Brain, Vol. 122, No. 1, 87-97,
January 1999
© 1999 Oxford University Press
Non-motor associative learning in patients with isolated degenerative cerebellar disease
1 Department of Neurology, University of Essen, Essen and 2 Institute of Physiology, University of Munich, Pettenkoferstrasse 12, Munich, Germany
Correspondence to:
J. Drepper, Department of Neurology, University of Essen, Hufelandstrasse 55, D-45122 Essen, Germany E-mail: johannes.drepper{at}uni-essen.de
 |
Abstract |
|---|
 Top Abstract IntroductionMethodResultsDiscussionReferences |
|---|
In recent decades it has become clear that the cerebellum is involved in associative motor learning, but its exact role in motor learning as such is still controversial. Recently, a contribution of the cerebellum to different cognitive abilities has also been considered, but it remains unclear whether the cerebellum contributes to cognitive associative learning. We compared nine patients with an isolated cerebellar degenerative disease in a cognitive associative learning task with 10 controls. Patients and controls were matched for age, sex, handedness, level of education, intelligence and capabilities of visual memory. The subjects were asked to learn the association between six pairs of colours and numerals by trial and error. Additionally, a simple reaction time and a visual scanning test were conducted in order to control for the influence of motor performance deficits in cerebellar patients. In comparison with the controls, it took the patients significantly longer to learn the correct associations between colours and numerals, and they were impaired in recognizing them later on. Two patients showed no associative learning effect at all. Neither the simple reaction time nor the visual scanning time correlated substantially with the results of associative learning. Therefore, motor-associated disabilities are unlikely to be the reason for the learning deficit in cerebellar patients. Our results suggest that the cerebellum might contribute to motor-independent processes that are generally involved in associative learning.
cerebellum; cognition; reaction times
RFT = recurring figures test; SPM = standard progressive matrices
|
Introduction |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
The cerebellum has long been known as a structure that contributes essentially to motor co-ordination and control (Holmes, 1939
). In recent decades, research in this area has focused increasingly on the role of the cerebellum in motor learning. Today there is a large number of studies in which a variety of procedures—including physiological studies with lesioned animals, the comparison of cerebellar patients with healthy subjects and imaging techniques—have been used that show an involvement of the cerebellum in motor learning (for recent reviews, see Thach, 1996; Thompson and Kim, 1996). However, the interpretation of the data is still a matter of debate. The controversial question is whether the cerebellum contributes rather to the motor or to the learning part of a motor learning task (Thompson and Krupa, 1994; Bloedel and Bracha, 1995). One of the problems is to dissociate motor functions from learning functions in a motor learning paradigm (Timmann and Diener, 1996).
Apart from this ongoing debate, another question, first posed by Leiner et al. (1986), has attracted the attention of many researchers: `Does the cerebellum contribute to mental skills?' The authors gave three reasons for their question: (i) there is a late enlargement of the cerebellum in the phylogeny of man, which parallels the development of the neocortex. It was argued that functions must reside in the newly evolved structures that are typical for higher primates and humans; (ii) cerebellar projections expand to almost all parts of the associative cortex, and especially the projections to the frontal and parietal association areas do not seem to serve a purely motor function (Glickstein, 1993); and (iii) some patients with cerebellar lesions show unexpected intellectual deficits (Leiner et al., 1986).
Ever since these early reflections on the contribution of the cerebellum to cognition, many empirical studies have been carried out to show the influence of cerebellar impairment on various intellectual functions (Schmahmann, 1991; Leiner et al., 1993; Fiez, 1996).
However, this accumulation of evidence in favour of a cerebellar contribution to various cognitive functions has led to new problems. For instance, in some studies a lower verbal IQ was observed in cerebellar patients (Bracke-Tolkmitt et al., 1989) whereas in others it was not (Fehrenbach et al., 1984; Wallesch and Horn, 1990; Fiez et al., 1992). Leiner et al. (1986) originally postulated an impairment of frontal lobe functions in cerebellar patients, which was confirmed by some researchers (Grafman et al., 1992) but not by others (Bracke-Tolkmitt et al., 1989).
Daum and her colleagues (Daum et al., 1993a; Daum and Ackermann, 1995) suggested a number of reasons for these controversial findings. One basic difficulty in this field of research is posed by the rare clinical occurrence of selective cerebellar damage. Therefore, many studies are based on single cases or very small samples which often include patients with extracerebellar damage. The interpretation of the various neuropsychological findings is further restricted by methodological shortcomings, such as inadequately matched control groups and an insufficient assessment of background variables. Consequently, Daum et al. (1993a) tested patients with pure cerebellar lesions and found no cognitive deficits, whereas impairments in frontal lobe functions could be demonstrated in patients with cerebellar and additional brainstem lesions. Also, Dimitrov et al. (1996) found that a number of cognitive functions were spared in cerebellar disease which were reported as impaired in earlier studies.
The purpose of the present study was to investigate critically whether the cerebellum is involved in cognitive associative learning. According to the methodological requirements set up by Daum and Ackermann (1995), we tested only patients with an isolated cerebellar disease and compared them with healthy controls who were matched not only for age but also for intelligence, visual memory, sex and level of education.
An experimental learning task was introduced in which associations of true cognitive representations had to be acquired. The latencies of the answers were measured and divided into motor and decision components. These choice reaction times are more sensitive indicators of associative learning performance than false responses alone. Additional tasks, including a simple reaction time task, were implemented to control all remaining motor demands of the experimental learning procedure.
|
Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Subjects
Nine patients with idiopathic cerebellar ataxia (Harding, 1993
) were compared with 10 healthy controls. The diagnosis of idiopathic cerebellar ataxia was based on a neurological examination, radiological diagnostics (CT or MRI), the absence of extracerebellar signs and the lack of a positive family history. The neurological examination included the ataxia rating scale from Trouillas et al. (1997) as well as a scoring of cerebellar ataxia according to Klockgether et al. (1990). The clinical data of the patients are summarized in
Table
1. Their average age was 63.9 ± 10.83 years and the mean duration of disease was estimated as 4.4 ± 3 years. Seven patients were female, two male, seven right-handed and two left-handed.
|
From an overall sample of 18 examined controls, five subjects were rejected after the neurological examination because of neurological symptoms. The 10 subjects who best matched the patients were then selected with regard to the following criteria: age, level of school education, IQ and visual memory capabilities. Seven of them were female, three male, eight right-handed and two left-handed. Their mean age was 61.8 ± 8.09 years. None of them had a history of neurological disease.
Neither control nor cerebellar subjects were taking any centrally acting drugs at the time of testing.
Procedure
All subjects gave written consent to testing and were instructed in accordance with the requirements of the ethical committee that they could discontinue participation at any stage of the procedure. In accordance with legal requirements, all subjects additionally gave written consent to the anonymous evaluation and publication of their data for scientific purposes. The local ethical committee of the University of Essen approved the study.
The mean duration of all tests was ~4 h and the whole neuropsychological battery was carried out over 2 days. Two patients completed the battery on one day, interrupted only by an extended lunch-break. Due to organizational requirements the tests carried out by all subjects were not in any specific order. All subjects completed all of the tests.
Neuropsychological tests
In addition to the standardized neurological examination described above, all subjects were tested with the colour plates described by Ishihara (1990) to rule out colour-blindness.
As a reliable measure of intelligence, the standard progressive matrices test (SPM) (Raven, 1956) without time restrictions was chosen, thus minimizing the potential influence on the IQ of the patients' motor performance deficits. A second advantage of this test is that it gives extensive information about intelligence in addition to that indicated by the level of education. In order to measure visual memory capabilities, the recurring figures test (RFT) (Kimura, 1963; Hartje and Rixecker, 1978) was administered.
Additionally, two subtests of a German computerized neuropsychological battery for evaluating attentional deficits (Zimmermann and Fimm, 1992) were carried out. The first subtest, `alertness', is a simple visual reaction time task where in some trials an acoustic warning occurs before the visual stimulus. The two conditions differentiate between phasic and tonic alertness deficits (Posner and Petersen, 1990). The test consisted of four blocks with 20 trials each. In blocks 2 and 3 the acoustic warning occurred. The second subtest was a visual scanning task in which the subject had to decide and respond as quickly as possible whether or not a specific stimulus occurred in a visually presented matrix. These two tasks served the purpose of controlling the influence of motor performance deficits, which are typical of cerebellar patients, in complex visual reaction time tasks.
Associative learning task
The associative learning task was presented on an x86-compatible PC with a 15-inch colour monitor placed 60 cm in front of the comfortably seated subject. During the whole experiment the background colour on the monitor was light grey. All written information on the screen was presented in dark, bold letters which were easy to read for all the subjects. Stimulus presentation and data collection were controlled by a computer program.
The subjects were asked to learn the associations between six colours and the numerals from 1 to 6 by trial and error. Colours were chosen as stimuli in order to make their visuospatial encoding unnecessary. The colours were different shades of green and blue, such that it was difficult to distinguish between them and describe them verbally. In this way the influence of subvocal strategies and verbal abilities could be minimized. Before the test, the six colours were first presented side by side as filled squares with a height and width of 4 cm to ensure that all subjects could differentiate between them. Then each numeral was shown in the correct colour for 4 s and the subjects were instructed to remember the link between the colour and the numeral. All numerals were presented in bold style with a height of 3 cm.
After this procedure, the subjects started each trial by using a specialized keyboard. The keyboard had three buttons, a central home button and two target buttons, one to the left and one to the right of the home button. Each button had a surface area of 1 cm2 and the distance between the buttons was 1 cm. If the central button was pressed long enough, a new trial was presented, consisting of a coloured square and a black numeral either to the side of or above the square. The configuration of the stimuli, i.e. whether the colour was on the right or left, above or below the numeral, changed randomly from trial to trial, but was identical for all subjects. The stimuli could be inspected as long as the central button was pressed. On releasing the button the stimuli vanished immediately. While pressing the central button and inspecting the presented pair of stimuli, the subjects had to decide if the association of colour and numeral was correct. For correct associations they had to move the finger to the right button and for false associations they had to move it to the left. Feedback was displayed on the monitor immediately after pressing one of the two answer buttons. All subjects were instructed to use only one finger of the dominant hand (tested in all subjects) for all keyboard operations. They were informed that all reaction times were measured and that speed and accuracy were equally important characteristics of the task.
This procedure allowed differentiation between decision time and moving time. Decision time was defined as the time between stimulus onset and releasing the central button, whereas moving time was the time between releasing the home button and pressing the correct target button.
The stimuli were presented in blocks of 12 trials. In each block, all six colours were presented with the correct numeral as well as with a false one. The sequence of trials was the same for all subjects. The procedure was successfully finished when a subject completed two successive blocks correctly. At the end of the test the six colours were presented again side by side on the monitor and the subjects were asked to name the correct number for each colour.
|
Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
The results were obtained in four steps. First, the patients and controls were compared with respect to the background variables of age, memory, intelligence, simple reaction time and visual scanning. Secondly, group differences under the experimental associative learning condition were analysed with respect to the correctness of the answers, and thirdly the differences were analysed with respect to reaction time. Fourthly, possible interactions between the results of the standardized tests and the experimental learning situation were investigated.
To evaluate the group differences, t tests for independent samples were calculated for interval data, and the Mann–Whitney U test was used for ordinal data. To determine the interactions between different independent measures, Pearson's correlation coefficient was used for interval data and Spearman's 
for ordinal data.
Background variables
Interesting background variables were the percentiles for the SPM as a measure of intelligence as well as the RFT, which indicates capabilities of visual memory. As a measure of simple reaction time, the mean reaction times of all four blocks in the alertness test were used. This test permits calculation of the difference in reaction time for trials with and without an alerting acoustic stimulus, indicating phasic alertness. Phasic alertness was calculated as the difference of the median reaction time of the blocks with and without alerting stimulus, divided by the median reaction time of all four blocks (Zimmermann and Fimm, 1992). In the visual scanning task we calculated the mean reaction time for those correctly answered trials in which no critical stimulus appeared in the matrix. This measure was used as an indicator of visual scanning speed, because the subjects had to scan the whole matrix in these trials in order to find the correct answer. Additionally, the number of false reactions was counted.
There were no significant group differences with regard to age (t = 0.48, P = 0.638), memory (t = 0.73, P = 0.476) or intelligence (t = –0.62, P = 0.543). The simple reaction time task revealed a longer mean reaction time in patients (284.58 ± 54.81 ms) than in controls (236.34 ± 27.24 ms; t = 2.47, P = 0.024). The ratio between reaction times with and without alerting stimulus showed no significant difference between the two groups (t = 0.88, P = 0.391) (Fig. 1A).
|
Compared with the controls, patients needed a prolonged visual scanning time for the whole of the presented matrix (patients, 8.03 ± 3.62 s; controls, 4.7 ± 1.39 s; t = 2.7, P = 0.015). There was no indication of different speed–accuracy trade-offs in the two groups. The patients made a few more errors than the controls, but the effect was not significant (patients, 9.22 ± 7.26; controls, 6.60 ± 5.13; t = 0.917, P = 0.372) (Fig. 1B
). The results are summarized in
Table
2.
|
Associative learning task: correctness of responses
Three patients and two control subjects discontinued the associative learning procedure before reaching the criterion (Fig. 2
: patients 7, 8 and 9; controls 9 and 10). Although the task had been evaluated with several students, it proved to be too difficult for the chosen sample. The mean time taken to complete the procedure was ~1 h.
|
In order to determine whether those subjects who did not reach the criterion showed any kind of learning effect in the experimental procedure, the linear correlation between the block and the number of correct responses per block was calculated. Two of the patients who did not succeed in the task failed to reach a significant correlation (P > 0.05) (Fig. 2
: patients 8 and 9). All other patients and control subjects reached a significant correlation (P < 0.01).
Among the subjects who did reach the criterion, the patients needed more blocks (50.7 ± 12.5) than the controls (31.5 ± 12.4; Mann–Whitney U = 4, P = 0.01) (Fig. 3A). These subsamples of six patients (four female, two male, five right-handed, one left-handed, five school level I, one school level II) and eight controls (five female, three male, five right-handed, two left-handed, seven school level I, one school level III; for explanation of school levels see footnote to Table 2) were still well matched with respect to age (patients, 64.3 ±12.7; controls, 61.6 ± 9.0), intelligence (patients, 80.0 ± 27.6; controls, 81.3 ± 13.3; percentiles of the SPM) and visual memory capabilities (patients, 63.0 ± 31.2; controls, 46.3 ± 32.3; percentiles of the RFT).
|
As a measure of performance for all subjects, including those who did not reach the criterion, the linear regression between number of block and number of correct reactions per block was calculated individually. The mean slope of this regression differed significantly between patients and controls (t = –2.24, P = 0.039) (Fig. 3B
). The number of correctly named associations uttered after the learning procedure was significantly higher for the controls (median = 6) than for the patients (median = 4, Mann–Whitney U = 18.5, P = 0.011; Fig. 3C). The results are summarized in
Table
2.
Associative learning task: latency of responses
If not otherwise indicated, only decision and moving times from correctly answered trials are considered here. For the calculation of the mean decision times per block, reaction times which differed from the mean per block by >2 SD were discarded. The mean number of outliers per block did not differ significantly between the two groups (patients, 0.35 ± 0.20; controls, 0.42 ± 0.15; t = –0.897, P = 0.382). Neither the mean decision time nor the mean moving time showed a significant difference between the two groups (Table 2). This result was unexpected. We had expected the moving time in particular to be longer for patients with motor deficits, in a similar way to the simple reaction time task. However, the decision and moving times of all subjects were highly variable.
The decision and movement times showed a linear decrease with the number of trials, which was significantly different from 0 (all P < 0.05, one-sample t test) for both groups. Whereas the slope of decision time versus number of trials was not significantly different between the two groups, the decrease in movement time was higher in control subjects (–22.14 ms) than in patients (–3.07 ms; P = 0.029) (Table 2). For the whole sample, however, the slopes for decision and moving times were highly dependent on the number of blocks used for the learning procedure. There was a significant correlation between the number of blocks and the decrease in both decision time (r = 0.585, P = 0.008) and moving time (r = 0.513, P = 0.025). This dependency is probably the result of a ceiling effect, i.e. after a given number of blocks the learning effect on reaction time disappeared. This could be due to increasing familiarity with the task and the keyboard. If the first 10 blocks are excluded, the difference in the decrease in movement time becomes smaller (patients, –1.99 ms; controls, –11.37 ms) and is no longer significant (P = 0.11). For the first 10 blocks the decrease in moving time of the patients was even higher (–49.27 ms) than that of the controls (–33.11 ms).
Almost all subjects spontaneously produced some correct verbal comments on the combination of colour and number actually seen while pressing the false answer button. The question is whether this phenomenon occurred more often in patients than in controls. If this was indeed the case, it could be due to the patients' greater difficulty in co-ordinating their correct knowledge about the answer with the correct movement direction on the keyboard. Short decision times and false responses were interpreted as an indication of this possibility. In this case the absolute correlation between decision times and correctness (0 = false, 1 = correct) should be higher for the controls than for the patients. In contrast, the variance of decision time for false responses should be higher in patients than in controls. The data, however, did not support this hypothesis. The correlations for decision time and correctness were very similar for the two groups (patients, r = –0.209; controls, r = –0.214) and the standard deviation of the decision time for false responses was even higher in controls (2586 ms) than in patients (2166 ms).
Interactions
Interaction of the background variables with those performance measures of the experimental learning task that differed significantly between patients and controls was tested. Because of the small size of the sample only performance measures that could be calculated for all subjects were considered.
The background variables age, intelligence, visual memory, visual scanning time and simple reaction time were not correlated significantly with the number of correctly named associations after the learning procedure or with the linear increase for correct responses per block (P > 0.2; Table 3).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
The present study revealed a specific deficit of cerebellar patients in a cognitive associative learning task. The patients needed more trials to reach a previously defined criterion when learning six associations between colours and numerals, and they had more difficulty in naming the correct associations after the learning phase than healthy control subjects. This result could not be linked to extracerebellar lesions, because only patients with lesions confined to the cerebellum were included in the tests. Nor was this deficit in cognitive associative learning attributable to any sample differences in intelligence, age, sex, schooling, visual memory or motor performance.
The background variables investigated in this study served the purpose of isolating some basic operations, in particular motor associated processes, which are possibly involved in the experimental associative learning task. For this reason simple reaction time and visual scanning were tested, which yielded data on two elementary operations involved in the associative learning task. As expected, both were affected by cerebellar impairment, but neither the higher simple reaction time nor the impaired visual scanning time of the patients could explain their poorer performance in the experimental learning task. No correlations between the results in these different tasks were found.
Our results differ from those of two studies (Daum et al., 1993a; Dimitrov et al., 1996) in which no difference was found between cerebellar patients and normal controls in cognitive associative learning. One reason for these different findings could be the difference in the testing procedures used. Daum et al. (1993a) used verbal paired associates and Dimitrov et al. (1996) used word-list learning, in which associative processes are presumably involved. We think that our experimental procedure will turn out to be a better measure for a pure cognitive associative process than previous procedures, because almost no verbal or spatial processes are involved in our task. A second reason may be the greater variability of the patient samples in the two previous studies. Whereas our patients are all suffering from idiopathic cerebellar ataxia (Harding, 1993), Daum et al. (1993a) included an additional three patients with various ischaemic lesions and one patient with a resection of parts of the cerebellum for astrocytoma removal. In the study by Dimitrov et al. (1996), some patients with olivopontocerebellar atrophy were tested. Greater intragroup variability always reduces the probability of finding a significant group difference. Finally, in both of these earlier studies the authors tried to demonstrate that there is no group difference with respect to learning measures between cerebellar patients and normal controls. But this is difficult to show. It is insufficient to simply demonstrate that there are no significant group differences. In neither study were all mean values, standard deviations and error probabilities specified. Therefore the strength of the indication of comparable performance of cerebellar patients and normal controls in learning tasks remains unclear.
Our results confirm those of earlier studies by Canavan et al. (1994) and Tucker et al. (Tucker et al., 1996), who also showed deficits in associative learning in cerebellar patients. The present study, however, goes beyond their approach in that we chose an alternative experimental design that takes better account of the difference between cognitive and motor associative learning. The problem becomes apparent if one compares the study by Canavan et al. (1994) with the experimental approach described here. They additionally introduced an associative learning reaction time paradigm in which the subjects had to learn the predictive value of an original neutral visual stimulus. This was followed by presentation of an arrow either to the left or to the right, which indicated to the subjects that they should press either the left or the right response button. This learning condition was compared with a simple reaction time condition without any predictive stimuli, where all task requirements except associative learning were identical. Patients with circumscribed cerebellar lesions had more difficulty in learning the predictive meaning of the abstract cues than healthy controls (Canavan et al., 1994).
However, the similarity of Canavan's experimental design to typical motor learning studies poses a problem. As the authors put it themselves, the task was `. . . to associate particular movements with particular visual cues' (Canavan et al., 1994, p. 2). This means that the task was more of a conditional motor learning paradigm than a genuinely cognitive associative learning task. The subjects had to learn to associate different visual contexts with a movement to the left or to the right. The association is merely a context response linkage; in other words, only the cognitive influence on motor responses is measured. This problem is also apparent in the study by Tucker et al. (1996), in which the association between six geometric shapes and six buttons had to be learned.
In order to ensure measurement of the learning of purely cognitive associations, a different experimental paradigm was chosen in the present study. The test was designed in such a way that the subject could not master the task by learning the association between one colour or one numeral and a moving direction for the index finger alone. Only information about the combination of colour and numeral was sufficient to determine the correct answer. Secondly, the changing configurations of colours and numerals on the screen made it more difficult for the subjects to learn solely the association between a special visual context and a moving direction, regardless of the actual association of colour and numeral.
One might argue that even if cognitive operations were being tested, deficits in cerebellar patients might be a consequence of the additional motor requirements of the task. After all, most cognitive tasks do imply motor performance (Adams, 1987). Consequently it has to be asked whether there are any motor processes involved in the associative learning task that are not controlled in the simple reaction task or the visual scanning task.
The choice of a left or right movement direction in the trials could be considered such a motor process. We had to take into account the possibility that patients had specific difficulties in co-ordinating their knowledge of the correct answer with the correct movement direction of the finger on the keyboard. However, the correlations of decision times and correctness of reactions did not support this. Short decision times predicted correct responses for both groups, and, as such, were also indicators of knowledge of the correct answer. Additionally, the visual scanning task was a choice reaction time task, and no differences were revealed between the two groups with respect to the above-mentioned co-ordinating ability. The number of errors made in the visual scanning task was similar for the two groups.
An even more subtle problem of additional motor requirements in cognitive tasks is generated by the potential interaction effect caused by attentional capacity limits. If, for example, cerebellar patients showed the same motor performance as controls, there would remain the possibility that they require more attentional resources for motor control than others. As a consequence they might display a poorer cognitive performance, although the cerebellum is not necessarily involved in cognitive operations as such. Different speed–accuracy trade-offs would be a special example of this general problem.
A weak indicator of such a motor deficit compensating mechanism may be the similar movement times shown by patients and controls in the learning task despite different simple reaction times. Alternatively, this phenomenon could be explained by the lower prominence of the speed factor for both groups in the learning task, due to its difficulty. The long and highly variable mean decision and movement times rather support the second hypothesis and rule out different speed–accuracy trade-offs as an exclusive explanation of the observed group differences. Nevertheless, until now such motor deficit compensating mechanisms have been completely neglected in the neuropsychological literature about cognitive deficits in cerebellar patients, and they deserve more attention in future studies.
However, the strongest argument against the assumption of a pure motor deficit is the fact that the patients could assign fewer colours correctly at the end of testing than the control subjects, even after a longer learning phase.
When exploring cognitive deficits of cerebellar patients, it is often difficult to control the involvement of verbal processes. Fiez et al. (1992) showed a possible involvement of the cerebellum in semantic processing. Thus, it could be argued that patients have more difficulty in establishing a useful verbal strategy for the task than control subjects. One possible strategy could be subvocal rehearsal of the numbers and colour labels. We tried to deal with this problem by minimizing the profit from a verbal strategy, thus avoiding the problem of the possible involvement of the cerebellum in subvocal speech. When at the end of the learning procedure the six colours were shown side by side on the screen, all subjects referred to the colours by pointing to them. Obviously, none of them had useful verbal labels to differentiate between them.
Another problem is posed by the choice of colours. Because of the high degree of similarity between the colours chosen, the two factors of associative learning and colour discrimination cannot be clearly distinguished in the experimental design. However, there are two reasons for favouring associative learning as the crucial factor in this experiment. First, the colour test (Ishihara, 1990) gave no indication of differences between the two groups, and secondly no reference to the involvement of the cerebellum in colour perception or colour discrimination was found in the current literature (Zeki, 1993).
One problem of the present study was the great difficulty the chosen sample of patients and control subjects experienced in performing the experimental task. Three patients and two controls out of 19 subjects were not able to reach the criterion in an acceptable period of time. Differences in motivation and concentration or differential fatigue effects might have had more influence on the results than expected. The mean duration of the chosen procedure was ~1 h. Although this is not much longer than the duration of the intelligence test (mean = 48 min), there remains a possibility that this task required a higher degree of sustained effort. As Appollonio et al. (1993) were able to show, it is possible that cerebellar patients have difficulty with response initiation and perseveration control, especially in effortful tasks. Whereas the use of perseverative strategies cannot be established with our experimental procedure, the similarity in decision time between patients and controls does not support the hypothesis of impaired response initiation in cerebellar patients. Additionally, it should be mentioned that Appollonio et al. (1993) included two patients with extracerebellar diseases in their study. The fact that not all subjects reached our special learning criterion could also be an indication of sampling differences between cerebellar patients and controls. With a sample size of 9 patients vs. 10 controls this possibility cannot be ruled out completely. We did show, however, that there was no substantial sample difference with respect to intelligence, age or visual memory between the two groups. Furthermore, the patients exhibited results similar to those of the controls in all additional tests without intense motor demands. This also indicates that there were no motivational differences between the groups.
In summary, we obtained three measures of a learning process: the number of blocks to reach the criterion, the increase in correct responses per block and the number of correctly named associations after the learning procedure. Each measure alone is a weak indicator of a real group difference. Because some subjects did not reach the criterion set, the first measure accounts only for smaller subsamples of patients and controls. Some subjects achieved only eleven correct answers per block, and one might argue that these subjects showed only a simple lack of concentration or a different speed–accuracy trade-off compared with other subjects. However, if this had been the case, the increases in correct responses per block and the number of correctly named associations at the end of the test would have been similar to the results of the subjects who reached the criterion.
The measures we obtained for all of the subjects are each arguable on their own. However, the results of the successful subjects show the same group difference as the pooled measures. Additionally, we were able to show that, including the discontinuing subjects, all subjects showed significant learning with the exception of two patients.
Recently, involvement of the human cerebellum in classical conditioning of motor responses has been shown in a number of studies (Thompson, 1991; Daum et al., 1993b; Topka et al., 1993; Timmann et al., 1996; Woodruff-Pak et al., 1996). Canavan et al. (1994) and Tucker et al. (1996) were able to extend these findings when they found that cerebellar patients were impaired in learning associations between visual cues and required particular movements. In the present study we showed that patients with lesions confined to the cerebellum are also impaired in learning purely cognitive associations.
One might assume that all these associative learning tasks share one common process for which cerebellar patients display a deficiency. This process could be the establishment of a link between two pieces of information or the preparation of sensory input for further associative processes. The notion that the cerebellum may actually be involved in co-ordinating the acquisition of sensory data and not in motor control or cognitive functions per se has recently been advocated by Bower and co-workers (Gao et al., 1996; Bower, 1997). Whether the cerebellar involvement in associative learning tasks is due to sensory specialization or key associative processes or both remains an important issue for further research in this field.
|
Acknowledgments |
|---|
We wish to thank Vanessa Borgmann for editorial help, Matthias Maschke for his readiness to administer some neurological examinations and Hans Gerd Elles for constructing the keyboard for the reaction time experiments. The study was supported by a grant from the Deutsche Forschungsgemeinschaft (Ti 239/2–1) to D.T.
|
References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Adams JA. Historical review and appraisal of research on the learning, retention, and transfer of human motor skills. Psychol Bull 1987; 101: 41–74.[Web of Science]
Appollonio IM, Grafman J, Schwartz V, Massaquoi SG, Hallett M. Memory in patients with cerebellar degeneration. Neurology 1993; 43: 1536–44.
Bloedel JR, Bracha V. On the cerebellum, cutaneomuscular reflexes, movement control and the elusive engrams of memory. [Review]. Behav Brain Res 1995; 68: 1–44.[Web of Science][Medline]
Bower JM. Is the cerebellum sensory for motor's sake, or motor for sensory's sake: the view from the whiskers of a rat? [Review]. In: de Zeeuw CI, Strata P and Voogd J, editors. Prog Brain Res 1997; 114: 463–96.[Web of Science][Medline]
Bracke-Tolkmitt R, Linden A, Canavan AGM, Rockstroh B, Scholz E, Wessel K, et al. The cerebellum contributes to mental skills. Behav Neurosci 1989; 103: 442–6.[Web of Science]
Canavan AG, Sprengelmeyer R, Diener HC, Hömberg V. Conditional associative learning is impaired in cerebellar disease in humans. Behav Neurosci 1994; 108: 475–85.[Web of Science][Medline]
Daum I, Ackermann H. Cerebellar contributions to cognition. [Review]. Behav Brain Res 1995; 67: 201–10.[Web of Science][Medline]
Daum I, Ackermann H, Schugens MM, Reimold C, Dichgans J, Birbaumer N. The cerebellum and cognitive functions in humans. Behav Neurosci 1993a; 107: 411–9.[Web of Science][Medline]
Daum I, Schugens MM, Ackermann H, Lutzenberger W, Dichgans J, Birbaumer N. Classical conditioning after cerebellar lesions in humans. Behav Neurosci 1993b; 107: 748–56.[Web of Science][Medline]
Dimitrov M, Grafman J, Kosseff P, Wachs J, Alway D, Higgins J, et al. Preserved cognitive processes in cerebellar degeneration. Behav Brain Res 1996; 79: 131–5.[Web of Science][Medline]
Fehrenbach RA, Wallesch CW, Claus D. Neuropsychologic findings in Friedreich's ataxia. Arch Neurol 1984; 41: 306–8.
Fiez JA. Cerebellar contributions to cognition. [Review]. Neuron 1996; 16: 13–5.[Web of Science][Medline]
Fiez JA, Petersen SE, Cheney MK, Raichle ME. Impaired non-motor learning and error detection associated with cerebellar damage. Brain 1992; 115: 155–78.
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: 545–547. Comment in: Science 1996; 272: 482–3.[Abstract]
Glickstein M. Motor skills but not cognitive tasks [comment]. Trends Neurosci 1993; 16: 450–451. Comment on: Trends Neurosci 1993; 16: 444–7.[Web of Science][Medline]
Grafman J, Litvan I, Massaquoi S, Stewart M, Sirigu A, Hallett M. Cognitive planning deficit in patients with cerebellar atrophy [see comments]. Neurology 1992; 42: 1493–6. Comment in: Neurology 1993; 43: 2153–4.
Harding AE. Clinical features and classification of inherited ataxias. [Review]. Adv Neurol 1993; 16: 1–14.
Hartje W, Rixecker H. Der Recurring-Figures-Test von Kimura. Normierung an einer deutschen Stichprobe. Nervenarzt 1978; 49: 354–6.[Web of Science][Medline]
Holmes G. The cerebellum of man. Brain 1939; 62: 1–30.
Ishihara S. Ishihara's design charts for colour-blindness of unlettered persons. Tokyo: Kanehara; 1990.
Kimura D. Right temporal-lobe damage. Arch Neurol 1963; 8: 264–71.
Klockgether T, Schroth G, Diener HC, Dichgans J. Idiopathic cerebellar ataxia of late onset: natural history and MRI morphology. J Neurol Neurosurg Psychiatry 1990; 153: 297–305.
Leiner HC, Leiner AL, Dow RS. Does the cerebellum contribute to mental skills? Behav Neurosci 1986; 100: 443–54.[Web of Science][Medline]
Leiner HC, Leiner AL, Dow RS. Cognitive and language functions of the human cerebellum [see comments]. Trends Neurosci 1993; 16: 444–7. Comment in: Trends Neurosci 1993; 16: 448–54.[Web of Science][Medline]
Posner MI, Petersen SE. The attention system of the human brain. [Review]. Annu Rev Neurosci 1990; 13: 25–42.[Web of Science][Medline]
Raven JC. Standard Progessive Matrices. London: H.K. Lewis; 1956.
Schmahmann JD. An emerging concept: the cerebellar contribution to higher function [see comments]. Arch Neurol. 1991; 48: 1178–87. Comment in: Arch Neurol 1992; 49: 1229–30.
Thach WT. On the specific role of the cerebellum in motor learning and cognition: clues from PET activation and lesion studies in man. Behav Brain Sci. 1996; 19: 411–31.
Thompson RF. Are memory traces localized or distributed? [Review]. Neuropsychologia 1991; 29: 571–82.[Web of Science][Medline]
Thompson RF, Kim JJ. Memory systems in the brain and localization of a memory. [Review]. Proc Natl Acad Sci USA 1996; 93: 13438–44.
Thompson RF, Krupa DJ. Organization of memory traces in the mammalian brain. Annu Rev Neurosci 1994; 17: 519–49.[Web of Science][Medline]
Timmann D, Diener HC. Limitations of PET and lesion studies in defining the role of the human cerebellum in motor learning [comment]. Behav Brain Sci 1996; 19: 477. Comment on: Behav Brain Sci 1996; 19: 411–31.
Timmann D, Kolb FP, Baier C, Rijntjes M, Mueller SP, Diener HC, et al. Cerebellar activation during classical conditioning of the human flexion reflex: a PET study. Neuroreport 1996; 7: 2056–60.[Web of Science][Medline]
Topka H, Valls Sole J, Massaquoi SG, Hallett M. Deficit in classical conditioning in patients with cerebellar degeneration. Brain 1993; 116: 961–9.
Trouillas P, Takayanagi T, Hallett M, Currier RD, Subramony SH, Wessel K, et al. International Cooperative Ataxia Rating Scale for pharmacological assessment of the cerebellar syndrome. J Neurol Sci 1997; 145: 205–11.[Web of Science][Medline]
Tucker J, Harding AE, Jahanshahi M, Nixon PD, Rushworth M, Quinn NP, et al. Associative learning in patients with cerebellar ataxia. Behav Neurosci 1996; 110: 1229–34.[Web of Science][Medline]
Wallesch CW, Horn A. Long-term effects of cerebellar pathology on cognitive functions. Brain Cogn 1990; 14: 19–25.[Web of Science][Medline]
Woodruff-Pak DS, Papka M, Ivry RB. Cerebellar involvement in eyeblink classical conditioning in humans. Neuropsychology 1996; 10: 443–58.
Zeki S. A Vision of the brain. Oxford: Blackwell Scientific; 1993.
Zimmermann P, Fimm B. Testbatterie zur Aufmerksamkeitsprüfung. Freiburg: 1992.
Received August 4, 1998. Accepted August 24, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  P. Ventura, A. Presicci, T. Perniola, M. G. Campa, and L. Margari Mental Retardation and Epilepsy in Patients With Isolated Cerebellar Hypoplasia J Child Neurol, September 1, 2006; 21(9): 776 - 781. [Abstract] [PDF]
|
|
|
|
 |
|
 M. Glickstein Thinking about the cerebellum Brain, February 1, 2006; 129(2): 288 - 290. [Full Text] [PDF]
|
|
|
|
 |
|
 B Gottwald, B Wilde, Z Mihajlovic, and H M Mehdorn Evidence for distinct cognitive deficits after focal cerebellar lesions J. Neurol. Neurosurg. Psychiatry, November 1, 2004; 75(11): 1524 - 1531. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Frings, R. Boenisch, M. Gerwig, H.-C. Diener, and D. Timmann Learning of Sensory Sequences in Cerebellar Patients Learn. Mem., May 1, 2004; 11(3): 347 - 355. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Clower, R. A. West, J. C. Lynch, and P. L. Strick The Inferior Parietal Lobule Is the Target of Output from the Superior Colliculus, Hippocampus, and Cerebellum J. Neurosci., August 15, 2001; 21(16): 6283 - 6291. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J S Fischer, R A Rudick, G R Cutter, S C Reingold, and National MS Society Clinical Outcomes Assessment T The Multiple Sclerosis Functional Composite measure (MSFC): an integrated approach to MS clinical outcome assessment Multiple Sclerosis, August 1, 1999; 5(4): 244 - 250. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||