Brain, Vol. 124, No. 11, 2188-2202,
November 2001
© 2001 Oxford University Press
Dissociation between intentional and incidental sequence learning in Huntington's disease
1 Department of Psychology, Institute of Psychiatry, King's College London, 2 Department of Psychology, University College London, UK and Departments of 3 Neurology and 4 Molecular Genetics, Hospital Virgen Macarena, Sevilla, Spain
Correspondence to:
Dr R. G. Brown, Department of Psychology, Institute of Psychiatry, King's College London, De Crespigny Park, London SE5 8AF, UK E-mailr.brown{at}iop.kcl.ac.uk
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
The ability to acquire and act upon serial order information is fundamental to almost all forms of adaptive behaviour. There is growing evidence that such knowledge may be acquired through a number of different means, each perhaps with its own neuronal substrate. One major distinction is between serial order information acquired intentionally and leading to explicit conscious knowledge of the sequence structure, and information acquired incidentally through experience. While this latter form of knowledge influences behaviour, it may do so without the participant being aware of the sequential information, i.e. it is acquired implicitly. Evidence from physiological and lesion studies in animals and imaging studies in humans suggests that these two forms of learning may have dissociable neuronal substrates. Specifically, the striato-thalamo-cortical circuit centred on the caudate nucleus is proposed to be involved in intentional sequence learning and that based on the putamen on incidental learning. The present study tested one part of this proposed dissociation by assessing patients with Huntington's disease on tasks of the two forms of learning. On the test of trial-and-error intentional learning there were marked deficits, which were closely related to disease progression and to measures of executive cognitive dysfunction. This finding was in contrast to the finding from the incidental learning task. Performance of the Huntington's disease group was essentially normal and unrelated to measures of disease progression and cognitive status. The results, although supportive of the proposed dual-system hypothesis, offer only partial confirmation. Further direct study is required using similar tasks in patients with putamenal disorder or lesions within the skeletomotor striato-thalamo-cortical circuit.
striatum; cognition; SRT task; implicit learning; trial-and-error learning
SMA = supplementary motor area; SRT = serial reaction time; RT = reaction time; UHDRS = Unified Huntington's Disease Rating Scale; MMSE = Mini-Mental State Examination; WCST = Wisconsin Card Sorting Test
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The ability to acquire, retain and utilize knowledge about serial order is fundamental to survival and adaptive behaviour. Recording neuronal activity in behaving animals has shown that many brain regions may encode information related to serial order. Neurones showing activation during the processing of serial order information have been identified in the lateral prefrontal cortex (Barone and Joseph, 1989
), premotor cortex (Mushiake et al., 1991), supplementary motor area (SMA) (Mushiake et al., 1990), anterior cingulate (Procyk et al., 2000) and even the primary motor cortex (Carpenter et al., 1999). Subcortically, such neurones have been identified in the globus pallidus (Mushiake and Strick, 1995), caudate nucleus (Kermadi and Joseph, 1995) and dentate nucleus (Mushiake and Strick, 1993). In animals, lesions or transient inactivations within the same cerebellar– and basal ganglia–thalamo-cortical circuits can interfere with the execution of sequential movements above and beyond the individual movement elements (e.g. Benita et al., 1979; Van den Bercken and Cools, 1982; Mushiake et al., 1990; Cromwell and Berridge, 1996). In humans also, lesions or pathophysiological changes to the same circuits can lead to significant impairment in the execution of action sequences (Dick et al., 1986; Benecke et al., 1987). Such combined evidence points to an important role for frontal cortical regions and associated subcortical structures in encoding serial order information and in the execution of skilled (i.e. pre-learned) sequential actions. However, it does not address the question of whether the same areas are involved in the acquisition of that information.
Animal studies of trial-and-error sequence learning
Following earlier studies of simple stimulus–response learning (for reviews, see Graybiel, 1995; Schultz, 1999) researchers have begun to address the learning of serial order information per se. Using a 2 x 5 hyperset task (Hikosaka et al., 1995), Nakamura and colleagues observed task-related activity in medial frontal neurones, particularly in the pre-SMA, in primates during trial-and-error sequence learning (Nakamura et al., 1998). Anatomically, the pre-SMA receives input from the prefrontal cortex and projects to the cortical and spinal motor pathways only through the SMA itself (Luppino et al., 1993). The pre-SMA also projects to the caudate nucleus, while the main striatal target of SMA output is the putamen (Parthasarathy et al., 1992). Miyachi and colleagues have shown that transient inactivation of the anterior striatum, which receives input from the the pre-SMA and dorsolateral prefrontal cortex, disrupts new sequence learning but not the performance of established sequences (Miyachi et al., 1997). The reverse pattern is observed after disruption of the middle posterior putamen, which receives its input from the SMA proper, the premotor cortex and the primary motor cortex.
Evidence from functional imaging
A number of studies have examined the neuroanatomical substrate of trial-and-error learning in humans (Ghilardi et al., 2000), including the use of a 2 x 10 hyperset task (Hikosaka et al., 1996; Sakai et al., 1998). In general, such studies support the involvement of the SMA/pre-SMA and dorsolateral prefrontal cortex in learning new sequences, often together with activation from other areas, including the cingulate cortex, the premotor cortex and the cerebellum. Evidence of basal ganglia activation, however, has been more elusive. When found, putamen activation tends to occur in both new learning and in the performance of learned sequences (Jenkins et al., 1994; Toni et al., 1998), although caudate activity has been related more clearly to trial-and-error learning (Jueptner et al., 1997; Toni et al., 1998).
In combination, the imaging and animal evidence offers strong support for a network of cortical and subcortical regions being involved in the acquisition of serial order information (specifically the anterior striatum, prefrontal cortex and pre-SMA) and with a separate parallel network being involved in the execution of serial tasks once they have become learned (posterior striatum, SMA and parietal cortex) (Graybiel, 1995). These circuits map closely on to the cognitive (dorsolateral prefrontal) and skeletomotor circuits described by Alexander (Alexander, 1994).
Incidental sequence learning
Trial-and-error paradigms encourage an active exploratory approach to learning. These draw heavily on executive processes and working memory, consistent with the involvement of prefrontal and associated cortical and subcortical regions. Although such processes may offer advantages for rapid learning, it is clear that serial order information can be acquired by other means. Evidence from lower species and even infant primates suggests that learning may occur through the repeated pairing (under conditions of reinforcement) of particular stimulus and response combinations. This notion is at the heart of the contemporary description of implicit or `incidental' learning in humans, which is manifest through gradual incremental changes at the behavioural level without (necessarily) any awareness of what is being learned or how it is learned.
Research in this area has been dominated by the serial reaction time (SRT) paradigm, first described by Nissen and Bullemer (Nissen and Bullemer, 1987). Participants ostensibly perform a choice reaction time (RT) task but, unknown to them, a particular series of stimuli is repeated. With such repetition, RT decreases reliably. When the subject is given a block of randomly ordered stimuli, RT increases again, providing indirect evidence that sequential knowledge is being acquired.
Although not informed of the sequential nature of the task, some participants nevertheless become aware of the repetition and may acquire explicit knowledge of some or all of the sequence. This leads to greater improvements in RT, with a more predictive response mode. It is presumed that awareness gives the participant the option to bring in alternative active strategies for acquiring the sequential knowledge, just as in trial-and-error learning.
The SRT task has been used widely in functional imaging studies (Grafton et al., 1995; Rauch et al., 1995, 1997; Doyon et al., 1996, 1997; Hazeltine et al., 1997; Jenkins et al., 1997; Honda et al., 1998), although procedural variations between studies, particularly in relation to the relative contributions of awareness and explicit processes to learning, make direct comparison across studies difficult.
Jenkins and colleagues scanned participants performing a task under incidental learning conditions and found that improvement in RT was positively correlated with changes in blood flow in contralateral motor and premotor areas and in the contralateral putamen, but negatively correlated with changes in the bilateral dorsolateral prefrontal cortex (Jenkins et al., 1997). Grafton and colleagues blocked the development of awareness by requiring subjects to perform a concurrent attention-demanding task. Incidental learning under these conditions was associated with increased blood flow in contralateral sensorimotor cortex and SMA, and putamen bilaterally (Grafton et al., 1995).
Evidence from other studies using variants of the SRT task has been broadly consistent. In general, performance under incidental learning conditions without awareness is associated with cortical activation in areas involved in motor circuitry; less involvement of the more anterior frontal structures is found when participants learn with awareness or intentionally by trial-and-error. Similarly, at the level of the striatum, a broad dissociation is observed between activity within the putamen and caudate nuclei under the two learning conditions. The role of the putamen in learning under incidental conditions is well illustrated by the study of Rauch and colleagues, who observed a strong linear association between the magnitude of the signal intensity change in the putamen and behavioural indices of incidental learning (Rauch et al., 1997).
As noted elsewhere (Brown, 1999; Hikosaka et al., 1999), the circuitry activated during incidental learning is the same as that involved in the execution of sequences learned previously by a process of trial-and-error. In other words, the circuitry involved in the relatively automatic execution of skilled action may itself be capable of slow incremental and `unconscious' learning.
Evidence from clinical studies
The final source of evidence to consider comes from the study of patients with focal or degenerative lesions within the circuits that appear to underlie the different forms of serial order learning. Despite the clear practical and theoretical distinction between intentional and incidental learning, only the latter has been the subject of any significant study to date. Parkinson's disease involves the loss of ascending nigrostriatal dopamine, with the most severe loss in the putamen. Pathophysiological changes in cortical function are observed in other parts of the motor circuit, particularly the SMA. Parkinson's disease therefore provides a useful model to study the role of the motor circuit in incidental sequence learning. Surprisingly, however, the evidence is far from conclusive. In two studies, very clear evidence of incidental learning deficit has been found in patients with Parkinson's disease (Jackson et al., 1995; Stefanova et al., 2000). In other studies, results have been more equivocal (Ferraro et al., 1993; Pascual-Leone et al., 1993; Sarazin et al., 1995; Westwater et al., 1998; Sommer et al., 1999; Helmuth et al., 2000). Patients have tended to show an apparent attenuation of learning compared with controls, the effect being significant in most, but not all, studies. In some studies, there is evidence for dissociation between subgroups of patients, some showing worse performance than others. For example, in a study by Jackson and colleagues, the most marked impairment was shown by patients with evidence of executive deficits (Jackson et al., 1995), suggesting that dysfunction outside the motor circuit may underlie the observed behavioural impairment.
The other disease relevant to the issues being addressed here is Huntington's disease, in which the primary pathology is degeneration of striatal medium spiny projection neurones, initially in the more dorsomedial regions and spreading ventrolaterally with progression of the disease. Thus, early pathological changes are most evident in the caudate nucleus, the putamen only being affected later and the ventral striatum largely spared (Hedreen and Folstein, 1995; Mitchell et al., 1999). Although cortical pathology occurs only late in the course of the disease, metabolic changes in the prefrontal cortex are observed from an early stage (Weinberger et al., 1988; Kuwert et al., 1990), perhaps as a result of disordered striatal outflow.
From the pathophysiological model presented earlier, such a pattern should not be associated with any marked impairment in incidental sequence learning, at least in the early stages of the disease. In fact, both studies to date have shown significant deficits (Knopman and Nissen, 1991; Willingham and Koroshetz, 1993). In the study by Knopman and Nissen, the results were reminiscent of those of studies in Parkinson's disease, some but not all patients showing impairment. The overall picture is one of attenuated rather than absent learning.
In contrast to the number of studies on incidental learning in Parkinson's disease and Huntington's disease, no study has assessed intentional or trial-and-error sequence learning in either disease. Given the equivocal nature of the incidental sequence learning deficit in Huntington's disease and the absence of any comparable evidence of intentional learning, a study was planned to compare these two forms of learning directly in a group of patients with Huntington's disease. On the basis of evidence from other sources, we predicted that the patients would show a marked impairment in intentional learning and that the impairment would be closely related to clinical indicators of disease pathology and associated with other aspects of cognition dependent upon the `cognitive' frontostriatal circuits. In contrast, we predicted only a minimal effect on incidental learning and that any impairment observed would be less consistently associated with clinical or neuropsychological indicators of caudate circuit pathology or pathophysiology. A counter-hypothesis is that the caudate circuit is involved in both intentional and incidental learning. In this case, we would expect to find impairments in both forms of learning in Huntington's disease, each form having similar relationships to indicators of disease progression.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Participants
Sixteen patients with Huntington's disease participated in the study. All had genetically confirmed Huntington's disease, with a mean CAG repeat length of 47.6 ± 4.5 (range 40–54). All showed clinical evidence of symptomatic disease at the time of testing. The patients were recruited through the neurology and genetics services in Sevilla, Badajoz, Cadiz and Jerez, Spain. Only those judged able to cope with two sessions of intensive assessment were asked to participate. The group comprised 12 males and four females [mean age 42.8 ± 12.2 years (range 23–62 years)]. The average approximate duration of Huntington's disease was 6.4 ± 3.0 years (range 1–12 years). Eleven were taking medication associated with their Huntington's disease. Of these, 10 were taking neuroleptics, six benzodiazepines and three selective serotonergic reuptake inhibitors. Those taking neuroleptics had more severe motor symptoms [t(13.7) = 2.63, P < 0.05] but were otherwise matched with the remaining sample.
Twelve of the patients were right-hand-dominant, two were left-handed and two had mixed hand dominance. The mean number of years of education was 9.3 ± 4.8 (range 1–18). This relative lack of education reflected the low level of formal provision in rural areas of Spain at the time of the participants' childhood. All were, or had been, employed in either unskilled manual or semi-skilled occupations.
All patients were assessed by the same investigator (L.R.-V.) using the Unified Huntington's Disease Rating Scale (UHDRS) (Huntington Study Group, 1996). The mean score for the motor section was 41.8 ± 19.6 (range 20–82). A five-point disease stage was assigned, based on the total functional capacity score (Shoulson, 1981), stage I representing functional independence and stage V the need for total care. Five of the 16 patients were in stage I, two were in stage II, seven were in stage 3 and one was in stage IV. In addition, three asymptomatic gene-positive participants were also assessed. With the exception of their disease status (all were in stage I with no significant motor, cognitive or behavioural signs), they were similar to the symptomatic patients. Selected aspects of their performance will be considered in the Results and Discussion sections.
Sixteen control participants were recruited from a number of sources, including hospital staff and genetically confirmed gene-negative members of the patients' families (CAG repeat length <32). Controls and patients were matched as closely as possible for age, sex, handedness, years of education and class of occupation. The control group comprised 12 males and four females, mean age 43.7 ± 12.6 years (range 23–68 years). Fifteen were right-hand-dominant and one was left-handed. The mean number of years of education was 8.5 ± 3.6 (range 4–16). As with the patients, all were employed in either unskilled manual or semiskilled occupations. All participants gave informed consent.
Apparatus
The experiments were controlled by a personal computer modified to enable millisecond timing. Response input was through a set of five keyboard-style keys (1 x 1 cm). In both tasks, the keys were arranged in a cross with the arms oriented up (U), down (D), left (L) and right (R). The fifth key, in the centre, served as a home key in the SRT task only. The centres of the four response keys were positioned 2 cm from the centre of the home key. The arrangement of response keys corresponded to the positions of the on-screen visual stimuli in both tasks.
Incidental learning: SRT test
The test was based on that described by Nissen and Bullemer (Nissen and Bullemer, 1987) and was presented to the participant as a test of RT. The participants were asked to press the central response key with the index or middle finger of their dominant hand. After a 500 ms delay, one of the four stimulus positions was clearly highlighted on the screen: up (U), down (D), left (L) or right (R). The participant was instructed to move their finger quickly and accurately to the appropriate response key, and then to return to the home key and wait for the next stimulus. Following instruction, participants were given a practice of 20 trials with randomly ordered stimuli. The main test comprised nine blocks of 120 trials each, with a short break between each if required. Blocks 3–7 and 9 consisted of 12 repetitions of a 10-item sequence (DULRDLUDRL), while Blocks 1, 2 and 8 each comprised a pseudo-random series of trials with the same zero-order probabilities of the positions used in the repeating sequence blocks.
RT and accuracy of response were noted for each trial, and the median RT and total number of errors were calculated for the block. Two summary measures were calculated to indicate changes in performance: (i) the difference in RT between the first repeating sequence block (Block 3) and the fifth (Block 7) (SRT1); and (ii) the difference between a pseudo-random block (Block 8) and the mean of the adjacent sequence Blocks 7 and 9 (SRT2).
Following the final block, participants were given a debriefing interview based on that described by Willingham and Koroshetz (Willingham and Koroshetz, 1993). Participants were asked if they were aware of any sequence and, if so, when they first became aware of it, when it occurred and to reproduce it. Finally, they were asked to make 120 key-presses reproducing the sequence or a series of responses that `felt similar' to what they had been making during the experiment (Free Generate task). We assessed explicit knowledge according to two criteria: the number of complete sequences produced and, more liberally, the production of eight out of 10 correct items in a row.
Intentional learning: trial-and-error learning test
Intentional learning was assessed on a subsequent occasion. Participants were asked to learn a sequence that would repeat throughout the test and would not change. The sequence consisted of eight items (URLDUDRL), but this information was not provided to the participants. The test commenced with the first position in the array being clearly highlighted on screen. Participants had to work out what position came next and to respond by pressing one of the keys corresponding to the three possible stimulus positions. If the response was incorrect, the participant received a brief auditory feedback tone and the highlighted stimulus remained in the same position. The participant was then asked to make another choice. This continued until the correct response was made, when the new position became highlighted. Participants continued in this way for each trial, working out the next item in the repeating sequence. Further verbal instruction and encouragement was given until the nature of the task was understood clearly. Each block comprised 10 repetitions of the eight-item sequence presented without a break. For each block, two statistics were calculated: Total Correct, the number trials in which the response was correct on the first attempt; and Maximum Length, the maximum number of consecutive correct responses made without error. The criterion for learning was set at 90% total correct responses in the block. In addition, two less stringent measures were calculated retrospectively from the numbers of blocks required to reach accuracies of 75 and 50% correct.
Testing ended when the participant reached criterion or 10 blocks were completed. When 90% criterion was reached before Block 10, subsequent blocks were credited with the participant's performance at the point when criterion was reached. It should be noted that this was a conservative strategy as it prevented performance measures improving further with practice.
Cognitive assessments
Neuropsychological assessments comprised a Spanish translation of the Mini-Mental State Examination (MMSE) (Folstein et al., 1975), verbal fluency for the letters F, A and S (Controlled Oral Word Association Test) (Benton and Hamsher, 1976), the Symbol Digit Modalities Test (Smith, 1968), the Stroop Interference Test (Stroop, 1935), the Wisconsin Card Sorting Test (WCST) (Nelson, 1976), the digit span subtest of the WAIS–R [Wechsler Adult Intelligence Scale (Wechsler, 1981)] and a computerized version of the Corsi blocks test (Milner, 1971). This last test used five non-symmetrical screen positions and corresponding response keys. The two span tasks were administered and scored in a similar manner for both forward and backward span. Participants were given two attempts at each span length, starting at 2. Maximum forward and backward spans were calculated.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The Huntington's disease and control groups did not differ in mean age [t (30) = –0.21, P > 0.10] or years of education [t (30) = 0.50, P > 0.10]. There were no differences in the distributions of sex [
2 (1) = 0, P = 1.0] and handedness [2 (1) = 2.7, P > 0.10].
Cognitive assessment
The mean performances of the two groups are shown in Table 1. The patient group showed an overall reduction in cognitive function as indexed by the MMSE. Six of the patients scored <25 (minimum, 19). All of the controls scored 
26, appropriately for their age and level of education (Crum et al., 1993).
|
Significant impairments were noted on all of the tests of psychomotor function, attention, executive function and working memory. The only exceptions were the WCST, in which the patients were unimpaired relative to the control group, and on forward digit span. Backward span was impaired for both the verbal and spatial task. On the Stroop task, patients were slower in each of the three conditions. The most marked slowing was in the reading of non-interfering colour words. There was little indication of a specific interference deficit. Differences between conditions suggested that the controls showed a greater decrement than patients in performance in the interference task compared with simple colour-naming. However, when the ratio of performances in the two tasks was examined, no significant difference was found.
Serial RT test
RT data
Figure 1A shows the mean RTs for the Huntington's disease and control groups, for the nine blocks of trials plotted against different 500 ms ranges. Repeated measures ANOVA (analysis of variance) using the GLM (general linear model) procedure from SPSS (Statistical Package for Social Sciences, 1998) revealed a main effect for group (Huntington's disease versus controls) [F(1,30) = 17.19, P < 0.001] and for block (1–9) [F(1.7,50.1) = 18.75, P < 0.001]. There was also a significant group x block interaction [F(1.7,50.1) = 4.62, P < 0.05]. Post hoc analyses revealed that both groups showed significant changes in RT across the nine blocks.
|
Subsequent analyses were carried out on the data from specific sets of blocks relating to learning effects. The patients remained significantly slower than the controls throughout the task, and so only the block effect and group x block interactions will be reported. Analysis of the data from Blocks 3–7 (SRT1) revealed the same pattern as the previous analysis, with significant main effects of block and group x block interaction. Separate analyses revealed a significant decrease in RT for both patients (122 ± 150 ms) and controls (37 ± 27 ms). However, the magnitude of the change in the patients is, in part, a reflection of the high initial value. Calculating the change in RT as a proportion of Block 3 RT revealed a comparable improvement of 15 ± 15% for the patients and 10 ± 7% for the controls [t(21.2) = 1.3, P > 0.10].
The next analysis concerned Blocks 7, 8 and 9 (SRT2). The mean data for these blocks are shown in Fig. 1B
on expanded y-axes. There was a significant main effect of block [F(2,60) = 8.91, P < 0.001] but no significant group x block interaction [F(2,60) = 1.31, P > 0.10]. The two groups showed similar increases in RT for the pseudo-random block (SRT2) (Huntington's disease 30.6 ± 50.4 ms, controls 33.0 ± 29.7 ms). Figure 2 shows the individual data for the 16 patients for SRT2, plus the data for the three gene-positive asymptomatic patients and the 16 controls. As can be seen, the performances of the large majority of patients lay within the normal range, most showing a normal increase in RT for the pseudo-random block. These data are plotted against the rating of total functional capacity and reveal a lack of any consistent relationship between incidental learning and disease progression (see Table 2 and correlational analysis below).
|
|
Error data
The mean error rate across the nine blocks was significantly higher in the Huntington's disease group (mean 9.9 ± 6.8 errors per block) than the controls (mean 2.4 ± 2.0 errors per block) [F(1,30) = 6.52, P < 0.05]. However, there was no significant effect of block [F(2.9,87.3) = 1.3, P > 0.10] or group x block interaction (F < 1). The high mean error rate was due largely to four patients. However, repeating the main analyses after excluding these participants did not alter the results.
Awareness
On the basis of the post-task interview, participants were classified as those who (i) reported that they were unaware of any systematic pattern in the order of the trials; (ii) claimed to have noticed some non-randomness but could give no further information, or who gave incorrect information; (iii) noticed some non-randomness and could indicate approximately but correctly where it occurred in the series of blocks; and (iv) were fully aware of the sequence by the end of the task. In practice, none of the participants fell into the last category, and this was confirmed by the poor performance of all participants on the Free Generate task. Most participants were classified in the first two categories (Huntington's disease, 14 out of 16; controls, 12 out of 16). The remainder, comprising two patients and four controls, fell into the third category (partial awareness). Excluding this last group of participants had no effect on the overall pattern of results. Regardless of subjective awareness, no participant in either group was able to reproduce the complete 10-item sequence in the Free Generate task, or could even reproduce eight out of 10 items in a row without error.
Correlations
For the purposes of correlational analysis, we used SRT2 as the main index of learning. Non-parametric (Spearman) correlation coefficients were calculated and significance was assessed using a one-tailed test adjusted for multiple analyses (
level, P < 0.002). The analysis was performed for the Huntington's disease group only, to assess the relationships between incidental learning and the main clinical and cognitive measures. To extend both the sample size (n = 19) and the clinical range, the three pre-symptomatic patients were included. As shown in Table 2, incidental learning was only weakly related to the main demographic and clinical variables. Less learning tended to be associated with increased age, disease duration and disease severity, although no association was significant. None of the cognitive measures showed any consistent or significant relationship with incidental learning, even at the more liberal level of P < 0.05.
Trial-and-error learning test
Block 1 performance
Data were collapsed within the 10 sets of the eight-item sequence and the total number of errorless responses (i.e. correct on first attempt) was computed. The data for the two groups are shown in Fig. 3. Repeated measures ANOVA revealed significant main effects of group [F(1,30) = 18.79, P < 0.001] and set [F(6.6,197.1) = 4.51, P < 0.001] and a significant group x set interaction [F(6.6,197.1) = 2.18, P < 0.05]. Post hoc analyses revealed a significant effect of block for the control group [F(5.3,79.3) = 5.19, P < 0.001] but not the Huntington's disease group [F(5.3,79.3) = 1.15, P > 0.10]. As shown in Fig. 2, the mean performance of the Huntington's disease group showed little variation from chance performance (2.67 correct per eight items) over the course of the first block. In contrast, the controls showed a marked improvement, at least up to sets 7 and 8.
|
Performance across Blocks 1–9
While all controls either reached criterion or completed 10 blocks, three of the Huntington's disease patients abandoned the task before the end. However, all completed at least five blocks. Because of this, two sets of analyses were performed: first, the performance over Blocks 1–5 with all participants; secondly, performance across 10 blocks in the reduced sample of 13 patients.
The mean percentage correct responses per block (Blocks 1–5) for the two groups are shown in Fig. 4A. Repeated measures ANOVA revealed significant main effects for group [F(1,30) = 27.1, P < 0.001] and block [F(2.8,85.4) = 28.3, P < 0.001] and a significant group x block interaction [F(2.8,85.4) = 5.8, P < 0.01]. Although the mean improvement in the Huntington's disease group was less than that shown by controls, post hoc analysis revealed significant change across the five blocks. However, the group data masks considerable variability between participants. Figure 7 shows individual patient data plotted against total functional capacity score. The horizontal dashed line indicates chance performance (133 correct responses in 400 trials). As can be seen, a number of patients failed to show any marked improvement above this level and a similar number showed performance within the normal range.
|
|
The results for participants who reached criterion or completed all 10 blocks (Fig. 4B
) revealed essentially the same pattern. An exception was the absence of a significant interaction effect [F(1,27) = 2.2, P < 0.10]. However, this finding is largely an artefact of the asymptotic performance of the control group as they reached criterion and stopped the test. The Huntington's disease group, in contrast, showed only slight improvement over the later blocks despite continuing with the test in the majority of cases.
The next variable to be analysed was the maximum number of consecutive correct responses per block. The data for Blocks 1–5 are shown in Fig. 5A. Analysis revealed a significant main effect for group [F(1,30) = 11.7, P < 0.001] and block [F(2.1,62.0) = 12.6, P < 0.001] and a significant group x block interaction [F(2.1,62.0) = 5.8, P < 0.01]. Post hoc analysis indicated that the Huntington's disease group failed to show a significant increase above chance in the maximum number of consecutive correct responses, in contrast to the controls.
|
The same pattern was revealed for participants who completed the task for Blocks 1–10 (Fig. 5B
). As with the percentage correct measure, it should be noted that the data for the control participants are an underestimate of learning potential. In contrast, the data for the Huntington's disease patients are a true reflection of their poor performance. Post hoc analysis revealed that, although their performance improved slightly across the 10 blocks, the effect was not significant.
Proportion of participants reaching criterion
Figure 6 shows the cumulative proportions of participants in the two groups reaching the experimental criterion of 90% correct and the more liberal criteria of 75 and 50% correct. Only two patients reached the 90% criterion compared with 13 controls. Of the participants who reached the 90% correct criterion, one patient reached criterion within five blocks compared with seven controls. Relaxing the criterion to 75% had little effect on the Huntington's disease group but led to a dramatic improvement in the performance profile of the controls. All but one of the controls achieved the 75% level within 10 blocks compared with only three of the patients. Only when the criterion was relaxed to 50% (chance level = 33%) did the Huntington's disease group show any appreciable success, 11 reaching this level by Block 10. Most of the controls achieved this level within the first block and all but one by the end of Block 2. The best discrimination between the groups was found by applying the 75% criterion with a cut-point of Block 6. This correctly classified 87.5% of Huntington's disease and controls, misclassifying two members of each group.
|
Correlations
For correlational analysis, the total number of correct responses made over the first five blocks was taken as a summary measure of learning. The results of the analysis are shown in Table 2
. A large number of significant correlations were found (one-tailed test, < 0.002). Impaired learning was associated with increased disease duration, motor symptom severity and reduced functional capacity. Figure 7 shows the clear deterioration in performance with disease progression. The figure also illustrates the normal performance of the asymptomatic participants and that of the more mildly affected patients relative to the controls. Learning was also associated with cognitive/motor decline as assessed by the symbol digit modalities test, the Stroop test and the categories measure of the WCST. The association between learning and verbal fluency and total errors on the WCST approached significance. In the span tests, learning performance was significantly associated only with backward span measures. |
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The present study is the first to compare directly the relative abilities of patients with striatal disease to acquire serial order information through incidental (SRT) and intentional (trial-and-error) procedures. The results for the Huntington's disease group support the hypothesis that the caudate nucleus and associated cortical circuits may have greater involvement in the intentional learning of sequential information than in incidental learning.
Incidental sequence learning
The primary index of learning was the increase in response time to a block of pseudo-random trials compared with adjacent blocks of the repeating 10-item sequence (SRT2). As a group, the Huntington's disease patients showed evidence of learning by a significant increase in RT (30.6 ms), which was similar to the 33.0 ms difference shown by the controls and the 28.3 ms effect shown by the three asymptomatic gene-positive participants. From the individual data (Fig. 2) some participants in all groups failed to show a reliable learning effect, and overall the size of the mean learning effect was small. For instance, Knopman and Nissen found a mean difference of >100 ms in their controls (Knopman and Nissen, 1991), Willingham and Koroshetz a difference of ~70 ms (Willingham and Koroshetz, 1993) and Jackson and colleagues one of 74 ms (Jackson et al., 1995). It is likely that the variation between studies reflects methodological or procedural differences. For example, the fact that subjects in the present study returned their finger to the home key after each response may have punctuated the serial behaviour and reduced the degree of learning.
The absence of a significant group difference in learning in the present study is in contrast with the two previous studies in Huntington's disease (Knopman and Nissen, 1991; Willingham and Koroshetz, 1993). However, in both of these studies the evidence for impairment and the absolute degree of impairment in the Huntington's disease patients is uncertain. In the first study, 13 Huntington's disease patients showed a mean learning effect of 78 ms compared with 103 ms in controls. When individual response patterns were analysed, the majority of patients (eight out of 13) showed the expected increase in RT with the pseudo-random block. In the study by Willingham and Koroshetz a more robust result was obtained. The precise magnitude of the learning effect is difficult to judge from the paper as the data were presented as log RT. However, estimating from the figure suggests an effect of ~68 ms for controls and 28 ms for the patients. Only one control reported awareness of any non-randomness. This low level of awareness and more evident impairment in the patients may be due to the use of a 12-item sequence compared with the 10-item sequence used in the present study and in the Knopman and Nissen study.
In combination, the present study and others in patients with striatal dysfunction indicate that the incidental learning of sequences using the SRT task is remarkably robust, and any impairment may be more a matter of degree rather than being absolute. A more convincing demonstration would be the abolition of a learning effect rather than just an attenuation. Such a deficit has been demonstrated in patients with cerebellar degeneration (Pascual-Leone et al., 1993), who show no evidence of learning even with short sequence lengths.
Another reason for the equivocal nature of the implicit learning deficit observed in striatal disease may relate to the subtotal nature of the `lesion', particularly in Parkinson's disease where striatal dysfunction is physiological rather than structural and a considerable degree of compensation may be possible. In contrast, there is some evidence that more complete lesions within the motor circuit may abolish incidental sequence learning effectively, either in the globus pallidus (Brown et al., 1998) or the SMA (Ackermann et al., 1996).
Can the same argument for a subtotal lesion be applied to Huntington's disease? Unlike Parkinson's disease, Huntington's disease causes striatal degeneration that increases in severity with disease progression (Aylward et al., 1997). If caudate lesion volume is a significant factor in the magnitude of any SRT learning deficit, one would expect to find a clear relationship between the learning effect and clinical indicators of disease severity. Knopman and Nissen found no significant association between motor or cognitive symptom severity and SRT learning (Knopman and Nissen, 1987). In the present study, total functional capacity, a widely used measure of such progression (Marder et al., 2000) was also unrelated to SRT learning (Fig. 2): even patients in advanced stages of the disease showed performance within the normal range. Also striking in these results was the absence of any significant association between incidental learning and cognitive dysfunction, either globally (MMSE) or on specific tests. The lack of association between SRT learning and indices of frontal lobe dysfunction is surprising, given the results of studies such as that of Jackson and colleagues, who found the most marked SRT impairment in a group of Parkinson's disease patients with executive deficits (Jackson et al., 1995). However, this may relate more to the ability of patients with awareness to use explicit strategies to learn the sequence. In the present study, awareness was low in both groups, making executive function largely irrelevant for any learning that took place. The relatively normal performance of the patients in SRT learning and the independence of this learning from clinical indicators of disease severity or progression together argue against the caudate nucleus or other structures that are implicated in Huntington's disease having a major role in incidental serial order learning.
Intentional sequence learning
A quite different picture emerged for trial-and-error learning. As a group, the patients were at chance level in the first block, when controls had already shown significant evidence of learning. In the patients, accuracy improved slowly but steadily over subsequent blocks but never approached the level shown by controls. However, such learning may indicate the ability to learn particular simple stimulus–stimulus associations without necessarily implying the acquisition of higher-order sequence knowledge. For instance, in the sequence used (URLDUDRL), a response to the right-hand button was invariably followed by a response to the left-hand button. Participants who learn this simple fact could quickly achieve a 20% improvement in accuracy, without this implying a significant degree of higher-order sequential learning. A better measure of this learning is the ability to produce a complete sequence of eight items without error. By this stricter criterion, the performance of patients with Huntington's disease failed to rise above baseline until well into the experiment (Fig. 5), whereas controls had already shown above-chance levels of sequence knowledge in the first block. The true measure of the deficit was reflected in the fact that only three of 16 patients reached criterion performance of 75% accuracy within 10 blocks (800 trials) compared with fifteen of 16 controls.
Also distinct from the results of incidental learning, summary measures of intentional learning were strongly associated with indices of disease severity/progression and various cognitive measures. In terms of total functional capacity, there was a strong linear association between decreasing functional capacity and decreasing learning (Fig. 7). Also evident from Fig. 7 is the fact that the patients with the mildest disease and the three asymptomatic gene-positive participants all had performance within the normal range. The possible implications of this for the role of non-caudate pathology in learning will be discussed below.
As noted in the Introduction, intentional trial-and-error sequence learning has not been studied previously in Huntington's disease, although marked impairments have been reported in another task involving trial-and-error (conditional associative learning) (Sanchez-Pernaute et al., 2000). The results of the present study are also consistent with anecdotal information provided by Willingham and Koroshetz (Willingham and Koroshetz, 1993). Following an SRT task, patients were given 40 trials of trial-and-error learning for the sequence presented previously. Although no data are presented, the authors state that they discontinued testing as patients showed no evidence of learning and became distressed.
Processes underlying impaired intentional sequence learning in Huntington's disease
The aim of cognitive neuroscience research is to determine the representations and processes underlying cognition, and their neuronal bases. Inevitably, the processes revealed will depend upon the nature of the tasks being employed. Thus, the processes underlying sequence learning through trial-and-error may be different from those involved in paradigms in which a participant learns a short sequence and gradually builds up elements to make a longer sequence. Interestingly, this latter form of incremental sequence learning also seems to be impaired in Huntington's disease, deficits being reported even in currently asymptomatic gene-positive individuals (Blackmore et al., 1995). In that study, supra-span sequence learning was the only one of a wide range of neuropsychological measures that distinguished gene-positive from gene-negative participants.
The imaging evidence and animal physiological evidence suggests that trial-and-error sequence learning involves several brain regions associated with attention, working memory and executive processes. This is consistent with the pattern of cognitive impairment that is seen in Huntington's disease and accompanied the sequence learning deficit in the present study. Executive function, however, is not a unitary processes but involves many components, each perhaps with its own neuronal substrate. We might expect working memory, particularly spatial working memory, to play a significant role in the task employed in the present study. Evidence reviewed by Lawrence and colleagues suggests that patients with Huntington's disease are impaired both in aspects of simple spatial span and in the short-term retention and manipulation of spatial information (Lawrence et al., 1998a). Data on the span test in the present study support this conclusion, although parallel deficits were found on the verbal test, suggesting that the problem is not modality-specific. A reduced spatial span would be expected to affect the ability to learn a complex sequence. However, even in controls, the mean spatial span (4.6) was considerably shorter than the sequence to be learned. This suggests that the ability of participants to learn supra-span sequences was more important than immediate span capacity. This is reflected in the non-significant correlation between spatial span and the main summary measure of intentional learning (Table 2). However, larger and significant correlations were obtained for the two backward span measures (verbal and spatial), indicating the importance of the working memory processes of on-line storage and manipulation to the sequence learning.
Another executive process that may be relevant to task performance is inhibitory control. With an eight-item sequence made up of only four possible responses, there is no simple one-to-one correspondence between one response and the following one. Thus, in the sequence used (URLDUDRL), a response to the Upper button is followed by a response to the Right and Down buttons at different points. A participant who perseverates the U–R pairing might have problems learning the whole sequence. However, although perseverative behaviour is commonly found in Huntington's disease (Lawrence et al., 1998a), there was little evidence of it in the present study. WCST performance was poor in both patients and closely matched controls, and appeared to be unrelated to sequence learning. The Stroop test also provides a measure of inhibitory control. However, rather than being specifically impaired in the interference task, the patients showed a general slowness across all conditions, with no difference in the ratio of interference/simple colour-naming between the two groups. The results of this and the other cognitive tests are in line with those reported previously on a large cohort of Huntington's disease patients (Huntington Study Group, 1996). This indicates that failure of inhibitory control does not seem to be a strong candidate for the sequence learning deficit in the present study. Therefore, in addition to the failure of any sequence-specific processes, impairment in working memory would appear to be the major executive process contributing to poor performance.
Anatomical substrate of intentional learning deficit in Huntington's disease
The purpose of studying patients with Huntington's disease was to test the hypothesis that the caudate nucleus and associated circuits provide the neuronal substrate for intentional trial-and-error serial order learning. Inasmuch as Huntington's disease provides a clinical model of caudate degeneration, the results confirm the hypothesis, and indicate that the same pathology is of less significance in the processes underlying incidental sequence learning. However, Huntington's disease is far from a pure model of a caudate lesion. While the caudate nucleus is the major site of degenerative change in the earliest stages of the disease, as the disease progresses pathological changes affect more widespread striatal and extrastriatal regions. The volumes of the caudate and putamen decrease as the disease progresses (Aylward et al., 1997) and additional pathology is observed in the substantia nigra, globus pallidus, subthalamic nucleus, thalamus and hypothalamus, and cortically in multiple regions (Hersch and Ferrante, 1997). What clues do we have that caudate pathology is critical for cognitive dysfunction in general and serial order learning in particular? Although there have been no functional imaging studies to date, several studies have sought to correlate cognitive function with measures of pathology derived from MRI and PET (e.g. Berent et al., 1988; Starkstein et al., 1992; Lawrence et al., 1998b; Sanchez-Pernaute et al., 2000). Where observed, any significant associations between cognitive dysfunction and structural or metabolic indicators of pathology tend not to be specific to the caudate. For example, Lawrence and colleagues found that D2 receptor binding in both caudate and putamen was associated with a range of cognitive test performances including verbal fluency, non-verbal fluency (sequence generation), pattern recognition and spatial span (Lawrence et al., 1998b). However, pathological changes in the caudate and other striatal and extrastriatal structures tend to occur in parallel, such that worsening caudate degeneration is associated with increased pathology in the other structures. It is therefore difficult to determine whether the cognitive decline is due to worsening caudate degeneration specifically, to the parallel changes in other structures, or to an effect of total lesion volume. More powerful evidence for caudate specificity is suggested when cognitive changes are found in the patients with the mildest disease, or even in pre-symptomatic patients. In the present study, there was no evidence for such an effect. Trial-and-error learning was normal in such patients and impairment became evident only with disease progression. While this does not argue against a role for the caudate, it suggests that caudate degeneration might have to reach a certain level before behavioural change occurs, and/or that later cortical pathology may play a critical role in the overall functioning of the distributed system underlying intentional sequence learning.
Implications for functional anatomical models of caudate and sequence learning
The results of the present study support the functional anatomical model of implicit and explicit serial order learning presented in the introduction (Graybiel, 1995; Brown, 1999). However, the study provides evidence only for a single dissociation, and we lack direct evidence that patients with dysfunction restricted more to the putamen–SMA circuit show the opposite pattern of impaired incidental learning and intact intentional learning. The data on incidental learning in Parkinson's disease reviewed earlier offers partial support (Jackson et al., 1995; Stefanova et al., 2000), particularly when the circuit is disrupted by a structural lesion (Ackermann et al., 1996; Brown et al., 1998). To date, intentional trial-and-error sequence learning has not been assessed directly in Parkinson's disease, although there is indirect evidence to suggest that such learning is relatively normal (Pascual-Leone et al., 1993; Doyon et al., 1997; Sommer et al., 1999; Stefanova et al., 2000).
The potential role(s) of the striatum in sequence learning
Recent attempts to integrate anatomical, physiological and behavioural data have led to significant theoretical advances in our understanding of basal ganglia function. In relation to the issue of serial order learning, the most relevant model is that of Houk and Wise (Houk and Wise, 1995). This model proposes that a major role of the striatal spiny neurones is in the detection of context, which is achieved through the integration of input from multiple cortical areas. Changes in the firing patterns of striatal neurones signalling a particular context modify frontal cortical regions concerned with goal-directed behaviour. Although this function is applicable to a wide range of adaptive behaviour, it is particularly relevant to detection of serial order. Such context-specific firing is well illustrated by the study of Kermadi and Joseph, which identified caudate neurones that coded the serial order of at least successive pairs of stimuli; these neurones could potentially code higher-order sequences (Kermadi and Joseph, 1995).
In the models of Houk and Wise (Houk and Wise, 1995) and others (e.g. Graybiel, 1995), dopamine plays a central role in the learning of contextual patterns through the ability of dopaminergic neurones to signal anticipated reward, unanticipated reward and novelty (Schultz et al., 1997). In support of this, Matsumoto and colleagues demonstrated recently how unilateral dopaminergic lesions within the substantia nigra pars compacta interfere with the ability of monkeys to learn sequential motor acts in a coordinated and predictive manner (Matsumoto et al., 1999).
What role might the striatum and its nigrostriatal afferents play in incidental and intentional sequence learning? In intentional learning, the explicit involvement of working memory and conscious predictive processes would provide the striatum, and particularly the caudate nucleus, with a range of cortical input through which to shape the firing patterns of any context detection units. Immediate task feedback of response accuracy would modify the firing of dopaminergic afferents synapsing on the same striatal neurones to rapidly reinforce the learning and recognition of the correct contextual pattern. In incidental learning, neither the cortically coded information nor a strong reinforcement signal is available to the striatal neurones. This being so, how might learning occur? One possibility is that relatively low-level context detection occurs in the same neurones that achieve more predictive ability with the benefit of additional cortical and reinforcement information. In other words, there is a single substrate for learning and a single process by which it occurs; all that differs is the availability or otherwise of extra information that is useful for rapid learning to occur. According to this hypothesis, context detection may be a common feature of the spiny neurones, whether in the caudate or in the putamen. It is simply the caudate neurones' access to prefrontal cortical input that enables them to achieve high levels of context detection rapidly. However, this hypothesis would predict that caudate pathology that affects intentional sequence learning would also interfere with lower-level learning occurring without such prefrontal cortical input. The results from the present study fail to support this prediction and suggest that the critical substrate for incidental learning lies outside the caudate nucleus and its associated circuits. Whether the putamen has a similar selective role for the form of sensorimotor learning found in the SRT task cannot be determined from the present data. However, it does seem that increasing disease severity, which is associated with pathological spread from the caudate to the neighbouring putamen, is not inevitably associated with worsening of incidental learning. As noted previously, incidental learning appears to be a more robust phenomenon than intentional learning. Striatal processing, possibly in the putamen, may facilitate such learning. However, if it occurs at a distributed level through the sensorimotor circuitry, it would be less sensitive to the effect of a single lesion than intentional learning and would be more difficult to abolish. Further research in patients with discrete lesions may tell us more about the role of the sensorimotor circuitry in incidental sequence learning and its relationship to intentional learning.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Ackermann H, Daum I, Schugens MM, Grodd W. Impaired procedural learning after damage to the left supplementary motor area (SMA). J Neurol Neurosurg Psychiatry 1996; 60: 94–7.
Alexander GE. Basal ganglia-thalamocortical circuits: their role in control of movements. [Review]. J Clin Neurophysiol 1994; 11: 420–31.[Web of Science][Medline]
Aylward EH, Li Q, Stine OC, Ranen N, Sherr M, Barta PE, et al. Longitudinal change in basal ganglia volume in patients with Huntington's disease. Neurology 1997; 48: 394–9.
Barone P, Joseph JP. Prefrontal cortex and spatial sequencing in macaque monkey. Exp Brain Res 1989; 78: 447–64.[Web of Science][Medline]
Benecke R, Rothwell JC, Dick JP, Day BL, Marsden CD. Disturbance of sequential movements in patients with Parkinson's disease. Brain 1987; 110: 361–79.
Benita M, Conde H, Dormont JF, Schmied A. Effects of caudate nucleus cooling on the performance of conditioned movements in cats. Neurosci Lett 1979; 14: 25–30.[Web of Science][Medline]
Benton AL, Hamsher KDS. Multilingual Aphasia Examination (Manual). Iowa City: University of Iowa; 1976.
Berent S, Giordani B, Lehtinen S, Markel D, Penney JB, Buchtel HA, et al. Positron emission tomographic scan investigations of Huntington's disease: cerebral metabolic correlates of cognitive function. Ann Neurol 1988; 23: 541–6.[Web of Science][Medline]
Blackmore L, Simpson SA, Crawford JR. Cognitive performance in UK sample of presymptomatic people carrying the gene for Huntington's disease. J Med Genet 1995; 32: 358–62.
Brown RG. The roles of cortico-striatal circuits in learning sequential information. In: Stern GM, editor. Parkinson's disease. Advances in neurology, Vol. 80. Philadelphia: Lippincott Williams & Wilkins; 1999. p. 31–9.
Brown RG, Jahanshahi M, Limousin P, Quinn NP, Rothwell JC. Incidental sequence learning in patients with Parkinson's disease (PD): the impact of stereotaxic posteroventral pallidotomy (PVP) [abstract]. Mov Disord 1998; 13 Suppl 2: 199.[Web of Science][Medline]
Carpenter AF, Georgopoulos AP, Pellizzer G. Motor cortical encoding of serial order in a context-recall task. Science 1999; 283: 1752–7.
Cromwell HC, Berridge KC. Implementation of action sequences by a neostriatal site: a lesion mapping study of grooming syntax. J Neurosci 1996; 16: 3444–58.
Crum RM, Anthony JC, Bassett SS, Folstein MF. Population-based norms for the Mini-Mental State Examination by age and educational level. JAMA 1993; 269: 2386–91.
Dick JP, Benecke R, Rothwell JC, Day BL, Marsden CD. Simple and complex movements in a patient with infarction of the right supplementary motor area. Mov Disord 1986; 1: 255–66.[Medline]
Doyon J, Owen AM, Petrides M, Sziklas V, Evans AC. Functional anatomy of visuomotor skill learning in human subjects examined with positron emission tomography. Eur J Neurosci 1996; 8: 637–48.[Web of Science][Medline]
Doyon J, Gaudreau D, Laforce R Jr, Castonguay M, Bedard PJ, Bedard F, et al. Role of the striatum, cerebellum, and frontal lobes in the learning of a visuomotor sequence. Brain Cogn 1997; 34: 218–45.[Web of Science][Medline]
Ferraro FR, Balota DA, Connor LT. Implicit memory and the formation of new associations in nondemented Parkinson's disease individuals and individuals with senile dementia of the Alzheimer type: a serial reaction time (SRT) investigation. Brain Cogn 1993; 21: 163–80.[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]
Ghilardi M, Ghez C, Dhawan V, Moeller J, Mentis M, Nakamura T, et al. Patterns of regional brain activation associated with different forms of motor learning. Brain Research 2000; 871: 127–45.[Web of Science][Medline]
Grafton ST, Hazeltine E, Ivry R. Functional mapping of sequence learning in normal humans. J Cogn Neurosci 1995; 7: 497–510.[Web of Science]
Graybiel AM. Building action repertoires: memory and learning functions of the basal ganglia. [Review]. Curr Opin Neurobiol 1995; 5: 733–41.[Web of Science][Medline]
Hazeltine E, Grafton ST, Ivry R. Attention and stimulus characteristics determine the locus of motor-sequence encoding. A PET study. Brain 1997; 120: 123–40.
Hedreen JC, Folstein SE. Early loss of neostriatal striosome neurons in Huntington's disease. J Neuropathol Exp Neurol 1995; 54: 105–20.[Web of Science][Medline]
Helmuth LL, Mayr U, Daum I. Sequence learning in Parkinson's disease: a comparison of spatial-attention and number-response sequences. Neuropsychologia 2000; 38: 1443–51.[Web of Science][Medline]
Hersch SH, Ferrante RJ. Neuropathology and neurophysiology of Huntington's disease. In: Watts RL, Koller WC, editors. Movement disorders: neurologic principles and practice. New York: McGraw-Hill; 1997. p. 503–18.
Hikosaka O, Rand MK, Miyachi S, Miyashita K. Learning of sequential movements in the monkey: process of learning and retention of memory. J Neurophysiol 1995; 74: 1652–61.
Hikosaka O, Sakai K, Miyauchi S, Takino R, Sasaki Y, Putz B. Activation of human presupplementary motor area in learning of sequential procedures: a functional MRI study. J Neurophysiol 1996; 76: 617–21.
Hikosaka O, Nakahara H, Rand MK, Sakai K, Lu X, Nakamura K, et al. Parallel neural networks for learning sequential procedures. [Review]. Trends Neurosci 1999; 22: 464–71.[Web of Science][Medline]
Honda M, Deiber M-P, Ibanez V, Pascual-Leone A, Zhuang P, Hallett M. Dynamic cortical involvement in implicit and explicit motor sequence learning. A PET study. Brain 1998; 121: 2159–73.
Houk JC, Wise SP. Distributed modular architectures linking basal ganglia, cerebellum, and cerebral cortex: their role in planning and controlling action. [Review]. Cereb Cortex 1995; 5: 95–110.
Huntington Study Group. Unified Huntington's Disease Rating Scale: reliability and consistency. Mov Disord 1996; 11: 136–42.[Web of Science][Medline]
Jackson GM, Jackson SR, Harrison J, Henderson L, Kennard C. Serial reaction time learning and Parkinson's disease: evidence for a procedural learning deficit. Neuropsychologia 1995; 33: 577–93.[Web of Science][Medline]
Jenkins IH, Brooks DJ, Nixon PD, Frackowiak RS, Passingham RE. Motor sequence learning: a study with positron emission tomography. J Neurosci 1994; 14: 3775–90.[Abstract]
Jenkins IH, Tarazona FJ, Pascual-Leone A, Brooks DJ. The functional anatomy of implicit and explicit motor learning [abstract]. Neurology 1997; 48(3 Suppl 2): A305.
Jueptner M, Frith CD, Brooks DJ, Frackowiak RSJ, Passingham RE. Anatomy of motor learning. II. Subcortical structures and learning by trial and error. J Neurophysiol 1997; 77: 1325–37.
Kermadi I, Joseph JP. Activity in the caudate nucleus of monkey during spatial sequencing. J Neurophysiol 1995; 74: 911–33.
Knopman DS, Nissen MJ. Implicit learning in patients with probable Alzheimer's disease. Neurology 1987; 37: 784–8.
Knopman D, Nissen MJ. Procedural learning is impaired in Huntington's disease: evidence from the serial reaction time task. Neuropsychologia 1991; 29: 245–54.[Web of Science][Medline]
Kuwert T, Lange HW, Langen KJ, Herzog H, Aulich A, Feinendegen LE. Cortical and subcortical glucose consumption measured by PET in patients with Huntington's disease. Brain 1990; 113: 1405–23.
Lawrence AD, Sahakian BJ, Robbins TW. Cognitive functions and corticostriatal circuits: insights from Huntington's disease. Trends Cogn Sci 1998a; 2: 379–88.[Web of Science]
Lawrence AD, Weeks RA, Brooks DJ, Andrews TC, Watkins LH, Harding AE, et al. The relationship between striatal dopamine receptor binding and cognitive performance in Huntington's disease. Brain 1998b; 121: 1343–55.
Luppino G, Matelli M, Camarda R, Rizzolatti G. Corticocortical connections of area F3 (SMA-proper) and area F6 (pre-SMA) in the macaque monkey. J Comp Neurol 1993; 338: 114–40.[Web of Science][Medline]
Marder K, Zhao H, Myers RH, Cudkowicz M, Kayson E, Kieburtz K, et al. Rate of functional decline in Huntington's disease. Huntington Study Group. Neurology 2000; 54: 452–8.
Matsumoto N, Hanakawa T, Maki S, Graybiel AM, Kimura M. Role of nigrostriatal dopamine system in learning to perform sequential motor tasks in a predictive manner. J Neurophysiol 1999; 82: 978–98.
Milner B. Interhemispheric differences in the localization of psychological processes in man. [Review]. Br Med Bull 1971; 27: 272–7.
Mitchell IJ, Cooper AJ, Griffiths MR. The selective vulnerability of striatopallidal neurons. [Review]. Prog Neurobiol 1999; 59: 691–719.[Web of Science][Medline]
Miyachi S, Hikosaka O, Miyashita K, Karadi Z, Rand MK. Differential roles of monkey striatum in learning of sequential hand movement. Exp Brain Res 1997; 115: 1–5.[Web of Science][Medline]
Mushiake H, Strick PL. Preferential activity of dentate neurons during limb movements guided by vision. J Neurophysiol 1993; 70: 2660–4.
Mushiake H, Strick PL. Pallidal neuron activity during sequential arm movements. J Neurophysiol 1995; 74: 2754–8.
Mushiake H, Inase M, Tanji J. Selective coding of motor sequence in the supplementary motor area of the monkey cerebral cortex. Exp Brain Res 1990; 82: 208–10.[Web of Science][Medline]
Mushiake H, Inase M, Tanji J. Neuronal activity in the primate premotor, supplementary, and precentral motor cortex during visually guided and internally determined sequential movements. J Neurophysiol 1991; 66: 705–18.
Nakamura K, Sakai K, Hikosaka O. Neuronal activity in medial frontal cortex during learning of sequential procedures. J Neurophysiol 1998; 80: 2671–87.
Nelson HE. A modified card sorting test sensitive to frontal lobe defects. Cortex 1976; 12: 313–24.[Web of Science][Medline]
Nissen MJ, Bullemer P. Attentional requirements of learning: evidence from performance measures. Cogn Psychol 1987; 19: 1–32.
Parthasarathy HB, Schall JD, Graybiel AM. Distributed but convergent ordering of corticostriatal projections: analysis of the frontal eye field and the supplementary eye field in the macaque monkey. J Neurosci 1992; 12: 4468–88.[Abstract]
Pascual-Leone A, Grafman J, Clark K, Stewart M, Massaquoi S, Lou JS, et al. Procedural learning in Parkinson's disease and cerebellar degeneration. Ann Neurol 1993; 34: 594–602.[Web of Science][Medline]
Procyk E, Tanaka YL, Joseph JP. Anterior cingulate activity during routine and non-routine sequential behaviors in macaques. Nat Neurosci 2000; 3: 502–8.[Web of Science][Medline]
Rauch SL, Savage CR, Brown HD, Curran T, Alpert NM, Kendrick A, et al. A PET investigation of implicit and explicit sequence learning. Hum Brain Mapp 1995; 3: 271–86.
Rauch SL, Whalen PJ, Savage CR, Curran T, Kendrick A, Brown HD, et al. Striatal recruitment during an implicit sequence learning task as measured by functional magnetic resonance imaging. Hum Brain Mapp 1997; 5: 124–32.[Web of Science][Medline]
Sakai K, Hikosaka O, Miyauchi S, Takino R, Sasaki Y, Putz B. Transition of brain activation from frontal to parietal areas in visuomotor sequence learning. J Neurosci 1998; 18: 1827–40.
Sanchez-Pernaute R, Kunig G, del Barrio Alba A, de Yebenes JG, Vontobel P, Leenders KL. Bradykinesia in early Huntington's disease. Neurology 2000; 54: 119–25.
Sarazin M, Deweer B, Pillon B, Agid Y, Dubois B. Procedural learning in Parkinson's disease [abstract]. Neurology 1995; 45Suppl 4: A264.
Schultz W. The primate basal ganglia and the voluntary control of behaviour. J Conscious Stud 1999; 6: 31–45.
Schultz W, Dayan P, Montague PR. A neural substrate of prediction and reward. [Review]. Science 1997; 275: 1593–9.
Shoulson I. Huntington disease: functional capacities in patients treated with neuroleptic and antidepressant drugs. Neurology 1981; 31: 1333–5.
Smith A. The Symbol Digit Modalities Test: a neuropsychological test for economic screening of learning and other cerebral disorders. Learn Disord 1968; 3: 83–91.
Sommer M, Grafman J, Clark K, Hallett M. Learning in Parkinson's disease: eyeblink conditioning, declarative learning, and procedural learning. J Neurol Neurosurg Psychiatry 1999; 67: 27–34.
SPSS. SPSS 8.0 for Windows. 1998. Chicago: SPSS; 1998.
Starkstein SE, Brandt J, Bylsma F, Peyser C, Folstein M, Folstein SE. Neuropsychological correlates of brain atrophy in Huntington's disease: a magnetic resonance imaging study. Neuroradiology 1992; 34: 487–9.[Web of Science][Medline]
Stefanova ED, Kostic VS, Ziropadja L, Markovic M, Ocic GG. Visuomotor skill learning on serial reaction time task in patients with early Parkinson's disease. Mov Disord 2000; 15: 1095–103.[Web of Science][Medline]
Stroop JR. Studies of interference in serial verbal reactions. J Exp Psychol 1935; 18: 643–62.[Web of Science]
Toni I, Krams M, Turner R, Passingham RE. The time course of changes during motor sequence learning: a whole-brain fMRI study. Neuroimage 1998; 8: 50–61.[Web of Science][Medline]
Van den Bercken JH, Cools AR. Evidence for a role of the caudate nucleus in the sequential organization of behavior. Behav Brain Res 1982; 4: 319–27.[Web of Science][Medline]
Wechsler D. The Wechsler Adult Intelligence Test—Revised. New York: Psychological Corporation; 1981.
Weinberger DR, Berman KF, Iadarola M, Driesen N, Zec RF. Prefrontal cortical blood flow and cognitive function in Huntington's disease. J Neurol Neurosurg Psychiatry 1988; 51: 94–104.
Westwater H, McDowall J, Siegert R, Mossman S, Abernethy D. Implicit learning in Parkinson's disease: evidence from a verbal version of the serial reaction time task. J Clin Exp Neuropsychol 1998; 20: 413–18.[Web of Science][Medline]
Willingham DB, Koroshetz WJ. Evidence for dissociable motor skills in Huntington's disease patients. Psychobiology 1993; 21: 173–82.[Web of Science]
Received March 26, 2001. Revised May 23, 2001. Accepted June 13, 2001.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. Muslimovic, B. Post, J. D. Speelman, and B. Schmand Motor procedural learning in Parkinson's disease Brain, November 1, 2007; 130(11): 2887 - 2897. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. R. Bailey and R. G. Mair The Role of Striatum in Initiation and Execution of Learned Action Sequences in Rats J. Neurosci., January 18, 2006; 26(3): 1016 - 1025. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Van Raamsdonk, J. Pearson, E. J. Slow, S. M. Hossain, B. R. Leavitt, and M. R. Hayden Cognitive Dysfunction Precedes Neuropathology and Motor Abnormalities in the YAC128 Mouse Model of Huntington's Disease J. Neurosci., April 20, 2005; 25(16): 4169 - 4180. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||