A review of medial temporal lobe and caudate contributions to visual category learning.

Here we review recent functional neuroimaging, neuropsychological and behavioral studies examining the role of the medial temporal lobe (MTL) and the caudate in learning visual categories either by verbalizeable rules or without awareness. The MTL and caudate are found to play dissociable roles in different types of category learning with successful rule-based (RB) categorization depending selectively on the MTL and non-verbalizeable information-integration (II) category learning depending on the posterior caudate. These studies utilize a combination of experimental cognitive psychology, mathematical modeling (Decision Bound Theory (DBT)) and cognitive computational modeling (the COVIS model of Ashby et al. [1998. A neuropsychological theory of multiple systems in category learning. Psychological Review 105, 442-481]) to enhance the understanding of data obtained via functional magnetic resonance imaging (fMRI). The combination of approaches is used to both test hypotheses of the cognitive model and also to incorporate hypotheses about the strategies used by participants to direct analysis of fMRI data. Examination of the roles of the MTL and caudate in visual category learning holds the promise of bridging between abstract cognitive models of behavior, systems neuroscience, neuropsychology, and the underlying neurophysiology of these brain regions.

Intact learning of artificial grammars and intact category learning by patients with Parkinson's disease.

Patients with Parkinson's disease (PD) have been shown to be impaired on some nondeclarative memory tasks that require cognitive skill learning (perceptual-motor sequence learning, probabilistic classification). To determine what other skill-based tasks are impaired, 13 patients with PD were tested on artificial grammar learning, artificial grammar learning with transfer to novel lettersets, and prototype learning. Patients with PD performed similarly to controls on all 3 tests. The intact learning exhibited by PD patients on these tests suggests that nondeclarative cognitive skill learning is not a single entity supported by the neostriatum. If learning the regularities among visual stimuli is the principal feature of artificial grammar learning and prototype learning, then these forms of skill learning may be examples of perceptual learning, and they may occur in early visual cortical processing areas.

Dissociable properties of memory systems: differences in the flexibility of declarative and nondeclarative knowledge.

Amnesic patients (n = 8), who have severely impaired declarative memory, learned a probabilistic classification task at the same rate as normal subjects (n = 16) but subsequently were impaired on transfer tests that required flexible use of their task knowledge. A second group of controls (n = 20) rated the questions on the transfer tests according to whether the questions simply reinstated the training conditions or required flexible use of task knowledge. The amnesic patients tended to be impaired on the same items that were rated as requiring indirect or flexible use of knowledge. Thus, control subjects acquired declarative knowledge about the task that could be applied flexibly to the transfer tests. The nondeclarative memory available to amnesic patients was relatively inflexible and available only in conditions that reinstantiated the conditions of training. These findings show that declarative memory has different operating characteristics than nondeclarative memory.

Action selection and refinement in subcortical loops through basal ganglia and cerebellum.

Subcortical loops through the basal ganglia and the cerebellum form computationally powerful distributed processing modules (DPMs). This paper relates the computational features of a DPM's loop through the basal ganglia to experimental results for two kinds of natural action selection. First, functional imaging during a serial order recall task was used to study human brain activity during the selection of sequential actions from working memory. Second, microelectrode recordings from monkeys trained in a step-tracking task were used to study the natural selection of corrective submovements. Our DPM-based model assisted in the interpretation of puzzling data from both of these experiments. We come to posit that the many loops through the basal ganglia each regulate the embodiment of pattern formation in a given area of cerebral cortex. This operation serves to instantiate different kinds of action (or thought) mediated by different areas of cerebral cortex. We then use our findings to formulate a model of the aetiology of schizophrenia.

Neural correlates of rule-based and information-integration visual category learning.

An emerging theory of the neurobiology of category learning postulates that there are separate neural systems supporting the learning of categories based on verbalizeable rules (RB) or through implicit information integration (II). The medial temporal lobe (MTL) is thought to play a crucial role in successful RB categorization, whereas the posterior regions of the caudate are hypothesized to support II categorization. Functional neuroimaging was used to assess activity in these systems during category-learning tasks with category structures designed to afford either RB or II learning. Successful RB categorization was associated with relatively increased activity in the anterior MTL. Successful II categorization was associated with increased activity in the caudate body. The dissociation observed with neuroimaging is consistent with the roles of these systems in memory and dissociations reported in patient populations. Convergent evidence from these approaches consistently reinforces the idea of multiple neural systems supporting category learning.

Parallel brain systems for learning with and without awareness.

A fundamental issue about memory and its different forms is whether learning can occur without the development of conscious knowledge of what is learned. Amnesic patients and control subjects performed a serial reaction time task, exhibiting equivalent learning of an imbedded repeating sequence as measured by gradually improving reaction times. In contrast, four tests of declarative (explicit) knowledge indicated that the amnesic patients were unaware of their knowledge. Moreover, after taking the tests of declarative memory, all subjects continued to demonstrate tacit knowledge of the repeating sequence. This dissociation between declarative and nondeclarative knowledge indicates that the parallel brain systems supporting learning and memory differ in their capacity for affording awareness of what is learned.

Encapsulation of implicit and explicit memory in sequence learning.

Contrasts between implicit and explicit knowledge in the serial reaction time (SJRT) paradigm have been challenged because they have depended on a single dissociation; intact implicit knowledge in the absence of corresponding explicit knowledge. In the SRT task, subjects respond with a corresponding keypress to a cue that appears in one of four locations. The cue follows a repeating sequence of locations, and subjects can exhibit knowledge of the repeating sequence through increasingly rapid performance (an implicit test) or by being able to recognize the sequence (an explicit test). In our study, amnesic patients were given extensive SRT training. Their implicit and explicit test performance was compared to the performance of control subjects who memorized the training sequence. Compared with control subjects, amnesic patients exhibited superior performance on the implicit task and impaired performance on the explicit task. This crossover interaction suggests that implicit and explicit knowledge of the embedded sequence are separate and encapsulated and that they presumably depend on different brain systems.

Relaxing decision criteria does not improve recognition memory in amnesic patients.

An important question about the organization of memory is whether information available in non-declarative memory can contribute to performance on tasks of declarative memory. Dorfman, Kihlstrom, Cork, and Misiaszek (1995) described a circumstance in which the phenomenon of priming might benefit recognition memory performance. They reported that patients receiving electroconvulsive therapy improved their recognition performance when they were encouraged to relax their criteria for endorsing test items as familiar. It was suggested that priming improved recognition by making information available about the familiarity of test items. In three experiments, we sought unsuccessfully to reproduce this phenomenon in amnesic patients. In Experiment 3, we reproduced the methods and procedure used by Dorfman et al. but still found no evidence for improved recognition memory following the manipulation of decision criteria. Although negative findings have their own limitations, our findings suggest that the phenomenon reported by Dorfman et al. does not generalize well. Our results agree with several recent findings that suggest that priming is independent of recognition memory and does not contribute to recognition memory scores.

Contrasting cortical activity associated with category memory and recognition memory.

We collected functional neuroimaging data while volunteers performed similar categorization and recognition memory tasks. In the categorization task, volunteers first studied a series of 40 dot patterns that were distortions of a nonstudied prototype dot pattern. After a delay, while fMRI data were collected, they categorized 72 novel dot patterns according to whether or not they belonged to the previously studied category. In the recognition task, volunteers first studied five dot patterns eight times each. After a delay, while fMRI data were collected, they judged whether each of 72 dot patterns had been studied earlier. We found strikingly different patterns of brain activity in visual processing areas for the two tasks. During the categorization task, the familiar stimuli were associated with decreased activity in posterior occipital cortex, whereas during the recognition task, the familiar stimuli were associated with increased activity in this area. The findings indicate that these two types of memory have contrasting effects on early visual processing and reinforce the view that declarative and nondeclarative memory operate independently.

Reaction time distributions across normal forgetting: searching for markers of memory consolidation.

Volunteers studied pictures of objects and were then tested for yes/no recognition at 10 min and 1 week after learning (experiment 1), or at 10 min and 4 months after learning (experiment 2). Because the gradual consolidation of long-term memory is thought to occur across this time scale (weeks and months), the reaction time distributions of successfully retrieved items were analyzed in an attempt to detect markers of consolidation. At each retention interval, reaction times for items retrieved successfully were well fit by a model that assumed a single underlying distribution. No evidence for a bimodal distribution of reaction times was observed. Furthermore, there was no evidence that some small subset of items was actually retrieved faster after a long retention interval than after a short interval. The results are consistent with the idea that consolidation works not to increase memory trace strength but to change the nature of memory storage. This process occurs during the course of normal forgetting and may not be observable in the behavior of normal memory.

Cortical areas supporting category learning identified using functional MRI.

Functional MRI was used to identify cortical areas involved in category learning by prototype abstraction. Participants studied 40 dot patterns that were distortions of an underlying prototype and then, while functional MRI data were collected, made yes-no category judgments about new dot patterns. The dot patterns alternated between ones mostly requiring a "yes" response and ones mostly requiring a "no" response. Activity in four cortical areas correlated with the category judgment task. A sizeable posterior occipital cortical area (BA 17/18) exhibited significantly less activity during processing of the categorical patterns than during processing of noncategorical patterns. Significant increases in activity during processing the categorical patterns were observed in left and right anterior frontal cortex (BA 10) and right inferior lateral frontal cortex (BA 44/47). Decreases in activation of visual cortex when categorical patterns were being evaluated suggest that these patterns could be processed in a more rapid or less effortful manner after the prototype had been learned. Increases in prefrontal activity associated with processing categorical patterns could be related to any of several processes involved in retrieving information about the learned exemplars.

The human perirhinal cortex and recognition memory.

The importance of the perirhinal cortex for visual recognition memory performance is undisputed. However, it has not been clear whether its contribution to performance is mainly perceptual, or mainly mnemonic, or whether the perirhinal cortex contributes to both perception and memory. We determined the effects of medial temporal lobe damage that includes complete damage to the perirhinal cortex in two amnesic patients by assessing recognition memory for complex visual stimuli across delays from 0 to 40 s. These patients, as well as six other amnesic patients with damage limited to the hippocampal formation or diencephalic structures, exhibited intact recognition memory at delays of 0-2 s and a delay-dependent memory impairment at delays of 6 s and longer. Additionally, the patients with damage to the perirhinal cortex performed worse than the other amnesic patients at delays of 25 s and longer. The findings suggest that the perirhinal cortex is not important for visual perception or immediate memory. In this respect, the findings for perirhinal cortex resemble the findings for other medial temporal lobe structures, including the hippocampus.

Correction of off resonance-related distortion in echo-planar imaging using EPI-based field maps.

We present, here, a simple method for measurement and correction of off-resonance related geometric distortion in echo-planar imaging (EPI). This method uses high signal-to-noise ratio (SNR) EPI-based field maps, rapidly acquired using a series of gradient recalled images collected across a range of TE values. This field map is distorted in the same manner as the EPI images to be unwarped, providing a direct look-up table for the correct location of each pixel of data. This method adds very little scan time and is robust and easy to implement.

Neural correlates of artificial grammar learning.

Artificial grammar learning (AGL) is a form of nondeclarative memory that involves the nonconscious acquisition of abstract rules. While data from amnesic patients indicate that AGL does not depend on the medial temporal lobe, the neural basis of this type of memory is unknown and was therefore examined using event-related fMRI. Prior to scanning, participants studied letter strings constructed according to an artificial grammar. Participants then made grammaticality judgments about novel grammatical and nongrammatical strings while fMRI data were collected. The participants successfully acquired knowledge of the grammar, as evidenced by correct identification of the grammatical letter strings (57.4% correct; SE 1.9). During grammaticality judgments, widespread increases in activity were observed throughout the occipital, posterior temporal, parietal, and prefrontal cortical areas, reflecting the cognitive demands of the task. More specific analyses contrasting grammatical and nongrammatical strings identified greater activity in left superior occipital cortex and the right fusiform gyrus for grammatical stimuli. Increased activity was also observed in the left superior occipital and left angular gyrus for correct responses compared to incorrect. Comparing activity during grammaticality judgments versus a matched recognition control task again identified greater activation in the left angular gyrus. The network of areas exhibiting increased activity for grammatical stimuli appears to have more in common with studies examining word-form processing or mental calculation than the fluency effects previously reported for nondeclarative memory tasks such as priming and visual categorization. These results suggest that a novel nondeclarative memory mechanism supporting AGL exists in the left superior occipital and inferior parietal cortex.