RMJ, Rawlins JNP. Effects of aspiration lesions of hippocampus or overlying neocortex on concurrent and configural object discrimination in rats. Behav Brain Res 1996;77:165 – 74. [3] Eichenbaum H, Otto T, Cohen NJ. Two functional components of the hippocampal memory system. Behav Brain Sci 1994;17:449 – 518. [4] Jarrard LE. Selective hippocampal lesions and behavior: implications for current research and theorizing. In: Isaacson RL, Pribram KH, editors. The Hippocampus, 4. New York: Plenum Press, 1986. [5] Jarrard LE. On the role of the hippocampus in learning and memory in the rat. Behav Neural Biol 1993;60:9 – 26. [6] Kimble DP. The effect of bilateral hippocampal lesions in rats. J Comp Physiol Psychol 1963;56:273 – 83. [7] Monaghan DT, Monaghan CW. Distribution of NMDA-sensitive, L-3[H]glutamate binding sites in rat brain as determined by quantitiative antoradiography. J Neurosci 1985;5:2909 –19. [8] Murray TK, Ridley RM. The effect of dizocilpine (MK801) on conditional discrimination learning in the rat. Behav Pharmacol 1997;8:383 – 8. [9] Murray TK, Ridley RM, Snape MF, Cross AJ. The effect of dizocilpine (MK-801) on spatial and visual discrimination tasks in the rat. Behav Pharmacol 1995;6:540 – 9.
T.K. Murray, R.M. Ridley / Beha6ioural Brain Research 99 (1999) 103–113 [10] Oliveira MGM, Bueno OFA, Pomarico AC, Gugliano EB. Strategies used by hippocampal- and caudate-putamen-lesioned rats in a learning task. Neurobiol Learn Mem 1997;68:32 – 41. [11] Olton DS, Becker JT, Handelmann GE. Hippocampus, space and memory. Behav Brain Sci 1979;2:313–22. [12] Olton DS, Papas BC. Spatial memory and hippocampal function. Neuropsychologia 1979;17:669–82. [13] Paxinos G, Watson C. The rat brain. Stereotaxic Coordinates. Australia: Academic Press, 1986. [14] Ridley RM, Baker HF. Evidence for a specific information processing deficit in monkeys with lesions of the septohippocampal system. Cortex 1997;33:167–76. [15] Ridley RM, Baker HF, Harder JA, Pearson C. Effects of lesions of different parts of the septo-hippocampal system in primates on learning and retention of information acquired before or after surgery. Brain Res Bull 1996;40:21–32. [16] Robinson GS, Crooks GB, Shinkman PG, Gallagher M. Behavioural effects of MK-801 mimic deficits associated with hippocampal damage. Psychobiology 1989;17:156–64. [17] Rudy JW, Sutherland RJ. The hippocampal formation is necessary for rats to learn and remember configural discriminations. Behav Brain Res 1989;34:97–109.
.
113
[18] Samuels I. Hippocampal lesions in the rat: effects on spatial and visual habits. Physiol Behav 1972;8:1093 – 8. [19] Schenk F, Morris RGM. Dissociation between components of spatial memory in rats after recovery from the effects of retrohippocampal lesions. Exp Brain Res 1985;58:11 – 28. [20] Seldon NRW, Everitt BJ, Jarrard LE, Robbins TW. Complementary roles for the amygdala and hippocampus in aversive conditioning to explicit and contextual cues. Neuroscience 1991;42:335 – 50. [21] Silveira JM, Kimble DP. Brightness discrimination and reversal in hippocampally-lesioned rats. Physiol Behav 1968;3:625–30. [22] Stevens R. Effet of amount of training on reversal learning in hippocampectomized rats. Physiol Psychol 1973;1:377 –9. [23] Sutherland RJ, Rudy JW. Configural association theory: the role of the hippocampal formation in learning, memory, and amnesia. Psychobiology 1989;17:129 – 44. [24] Wible CG, Shiber JR, Olton DS. Hippocampus, fimbria-fornix, amygdala, and memory: object discrimination. Behav Neurosci 1992;106:751 – 61. [25] Winocur G. Effects of interference on discrimination learning and recall by rats with hippocampal lesions. Physiol Behav 1979;22:339 – 45.