Brain, Vol. 122, No. 6, 1063-1068,
June 1999
© 1999 Oxford University Press
Classification learning in Alzheimer's disease
1 Departments of Psychiatry and 2 Physiology, Albert Szent-Györgyi Medical University, Szeged, Hungary and 3 Neurology Section, VA Medical Center and 4 Department of Neurology, University of Arizona, Tucson, Arizona, USA
Correspondence to:
Szabolcs Kéri, MD, Albert Szent-Györgyi Medical University, Department of Psychiatry, Szeged, 6701, Pf.397, Hungary E-mail: SZKERI{at}Phys.SZOTE.U.-Szeged.HU
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Previous research has demonstrated that explicit recognition of dot patterns is impaired in amnesic patients with damage to the limbic–diencephalic memory system, while implicit categorization of the same kind of stimuli is preserved. The aim of the present study was to investigate the relationship between recognition and categorization performances in patients with Alzheimer's disease. Consistent with the findings in amnesic subjects, our results revealed that the explicit recognition of dot patterns was significantly impaired in Alzheimer's disease. However, implicit categorization functions were also disrupted. This was selective for the prototype stimuli; the categorization of non-prototype dot patterns was spared. The impaired category learning is likely to reflect the damage of modality-specific neocortical areas in Alzheimer's disease.
Alzheimer's disease; classification learning; non-declarative memory; prototype
ANOVA = analysis of variance
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The impairment of conscious and controlled declarative memory is one of the most important symptoms of Alzheimer's disease. Neuroimaging studies revealed a relationship between this memory impairment and structural/metabolic abnormalities in the medial temporal regions (Ishii et al., 1996
; Detoledo-Morrell et al., 1997; Mori et al., 1997). In contrast, unconscious and automatic non-declarative memory functions, such as perceptual repetition priming, seem to be spared until the later stages of Alzheimer's disease (Gabrieli, 1996). Recently, it has been suggested that while explicit recognition memory for dot patterns is mediated by the limbic–diencephalic declarative memory system, the implicit categorization of the same patterns is associated with the neocortical non-declarative memory system (Knowlton and Squire, 1993; Squire and Knowlton, 1995). In a categorization task, distorted versions of a prototype dot pattern were presented to amnesic patients (Fig. 1). Although the patients had severely impaired declarative memory, they were able to extract the prototype from the studied category exemplars. In the testing phase, the amnesic subjects categorized new exemplars successfully; moreover, their performance was higher for the prototype stimuli compared with the distorted items (Knowlton and Squire, 1993; Squire and Knowlton, 1995). These results have demonstrated that prototype-based category learning is independent of the limbic–diencephalic structures. However, the precise anatomical localization and functional organization of the neocortical memory system, mediating category learning, remain to be elucidated.
|
In Alzheimer's disease, in addition to medial temporal lobe pathology, the degenerative process also involves the visual association areas, whereas lower-order sensory neocortex is believed to be relatively spared until the later stages of the disease (Braak and Braak, 1991
; Detoledo-Morrell et al., 1997). In fact, it has been proposed that the sparing of early visual areas may be responsible for the apparent preservation of certain non-declarative memory functions such as the above-mentioned perceptual repetition priming (Ungerleider, 1995; Gabrieli, 1996). In the present study, we investigated the recognition and categorization of dot patterns in 25 patients with Alzheimer's disease. Specifically, we intended to determine whether or not the additional involvement of neocortical areas might interfere with implicit category learning. If not, the performance of Alzheimer's disease patients may be similar to that of amnesic patients, which is characterized by the deficit of recognition and the preservation of categorization functions (Knowlton and Squire, 1993; Squire and Knowlton, 1995; Sinha et al., 1997). |
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Subjects
The patient group comprised 25 subjects (20 female and five male) who met the DSM III-R (Diagnostic and Statistical Manual of Mental Disorders, American Psychiatric Association, 1987) criteria of senile dementia of the Alzheimer type (mean age 71.8 years, SD = 5.6; mean level of education 12.3 years, SD = 9.7). In addition, 20 age-matched (mean 72.3 years, SD = 9.2) and education-matched (mean duration 11.9 years, SD = 9.8) control subjects (17 female and three male) participated in the study. The visual acuities were normal or corrected to normal in both groups. The cognitive functions of the patients were evaluated with the Mini-Mental State Examination (Folstein et al., 1975
) (mean 20.9, SD = 5.2) and the Enhanced Cued Recall Test (Grober et al., 1988) (means 1.9/4.1, 2.3/3.4 and 2.0/2.9, SD = 0.9/1.7, 1.1/2.6 and 0.7/0.9). All patients underwent CT, MRI and SPECT studies in order to exclude any other neurological disorder. Informed consent was obtained from each patient and control subject. The study was approved by The Human Biomedical Research Ethics Committee of The Albent-Györgyi Medical University, Szeged, Hungary.
Stimuli
Stimuli consisted of nine dots randomly placed within a 12 x 12 cm area of the computer screen. Four types of stimuli were used. First, a particular pattern of dots was created and was designated as the prototype. The `low' and `high' distortions of the prototype were generated by various degrees of displacement of the nine dots. The difference between `low' and `high' distortions was only quantitative. Random items with a pattern independent of the prototype were also included in the experiment (Fig. 1
) [for methodological details, see Posner and Keele (Posner and Keele, 1968)].
Procedure
Categorization
In the training phase, 20 `low' and 20 `high' distortions were presented successively, each for 5 s. The interstimulus interval was 1 s. With each presentation, the subject was asked to point to the dot closest to the centre. After an unfilled delay period of 5 min, participants were told that the previously seen patterns all belonged in the same category, in the same sense that if a series of dogs had been presented, they would comprise the members of the category `dog' (Squire and Knowlton, 1995). The experimenter ensured that the patients had understood the task by asking them to repeat the instructions. In the testing phase, six prototypes, 20 new `low' distortions, 20 new `high' distortions and 40 random patterns were presented in a pseudorandomized order. The participants were asked to respond `yes' if the testing item belonged in the category learned in the training phase and `no' if it did not. The category judgement task was not strictly time-limited. Categorization performance was defined as the percentage of successful classification judgements, including the endorsement of the prototypes/distortions and the rejection of random patterns.
Recognition
In the training phase, a single prototype pattern was presented 40 times. The instructions and the exposure time were the same as in the categorization procedure. Five minutes later, six previously seen dot patterns (prototypes), 20 `low' and 20 `high' distortions of the prototype (`near' and `far' targets, respectively) and 40 random patterns were presented in a pseudorandomized order. The participants were asked to respond `yes' only if the presented item was identical to the original pattern. Again, decision making was not strictly time-limited. Recognition performance was defined as the percentage of correct stimulus identifications, including the endorsement of the prototype and the rejection of the distorted versions and random patterns. The sequence of categorization and recognition tests was counterbalanced across subjects and no order effect was observed.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Categorization
A 2 (group) x 4 (stimulus type) analysis of variance (ANOVA) on the categorization performance revealed a marginally significant main effect of group [F(1,43) = 4.05, P = 0.05], a significant main effect of stimulus type [F(3,129) = 3.15, P < 0.03] and a two-way interaction between group and stimulus type [F(3,129) = 4.74, P < 0.004]. This indicated a marginal but stimulus-specific categorization deficit in the Alzheimer's disease patients. Newman–Keuls tests and separate one-way ANOVAs demonstrated a significant and selective categorization deficit for the prototype stimuli in the Alzheimer's disease patients. The post hoc analyses also indicated that `low' distortions had a significant advantage over `high' distortions only in the control group (P < 0.05) (Table 1
).
|
Recognition
The same two-way ANOVA as in the categorization session was conducted on the recognition performance. In this case, significant main effects of group [F(1,43) = 275.99, P < 0.0001] and stimulus type [F(3,129) = 3.50, P < 0.02] were obtained. The group x stimulus type interaction was not significant. Newman–Keuls tests and separate one-way ANOVAs yielded lower recognition performances for each stimulus type in the Alzheimer's disease group (Table 2
).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The present study has demonstrated that besides the prominent explicit recognition memory deficit, there was a specific disorder of implicit category learning in the Alzheimer's disease group: a selective loss of prototype acquisition.
In abstraction, similar particulars of reality are grouped as members of the same category and dissimilar items are separated as members of distinct categories (Smith and Medin, 1981). There are two main approaches to the classification process. According to the prototype-based model, an averaged central tendency, a prototype of the category is extracted from different exemplars (`family resemblance') (Posner and Keele, 1968; Homa and Vosburgh, 1976; Hayes-Roth and Hayes-Roth, 1977; Omohundro, 1981; Shepard, 1987). In further tasks, categorization performance is better for the prototype than for other novel non-prototype patterns. In contrast, exemplar-based or `store-instances-only' models suggest that categories are encoded as groups of distinct exemplars without the representation of a single central prototype (Medin and Schaffer, 1978; Gillund and Shiffrin, 1984; Estes, 1986; Nosofsky, 1992). According to this approach, an item belongs in the category that includes the most similar exemplar. Classification learning arises as an emergent property of a neural network that stores only item-specific information about the individual exemplars.
Consistent with previous studies (Posner and Keele, 1968), categorization performance was highest for the prototype items in the control group, despite the fact that the subjects were not exposed to these items in the training phase. By contrast, this pattern was not observed in the Alzheimer's disease patients who appeared to have a selective deficit for the prototype patterns. In the case of novel distortions, the controls had relatively lower levels of performance because these stimuli were less similar to the stored prototype. On the other hand, the Alzheimer's disease patients exhibited relatively high levels of performance for both the `low' and `high' distortions. Moreover, post hoc analyses revealed a significant advantage of the `low' distortions over `high' distortions only in the control group, presumably because `low' distortions were more similar to the prototype. Again, this was not the case in the Alzheimer's disease group. In addition, there was a tendency to better performance for `high' distortions compared with the normal controls (Table 1). Overall, our results indicate that the controls learned about categories in a prototype-based manner, while the Alzheimer's disease patients did so in an exemplar-based manner, which apparently did not result in the induction of a central prototype.
The Alzheimer's disease patients in this study had clinically significant memory impairment and they performed poorly on the Enhanced Cued Recall Test investigating explicit memory functions. Furthermore, their recognition performance for the dot patterns was comparable with that of the amnesic patients described by Knowlton and Squire (Knowlton and Squire, 1993). Therefore, it is unlikely that our results reflect partially preserved declarative memory for the individual training items. Instead, our findings, together with the results of investigations of category learning in amnesic patients (Knowlton and Squire, 1993; Squire and Knowlton, 1995), suggest that prototype-based and exemplar-based category learning can both be observed in the absence of normal declarative memory.
One could assume that our data reflect a general cognitive decline rather than a specific deficit of categorization. However, this explanation is unlikely. Concerning the categorization performance, there was a significant group x stimulus type interaction, indicating a specific impairment of prototype induction in the Alzheimer's disease group. The difference between the overall categorization performances was only marginally significant, and in the recognition task the Alzheimer's disease patients performed much worse than the controls. This demonstrates a mild and specific deficit of the non-declarative category knowledge and a robust impairment of the declarative memory for item knowledge.
Theoretically, it is possible to use brain–behaviour parallels from Alzheimer's disease patients in order to constrain and refine models of normal cognition (Penniello et al., 1995; Grossman et al., 1997). At the same time, it is difficult to draw precise anatomical conclusions about the cortical localization of prototype-based versus exemplar-based learning due to the diffuse nature of the degenerative process, which may show considerable variations across different patient populations. Therefore, further studies should extend the present results, using different behavioural and functional imaging methods in patients with different stages of dementia and subjects with circumscribed cortical lesions.
Previous research in amnesic patients has demonstrated that category learning is independent of the limbic–diencephalic memory system and may be mediated by modality-specific sensory cortex (Knowlton and Squire, 1993; Squire and Knowlton, 1995). In Alzheimer's disease, both medial temporal lobe and neocortical memory systems are impaired, with particular emphasis on the association areas (Wright et al., 1987; Braak and Braak, 1991; Detoledo-Morrell et al., 1997). Single-cell recordings from monkeys demonstrate that along the occipitotemporal (ventral) visual pathway, responses of more posteriorly located neurons are determined by particular stimulus features such as position, orientation, size and surface characteristics. In contrast, neurons located in more anterior regions may code abstract properties of stimuli. For example, cells in the inferotemporal cortex maintain their selectivity for global shapes across various stimulus transformations, including retinal position, orientation, size and surface characteristics (Oram and Perret, 1994; Logothetis and Sheinberg, 1996). Thus, it is possible that prototype formation and storage take place in the inferotemporal cortex (Weiskrantz, 1990), whereas posterior areas are involved in the processing of visual information that pertains to features of individual exemplars. Damage to the higher-order visual association cortex may account for the loss of prototype learning, while the preservation of lower-order visual cortex may explain why exemplar-based classification learning was spared in the Alzheimer's disease patients (Braak and Braak, 1991; Detoledo-Morrell et al., 1997).
However, the results of a recent functional imaging study suggest an alternative hypothesis. In healthy subjects, there was a relative occipital deactivation (Brodmann areas 17 and 18) for dot patterns belonging in the same category, suggesting that category exemplars were processed more easily or quickly than patterns outside the category (Reber et al., 1998). Although blood flow changes associated with the processing of prototype relative to non-prototype items were not investigated, it is possible that both prototype-based and exemplar-based category acquisitions are mediated by the early visual areas which are also implicated in perceptual priming (Ungerleider, 1995).
Theoretically, different types of classification learning may reflect distinct functional states of the same neural network located in the early visual cortex. For instance, one can propose a three-level model of category induction. Firstly, the neuronal representations of individual dot patterns are formed. Second, the individual patterns are grouped on the basis of their common features without the induction of a prototype. This may correspond to the exemplar-based category learning that was found to be relatively preserved in our Alzheimer's disease patients. Third, in a subsequent phase of integration, the prototype is extracted from the individual exemplars. This level of category learning is impaired in Alzheimer's disease. Alternatively, a similar deficit may result from the disconnection between the exemplar and prototype levels of the model. However, the above-described three stages refer to cognitive domains and not necessarily to separate cortical areas. Hypothetically, distinct functional states of the same or closely related neuronal networks may perform all these levels of information processing (Moscovitch, 1995). Different vulnerabilities of neurons in the visual cortex may be responsible for the loss of prototype extraction in the Alzheimer's disease patients. We hypothesize that neurons in layer III, which provide cortico-cortical connections, may play a crucial role in the integrative process of prototype induction. These neurons have been shown to be markedly vulnerable in Alzheimer's disease (Khachaturian, 1985; Hof and Bouras, 1991; Hof et al., 1997). With further progression of the disease process, however, we expect that exemplar-based category learning will also become compromised.
In summary, two hypotheses arise regarding the neuropsychological mechanisms of defective category learning in Alzheimer's disease. The first is based on the classical view that occipital cortex is relatively spared in Alzheimer's disease and the deficit of prototype extraction reflects damage to the higher-order visual association areas. According to the second hypothesis, early visual functions are not preserved and different functional states of a neuronal network may serve as substrates of different cognitive operations (Moscovitch, 1995).
The question of additional neural structures in category learning is important. The functional imaging study of Reber and colleagues (Reber et al., 1998) found bilateral activation in the anterior frontal cortex (Brodmann area 10) and in the right inferior lateral frontal cortex (Brodmann areas 44 and 47) for the category items compared with the random patterns. This may refer to intentional processes to recall the prototype or working memory needed to match the prototype and the actual pattern for category decision. The amnesic patient described by Squire and Knowlton (Squire and Knowlton, 1995) had abnormally low scores in the Wisconsin Card Sorting Test, indicating executive dysfunction. In spite of this, the categorization performance was within the normal range, suggesting that these prefrontal functions are not indispensable for perceptual categorization. However, on the basis of available data, the role of the prefrontal cortex cannot be excluded entirely. Recently, it has been shown that non-declarative processes, which are important in abstraction and problem solving, are related to prefrontal working memory functions (Reber and Kotovsky, 1997).
Participation of the subcortical habit-learning system is also possible (Saint-Cyr et al., 1988), since Parkinson's disease patients with neostriatal disturbances (but not amnesic subjects with medial temporal lobe damage) are impaired in the probabilistic classification learning test (Knowlton et al., 1996). In this respect, the prefrontal cortex may provide working memory modulation of information via striatal–thalamic–cortical loops (Gabrieli, 1998). Cerebellum can be another potential neuronal structure, participating in sensory acquisition and procedural learning functions (Gao et al., 1996; Molinari et al., 1997). Although Reber et al. (Reber et al., 1998) have not reported activation in the neostriatum and cerebellum, their study has evaluated only the recall phase and not the study phase of perceptual categorization. Thus, further functional imaging studies should investigate the role of these neuronal structures. Inclusion of patients with cerebral lesions may also help to determine the neuronal mechanisms of prototype-based and exemplar-based category representations.
|
Acknowledgments |
|---|
We wish to thank P. Aszalós, G. Kovács and A. Lörincz for useful comments and their technical contribution. We also thank K. Mazurák for research assistance. This work was supported by the Hungarian–Flanders Joint Project B-4/97 and the grants ETT 5809604/OTKA T025160.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Braak H, Braak E. Neuropathological staging of Alzheimer-related changes. [Review]. Acta Neuropathol (Berl.) 1991; 82: 239–59.[Medline]
Detoledo-Morrell L, Sullivan MP, Morrell F, Wilson RS, Bennett DA, Spencer S. Alzheimer's disease: in vivo detection of differential vulnerability of brain regions. Neurobiol Aging 1997; 18: 463–8.[Web of Science][Medline]
Estes WK. Array models for category learning. Cogn Psychol 1986; 18: 500–49.[Web of Science][Medline]
Folstein MF, Folstein SE, McHugh PR. `Mini-mental state'. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12: 189–98.[Web of Science][Medline]
Gabrieli JD. Memory system analyses of mnemonic disorders in aging and age-related diseases. [Review]. Proc Natl Acad Sci USA 1996; 93: 13534–40.
Gabrieli JD. Cognitive neuroscience of human memory. [Review]. Annu Rev Psychol 1998; 49: 87–115.[Web of Science][Medline]
Gao JH, Parsons LM, Bower JM, Xiong J, Li J, Fox PT. Cerebellum implicated in sensory acquisition and discrimination rather than motor control [see comments]. Science 1996; 272: 545–7. Comment in: Science 1996; 272: 482–3.[Abstract]
Gillund G, Shiffrin RM. A retrieval model for both recognition and recall. Psychol Rev 1984; 91: 1–67.[Web of Science][Medline]
Grober E, Buschke H, Crystal H, Bang S, Dresner R. Screening for dementia by memory testing. Neurology 1988; 38: 900–3.
Grossman M, Payer F, Onishi K, White-Devine T, Morrison D, D'Esposito M, et al. Constraints on the cerebral basis for semantic processing from neuroimaging studies of Alzheimer's disease. J Neurol Neurosurg Psychiatry 1997; 63: 153–8.
Hayes-Roth B, Hayes-Roth F. Concept learning and the recognition and classification of exemplars. J Verb Learn Verb Behav 1977; 16: 321–38.
Hof PR, Bouras C. Object recognition deficit in Alzheimer's disease: possible disconnection of the occipito-temporal component of the visual system. Neurosci Lett 1991; 122: 53–6.[Web of Science][Medline]
Hof PR, Vogt BA, Bouras C, Morrison JH. Atypical form of Alzheimer's disease with prominent posterior cortical atrophy: a review of lesion distribution and circuit disconnection in cortical visual pathways. [Review]. Vision Res 1997; 37: 3609–25.[Web of Science][Medline]
Homa D, Vosburgh R. Category breadth and the abstraction of prototypical information. J Exp Psychol Hum Learn Mem 1976; 2: 322–30.
Ishii K, Kitagaki H, Kono M, Mori E. Decreased medial temporal lobe oxygen metabolism in Alzheimer's disease shown by PET. J Nucl Med 1996; 37: 1159–65.
Khachaturian ZS. Diagnosis of Alzheimer's disease. Arch Neurol 1985; 42: 1097–105.
Knowlton BJ, Squire LR. The learning of categories: parallel brain systems for item memory and category knowledge. Science 1993; 262: 1747–9.
Knowlton BJ, Mangels JA, Squire LR. A neostriatal habit learning system in humans [see comments]. Science 1996; 273: 1399–402. Comment in: Science 1996; 273: 1353–4.[Abstract]
Logothetis NK, Sheinberg DL. Visual object recognition. [Review]. Annu Rev Neurosci 1996; 19: 577–621.[Web of Science][Medline]
Medin DL, Schaffer MM. Context theory of classification learning. Psychol Rev 1978; 85: 207–38.[Web of Science]
Molinari M, Leggio MG, Solida A, Ciorra R, Misciagna S, Silveri MC, et al. Cerebellum and procedural learning: evidence from focal cerebellar lesions. Brain 1997; 120: 1753–62.
Mori E, Yoneda Y, Yamashita H, Hirono N, Ikeda M, Yamadori A. Medial temporal structures relate to memory impairment in Alzheimer's disease: an MRI volumetric study. J Neurol Neurosurg Psychiatry 1997; 63: 214–21.
Moscovitch M. Models of consciousness and memory. In: Gazzaniga MS, editor. The cognitive neurosciences. Cambridge (MA): MIT Press; 1995. p. 1341–56.
Nosofsky RM. Exemplar-based approach to relating categorization, identification, and recognition. In: Ashby FG, editor. Multidimensional models of perception and cognition. Hillsdale (NJ): Lawrence Erlbaum; 1992. p. 363–93.
Omohundro J. Recognition vs. classification of ill-defined category exemplars. Mem Cogn 1981; 9: 324–31.[Web of Science]
Oram MW, Perret DI. Modelling visual recognition from neurobiological constraints. [Review]. Neural Networks 1994; 7: 945–72.[Web of Science]
Penniello M-J, Lambert J, Eustache F, Petit-Taboue MC, Barre, Viader F, et al. A PET study of the functional neuroanatomy of writing impairment in Alzheimer's disease: the role of the left supramarginal and left angular gyri. Brain 1995; 118: 697–706.
Posner MI, Keele SW. On the genesis of abstract ideas. J Exp Psychol 1968; 77: 353–63.[Web of Science][Medline]
Reber PJ, Kotovsky K. Implicit learning in problem solving: the role of working memory capacity. J Exp Psychol Gen 1997; 126: 178–203.[Web of Science]
Reber PJ, Stark CE, Squire LR. Cortical areas supporting category learning identified using functional MRI. Proc Natl Acad Sci USA 1998; 95: 747–50.
Saint-Cyr JA, Taylor AE, Lang AE. Procedural learning and neostriatal dysfunction in man. Brain 1988; 111: 941–59.
Shepard RN. Toward a universal law of generalization for psychological science. Science 1987; 237: 1317–23.
Sinha R, Heindel W, Ott B. Classification learning and recognition memory for prototype dot patterns in Alzheimer's disease [abstract]. Soc Neurosci Abstr 1997; 32: 1580.
Smith EE, Medin DL. Categories and concepts. Cambridge (MA): Harvard University Press; 1981.
Squire LR, Knowlton BJ. Learning about categories in the absence of memory. Proc Natl Acad Sci USA 1995; 92: 12470–4.
Ungerleider LG. Functional brain imaging studies of cortical mechanisms for memory. [Review]. Science 1995; 270: 769–75.
Weiskrantz L. Visual prototypes, memory, and the inferotemporal cortex. In: Iwai E, Mishkin M, editors. Vision, memory, and the temporal lobes. New York: Elsevier Science; 1990. p. 13–27.
Wright CE, Drasdo N, Harding GF. Pathology of the optic nerve and visual association areas. Information given by the flash and pattern visual evoked potential, and the temporal and spatial contrast sensitivity function. Brain 1987; 110: 107–20.
Received July 24, 1998. Revised January 2, 1999. Accepted January 18, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  P. Koenig, E. E. Smith, V. Troiani, C. Anderson, P. Moore, and M. Grossman Medial Temporal Lobe Involvement in an Implicit Memory Task: Evidence of Collaborating Implicit and Explicit Memory Systems from fMRI and Alzheimer's Disease Cereb Cortex, December 1, 2008; 18(12): 2831 - 2843. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. B. CASALE and F. G. ASHBY A role for the perceptual representation memory system in category learning Atten Percept Psychophys, August 1, 2008; 70(6): 983 - 999. [Abstract] [PDF]
|
|
|
|
 |
|
 E. E. SMITH The case for implicit category learning Cogn Affect Behav Neurosci, March 1, 2008; 8(1): 3 - 16. [Abstract] [PDF]
|
|
|
|
 |
|
 F. G. Ashby and B. J. Spiering The Neurobiology of Category Learning Behav Cogn Neurosci Rev, June 1, 2004; 3(2): 101 - 113. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||