- Email: [email protected]
Ontogenetic dissociation between habit learning and recognition memory in capuchin monkeys (Cebus apella)
Neurobiology of Learning and Memory 79 (2003) 19–24 www.elsevier.com/locate/ynlme
Ontogenetic dissociation between habit learning and recognition memory in capuchin monkeys (Cebus apella) Michelline C. Resende, Maria Clotilde H. Tavares, and Carlos Tomaz* Primate Center and Department of Physiological Sciences, University of Brasılia, CEP 70910-900 Brasılia, DF, Brazil
Abstract The performance of young and adult capuchin monkeys (Cebus apella) on a Concurrent Discrimination Learning (CDL) test and a Delayed Non-Matching to Sample (DNMS) task were investigated. Results indicate that all subjects were able to learn the CDL test with 20-pairs simultaneously and retain this stimulus/reward association within 24-h interval. In contrast, young subjects did not perform the DNMS task with the same proficiency as adults. While adultsÕ scores were above chance across all memory test delays, the young capuchin monkeys performed the test by chance level. These results support the hypothesis that these two tasks require different cognitive processes mediated by two independent neural systems with a differentiated ontogenetic development. Moreover, they provide evidence that this dissociation occurs not only in humans and Old World monkeys but also in the New World capuchin monkeys indicating that this species can be a valuable alternative model for investigations of the neurobiological basis of memory. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: Habit learning; Concurrent discrimination; Recognition memory; Delayed non-matching to sample; Ontogenetic dissociation; Cebus apella
1. Introduction Several works have proposed the existence of two ontogenetic dissociable memory systems, each served by different neural circuits (e.g., Bachevalier & Mishkin, 1984; Bachevalier, Brickson, Hagger, & Mishkin, 1990; Diamond, 1990; Malamut, Saunders, & Mishkin, 1984; Overman, 1990). The cognitive memory system is thought to use a corticolimbic circuit to store sensory representations, serving both recognition and recall. On the other hand, the second system known as the habit system seems to employ a corticostriatal circuit to mediate stimulus–response retention, skill and habit learning. Moreover, it has been proposed that these two systems have a differentiated ontogenetic development (Mishkin, Malamut, & Bachevalier, 1984). Whereas the corticostriatal system is believed to be functionally mature at the birth, the corticolimbic circuit seems to develop later in life (Alvarado & Bachevalier, 2000; Bachevalier & Mishkin, 1984).
* Corresponding author. Fax: +55-61-274-1251. E-mail address: [email protected] (C. Tomaz).
Recently, there has been an expressive increase of investigations attempting to broach the development and functions of these two neural systems. In order to assess the cognitive system, researches have used the delayed non-matching to sample test (DNMS). In this task, the subject should discriminate between one object that has been seen before and one new. The subject is rewarded if chosen the new stimulus. The recognition memory can be assessed by lengthening the intervals between sample and the choice trial or the list of samples presented. The performance in this task is impaired in amnesic subjects with diencephalic lesions or in monkeys with damage in the medial temporal lobe (MTL) (Bachevalier & Mishkin, 1984; Malamut et al., 1984; Mishkin et al., 1984; Zola et al., 2000). In contrast, the non-cognitive memory system has been studied using a habit learning tasks. In monkeys, the concurrent discrimination learning task (CDL) is believed to be suitable for evaluating this type of learning, as subjects solve the test progressively over many trials (Bachevalier & Mishkin, 1984; Bachevalier et al., 1990; Malamut et al., 1984). This test involves a list of several pairs of objects presented concurrently. In order to solve the task, subjects are required to learn
1074-7427/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. PII: S 1 0 7 4 - 7 4 2 7 ( 0 2 ) 0 0 0 1 5 - 1
20
M.C. Resende et al. / Neurobiology of Learning and Memory 79 (2003) 19–24
that only one member in each pair of stimuli is rewarded. The ability to solve CDL tasks is preserved in MTL lesion subjects (Bachevalier & Mishkin, 1984; Gaffan & Murray, 1992; Malamut et al., 1984). There are evidences that the ontogenetic dissociation of these systems occurs not only in humans, but also in Old World rhesus monkeys. Bachevalier and Mishkin (1984) have pointed differences between young and adult monkeys performing CDL and DNMS tasks. As early as 3-month-old, rhesus monkeys were able to perform CDL task with 24-h intertrial interval as quickly as adult were. By contrast, they failed to master DNMS task until 4 month of age. The subjects reached the same proficiency level as adults only at about 2 years of age. The same pattern of proactive development in these two memory systems was found in human infants (Diamond, 1990; Overman, 1990; Overman, Bachevalier, Turner, & Peuster, 1992). For instance, Overman (1990) tested children in DNMS task using a rate of 15 trails/ day to a criterion of 87% correct response for two consecutive test days. He found that the youngest group of children (12–15 months old) only attained the criterion after 20 weeks of continuous training and it had the lowest rate of learning compared with the other four studied groups. By the other side, the performance of the two oldest groups (18–20-month-old and 22–32-monthold) was above chance since the beginning of the test. However, both young and old groups were able to solve a visual discrimination task. The 12-month-old children need 8.7 days to reach the learning criterion while 18-month-old attained the criterion in 6 days (Overman, 1990). In sum, the studies found that the ability to perform on DNMS problems improves systematically with age in both children and infant monkeys indicating that a further maturation of the neural circuitry is required to solve this task (Alvarado & Bachevalier, 2000; Bachevalier, 1990; Diamond, 1990; Overman et al., 1992). Nevertheless, CDL task can be solved as early as 12month-old by children and 3-month-old by rhesus monkeys (Bachevalier & Mishkin, 1984; Overman, 1990). Neurobehavioral studies have showed that early or late damage to limbic structures (hippocampus plus amygdala) produces severe impairment on DNMS test performed by monkeys (Malamut & Mishkin, 1981; Mishkin et al., 1984). However, recent findings demonstrated that the poor performance on visual recognition test may be due to unintentional damage to surrounded cortical areas. Recognition memory is affected after selective lesions to the perirhinal and entorhinal cortex (Gaffan & Murray, 1992) and hippocampus and parahippocampal area (Zola-Morgan, Squire, & Amaral, 1989). Additionally, selective damage to the hippocampal formation (Alvarez, Zola-Morgan, & Squire, 1995) did not abolish preference for the novelty in the DNMS test.
Bachevalier (1990) showed that late TE damage resulted in a drop of the adult monkeysÕ performance when they were tested on DNMS task. In contrast, when the injury occurred early in infancy, the visual recognition ability was spared. Taken together, these results indicate that the protracted development of the cognitive memory system in monkeys is due, in part, to a slow maturation of area TE instead of the limbic structures as amygdala and hippocampus per se (Alvarado & Bachevalier, 2000; Bachevalier, 1990). Although several studies have examined aspects of animal cognitive processes through CDL and DNMS tasks, especially employing Old WorldÕs monkeys (Bachevalier, 1990; Bachevalier & Mishkin, 1984; Bachevalier et al., 1990; Malamut et al., 1984), New World primate data on cognitive processes are presently scarce. The capuchin monkey (Cebus apella) has been the focus of several researches due to its behavioral similarities with apes (Antinucci, 1990). Moreover, capuchins are shown to display tool-using capacities (Visalberghi, 1993) and to readily solve DNMS memory tasks with sample/test trials delays up to 10 min (Tavares & Tomaz, 2002). Together, these features make capuchin monkeys a promising animal model for studying different aspects of memory and cognition. To our knowledge, no attempts have been made in the ambit of investigating the existence of two ontogenetic dissociable (cognitive and habit) memory systems in capuchin monkeys. Therefore, the aim of this study was to investigate the capabilities of young and adult capuchin monkeys in performing a concurrent object discrimination learning task with 24-h intertrial interval and a DNMS test.
2. Methods 2.1. Procedures The study was performed in the Primate Center— University of Brasılia. Animals were fed twice daily with a variety of foods and water was available ad libitum. Subjects were tested in their own home cages to avoid disruption of behavior by capture and transport to a novel environment. The animals were not food or water deprived, however during the experimental session food was removed from the cage. The experiment was carried out in a version of the Wisconsin General Test Apparatus (WGTA) containing a mobile test tray with two food wells of 5 cm in diameter, 12 cm apart. The experimental procedures and housing conditions were in accordance with the principles and regulation of the Ethics Committee of the University of Brasılia for the care and the use of animals for scientific purposes.
M.C. Resende et al. / Neurobiology of Learning and Memory 79 (2003) 19–24
2.2. Concurrent object discrimination learning test (CDL) Ten capuchin monkeys (C. apella), separated in two groups, served as subjects. One group was compost of five young monkeys (two females and three males) with age estimates to range from 1 to 312 years. The other group had five adult monkeys (three females and two males) with estimated age between 6 and 10 years. One young and one adult subject had some experience in visual discrimination tasks. The stimuli were ‘‘junk’’ objects that differed in shape, size, color, and texture. Pieces of grape were used as reward. CDL testing were as following. A set of 20 different pairs of objects, Set A, was presented to the subjects. In each pair only one object was baited with a food reward. Each object of the pair was present simultaneously over the lateral wells of the test tray. After the monkey had made its choice by displacing one of the objects, the second pair of stimuli was presented and so on until all 20 pairs had been displaced. The same list of objects was then repeated after a 24-h intertrial interval (24-h ITI). The objects of the pairs and the serial order of the pairs remained constant across the sessions. However, the left-right position of the baited object varied pseudorandomly in order to avoid position bias. The test was performed every day (7 days a week) until the subject attained the learning criterion setting at 90% (i.e., 18 correct response in 20 trials). Nine days after completing Set A, the monkeys were tested in the same way on a new set of objects, Set B. 2.3. Delayed non-matching to sample Eleven C. apella monkeys divided in two groups were used in this experiment. The young group consisted of the same five subjects used in the CDL experiment. The adult group was composed by six subjects (three female and three male) with estimated age rages from 4 to 8 years. These adult subjects had experience in Delayed Matching to Sample (DMS) task. However, previous study showed that the performance in DNMS test by subjects that had experience in DMS task did not differ from na€ıve animals (Tavares & Tomaz, 2002). The subjects were tested in a trial-unique version of DNMS task. The test begun showing to the subject a single object placed in the central food well (sample phase). A wooden screen was then raised, and the subject was able to see the object (sample trial). After an observing response was made towards the object, the screen was lowered to prevent the subjectÕs view of the food wells. Eight seconds later, the screen was raised once again and the animal was faced with two objects, one was the same presented before and the other a new one (test trial). The subject had to displace the new object to obtain the reward. The non-correction method
21
was used, i.e., if the subject displaced the familiar object, the test tray was moved back and the screen lowered. Fifteen to twenty seconds later (intertrial interval) a new trial begun. The left or right position of the baited object varied pseudorandomly. New pairs of object were used on every trial. The experimental session lasted about one hour and testing occurred three days a week. Training at 8 s delay condition continued until the animal reached the learning criterion of nine correct responses out of 10 consecutive trials. After reaching the criterion, subjectsÕ memory performance was assessed at delays of 15, 60, 120 (100 trials to each delay), and 600 s (50 trials). 2.4. Statistical analysis Statistical analyses were carried out using Wilcoxon Signed Rank test for intra-group comparisons and Mann–Whitney U test for inter-group comparisons in the 24-h ITI task. Based on previous studies (Bachevalier & Mishkin, 1984; Bachevalier et al., 1990) one-tailed probabilities were employed as a significant difference between Sets A and B were expected for young and adult subjects. In order to analyze the performance of the subjects in the DNMS task, a binomial test was used to establish a 95% confidence limits around chance performance based on the number of test trials. Thus, the upper limit was calculated as 60% for 100 trials and 64% for 50 trials (p < :05). The mean percentage of correct responses was compared to these confidence limits and any performance above the upper limit was considered significant. To determine whether the subjectsÕ performance differ across delays we used a one-way ANOVA. To analyze the number of trials to criterion that group required a paired t test was used.
3. Results Table 1 shows the number of daily sessions, total number of errors and trials required for each subject to learn Set A and Set B in the CDL test. The statistical analysis showed that young group required significantly fewer sessions and trials, and took less errors on Set B than Set A to attained the learning criterion (Wilcoxon; Z ¼ 2:060 for session; Z ¼ 2:023 for trials; Z ¼ 2:023 for errors, all ps < :05). However, there were not significant decrease by adult subjects in the number of sessions, trials and errors from Set A to Set B (Wilcoxon; Z ¼ 1:483 p ¼ :094 for session; Z ¼ 1:214 p ¼ :156 for trials; Z ¼ 1:483 p ¼ :094 for errors). Comparison between young and adult performances on Set A indicated no significant differences (Mann– Whitney test; U ¼ 10:5, p ¼ :369; U ¼ 11, p ¼ :421; U ¼ 10, p ¼ :325, for sessions, errors and trials, respectively). The analysis also showed no differences
22
M.C. Resende et al. / Neurobiology of Learning and Memory 79 (2003) 19–24
Table 1 Number of daily sessions (S), total number of errors (E), trials (T), and mean (M SEM) preceding criterion on the Concurrent Object Discrimination Learning test by young and adult capuchin monkeys Group
Young 1 2 3 4 5 M Adult 1 2 3 4 5 M *
Set A
Set B
S
E
T
S
E
20 21 25 24 23
133 156 192 200 164
460 449 460 418 400
9 10 7 13 13
22:6 0:93
169 12:20
437:4 12:11
10.4 1.17
71.6 7.85
207.8 23.21
5 36 15 24 22
39 284 110 239 148
100 720 295 457 440
12 23 13 16 11
68 172 82 108 74
240 441 260 320 220
20:4 5:12
164 44:06
402:4 102:04
15 2:17
100:5 19:06
296:2 39:88
59 62 56 87 94
T 140 259 260 200 180
Wilcoxon Signed Rank test, p < :05.
Table 2 Trials to criterion and percentage of correct response across delays (15, 60, 120, and 600 s) on DNMS test by young and adult capuchin monkeys Group
Young 1 2 3 4 5 M Adult 1 2 3 4 5 6 M
Trials to criterion
Delays (s) 15
60
120
600
265 82 224 173 295
50 61 56 51 62
50 59 55 52 72
50 70 54 59 55
58 46 70 62 60
207:8 41:56
56 2:47
57:6 3:90
57:6 3:41
59:2 3:88
71 184 357 366 54 10
53 63 78 85 61 81
60 54 56 83 67 71
67 46 48 72 77 73
78 65 52 70 72 80
173:67 28:94
70:2 5:25
65:2 4:44
63:8 5:48
69:5 4:14
M ¼ mean SEM.
among number of sessions, errors and trials took by young and adult subjects on Set B (Mann–Whitney test; U ¼ 5, p ¼ :075 for sessions; U ¼ 6, p ¼ :111 for errors; U ¼ 4:5, p ¼ :056 for trials). Table 2 summarizes the results obtained for DNMS experiment. All subjects reached criterion in the 8 s delay condition. The young needed 207:8 41:56 trials to attain the learning criterion, while the adult group needed 173:67 28:94 trials. Paired comparisons indicated no significant difference in the number of trials required by young and adult subjects to reach the learning criterion (t ¼ :436, p ¼ :673). An analysis of variance involving the 4 delays (15, 60, 120, and 600 s) revealed no difference across delays for
young (F ¼ :142, p ¼ :933) and adult groups (F ¼ :417, p ¼ :743, for adults). Memory tests indicated that adults, but not young subjects, perform above the upper confidence limits around chance performance (above 60% for 15, 60, and 120 s; above 64% for 600 s) in all sample-choice test intervals.
4. Discussion The results of the present study shows that young capuchin monkeys are able to learn a concurrent discrimination learning task with 20-pairs of stimuli
M.C. Resende et al. / Neurobiology of Learning and Memory 79 (2003) 19–24
simultaneously, and retain this stimulus/reward association based on single trials within 24-h interval. Moreover, the significant reduction in the number of sessions, errors and trials observed from Set A to Set B in young capuchin monkeys indicate learning set formation, which is considered a characteristic of habit memory system that subserves problem solve skills. These results are similar to that observed in a previous study using rhesus monkeys (Bachevalier, Hagger, & Bercu, 1989). While our young capuchin monkeys took in average 22 sessions on Set A and 13 on Set B, 3-month-old rhesus monkeys required an average of 18 sessions to attained the criterion on Set A and 6 sessions at the Set B (Bachevalier et al., 1989). The adult capuchin monkeys needed an average of 20 sessions to attain the criterion on Set A and 15 sessions on Set B. In two other studies adult rhesus monkeys needed 10 sessions to reach the learning criterion in both Sets (Bachevalier et al., 1989; Malamut et al., 1984), also showing no significant improvement from Set A to Set B. Although the young capuchins learned Set B quicker than Set A and adults did not, it was not observed differences between young and adultsÕ scores in both sets. Bachevalier and Mishkin (1984) found only a initial delay in learning of infant group compared with adults on Set A, but they did not find differences in infantÕ and adultÕ scores on Set B either. One possible explanation suggested by Bachevalier and co-workers (Bachevalier & Mishkin, 1984; Bachevalier et al., 1989; Bachevalier et al., 1990) is that, although the habit and the cognitive systems are present in adult subjects, they could be promoting a competition in the resolution of the problems. In young subjects only the habit memory is operating, which could enable them to perform Set B quicker than adults do. The present study also shows that young capuchin monkeys did not perform the DNMS task with the same proficiency as adult. While the adultÕ scores were above of the upper limit of confidence across all tested delays, the young subjects performed the DNMS memory test by chance level. It is important to mention that the young group was first trained on CDL task what could potentially have a negative transfer effect on performance of DNMS; i.e., the difference between the two groupsÕ DNMS performance could have resulted from their different training histories rather than their different ages. However, data obtained for humans showed the same developmental progression of the performance on DNMS task. Overman et al. (1992) found that the youngest group (12–15-month-old) only attained the learning criterion with 19 month of age, while 26month-old group reached the criterion in an average of 11 days after the beginning of the test. Another similarity between our results and that found by other researches is the steady percentage of
23
correct responses of humans and rhesus monkeys at DNMS task even at the longest delay (10 min) (Bachevalier & Mishkin, 1984; Meunier, Bachevalier, & Mishkin, 1997; Overman et al., 1992). On the other hand, Zola and colleagues (2000) found a drop in the performance of the control group at 10 and 40 min delays. However, differently from our experimental procedures, during the long delays the animals were removed from the test apparatus and placed back in their home cage. Therefore, it is possible that these methodological aspects may contribute for the decrease in performance observed in the Zola et al.Õs study (2000). Some differences were observed in DNMS task performed by humans, C. apella and rhesus monkeys. Young capuchin monkeys needed fewer trials than children and infant rhesus monkeys to attained the learning criterion at 8 s delay. While our young monkeys took in average 208 trials, 12–15-month-old children only attained the criterion in 1845 trials and monkeys with 3 months of age needed 720 trials. However, when we compared the adult groups of these species, humans were faster than monkeys. Capuchins and rhesus monkeys needed a similar number of trials (174 and 160 trials, respectively) to reach the criterion, while humans required only 90 trials to criterion (Bachevalier & Mishkin, 1984; Overman et al., 1992). Since it was found that human infants and infant monkeys have a good performance in the CDL task, their poor scores on DNMS test cannot be attributed to deficits in attention, motivation, stimulus-reward association or perceptual abilities (Bachevalier, 1990; Overman, 1990). The more acceptable explanation for differences in performing CDL and DNMS by young subjects is that these two tasks require different cognitive processes and that these processes are mediated by two independent neural systems that mature at different times (Alvarado & Bachevalier, 2000; Bachevalier, 1990; Bachevalier & Mishkin, 1984; Bachevalier et al., 1990). The CDL task requires that the subject retain a representation of stimuli, compare their features and associate objectÕs quality with reward. However, in trial unique non-matching to sample task, stimuli vary from trial to trial, such that no single stimulus is associated with the reinforcement. Success in this task has been interpreted as ÔconceptÕ learning, since the rule that allows subjects to make a correct choice regardless of the physical properties of the various stimuli presented in each problem, is the concept of novelty. Thus, DNMS becomes a more demand cognitive task than object discrimination task. According, it has been proposed that the performance in DNMS tasks requires the participation of high-order cerebral areas that seem do not be functionally mature early in the life (Bachevalier, 1990; Mishkin et al., 1984; Overman, 1990). Although no lesion data are available for Cebus genus, the outcomes from our behavioral study are similar
24
M.C. Resende et al. / Neurobiology of Learning and Memory 79 (2003) 19–24
to those obtained with the genus Macaca. It is known that the CebusÕ cerebral cortex citoarchitecture resembles that of Macaca (Falk, 1981). Therefore, it seems reasonable to speculate that in capuchin monkeys, CDL and DNMS tasks are mediated by similar neuroanatomical substrates as those found for macaques. Taken together, these results suggest that humans, New World capuchin monkeys and Old World rhesus monkeys may share a common neural substrate underlying the two memory systems and that the recognition memory system has a protracted development in all of these species. However, more systematic comparative researches using macaques and C. apella are necessary to explore this issue. To our knowledge, this study is the first report examining the possibility of ontogenetic dissociation of habit learning and declarative (explicit) memory systems in the New World capuchin monkey. The results indicate the existence of two distinct memory systems with different ontogenetic development: (1) a non-cognitive learning system based on automatic associations between stimulus and response, and (2) a cognitive system based on abstract rule learning and its correlation with the reward. The first been present in young C. apella, at least, as early as one year old, and the later seems to be not completed functionally mature until the age of 312 years. In addition, the results of the present study, together with those reporting other cognitive abilities in capuchin monkeys, suggest that this specie can be a useful model in investigations of neurobiological basis of memory. Further studies using different drug/neurotransmitters manipulations as well as brain lesion approaches would be desirable to explore the potential of this animal model.
Acknowledgments This research was partially supported by FINATEC grants to C.T. and M.C.H.T. M.C.R. is the recipient of a Ph.D. fellowship from CAPES (Brazil). We are thankful to Dr. R. de Oliveira for animal care and M. Barros for helping in statistical analysis.
References Alvarado, M. C., & Bachevalier, J. (2000). Revisiting the maturation of medial temporal lobe memory functions in primates. Learning & Memory, 7, 244–256. Alvarez, P., Zola-Morgan, S., & Squire, L. R. (1995). Damage limited to the hippocampla region produces long-lasting memory impairment in monkeys. Journal of Neuroscience, 15, 3796– 3807. Antinucci, F. (1990). The comparative study of cognitive ontogeny in four primate species. In S. T. Parker & K. R. Gilson (Eds.),
‘‘Language’’ and intelligence in monkeys and apes (pp. 157–171). Cambridge: Cambridge University Press. Bachevalier, J. (1990). Ontogenetic development of habit and memory formation in primates. In A. Diamond (Ed.), The development and neural basis of higher cognitive functions (pp. 1–19). New York: Academy of Sciences Press. Bachevalier, J., Brickson, M., Hagger, C., & Mishkin, M. (1990). Age and sex differences in the effects of selective temporal lobe lesion on the formation of visual discrimination habits in rhesus monkeys (Macaca mulatta). Behavioral Neuroscience, 104, 885–899. Bachevalier, J., Hagger, C., & Bercu, B. B. (1989). Gender differences in visual habit formation in 3-month-old rhesus monkeys. Developmental Psychobiology, 22, 585–599. Bachevalier, J., & Mishkin, M. (1984). An early and a late developing system for learning and retention in infant monkeys. Behavioral Neuroscience, 98,770–778. Diamond, A. (1990). Rate of maturation of the hippocampus and the developmental progression of childrenÕs performance on the delayed non-matching to sample and visual paired comparison tasks. In A. Diamond (Ed.), The development and neural basis of higher cognitive functions (pp. 394–426). New York: Academy of Sciences Press. Falk, D. (1981). Comparative study of the endocranial casts of new and old world monkeys. In R. L. Ciochon & A. B. Chiarell (Eds.), Evolutionary biology of the new world monkeys and continental drift (pp. 157–201). San Diego: Plenum Press. Gaffan, D., & Murray, E. A. (1992). Monkeys (Macaca fascicularis) with rhinal cortex ablations succeed in object discrimination learning despite 24-h intertrial intervals and fail at matching to sample double sample presentations. Behavioral Neuroscience, 106, 30–38. Malamut, B. L., & Mishkin, M. (1981). Differences between limbic and nonlimbic retention process. Society for Neuroscience Abstracts, 7, 236. Malamut, B. L., Saunders, R. C., & Mishkin, M. (1984). Monkeys with combined amygdalo-hippocampal lesions succeed in object discrimination learning despite 24-h intertrial intervals. Behavioral Neuroscience, 98, 759–769. Meunier, M., Bachevalier, J., & Mishkin, M. (1997). Effects of orbital frontal and anterior cingulate lesions on object and spatial memory in rhesus monkeys. Neuropsychologia, 35, 999–1015. Mishkin, M., Malamut, B. L., & Bachevalier, J. (1984). Memories and habits: Two neural systems. In G. Lynch, L. McGaugh, & N. M. Weinberger (Eds.), Neurobiology of learning and memory (pp. 65– 77). New York: Guilford Press. Overman, W. H. (1990). Performance on traditional match-to-sample, non-match-to-sample, and object discrimination task by 12–32 month-old children: A developmental progression. In A. Diamond (Ed.), The development and neural basis of higher cognitive functions (pp. 365–383). New York: Academy of Sciences Press. Overman, W. H., Bachevalier, J., Turner, M., & Peuster, A. (1992). Object recognition versus object discrimination: Comparison between human infants and infant monkeys. Behavioral Neuroscience, 106, 15–29. Tavares, M. C. H., & Tomaz, C. (2002). Working memory in capuchin monkeys (Cebus apella). Behavioural Brain Research, 131, 131–137. Visalberghi, E. (1993). Capuchin monkeys: A window into tool use in apes and humans. In P. Gibson & G. Ingold (Eds.), Tools, language, and cognition in human evolution (pp. 138–150). Cambridge: Cambridge University Press. Zola-Morgan, S., Squire, L. R., & Amaral, D. G. (1989). Lesions of the hipocampal formation but not lesions of the fornix or the mamillary nuclei produce long-lasting memory impairment in monkeys. Journal of Neuroscience, 9, 898–913. Zola, S. M., Squire, L. R., Teng, E., Stefanacci, L., Buffalo, E. A., & Clark, R. E. (2000). Impairment recognition memory in monkeys after damage limited to the hippocampal region. Journal of Neuroscience, 20, 451–463.
Ontogenetic dissociation between habit learning and recognition memory in capuchin monkeys (Cebus apella) Michelline C. Resende, Maria Clotilde H. Tavares, and Carlos Tomaz* Primate Center and Department of Physiological Sciences, University of Brasılia, CEP 70910-900 Brasılia, DF, Brazil
Abstract The performance of young and adult capuchin monkeys (Cebus apella) on a Concurrent Discrimination Learning (CDL) test and a Delayed Non-Matching to Sample (DNMS) task were investigated. Results indicate that all subjects were able to learn the CDL test with 20-pairs simultaneously and retain this stimulus/reward association within 24-h interval. In contrast, young subjects did not perform the DNMS task with the same proficiency as adults. While adultsÕ scores were above chance across all memory test delays, the young capuchin monkeys performed the test by chance level. These results support the hypothesis that these two tasks require different cognitive processes mediated by two independent neural systems with a differentiated ontogenetic development. Moreover, they provide evidence that this dissociation occurs not only in humans and Old World monkeys but also in the New World capuchin monkeys indicating that this species can be a valuable alternative model for investigations of the neurobiological basis of memory. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: Habit learning; Concurrent discrimination; Recognition memory; Delayed non-matching to sample; Ontogenetic dissociation; Cebus apella
1. Introduction Several works have proposed the existence of two ontogenetic dissociable memory systems, each served by different neural circuits (e.g., Bachevalier & Mishkin, 1984; Bachevalier, Brickson, Hagger, & Mishkin, 1990; Diamond, 1990; Malamut, Saunders, & Mishkin, 1984; Overman, 1990). The cognitive memory system is thought to use a corticolimbic circuit to store sensory representations, serving both recognition and recall. On the other hand, the second system known as the habit system seems to employ a corticostriatal circuit to mediate stimulus–response retention, skill and habit learning. Moreover, it has been proposed that these two systems have a differentiated ontogenetic development (Mishkin, Malamut, & Bachevalier, 1984). Whereas the corticostriatal system is believed to be functionally mature at the birth, the corticolimbic circuit seems to develop later in life (Alvarado & Bachevalier, 2000; Bachevalier & Mishkin, 1984).
* Corresponding author. Fax: +55-61-274-1251. E-mail address: [email protected] (C. Tomaz).
Recently, there has been an expressive increase of investigations attempting to broach the development and functions of these two neural systems. In order to assess the cognitive system, researches have used the delayed non-matching to sample test (DNMS). In this task, the subject should discriminate between one object that has been seen before and one new. The subject is rewarded if chosen the new stimulus. The recognition memory can be assessed by lengthening the intervals between sample and the choice trial or the list of samples presented. The performance in this task is impaired in amnesic subjects with diencephalic lesions or in monkeys with damage in the medial temporal lobe (MTL) (Bachevalier & Mishkin, 1984; Malamut et al., 1984; Mishkin et al., 1984; Zola et al., 2000). In contrast, the non-cognitive memory system has been studied using a habit learning tasks. In monkeys, the concurrent discrimination learning task (CDL) is believed to be suitable for evaluating this type of learning, as subjects solve the test progressively over many trials (Bachevalier & Mishkin, 1984; Bachevalier et al., 1990; Malamut et al., 1984). This test involves a list of several pairs of objects presented concurrently. In order to solve the task, subjects are required to learn
1074-7427/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. PII: S 1 0 7 4 - 7 4 2 7 ( 0 2 ) 0 0 0 1 5 - 1
20
M.C. Resende et al. / Neurobiology of Learning and Memory 79 (2003) 19–24
that only one member in each pair of stimuli is rewarded. The ability to solve CDL tasks is preserved in MTL lesion subjects (Bachevalier & Mishkin, 1984; Gaffan & Murray, 1992; Malamut et al., 1984). There are evidences that the ontogenetic dissociation of these systems occurs not only in humans, but also in Old World rhesus monkeys. Bachevalier and Mishkin (1984) have pointed differences between young and adult monkeys performing CDL and DNMS tasks. As early as 3-month-old, rhesus monkeys were able to perform CDL task with 24-h intertrial interval as quickly as adult were. By contrast, they failed to master DNMS task until 4 month of age. The subjects reached the same proficiency level as adults only at about 2 years of age. The same pattern of proactive development in these two memory systems was found in human infants (Diamond, 1990; Overman, 1990; Overman, Bachevalier, Turner, & Peuster, 1992). For instance, Overman (1990) tested children in DNMS task using a rate of 15 trails/ day to a criterion of 87% correct response for two consecutive test days. He found that the youngest group of children (12–15 months old) only attained the criterion after 20 weeks of continuous training and it had the lowest rate of learning compared with the other four studied groups. By the other side, the performance of the two oldest groups (18–20-month-old and 22–32-monthold) was above chance since the beginning of the test. However, both young and old groups were able to solve a visual discrimination task. The 12-month-old children need 8.7 days to reach the learning criterion while 18-month-old attained the criterion in 6 days (Overman, 1990). In sum, the studies found that the ability to perform on DNMS problems improves systematically with age in both children and infant monkeys indicating that a further maturation of the neural circuitry is required to solve this task (Alvarado & Bachevalier, 2000; Bachevalier, 1990; Diamond, 1990; Overman et al., 1992). Nevertheless, CDL task can be solved as early as 12month-old by children and 3-month-old by rhesus monkeys (Bachevalier & Mishkin, 1984; Overman, 1990). Neurobehavioral studies have showed that early or late damage to limbic structures (hippocampus plus amygdala) produces severe impairment on DNMS test performed by monkeys (Malamut & Mishkin, 1981; Mishkin et al., 1984). However, recent findings demonstrated that the poor performance on visual recognition test may be due to unintentional damage to surrounded cortical areas. Recognition memory is affected after selective lesions to the perirhinal and entorhinal cortex (Gaffan & Murray, 1992) and hippocampus and parahippocampal area (Zola-Morgan, Squire, & Amaral, 1989). Additionally, selective damage to the hippocampal formation (Alvarez, Zola-Morgan, & Squire, 1995) did not abolish preference for the novelty in the DNMS test.
Bachevalier (1990) showed that late TE damage resulted in a drop of the adult monkeysÕ performance when they were tested on DNMS task. In contrast, when the injury occurred early in infancy, the visual recognition ability was spared. Taken together, these results indicate that the protracted development of the cognitive memory system in monkeys is due, in part, to a slow maturation of area TE instead of the limbic structures as amygdala and hippocampus per se (Alvarado & Bachevalier, 2000; Bachevalier, 1990). Although several studies have examined aspects of animal cognitive processes through CDL and DNMS tasks, especially employing Old WorldÕs monkeys (Bachevalier, 1990; Bachevalier & Mishkin, 1984; Bachevalier et al., 1990; Malamut et al., 1984), New World primate data on cognitive processes are presently scarce. The capuchin monkey (Cebus apella) has been the focus of several researches due to its behavioral similarities with apes (Antinucci, 1990). Moreover, capuchins are shown to display tool-using capacities (Visalberghi, 1993) and to readily solve DNMS memory tasks with sample/test trials delays up to 10 min (Tavares & Tomaz, 2002). Together, these features make capuchin monkeys a promising animal model for studying different aspects of memory and cognition. To our knowledge, no attempts have been made in the ambit of investigating the existence of two ontogenetic dissociable (cognitive and habit) memory systems in capuchin monkeys. Therefore, the aim of this study was to investigate the capabilities of young and adult capuchin monkeys in performing a concurrent object discrimination learning task with 24-h intertrial interval and a DNMS test.
2. Methods 2.1. Procedures The study was performed in the Primate Center— University of Brasılia. Animals were fed twice daily with a variety of foods and water was available ad libitum. Subjects were tested in their own home cages to avoid disruption of behavior by capture and transport to a novel environment. The animals were not food or water deprived, however during the experimental session food was removed from the cage. The experiment was carried out in a version of the Wisconsin General Test Apparatus (WGTA) containing a mobile test tray with two food wells of 5 cm in diameter, 12 cm apart. The experimental procedures and housing conditions were in accordance with the principles and regulation of the Ethics Committee of the University of Brasılia for the care and the use of animals for scientific purposes.
M.C. Resende et al. / Neurobiology of Learning and Memory 79 (2003) 19–24
2.2. Concurrent object discrimination learning test (CDL) Ten capuchin monkeys (C. apella), separated in two groups, served as subjects. One group was compost of five young monkeys (two females and three males) with age estimates to range from 1 to 312 years. The other group had five adult monkeys (three females and two males) with estimated age between 6 and 10 years. One young and one adult subject had some experience in visual discrimination tasks. The stimuli were ‘‘junk’’ objects that differed in shape, size, color, and texture. Pieces of grape were used as reward. CDL testing were as following. A set of 20 different pairs of objects, Set A, was presented to the subjects. In each pair only one object was baited with a food reward. Each object of the pair was present simultaneously over the lateral wells of the test tray. After the monkey had made its choice by displacing one of the objects, the second pair of stimuli was presented and so on until all 20 pairs had been displaced. The same list of objects was then repeated after a 24-h intertrial interval (24-h ITI). The objects of the pairs and the serial order of the pairs remained constant across the sessions. However, the left-right position of the baited object varied pseudorandomly in order to avoid position bias. The test was performed every day (7 days a week) until the subject attained the learning criterion setting at 90% (i.e., 18 correct response in 20 trials). Nine days after completing Set A, the monkeys were tested in the same way on a new set of objects, Set B. 2.3. Delayed non-matching to sample Eleven C. apella monkeys divided in two groups were used in this experiment. The young group consisted of the same five subjects used in the CDL experiment. The adult group was composed by six subjects (three female and three male) with estimated age rages from 4 to 8 years. These adult subjects had experience in Delayed Matching to Sample (DMS) task. However, previous study showed that the performance in DNMS test by subjects that had experience in DMS task did not differ from na€ıve animals (Tavares & Tomaz, 2002). The subjects were tested in a trial-unique version of DNMS task. The test begun showing to the subject a single object placed in the central food well (sample phase). A wooden screen was then raised, and the subject was able to see the object (sample trial). After an observing response was made towards the object, the screen was lowered to prevent the subjectÕs view of the food wells. Eight seconds later, the screen was raised once again and the animal was faced with two objects, one was the same presented before and the other a new one (test trial). The subject had to displace the new object to obtain the reward. The non-correction method
21
was used, i.e., if the subject displaced the familiar object, the test tray was moved back and the screen lowered. Fifteen to twenty seconds later (intertrial interval) a new trial begun. The left or right position of the baited object varied pseudorandomly. New pairs of object were used on every trial. The experimental session lasted about one hour and testing occurred three days a week. Training at 8 s delay condition continued until the animal reached the learning criterion of nine correct responses out of 10 consecutive trials. After reaching the criterion, subjectsÕ memory performance was assessed at delays of 15, 60, 120 (100 trials to each delay), and 600 s (50 trials). 2.4. Statistical analysis Statistical analyses were carried out using Wilcoxon Signed Rank test for intra-group comparisons and Mann–Whitney U test for inter-group comparisons in the 24-h ITI task. Based on previous studies (Bachevalier & Mishkin, 1984; Bachevalier et al., 1990) one-tailed probabilities were employed as a significant difference between Sets A and B were expected for young and adult subjects. In order to analyze the performance of the subjects in the DNMS task, a binomial test was used to establish a 95% confidence limits around chance performance based on the number of test trials. Thus, the upper limit was calculated as 60% for 100 trials and 64% for 50 trials (p < :05). The mean percentage of correct responses was compared to these confidence limits and any performance above the upper limit was considered significant. To determine whether the subjectsÕ performance differ across delays we used a one-way ANOVA. To analyze the number of trials to criterion that group required a paired t test was used.
3. Results Table 1 shows the number of daily sessions, total number of errors and trials required for each subject to learn Set A and Set B in the CDL test. The statistical analysis showed that young group required significantly fewer sessions and trials, and took less errors on Set B than Set A to attained the learning criterion (Wilcoxon; Z ¼ 2:060 for session; Z ¼ 2:023 for trials; Z ¼ 2:023 for errors, all ps < :05). However, there were not significant decrease by adult subjects in the number of sessions, trials and errors from Set A to Set B (Wilcoxon; Z ¼ 1:483 p ¼ :094 for session; Z ¼ 1:214 p ¼ :156 for trials; Z ¼ 1:483 p ¼ :094 for errors). Comparison between young and adult performances on Set A indicated no significant differences (Mann– Whitney test; U ¼ 10:5, p ¼ :369; U ¼ 11, p ¼ :421; U ¼ 10, p ¼ :325, for sessions, errors and trials, respectively). The analysis also showed no differences
22
M.C. Resende et al. / Neurobiology of Learning and Memory 79 (2003) 19–24
Table 1 Number of daily sessions (S), total number of errors (E), trials (T), and mean (M SEM) preceding criterion on the Concurrent Object Discrimination Learning test by young and adult capuchin monkeys Group
Young 1 2 3 4 5 M Adult 1 2 3 4 5 M *
Set A
Set B
S
E
T
S
E
20 21 25 24 23
133 156 192 200 164
460 449 460 418 400
9 10 7 13 13
22:6 0:93
169 12:20
437:4 12:11
10.4 1.17
71.6 7.85
207.8 23.21
5 36 15 24 22
39 284 110 239 148
100 720 295 457 440
12 23 13 16 11
68 172 82 108 74
240 441 260 320 220
20:4 5:12
164 44:06
402:4 102:04
15 2:17
100:5 19:06
296:2 39:88
59 62 56 87 94
T 140 259 260 200 180
Wilcoxon Signed Rank test, p < :05.
Table 2 Trials to criterion and percentage of correct response across delays (15, 60, 120, and 600 s) on DNMS test by young and adult capuchin monkeys Group
Young 1 2 3 4 5 M Adult 1 2 3 4 5 6 M
Trials to criterion
Delays (s) 15
60
120
600
265 82 224 173 295
50 61 56 51 62
50 59 55 52 72
50 70 54 59 55
58 46 70 62 60
207:8 41:56
56 2:47
57:6 3:90
57:6 3:41
59:2 3:88
71 184 357 366 54 10
53 63 78 85 61 81
60 54 56 83 67 71
67 46 48 72 77 73
78 65 52 70 72 80
173:67 28:94
70:2 5:25
65:2 4:44
63:8 5:48
69:5 4:14
M ¼ mean SEM.
among number of sessions, errors and trials took by young and adult subjects on Set B (Mann–Whitney test; U ¼ 5, p ¼ :075 for sessions; U ¼ 6, p ¼ :111 for errors; U ¼ 4:5, p ¼ :056 for trials). Table 2 summarizes the results obtained for DNMS experiment. All subjects reached criterion in the 8 s delay condition. The young needed 207:8 41:56 trials to attain the learning criterion, while the adult group needed 173:67 28:94 trials. Paired comparisons indicated no significant difference in the number of trials required by young and adult subjects to reach the learning criterion (t ¼ :436, p ¼ :673). An analysis of variance involving the 4 delays (15, 60, 120, and 600 s) revealed no difference across delays for
young (F ¼ :142, p ¼ :933) and adult groups (F ¼ :417, p ¼ :743, for adults). Memory tests indicated that adults, but not young subjects, perform above the upper confidence limits around chance performance (above 60% for 15, 60, and 120 s; above 64% for 600 s) in all sample-choice test intervals.
4. Discussion The results of the present study shows that young capuchin monkeys are able to learn a concurrent discrimination learning task with 20-pairs of stimuli
M.C. Resende et al. / Neurobiology of Learning and Memory 79 (2003) 19–24
simultaneously, and retain this stimulus/reward association based on single trials within 24-h interval. Moreover, the significant reduction in the number of sessions, errors and trials observed from Set A to Set B in young capuchin monkeys indicate learning set formation, which is considered a characteristic of habit memory system that subserves problem solve skills. These results are similar to that observed in a previous study using rhesus monkeys (Bachevalier, Hagger, & Bercu, 1989). While our young capuchin monkeys took in average 22 sessions on Set A and 13 on Set B, 3-month-old rhesus monkeys required an average of 18 sessions to attained the criterion on Set A and 6 sessions at the Set B (Bachevalier et al., 1989). The adult capuchin monkeys needed an average of 20 sessions to attain the criterion on Set A and 15 sessions on Set B. In two other studies adult rhesus monkeys needed 10 sessions to reach the learning criterion in both Sets (Bachevalier et al., 1989; Malamut et al., 1984), also showing no significant improvement from Set A to Set B. Although the young capuchins learned Set B quicker than Set A and adults did not, it was not observed differences between young and adultsÕ scores in both sets. Bachevalier and Mishkin (1984) found only a initial delay in learning of infant group compared with adults on Set A, but they did not find differences in infantÕ and adultÕ scores on Set B either. One possible explanation suggested by Bachevalier and co-workers (Bachevalier & Mishkin, 1984; Bachevalier et al., 1989; Bachevalier et al., 1990) is that, although the habit and the cognitive systems are present in adult subjects, they could be promoting a competition in the resolution of the problems. In young subjects only the habit memory is operating, which could enable them to perform Set B quicker than adults do. The present study also shows that young capuchin monkeys did not perform the DNMS task with the same proficiency as adult. While the adultÕ scores were above of the upper limit of confidence across all tested delays, the young subjects performed the DNMS memory test by chance level. It is important to mention that the young group was first trained on CDL task what could potentially have a negative transfer effect on performance of DNMS; i.e., the difference between the two groupsÕ DNMS performance could have resulted from their different training histories rather than their different ages. However, data obtained for humans showed the same developmental progression of the performance on DNMS task. Overman et al. (1992) found that the youngest group (12–15-month-old) only attained the learning criterion with 19 month of age, while 26month-old group reached the criterion in an average of 11 days after the beginning of the test. Another similarity between our results and that found by other researches is the steady percentage of
23
correct responses of humans and rhesus monkeys at DNMS task even at the longest delay (10 min) (Bachevalier & Mishkin, 1984; Meunier, Bachevalier, & Mishkin, 1997; Overman et al., 1992). On the other hand, Zola and colleagues (2000) found a drop in the performance of the control group at 10 and 40 min delays. However, differently from our experimental procedures, during the long delays the animals were removed from the test apparatus and placed back in their home cage. Therefore, it is possible that these methodological aspects may contribute for the decrease in performance observed in the Zola et al.Õs study (2000). Some differences were observed in DNMS task performed by humans, C. apella and rhesus monkeys. Young capuchin monkeys needed fewer trials than children and infant rhesus monkeys to attained the learning criterion at 8 s delay. While our young monkeys took in average 208 trials, 12–15-month-old children only attained the criterion in 1845 trials and monkeys with 3 months of age needed 720 trials. However, when we compared the adult groups of these species, humans were faster than monkeys. Capuchins and rhesus monkeys needed a similar number of trials (174 and 160 trials, respectively) to reach the criterion, while humans required only 90 trials to criterion (Bachevalier & Mishkin, 1984; Overman et al., 1992). Since it was found that human infants and infant monkeys have a good performance in the CDL task, their poor scores on DNMS test cannot be attributed to deficits in attention, motivation, stimulus-reward association or perceptual abilities (Bachevalier, 1990; Overman, 1990). The more acceptable explanation for differences in performing CDL and DNMS by young subjects is that these two tasks require different cognitive processes and that these processes are mediated by two independent neural systems that mature at different times (Alvarado & Bachevalier, 2000; Bachevalier, 1990; Bachevalier & Mishkin, 1984; Bachevalier et al., 1990). The CDL task requires that the subject retain a representation of stimuli, compare their features and associate objectÕs quality with reward. However, in trial unique non-matching to sample task, stimuli vary from trial to trial, such that no single stimulus is associated with the reinforcement. Success in this task has been interpreted as ÔconceptÕ learning, since the rule that allows subjects to make a correct choice regardless of the physical properties of the various stimuli presented in each problem, is the concept of novelty. Thus, DNMS becomes a more demand cognitive task than object discrimination task. According, it has been proposed that the performance in DNMS tasks requires the participation of high-order cerebral areas that seem do not be functionally mature early in the life (Bachevalier, 1990; Mishkin et al., 1984; Overman, 1990). Although no lesion data are available for Cebus genus, the outcomes from our behavioral study are similar
24
M.C. Resende et al. / Neurobiology of Learning and Memory 79 (2003) 19–24
to those obtained with the genus Macaca. It is known that the CebusÕ cerebral cortex citoarchitecture resembles that of Macaca (Falk, 1981). Therefore, it seems reasonable to speculate that in capuchin monkeys, CDL and DNMS tasks are mediated by similar neuroanatomical substrates as those found for macaques. Taken together, these results suggest that humans, New World capuchin monkeys and Old World rhesus monkeys may share a common neural substrate underlying the two memory systems and that the recognition memory system has a protracted development in all of these species. However, more systematic comparative researches using macaques and C. apella are necessary to explore this issue. To our knowledge, this study is the first report examining the possibility of ontogenetic dissociation of habit learning and declarative (explicit) memory systems in the New World capuchin monkey. The results indicate the existence of two distinct memory systems with different ontogenetic development: (1) a non-cognitive learning system based on automatic associations between stimulus and response, and (2) a cognitive system based on abstract rule learning and its correlation with the reward. The first been present in young C. apella, at least, as early as one year old, and the later seems to be not completed functionally mature until the age of 312 years. In addition, the results of the present study, together with those reporting other cognitive abilities in capuchin monkeys, suggest that this specie can be a useful model in investigations of neurobiological basis of memory. Further studies using different drug/neurotransmitters manipulations as well as brain lesion approaches would be desirable to explore the potential of this animal model.
Acknowledgments This research was partially supported by FINATEC grants to C.T. and M.C.H.T. M.C.R. is the recipient of a Ph.D. fellowship from CAPES (Brazil). We are thankful to Dr. R. de Oliveira for animal care and M. Barros for helping in statistical analysis.
References Alvarado, M. C., & Bachevalier, J. (2000). Revisiting the maturation of medial temporal lobe memory functions in primates. Learning & Memory, 7, 244–256. Alvarez, P., Zola-Morgan, S., & Squire, L. R. (1995). Damage limited to the hippocampla region produces long-lasting memory impairment in monkeys. Journal of Neuroscience, 15, 3796– 3807. Antinucci, F. (1990). The comparative study of cognitive ontogeny in four primate species. In S. T. Parker & K. R. Gilson (Eds.),
‘‘Language’’ and intelligence in monkeys and apes (pp. 157–171). Cambridge: Cambridge University Press. Bachevalier, J. (1990). Ontogenetic development of habit and memory formation in primates. In A. Diamond (Ed.), The development and neural basis of higher cognitive functions (pp. 1–19). New York: Academy of Sciences Press. Bachevalier, J., Brickson, M., Hagger, C., & Mishkin, M. (1990). Age and sex differences in the effects of selective temporal lobe lesion on the formation of visual discrimination habits in rhesus monkeys (Macaca mulatta). Behavioral Neuroscience, 104, 885–899. Bachevalier, J., Hagger, C., & Bercu, B. B. (1989). Gender differences in visual habit formation in 3-month-old rhesus monkeys. Developmental Psychobiology, 22, 585–599. Bachevalier, J., & Mishkin, M. (1984). An early and a late developing system for learning and retention in infant monkeys. Behavioral Neuroscience, 98,770–778. Diamond, A. (1990). Rate of maturation of the hippocampus and the developmental progression of childrenÕs performance on the delayed non-matching to sample and visual paired comparison tasks. In A. Diamond (Ed.), The development and neural basis of higher cognitive functions (pp. 394–426). New York: Academy of Sciences Press. Falk, D. (1981). Comparative study of the endocranial casts of new and old world monkeys. In R. L. Ciochon & A. B. Chiarell (Eds.), Evolutionary biology of the new world monkeys and continental drift (pp. 157–201). San Diego: Plenum Press. Gaffan, D., & Murray, E. A. (1992). Monkeys (Macaca fascicularis) with rhinal cortex ablations succeed in object discrimination learning despite 24-h intertrial intervals and fail at matching to sample double sample presentations. Behavioral Neuroscience, 106, 30–38. Malamut, B. L., & Mishkin, M. (1981). Differences between limbic and nonlimbic retention process. Society for Neuroscience Abstracts, 7, 236. Malamut, B. L., Saunders, R. C., & Mishkin, M. (1984). Monkeys with combined amygdalo-hippocampal lesions succeed in object discrimination learning despite 24-h intertrial intervals. Behavioral Neuroscience, 98, 759–769. Meunier, M., Bachevalier, J., & Mishkin, M. (1997). Effects of orbital frontal and anterior cingulate lesions on object and spatial memory in rhesus monkeys. Neuropsychologia, 35, 999–1015. Mishkin, M., Malamut, B. L., & Bachevalier, J. (1984). Memories and habits: Two neural systems. In G. Lynch, L. McGaugh, & N. M. Weinberger (Eds.), Neurobiology of learning and memory (pp. 65– 77). New York: Guilford Press. Overman, W. H. (1990). Performance on traditional match-to-sample, non-match-to-sample, and object discrimination task by 12–32 month-old children: A developmental progression. In A. Diamond (Ed.), The development and neural basis of higher cognitive functions (pp. 365–383). New York: Academy of Sciences Press. Overman, W. H., Bachevalier, J., Turner, M., & Peuster, A. (1992). Object recognition versus object discrimination: Comparison between human infants and infant monkeys. Behavioral Neuroscience, 106, 15–29. Tavares, M. C. H., & Tomaz, C. (2002). Working memory in capuchin monkeys (Cebus apella). Behavioural Brain Research, 131, 131–137. Visalberghi, E. (1993). Capuchin monkeys: A window into tool use in apes and humans. In P. Gibson & G. Ingold (Eds.), Tools, language, and cognition in human evolution (pp. 138–150). Cambridge: Cambridge University Press. Zola-Morgan, S., Squire, L. R., & Amaral, D. G. (1989). Lesions of the hipocampal formation but not lesions of the fornix or the mamillary nuclei produce long-lasting memory impairment in monkeys. Journal of Neuroscience, 9, 898–913. Zola, S. M., Squire, L. R., Teng, E., Stefanacci, L., Buffalo, E. A., & Clark, R. E. (2000). Impairment recognition memory in monkeys after damage limited to the hippocampal region. Journal of Neuroscience, 20, 451–463.