Roles of the cerebellum and basal ganglia in temporal integration: insights gained from the synchronized tapping task

doi:10.21203/rs.3.rs-1964487/v1

Download PDF

Research Article

Roles of the cerebellum and basal ganglia in temporal integration: insights gained from the synchronized tapping task

https://doi.org/10.21203/rs.3.rs-1964487/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

The aim of this study was to clarify the roles of the cerebellum and basal ganglia for temporal integration. We studied 39 patients with spinocerebellar ataxia (SCA), comprising SCA6, SCA31, Machado–Joseph disease (MJD, also called SCA3), and multiple system atrophy (MSA). Thirteen normal subjects participated as controls. Participants were instructed to tap on a button in synchrony with isochronous tones. We analyzed the inter-tap interval (ITI), synchronizing tapping error (STE), negative asynchrony, and proportion of delayed tapping as indicators of tapping performance. The ITI coefficient of variation was increased only in MSA patients. The standard variation of STE was larger in SCA patients than in normal subjects, especially for MSA. Negative asynchrony, which is a tendency to tap the button before the tones, was prominent in SCA6 and MSA patients, with possible basal ganglia involvement. SCA31 patients exhibited normal to supranormal performance in terms of variability STE, which was surprising. In conclusion, cerebellar patients generally showed greater STE variability, except for SCA31. The pace of tapping was affected in patients with possible basal ganglia pathology. Our results suggest that interaction between the cerebellum and the basal ganglia is essential for temporal processing. The cerebellum and basal ganglia together with their interaction regulates synchronized tapping, resulting in distinct tapping patterns among different SCA subtypes.

cerebellum

basal ganglia

temporal processing

spinocerebellar ataxia

synchronized tapping task

Temporal processing in the brain is intricately associated with the sensorimotor system, spanning from motor-related cortical regions such as the supplementary motor area, premotor area, the inferior parietal cortex, and the premotor cortex to subcortical structures such as the basal ganglia and cerebellum [1–3]. Among them, the basal ganglia and the cerebellum play pivotal roles in coordinating and regulating the entire network function for motor control and temporal processing [4, 5], especially in motor timing mechanisms.

The basal ganglia are considered to be a core structure for processing explicit timing, especially in setting the pace of the internal clock, i.e., serving as a “core timer” for temporal processing [6–9], although this notion is currently controversial and has been modified (see Jones and Jahanshahi [10] for review). Early animal studies suggested that dopamine neurotransmission increases the presumed pacemaker speed. In humans, dopamine deficiency slows down the clock, while dopamine supplementation ameliorates the slowed pace, which has been corroborated by clinical studies in patients with Parkinson’s disease (PD) [11–16].

There is currently no consensus on the distinct role of the cerebellum compared with that of the basal ganglia in the timing function. Because its input and output flows can perform computations with millisecond precisions [17–22], the cerebellum is important for precise motor timing in the sub-second range [23–26]. However, the basal ganglia process timing in the seconds to minutes range, which requires more cognitive or explicit processing. Recent studies, however, indicate that supra-second timing requires both cerebellar and basal ganglia networks, where the cerebellum may not be viewed solely as a sub-second motor timing network, but rather a component of a timing network that cooperates with the basal ganglia for processing sub- and supra-second timing [27–30, see 31–33 for review]. The temporal information on the duration supplied to the cerebellum may then be sent back allowing the cortical and subcortical circuits to be adjusted.

The cerebellum is known to be involved in motor timing and in time estimation [23, 34, 20, 35]. Neuroimaging studies have shown cerebellum activation during the temporal discrimination of short intervals especially in the sub-second range, which can be processed automatically [36]. Neuropsychologically, cerebellar lesions increase discrimination thresholds rather than produce a distortion of time processing [37, 38, 29] and rather than speed up or slow down the pace of the internal clock. Coull and Noble [39] proposed that the roles played by the basal ganglia and the cerebellum depend on the context in which the timing tasks are performed, i.e., whether the timing required in the task is performed as part of a motor behavior or independent of movement or whether timing is explicit or implicit [40, 41]. Later studies, however, have not necessarily supported the absolute dichotomy of motor versus non-motor versus explicit/implicit (automatic/cognitive) timing because either can be associated distinctively with cerebellum or basal ganglia participation [24, 26, 42].

The discussions above would place the cerebellum’s role in temporal prediction in context, i.e., predicting event time in future motor action or predictive timing, as predicted from the forward model of the cerebellum [17], which is supported by by human and primate study results.

In humans, the cerebellum is implicated in monitoring ongoing timing and adjustment based on temporal predictions for shorter time intervals [25, 39, 41]. Avanzino et al. [43] showed that the cerebellum is engaged when temporal information is used to predict the temporal outcome of a motor act. Conversely, patients with cerebellar degeneration show reduced ability to use temporal regularity to facilitate motor preparation [44–48]. Coull et al. [41] investigated the neural substrate for predicting time intervals using functional magnetic resonance imaging (fMRI), and they found that activation of the inferior parietal cortex, left premotor cortex, and cerebellum was contrasted with a time reproduction task.

In primates, neural correlates of trial-by-trial variation in self-timing have been recorded in the cerebellum and the striatum, which are more active when a target event can be anticipated to occur at a specific moment in time relative to some preparatory cue or when the task involves temporal adjustments of behavior to the timing of these events [3, 49–53]. For example, when monkeys make a saccade in response to a single omission of periodically presented visual stimuli, neurons in the cerebellar dentate nucleus and basal ganglia exhibit a transient decrease in activity followed by a gradual increase in the firing rate (ramping neural activities) that peaks around the time of the next stimulus. In the dentate nucleus, the latency of self-timed saccades shows a good correlation with the rate of increase in preparatory activity when the delay period was 1 s or shorter, whereas it correlated with the onset time rather than the rate of rise of the preparatory activity when the delay period was longer than 2 s [30, 54–56]. Conversely, increasing basal ganglia activity correlated with the delay period for all the delay periods [52, 57]. These findings led to the hypothesis that the basal ganglia monitors the passage of time on a longer scale and codes the time from the previous event, measuring elapsed time relative to the intended interval, primarily in the supra-second range. However, the cerebellum represents the neuronal code for temporal prediction of the next stimulus, and it is engaged in a finer timing adjustment within a shorter time ahead in the range of 100–2000 ms just before the movement was made [21, 52, 53].

In patients, temporal prediction can be addressed using the synchronized tapping task [58–62], in which subjects are required to press a button or to tap a keyboard in synchrony with repetitive tones presented at fixed interstimulus intervals (ISIs). Using this task, we can also investigate the pace (how fast or slow) with which subjects can synchronize, and this indicates the pace of the internal clock. Additionally, temporal adjustments made based on the motor timing error, i.e., how subjects will adjust the timing of tapping according to the feedback on the difference in pace and the error signal for the timing of tapping relative to the presented tone on a trial-by-trial basis.

The slowest pace with which the subjects can keep pace when tapping in synchrony may indicate the span of temporal integration [63] and the duration that the subjects can perceive as one unit of time. As long as the ISI of the presented tones lies within this limit, subjects can synchronize their tapping accurately, whereas once the ISI exceeds this limit, the synchronization becomes disrupted, such that subjects no longer can tap in synchrony and become delayed in their tapping, responding reactively to the tones [61, 62]. This ISI limit is approximately 3 s in normal subjects, and above this limit, they have difficulty processing durations of time [58]. The cerebellum was activated in positron emission tomography (PET) and fMRI studies of sensorimotor synchronization at ISIs both in the millisecond and second ranges [64–67], and cerebellar lesions impair performance of this task [61, 68]. Compared with normal subjects, synchronized tapping to the tones in SCA patients fails at shorter ISIs, indicating a shortened span of temporal integration in SCA patients; these patients can predict a shorter time than normal subjects to maintain synchronous tapping in the seconds range.

Studying the task performance in neurological patients, in whom one or more of these structures, especially the basal ganglia and the cerebellum, are functionally affected in a relative homogeneous systematic manner within each group will reveal patterns that deviate from the normal pattern. This will help to dissect the contribution and the precise roles of the cerebellar and striato–thalamo–cortical networks among the motor-related systems for this timing task. In the present study, we compared the performance of synchronized tapping tasks in various types of the spinocerebellar ataxia (SCA), contrasting pure cerebellar types of SCD with those having a combination of ataxia and basal ganglia and brainstem pathology. The patients would show impairments specific to the pathology involved in each subtype, and by contrasting these different types of SCD, we can determine how the affected brain structures would lead to a specific pattern of tapping performance. We predicted that timed tapping tasks in patients with cerebellar disorders (e.g., [25, 43, 45, 46, 69]) would show deficits in automatic timing adjustments, whereas patients with basal ganglia disorders would present with impairment of the continuation mechanisms or pace adjustment in accordance with the presented tone. We also discussed how information about these two systems interacts and is integrated, examining how the task performance is disrupted in each subgroup of cerebellar ataxia patients.

Subjects

Thirty-nine non-demented SCA patients and 13 age-matched normal subjects participated in the present study. Written informed consent to participate in the study was obtained from all participants before the experiments. The experimental procedures were approved by the Ethics Committee at the University of Tokyo (No. 10889) and Kyorin University (H29-040-10), which was conducted in accordance with the ethical standards of the Declaration of Helsinki.

The SCA patients in the present study comprised 11 SCA type 6 (SCA6) patients (age, 64.7 ± 12.5 years; range, 44–88 years), seven SCA type 31 (SCA31) patients (age, 64.7 ± 10.1 years; range, 48–77 years), 11 Machado–Joseph disease (MJD) patients (age, 58.9 ± 9.0 years; range, 40–70 years), and ten patients with multiple system atrophy (age, 60.6 ± 9.0 years; range, 48–77 years). There was no statistically significant difference in age between normal subjects and SCD patients (Mann–Whitney U test, p = 0.830). The median duration of illness in SCA patients was 8 years (range, 1.5–35 years). Eleven SCA patients were genetically confirmed as having SCA6, while seven SCA patients had SCA31 and 11 SCA patients had MJD or SCA type 3 (SCA3). The patient characteristics are shown in Table 1.

Table 1

Subjects’ demographic information.
	Normal	SCA6	SCA31	MJD	MSA
Number of subjects	13	11	7	11	10
Men / Women	6 / 7	4 / 7	5 / 2	6 / 5	6 / 4
Age (years, mean ± SD)	60.8 ± 11.8	64.7 ± 12.5	64.7 ± 10.1	58.9 ± 9.0	60.6 ± 9.0
Disease duration (years, mean ± SD)	-	12.8 ± 11.1	10.3 ± 4.5	10.1 ± 3.4	2.9 ± 1.1
MMSE score (mean ± SD)	29.3 ± 1.1	28.1 ± 2.0	29.1 ± 1.9	27.7 ± 2.6	28.0 ± 2.4
SARA score (mean ± SD)	-	12.8 ± 9.1	12.0 ± 5.4	15.1 ± 4.8	13.5 ± 5.0
SCA, spinocerebellar ataxia; MJD, Machado–Joseph disease; MSA, multiple system atrophy; MMSE, Mini–Mental Status Examination; SARA; Scale for the Assessment and Rating of Ataxia

Ataxia severity in each patient was scored using the Scale for the Assessment and Rating of Ataxia (SARA) [70, 71]. The median SARA score for SCA patients was 12 (range, 2–25). We excluded subjects with cognitive or sensory impairment or other neurological disorders. For this purpose, the Mini–Mental Status Examination (MMSE) [72] was used to exclude SCA patients and normal subjects with dementia from this study. The median MMSE score of SCA patients was 29 (range, 22–30) and that of normal subjects was 30 (range, 26–30). There were no significant differences between subject groups (Mann–Whitney U test, p = 0.118).

Task procedure

Synchronized tapping task

The synchronized tapping task used in this study is illustrated in Fig. 1A. The subjects’ task was to tap the button exactly in synchrony with the tones, which were presented through a speaker, at a fixed ISI. One of the buttons on a USB response pad (Response pad RB-840, Cedrus Corporation, Seattle, WA, USA) served as the response key. Each tone was a simple tone with a frequency of 500 Hz and a duration of 50 ms. The experiment was performed using commercial software (Experiment Builder, SR Research, Ottawa, CA), which allowed presentation of the tones, recording of the button that was pressed, and the time at which the button was pressed with high temporal resolution, as described in Tokushige et al. [62]. The subjects were encouraged to tap in anticipation of the time of the oncoming tone and not to press the button “in reaction to” the tones. Rhythmic stimulus templates presented were quickly acquired even by the patients, and repetitive rhythmic patterns were usually established within several taps, although there was some variability.

In each session, the ISI was kept constant (isochronous sequence), and it was selected from among the following six different ISIs: 600, 900, 1200, 1800, 2400, and 3600 ms. Approximately 50 trials (taps) were performed for each ISI, but trials for ISIs 1200 and 1800 ms were divided into two separate sessions (25 taps × 2), and those for ISIs 2400 and 3600 ms were divided into three separate sessions (17 taps × 3). For all sessions, ten extra taps were added to the first part for training and ignored during data analysis because the patients’ responses tended to be unstable at the beginning of each session. The order of ISIs selected for consecutive sessions was randomized but counterbalanced across the subjects, such that each subject eventually completed each of the sessions. During and after each session, no feedback was provided to the subjects regarding their responses. The inter-tap interval (ITI) refers to the time between successive taps.

Simple reaction task

In addition to the synchronized tapping task, we also implemented the simple reaction time (SRT) to auditory tones; the subject was instructed to press the response key as quickly as possible after listening to the tone. Sixty tones were presented at intervals of 2400–4800 ms. The subsequent interval was randomly selected each time the subject pressed the button so that the subject could not predict the time of the upcoming tone. Among the 60 taps, the first ten taps and SRTs over 2000 ms were ignored. The SRT ensured that all patients could move their fingers quickly to perform the tapping task properly.

Data collection and statistical analysis

The data collected were the button-press times and tone presentation times. From these data, we derived the time difference between each tap onset and the tone (synchronizing tapping error [STE]), in accordance with the methods described by Mates et al. [58]. STE was calculated by subtracting the tone time from the actual measured tapping onset, where a positive value indicated that the tap was delayed relative to the tone, while a negative value indicated that the tap preceded the tone (Fig. 1A).

Even for normal subjects, performance of the synchronized tapping task showed negative asynchrony. The tap time preceded the tone time by approximately by 30–50 ms because achieving exact synchrony of the tapping timed with the tones requires that the subjects begin before the timing of the tone (see Discussion; [58, 73]). As the ISI becomes longer (typically over 2000 ms), however, synchronization fails and subjects’ taps begin to lag behind the tones [58, 61, 62].

The patients participating in this study almost always produced only one tap per tone as instructed. Thus, each tap could be reliably associated with the corresponding tone by determining the closest tone in relative time, starting from the time preceding the tone by half the ISI to the time following the tone by half the ISI.

The variability of STE and ITI should each reveal different aspects of abnormal temporal processing. STE variability would be expected to reflect the deviation of tapping time relative to the time of the tone and its trial-to-trial adjustment as tapping proceeds repetitively. Additionally, the ITI variability would serve as a measure of the abnormal internal clock speed, which is the variability in the tapping pace deviation with respect to the tone pace (see Discussion).

For deriving the proportion of delayed tapping, the tapping data in each session was first depicted as a histogram at each ISI (Fig. 1B), with the proportion of trials plotted against each STE bin on the abscissa. Based on the STE, the temporal integration limit was derived for each subject as described below. At shorter ISIs, subjects are able to tap approximately in synchrony with the tones, as shown by the narrow distribution clustered around 0 ms (first histogram for STE in Fig. 1B). Above a certain ISI, it becomes more difficult for the subjects to tap in synchrony with the tones, and the STE begins to show a broader and delayed distribution. As the ISI gets longer, the subject has even more difficulty synchronizing the taps with the tones, and the subject instead begins to tap after the tones rather than with accurate synchronization. We defined “delayed tapping” as taps with latency exceeding 100 ms compared with the tone (Fig. 1B). Taps that lag behind occur at a latency of 100–200 ms after the tone, forming a relatively sharp peak in the histogram, which corresponds to the reaction time range that was measured in the SRT task [61, 62]. Therefore, when temporal integration is disrupted, there is a transition from a state in which subjects predict the time of the next tone and tap in anticipation to a state in which subjects are reacting to the tones. The ISI at which this disruption takes place is called the limit of temporal integration, which was defined as the ISI where the proportion of delayed tapping exceeded 10% of the total trials. In the example shown in Fig. 1B and C, we estimated the limit of temporal integration to be 1500 ms, i.e., the midpoint between 1200 ms and 1800 ms.

In the following analysis, STEs over 100 ms were excluded from the statistics comparing each ISI to the disease type because these large STEs indicate delayed tapping, which means that the intended synchronized tapping has failed to occur.

Statistical assessment

Statistical analyses were performed using the R statistical software package (ver. 3.6.1, The comprehensive R Archive Network). We assessed how the tapping performance varied with ISI in terms of the STE, ITI, and proportion of delayed tapping as well as the variability of the respective parameters and how these differed among different subject groups. These parameters were analyzed using repeated measures two-way analysis of variance (ANOVA), where the within-subject factor was ISI and the between-subject factor was the subject group. When necessary, the Greenhouse–Geisser correction was used when Mauchly’s sphericity test showed non-sphericity. To compare the limit of temporal integration among different groups, a one-way ANOVA was performed where the between-subject factor was the subject group. If the analysis results were significant, Tukey’s post hoc test was performed to identify which of the pairs showed a significant difference. The significance threshold was set at p < 0.05 for all analyses.

We also investigated the STE time-course for each session, i.e., how the STE varied with the increasing tapping count. The time course of the tapping performance without excluding the initial ten taps and STEs over 100 ms was observed by averaging the STE at each tapping count for all subjects in each group. Here, STEs that exceeded the range of the mean ± two standard deviations (2SDs) for each subject at each tap count were excluded from analysis. We performed Pearson’s rank correlation analysis between the average STE at each tap and the tapping count.

To investigate the influence of the cerebellar motor dysfunction (ataxia) on SCA patients’ tapping and to investigate how age, disease duration, or ataxia scores (SARA) affected task performance and the above-mentioned parameters, multiple linear regression analyses were conducted for the SCA patient groups. We used the limit of temporal integration as the outcome variable, and age, disease duration, and ataxia scores were the predictor variables. The partial regression coefficients of predictor variables were expressed as β, and the p value from the t-test for the regression slope of predictor variables was used to determine the probability.

Figure 2 shows the STE histogram for each subject group’s ISIs from 600 ms to 3600 ms. In normal subjects, tapping begins to lag behind the tones as the ISI length increases, and tapping time distribution relative to the tones formed a narrow single STE peak approximately 200 ms after the tones, which corresponds to the timing of the “delayed tapping”. This phenomenon was also observed in other SCA groups, such as SCA6, MJD, and MSA. However, in the SCA31 group, this delayed tapping was rarely observed, showing no peak of STE 200 ms after the tones.

Overall, the proportion of delayed tapping apparently increased along with the ISI length in SCA patients, which was also affected by disease types (Fig. 3A). A repeated measures two-way ANOVA showed that the effects of ISI (F [5, 235] = 36.772, p < 0.001, ε = 0.474) and disease type (F[4, 47] = 5.179, p = 0.00154) were both significant. The interaction between ISI and disease type was also significant (F[20, 235] = 2.09, p = 0.0333, ε = 0.474), suggesting that the effect of ISI on delayed tapping depended on the disease types. In SCA31, however, the delayed tapping stayed at a low level even when the ISI became longer, while in the other groups, the occurrence of delayed tapping increased as the ISI duration increased.

The mean STE tended to shift toward the negative direction as the ISI increased (Fig. 3B). Using a repeated measures two-way ANOVA, the effect of ISI was significant (F[5, 215] = 29.573, p < 0.001, ε = 0.466) while the effect of disease type was not significant (F[4, 43] = 2.144, p = 0.0917), and the interaction between these factors was significant (F[20, 215] = 2.180, p = 0.0279, ε = 0.466). These results demonstrate that the extent of STE mainly depended on the ISI, and this dependency varied between disease types. SCD patients, especially SCA6 and MSA patients, tended to show a strongly negative STE as the ISI duration increased.

The SD of STE increased linearly as the ISI increased in all subject groups (Fig. 3C), conforming to the scalar property. A repeated measures two-way ANOVA indicated that the effect of ISI (F [5, 215] = 126.668, p < 0.001, ε = 0.552) and disease type (F [4, 43] = 5.542, p = 0.00224) were both significant, while the interaction between these factors was not significant (F[20, 215] = 1.476, p = 0.149, ε = 0.552). These results show that the STE linearity between ISI and SD was maintained across all subject groups. SCD patients, especially those of SCA6 and MSA, showed a longer SD for STE than normal subjects and SCA31 patients, reflecting their inaccurate and irregular tapping relative to the tones.

As mentioned above, the relationship between the SD of STE and ISI was approximately linear. To visualize this linearity, we divided the SD of STE by ISI (Fig. 3D). The SD curves of STE and ISI were almost level in the normal and SCA31 groups. However, in the other SCD groups, the SD of STE was large at shorter ISIs such as 600 ms and 900 ms and decreased at longer ISIs. Using a repeated measures two-way ANOVA, the effects of ISI (F [5, 215] = 9.907, p < 0.001, ε = 0.695) and disease type (F [4, 43] = 5.542, p = 0.00109) were significant, while the interaction between these factors was not significant (F [20, 215] = 1.18, p = 0.299, ε = 0.695). These results showed that the linearity of SD for STE, corresponding to the scalar property, was maintained in normal and SCA31 patients through all ISIs, but not in SCA6, MJD, and MSA especially at shorter ISIs. This is because patients with SCA6, MJD, and MSA were unable to keep pace with fast tapping rhythms, while this was not a problem for normal and SCA31 patients.

The coefficient of variation (CV) for ITI was stable with an increasing ISI, except for MSA patients (Fig. 3E). A repeated measures two-way ANOVA showed that both ISI (F [5, 235] = 15.117, p < 0.001) and disease type (F [4, 47] = 3.414, p = 0.0157) significantly affected the CV of ITI, and the interaction between them was also significant (F [20, 235] = 1.986, p = 0.00869). Although the effect of ISI was significant, the CV of ITI curves tended to show negative peaks for ISIs of 900 ms or 1800 ms, and the overall curves did not show a monotonous increase or decrease. This tendency mostly fits the scalar property similar to the SD of STE (Fig. 3C). Among all disease subtypes, MSA patients showed an increased CV for ITI especially for short ISIs.

Figure 3F compares the limits of temporal integration among different subject groups. A one-way ANOVA showed that the effect of disease types was significant (p = 0.00197). Post-hoc Tukey multiple comparisons revealed that the differences between SCA31 and MJD (p = 0.00725), between SCA31 and MSA (p < 0.001), and between SCA31 and SCA6 (p = 0.0418) were significant, indicating that SCA31 patients have a longer limit of temporal integration than the other groups, while the other groups were comparable. The difference between the normal subjects and each of the other SCA groups was not significant (p > 0.05).

Figure 3G compares the SRTs in each group. A one-way ANOVA showed that there was a significant difference among these groups (F [4, 18.10] = 7.427, p = 0.0010). Post-hoc Tukey multiple comparisons revealed that SCA6 patients have significantly longer SRTs than normal subjects (p = 0.0147).

Supplementary Fig. 1 shows the relationships between ISI and the mean + 2SD of STE, which would approximate the upper limit of the STE distribution, and this includes data from Fig. 3B and 3C. Both the effects of ISI (F [5, 215] = 69.629, p < 0.001, ε = 0.746) and disease type (F [4, 43] = 5.526, p = 0.00111) were significant, indicating that the mean + 2SD, which is the upper limit of the STE distribution, increases as the ISI duration increases, and it is also affected by the disease type. The interaction between ISI and disease type was not significant (F [20,215] = 1.669, p = 0.0624, ε = 0.746), showing that the effect of ISI on the upper limit of STE distribution is maintained in all disease types.

To clarify which disease group has an abnormally large STE, one-to-one comparisons between normal subjects and each disease group were performed using repeated measures two-way ANOVA, with a within-subject factor of ISI and a between-subject factor of subject group (Normal vs. MJD, Normal vs. SCA6, Normal vs. SCA31, and Normal vs. MSA). The effect of subject group became significant only when the Normal vs. MSA group was selected (p = 0.0135, Bonferroni correction), while there was no significant difference in the other comparisons (p > 0.05). These results show that MSA patients have a larger STE than other groups.

Figure 4 shows the STE time-courses in each session plotted against the tap count, where the ISI was 900 ms, which were averaged for all subjects in each group. Initially, the STE decreased dramatically with the trial number up to approximately the tenth tap in all groups. After this initial decrease, the STE continued to decrease slowly. After the tenth tap, the correlation between the tap count and STE was significant in normal subjects (R = − 0.633, p < 0.001) and the SCA31 (R = − 0.388, p = 0.00536) group, but not in the SCA6 (R = − 0.202, p = 0.159), MJD (R = − 0.155, p = 0.283) and MSA (R = − 0.259, p = 0.0694) groups. These correlations were strongly negative only in the normal and SCA31 groups, which may be because the noise levels in the STE curves was relatively small in these subjects, reflecting accurate and regular tapping.

In Supplementary Fig. 2, the STE time-course where the ISI was 3600 ms was plotted in the same way as in Fig. 4. Because these sessions are divided into three parts rather than continuing with 60 taps in one session, the maximal tap count is 25 in this plot. There was no significant correlation between the tap count and STE (p > 0.05).

Table 2 shows the result of multiple linear regression analysis using the limit of temporal integration as the outcome variable, and age, disease duration, and ataxia scores (SARA) as predictor variables. None of these predictor values significantly correlated with the limit of temporal integration.

Table 2

Results of multiple linear regression analysis where the outcome variable is the limit of temporal integration (ms).
Explanatory variable	p value	β
Age (years)	0.207	−24.0
Disease duration (years)	0.525	15.9
SARA score	0.374	−26.8
SARA; Scale for the Assessment and Rating of Ataxia

Here, we investigated the accuracy of synchronized tapping and the span of temporal integration in patients with four types of SCA (SCA6, SCA31, MJD, and MSA) and compared their performance with that of normal subjects. These groups included different types of SCA patients, with varying degrees of basal ganglia involvement, allowing us to investigate the contribution of the cerebellum and basal ganglia to sensorimotor integration from the different performance patterns. Overall, SCA patients showed increased variability in tapping on the basis of an increased SD of STE. Furthermore, the limit of temporal integration was reduced in SCA compared with normal subjects for most SCA subtypes, while delayed tapping was promoted prematurely at shorter ISIs. Patients with cerebellar diseases show variable timing for tapping around the tones (Fig. 3C, D) and a shortened limit of temporal integration (Fig. 3F), and these abnormalities especially stood out in MSA patients, who have basal ganglia dysfunction. Consistent with both the hypothesis that the cerebellum is involved in motor timing and with our previous study results in patients with pure cerebellar ataxia [61], cerebellar dysfunction leads to disruption of temporal integration, with prominent delayed tapping. Furthermore, these results suggest that cooperation between the cerebellum and basal ganglia is essential for successful synchronized tapping, and diseases in both structures generate a synergistic effect on the difficulty of performing this task.

However, we also found significant differences among different SCA subtypes. Whereas the deterioration of tapping performance is especially prominent in patients with MSA, SCA31 showed low tapping time variability, a smaller proportion of delayed responses, and a longer temporal integration limit, which is comparable to normal conditions or even comparable to a supernormal level. SCA6 and MJD showed intermediate performances. We analyzed how the extent and distribution of pathophysiology in the cerebellum and other timing systems including the basal ganglia contributed to the performance of the synchronized tapping task. We also discussed how the cerebellum and basal ganglia may interact to adjust the tapping time in synchrony with the tones and how this is affected by their dysfunction, leading to disruption of temporal integration [62, 74].

Correction mechanisms for synchronized tapping

To explain the behavior during a synchronized tapping task in normal subjects [75, 76], a process model of synchronization was proposed for the synchronized tapping task. When subjects perceive the mismatch between the presented and actual tapping times and adjust the taps to synchronize with the tones, they may use one of two different strategies to adjust the timing of the taps to occur simultaneously with the tones [77–79]. The first strategy is a period correction mechanism, in which the subject adjusts their overall pace of tapping or pace of the internal clock reflected in their internal representation of the interval to match the pace of the tones, i.e., to approximate the pace of the presented tones through a tempo generation process (beat-based correction [80]). This mechanism detects the difference in the paces of tapping and the tones and adjusts the deviation by slowly adjusting the pace of tapping (ITI). The other strategy is a phase correction mechanism in which the subjects adjust their tapping through a trial-and-error correction process, i.e., STE is corrected to ensure that the instantaneous deviation of the tapping times from those of the corresponding tones does not accumulate over time, which would lead to a loss of phase stability between the tones and the tapping responses (duration-based correction [80]). Disruption of either mechanism would result in an increased deviation of the tapping time from the tone time.

As a working hypothesis, we postulated that the period correction mechanism is affected mainly through the basal ganglia, while the phase correction mechanism is the function of the cerebellar system. To this end, we analyzed the tapping performance in different SCA subtypes using the stability of ITI across trials as an indicator of period correction and the trial-to-trial variability of STE as an indicator of phase correction [58, 81].

Roles of the cerebellum and basal ganglia in synchronized tapping

We first compared the trial-to-trial variability in tapping time around the tone time (STE) among different SCA subtypes. As shown in Figs. 2 and 3C, the overall inter-trial variation in STE (SD of STE) was larger in most SCA subtypes at the ISIs studied relative to normal subjects. This suggests that, as discussed above, most cerebellar subtypes (excluding SCA31) show a larger STE variability compared with that of normal subjects, which is consistent with the notion that the cerebellum is involved in motor timing, i.e., variability of timing around the pacing tones [23, 34, 82]. The STE variability increased from SCA31 to MJD to SCA6 in this order, possibly reflecting the distribution or extension of pathology within the cerebellum in these orders [83]. Although the cerebellar vermis is affected from early stages in both SCA6 and SCA31 [84], the slower progression of degenerative processes in SCA31 compared to those in SCA6 may allow a more functional compensation to take place. SCA31 patients were able to adjust their tapping time relative to the tone time at near-normal levels. Due to their restricted cerebellar pathology, SCA31 patients were capable of effectively using the detected error signal between the motor action timing and that of the tone (see below).

In contrast to the SD of STE, we found that the variation in ITI (CV of ITI, Fig. 3E) was stable in most SCA subtypes, almost to a similar level as that in normal subjects, with the exception of MSA patients. If we take this measure as the period correction mechanism, the results suggest that the tapping pace in most SCA patients was similarly stable compared with that of normal subjects as long as they tapped “at their own pace”; the central timekeeper component, subserved by the cerebellar hemisphere, was relatively spared. The CV of ITI in MSA patients was large, especially at shorter ISIs, suggesting disruption of the period correction mechanism and that MSA patients could not sustain a stable pacing rhythm at their own pace. Because of cerebellar afferent pathway involvement in MSA, a cerebellar afferent system deficit may explain the greater disruption of the period correction system in MSA.

Previous studies identified some differences in the roles of cerebellum and basal ganglia in performing timing tasks. Kunimatsu et al. [52] recorded activation of the caudate nucleus and cerebellar dentate nucleus in monkeys performing repetitive saccades using instructed timing following a visual cue, and they found a distinct activation patterns in these structures. The caudate was activated throughout the delay period, and the dentate nucleus was recruited only during the last part of the delay period. Their results suggested that basal ganglia measure elapsed time similar to that of a pacemaker and that the cerebellum is responsible for the fine adjustment of motor timing. This is consistent with our results, which showed that overall adjustment of the tapping pace to the tone pace is made by the basal ganglia (period correction mechanism), whereas fine adjustment of the tapping time around the tone time is performed by the cerebellum (phase correction mechanism). The sensitivity of the cerebellum to stimulus dynamics and subsequent integration with motor control allows it to accurately measure intervals within a range of 100–2000 ms [42, 53], but the basal ganglia may take this over for longer time spans. Thus, the large SD of STE in SCD patients may be caused by the poor adjustment of finger movement timing relative to the tones due to cerebellar lesions (enhancement of STE). MSA patients showed the poorest tapping performance possibly because cerebellar lesions and the pacemaker of finger movement was also disabled due to basal ganglia lesions.

Integration of cerebellar and basal ganglia systems for temporal integration

We have thus far discussed the tapping performance in cerebellar ataxia mainly in terms of the SD or average of STE, with the basal ganglia determining the overall tapping pace. Previous studies showed that olivocerebellar and striato–thalamo–cortical system circuits are connected by the inferior olivary nucleus, which receives inputs from the cerebral cortex (Fig. 5). How is the information integrated from the cerebellar and basal ganglia systems for successful and accurate tapping synchronized with the tones able to continue? This is determined by the interaction between the difference in the predicted and actual time of oncoming tones (asynchrony) and the trial-to-trial tapping variability (STE).

As mentioned in the introduction, the synchronized tapping task may be construed as a task of temporal prediction because for the subjects to tap in synchrony with the tones, they would have to predict the time of the oncoming tone and prepare a tapping response in synchrony with it, especially when the ISI is shorter than 2 s. At longer ISIs, when the ISI becomes longer than the limit of temporal integration, i.e., when successful interaction between the basal ganglia and cerebellum fails, it becomes more difficult for the subjects to perceive the pace of the tones automatically as a stable rhythm. The tapping then becomes more reactive to the tones than predictive of them, which means that they perform the tap after listening to the oncoming tones rather than predicting them.

When synchronized tapping is successfully performed, even normal subjects tend to tap ahead of the synchronized tapping task (negative synchrony; [73, 77, 78]). Although the mechanism remains uncertain and many possibilities have been proposed for its generation [59, 60, 73, 79, 85–88], negative asynchrony most likely occurs through temporal prediction of the oncoming tone relative to the veridical timing of the tone. More specifically, perceptual underestimation of the duration of empty ISIs occurs [81], and thus, the tap is performed before the tone occurs. Our data showed that this underestimation occurred for broad ISI ranges because STE variability (SD) increases approximately proportionally with increasing ISI within the range of 1800–3600 ms (Figs. 3C and D).

Overall, negative asynchrony was more prominent in cerebellar patients than in normal subjects. It also differed among different patient groups, with negative asynchrony occurring most frequently in patients with MSA and SCA6, with intermediate frequency in patients with MJD and SCA31, and least frequently in normal subjects (Fig. 3B). While the smallest extent and degree of pathological changes in SCA31 might explain the smaller negative asynchrony, patients with MJD, a multisystem disorder involving the cerebellum, the basal ganglia, and afferent pathways from the spinal cord, unexpectedly exhibited low negative asynchrony that was comparable to SCA31 patients. Conversely, SCA6 patients who presented with a similarly pure cerebellar ataxia, showed a negative asynchrony that was almost comparable to that of MSA patients who have a multisystem disorder involving the afferent pathways into the cerebellum from the basal ganglia and the cerebral cortex. Thus, the segregation between MSA and SCA6 compared with that of MJD and SCA31 cannot be readily explained by the extension and distribution of pathology within the cerebellum alone.

The present study showed disruption of temporal integration in most cerebellar ataxia subtypes (Fig. 3B). Both normal subjects and SCA patients tapped well ahead of the tones; the STE distribution showed a negative value across trials, which increased with the ISI in the range of 1200–3600 ms and was more prominent for SCA patients (negative [motor] asynchrony; Fig. 3B). There was normal to supranormal performance of temporal integration in SCA 31.

Using the same task, Matsuda et al. [61] showed that the limit of temporal integration is around 1800 ms in pure cerebellar ataxia patients, compared with 2–3 s in normal subjects. Psychologically, the results can be interpreted as a shorter span of temporal prediction, i.e., they can predict only a shorter time in the future compared with normal subjects. This is consistent with the report showing impaired spatio-temporal predictive motor timing associated with SCA6 [89].

The limit of temporal integration is influenced by the distribution of tapping timing around the tone time, which was determined using the STE variability and the negative asynchrony level (Fig. 6). If the distribution extends at its positive end into the reaction time range (i.e., the range of button-press time if the tap would have been made in response to the tone presentation), reactive responses to the tones would be favored over those timed to the predicted time of the oncoming tone, and thus, increase the proportion of delayed tapping (Fig. 6A). Patients with cerebellar ataxia generally show a larger SD of STE than normal subjects, and it increases with larger ISIs (Fig. 2C). If negative asynchrony becomes more negative, the overall distribution of tapping times around the tone should also shift towards a negative range and reactive responses then become unnecessary; thus, the proportion of delayed tapping would be smaller (Fig. 6C). If negative asynchrony becomes more positive, the opposite situation occurs (Fig. 6D). Tokushige et al. [62] showed that the span of temporal integration is prolonged in patients with early PD and speculated that a larger negative asynchrony prevented the delayed tapping from appearing prematurely because the tap time distribution shifted negatively around the tone time, although the tap time was slightly more variable than in normal subjects.

The SD of STE increased from SCA31 to SCA6, MJD, and MSA in this order, while negative asynchrony was at a similar level. The approximate distribution of STE at its positive end that was reflected in the mean negative asynchrony + 2 × (SD of STE) was the highest in MSA and the lowest in SCA31 patients (Supplementary Fig. 1), showing an intermediate value in normal subjects, SCA6, and MJD patients. This tap time distribution around the tone time (STE) plus negative asynchrony approximately explained the relative span of temporal integration. Temporal integration was most reduced in MSA patients and most preserved in SCA31 patients, while MJD and SCA6 showed intermediate values. Similarly, the proportion of delayed tapping (reactive responses) was smallest in SCA31, followed by MJD and SCA6, and it was largest in MSA.

Superior tapping performance in SCA31

One striking finding in the present study was that SCA31 patients showed the most accurate tapping performance, even to a normal or supranormal level, whereas SCA6 patients showed a significantly larger variability in STE, with a significantly larger negative asynchrony. In SCA31, the STE variability around the tone time (SD of STE) was significantly smaller compared with other SCA subtypes, and it was almost comparable to that of normal subjects. Thus, in SCA31 patients, the distribution of tap times around the tone time was smaller and did not reach the time range at the larger end of distribution corresponding to the reaction time range, which would cause delayed tapping.

This result was unexpected because the clinical manifestation and underlying pathology of both SCA6 and SCA31 are considered to be purely cerebellar. Additionally, the distribution of cerebellar atrophy does not differentiate between these two types of cerebellar ataxia because midline cerebellar structures such as the vermis, especially the anterior lobe that is implicated in temporal processing, are affected both in SCA subtypes from early stages of the disease [90]. However, in SCA6, pathological extension into the dentate nucleus and extracerebellar structures is common including the basal ganglia, cerebral cortex, pontocerebellar structures such as middle cerebellar peduncle, olivary nucleus, or even the spinal cord [83, 91, 92]. Conversely, extracerebellar extension of pathology is virtually absent in SCA31, and the pathology is almost completely confined to the Purkinje cells in SCA31 [83, 93]. At the cellular level, the granule cells and the dentate nucleus are also affected in SCA6 [83]. Furthermore, the slower progression of degenerative processes in SCA31 compared with SCA6 may allow more functional compensation and larger degeneracy to take place. Thus, the near-normal performance of the synchronized tapping task may indicate that the dysfunction of Purkinje cells in SCA31 is not sufficient to cause the trial-to-trial STE variance in tapping.

Although the SD of STE was comparable between SCA31 patients and normal subjects (Fig. 3C), the span of temporal integration was larger for the former group. This may be because of the smaller SD of STE in SCA31 patients than in the other patients. The STE distribution was approximated by the negative asynchrony ± 2SD of STE, which was mostly below that of the reaction time range in response to the tones (Supplementary Fig. 1). The smaller SD of STE made the upper limit of STE smaller than normal (Fig. 6B), leading to the reduction of delayed tapping and finally enlarging the span of temporal integration. In this situation, the upper limit of STE becomes smaller than normal, and it was also observed in early PD patients in our previous study [62]. In early PD patients, the span of temporal integration was larger than for normal subjects due to the larger negative asynchrony. Their upper limit of STE (i.e., negative asynchrony + 2SD of STE) became smaller than normal because of the prominent negative asynchrony (Fig. 6C). Therefore, in both SCA31 and early PD, the span of temporal integration became longer than normal because the upper limit of STE (i.e., negative asynchrony + 2SD of STE) became smaller than normal.

Another possibility for the difference in performance between SCA6 and SCA31 may be that the latter patients may have a particularly efficient timer for stimulus timing. This may be due to the pathological difference between SCA6 and SCA31 that affects the inferior olive, which subserves the period correction timing mechanism, in the former but it may be preserved in the latter. The superior performance of tapping in SCA31 patients on the basis of the trial-by-trial variability in reaction time may reflect the function of the relatively preserved olivocerebellar system, which is consistent with the report that the inferior olive and the climbing fiber system mediate temporal information encoding that is required for both motor and nonmotor cognitive processes and to send teaching signals representing errors to Purkinje cells via the climbing fibers [17, 94]. This may play an important role in the interaction between the basal ganglia and cerebellar signals in temporal integration [80]. The contrast may also apply to MSA and MJD. The cerebellar afferent pathway through the pontine nucleus is affected in both of these two cerebellar subtypes; the olivary nucleus is affected in the former but preserved in the latter [83], which may explain the smaller trial-by-trial variability in the latter group.

The performance in the synchronized tapping task was significantly more affected in MSA patients than in other SCA subtypes, on the basis of the SD of STE and negative asynchrony at longer ISIs and the CV of ITI at shorter ISIs. This may be because cerebellar afferent fibers from the cerebral cortex are affected in MSA but are more preserved in other SCA subtypes. For a sensorimotor integration task such as that used in the present investigation, cerebrocerebellar connectivity should play an important role [95–99], and in MSA, this interaction may be disrupted. The spinocerebellar afferent pathway is affected in MJD but the cerebrocerebellar afferent pathways may be relatively preserved [83], and this may explain the relatively smaller SD of STE in this group.

Limitations

The first limitation is the number of subjects in some patient subgroups, which was relatively small although hospitals throughout Japan participated in this study. A larger cohort would be required to clarify the relationship between age and temporal integration in various cerebellar subtypes. Second, we postulated that the pathological extension of cerebellar and extracerebellar structures would be relatively uniform within each cerebellar subtype based on literature about genetic classification and clinical presentations, but this was not pathologically confirmed in our cases, and there could be variation across cases. Taken together, these limitations suggest that the discussion should be interpreted with caution regarding the possible neural substrate suggested for distinct patterns of temporal integration in different SCA subtypes.

On the basis of the model developed by Vorberg et al., we postulated that the cerebellum controls the phase correction mechanism, while the basal ganglia control the period correction mechanism. The present study showed that cerebellar dysfunction in SCA patients overall exhibits increased trial-by-trial variation for STE (SD of STE) by disrupting the phase correction system and comprising the fine adjustment of the motor timing. This variation increases in approximate proportion to ISI, supporting the scalar property. Comparing cerebellar subtypes, the defect of the phase correction may largely reflect the extension of pathology within each cerebellar subtype.

However, the period correction mechanism uses the regularity of tone sequence to predict and prepare the tap in synchrony with the oncoming tone. The mechanism was especially disrupted in cerebellar subtypes with concomitant basal ganglia involvement, consistent with the role of the basal ganglia in setting the rate of the internal clock, in accordance with the striatal beat frequency model [15, 100, 101].

Ethical Approval

The experiments were approved by the Ethics Committee at the University of Tokyo (No. 10889) and Kyorin University (H29-040-10), in accordance with the ethical standards of the Declaration of Helsinki.

Consent to Participate

Written informed consent to participate in the study was obtained from all participants.

Consent to Publish

Written informed consent to participate in the study was obtained from all participants.

Authors Contributions

S.T. and Y.T. wrote the main manuscript text. S.T. prepared the figures. All authors reviewed the manuscript.

Funding

Dr. Tokushige was supported by a Research Project Grant-in-aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (Nos. 19K17046, 21K15687).

Dr. Ugawa received grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan (Nos. 25293206, 15H05881,16H05322, and 18K10821), The Research Committee on the Medical Basis of Motor Ataxias, Health and Labor Science Research Grants, the Ministry of Health, Labour and Welfare of Japan, the Support Center for Advanced Telecommunications Technology Research, the Association of Radio Industries Businesses, and the Novartis Foundation (Japan) for the Promotion of Science, Nihon Kohden, Ltd., Takeda Pharmaceutical Company Limited, Nippon Boehringer Ingelheim Co., Ltd., and Mitsubishi Tanabe Pharma Corporation.

Dr. Terao was supported by a Research Project Grant-in-aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (18H05523).

Competing Interests

Dr. Tokushige has received a speaker’s honorarium from Otsuka Pharmaceutical Co., Ltd., Daiichi Sankyo Co., Ltd., Dainippon Sumitomo Pharma Co., Ltd., Takeda Pharmaceutical Co., Ltd., Mitsubishi Tanabe Pharma Corporation, and AbbVie Inc.

Dr. Ugawa received honoraria from the Taiwan Movement Disorders Society, Chinese Neurology Society, Astellas Pharma Inc., Eisai Co., Ltd., FP Pharmaceutical Corporation, Otsuka Pharmaceutical Co., Ltd., Elsevier Japan K.K., Kissei Pharmaceutical Co., Ltd., Kyorin Pharmaceutical Co., Ltd., Kyowa Hakko Kirin Co., Ltd., GlaxoSmithKline K.K., Sanofi-Aventis K.K., Daiichi Sankyo Co., Ltd., Dainippon Sumitomo Pharma Co., Ltd., Takeda Pharmaceutical Co., Ltd., Mitsubishi Tanabe Pharma Corporation, Teijin Pharma Ltd., Nippon Chemiphar Co., Ltd., Nihon Pharmaceutical Co., Ltd., Nippon Boehringer Ingelheim Co., Ltd., Novartis Pharma K.K., Bayer Yakuhin, Ltd., and Mochida Pharmaceutical Co., Ltd. and received royalties from Chugai-Igakusha, Igaku-Shoin Ltd., Medical View Co. Ltd., and Blackwell Publishing K. K.

Availability of data and materials

Not applicable.

Acknowledgments

We thank all the healthy volunteers and patients for their participation.

Rao SM, Harrington DL, Haaland KY, Bobholz JA, Cox RW, Binder JR. Distributed neural systems underlying the timing of movements. J Neurosci. 1997;17:5528–35.
Mauk MD, BuonomanoDV. The neural basis of temporal processing. Annu Rev Neurosci. 2004;27:307–40.
Coull, JT, Cheng RK, Meck WH. Neuroanatomical and neurochemical substrates of timing. Neuropsychopharmacology. 2011;36:3–25.
Bostan AC, Dum RP, Strick PL. The basal ganglia communicate with the cerebellum. Proc Natl Acad Sci U S A. 2010;107:8452–6.
Bostan AC, Strick PL. The basal ganglia and the cerebellum: nodes in an integrated network. Nat Rev Neurosci. 2018;19:338–50.
Harrington DL, Haaland KY, Hermanowicz N. Temporal processing in the basal ganglia. Neuropsychology. 1998;12:3–12.
Harrington DL, Rao SM. Timing in neurodegenerative disorders of the basal ganglia. in Vatakis A, Allman M, (Eds.), Time Distortions in Mind: Temporal Processing in Clinical Population. Leiden (The Netherlands) / Boston (The United States): Brill; 2015. 190–225.
Jones CR, Jahanshahi M. (2011). Dopamine Modulates Striato-Frontal Functioning during Temporal Processing. Front Integr Neurosci. 2011;5:70.
Jones CR, Jahanshahi M. Motor and perceptual timing in Parkinson's disease. Adv Exp Med Biol. 2014;829:265–90.
Jones CRG, Jahanshahi M. Striatal and Frontal Pathology: Parkinson’s Disease and Patients with Lesions of the Basal Ganglia and Frontal Cortex. in Vatakis A, Allman M, (Eds.), Time Distortions in Mind: Temporal Processing in Clinical Population. Leiden (The Netherlands) / Boston (The United States): Brill; 2015. 250 – 80.
Buhusi CV, Meck WH. Differential effects of methamphetamine and haloperidol on the control of an internal clock. Behav Neurosci. 2002;116:291–7.
Buhusi CV, Meck WH. What makes us tick? Functional and neural mechanisms of interval timing. Nat Rev Neurosci. 2005;6:755–765.
Maricq AV, Church RM. The differential effects of haloperidol and methamphetamine on time estimation in the rat. Psychopharmacology (Berl). 1983;79:10–5.
Meck WH. Neuropharmacology of timing and time perception. Brain Res Cogn Brain Res. 1996;3:227–42.
Matell MS, Meck WH. Cortico-striatal circuits and interval timing: coincidence detection of oscillatory processes. Brain Res Cogn Brain Res. 2004;21:139–70.
Matell MS, King GR, Meck, WH. Differential modulation of clock speed by the administration of intermittent versus continuous cocaine. Behav Neurosci. 2004;118:150–6.
Ito M. (Ed.). The cerebellum and neural control. New York (The United States): Raven Press; 1984.
Perrett SP, Ruiz BP, Mauk MD. Cerebellar cortex lesions disrupt learning-dependent timing of conditioned eyelid responses. J Neurosci. 1993;13:1708–18.
Apps R, Garwicz M. Anatomical and physiological foundations of cerebellar information processing. Nat Rev Neurosci. 2005;6:297–311.
Ivry RB, Schlerf JE. Dedicated and intrinsic models of time perception. Trends Cogn Sci. 2008;12:273–80.
Yamazaki T, Nagao S. A computational mechanism for unified gain and timing control in the cerebellum. PLoS One. 2012;7:e33319.
Finnerty GT, Shadlen MN, Jazayeri M, Nobre AC, Buonomano DV. Time in Cortical Circuits. J Neurosci. 2015;35:13912–6.
Ivry RB, Keele SW. Timing functions of the cerebellum. J Cogn Neurosci. 1989;1:136–52.
Breska A, Ivry RB. Taxonomies of Timing: Where Does the Cerebellum Fit In? Curr Opin Behav Sci. 2016;8:282–8.
Bareš M, Lungu OV, Liu T, Waechter T, Gomez CM, Ashe J. (2011). The neural substrate of predictive motor timing in spinocerebellar ataxia. Cerebellum. 2011;10:233 – 44.
Bareš M, Apps R, Avanzino L, Breska A, D'Angelo E, Filip P, et al. Consensus paper: Decoding the Contributions of the Cerebellum as a Time Machine. From Neurons to Clinical Applications. Cerebellum. 2019;18:266–86.
Merchant H, Zarco W, Prado L. Do we have a common mechanism for measuring time in the hundreds of millisecond range? Evidence from multiple-interval timing tasks. J Neurophysiol. 2008;99:939–49.
Zarco W, Merchant H, Prado L, Mendez JC. Subsecond timing in primates: comparison of interval production between human subjects and rhesus monkeys. J Neurophysiol. 2009;102:3191–202.
Gooch CM, Wiener M, Hamilton AC, Coslett HB. Temporal discrimination of sub- and suprasecond time intervals: a voxel-based lesion mapping analysis. Front Integr Neurosci. 2011;5:59.
Ohmae S, Kunimatsu J, Tanaka M. Cerebellar Roles in Self-Timing for Sub- and Supra-Second Intervals. J Neurosci. 2017;37:3511–22.
Lewis PA, Miall RC. Brain activation patterns during measurement of sub- and supra-second intervals. Neuropsychologia. 2003;41:1583–92.
Lewis PA, Miall RC. Distinct systems for automatic and cognitively controlled time measurement: evidence from neuroimaging. Curr Opin Neurobiol. 2003;13:250–5.
Lewis PA, Miall RC. The precision of temporal judgement: milliseconds, many minutes, and beyond. Philos Trans R Soc Lond B Biol Sci. 2009; 364:1897–905.
Ivry RB. The representation of temporal information in perception and motor control. Curr Opin Neurobiol. 1996;6:851–7.
Spencer RM. (Ed.). Timing in the Cerebellum and Cerebellar Disorders. Leiden (The Netherlands) / Boston (The United States): Brill; 2015.
Merchant H, Harrington DL, Meck WH. Neural basis of the perception and estimation of time. Annu Rev Neurosci. 2013;36:313–36.
Nawrot M, Rizzo M. Motion perception deficits from midline cerebellar lesions in human. Vision Res. 1995;35:723–31.
Nichelli P, Alway D, Grafman J. Perceptual timing in cerebellar degeneration. Neuropsychologia. 1996;34:863–71.
Coull JT, Nobre AC. Dissociating explicit timing from temporal expectation with fMRI. Curr Opin Neurobiol. 2008;18:137–44.
Coull JT, Nobre AC. Where and when to pay attention: the neural systems for directing attention to spatial locations and to time intervals as revealed by both PET and fMRI. J Neurosci. 1998;18:7426–35.
Coull JT, Davranche K, Nazarian B, Vidal F. Functional anatomy of timing differs for production versus prediction of time intervals. Neuropsychologia. 2013;51:309–19.
Petter EA, Lusk NA, Hesslow G, Meck WH. Interactive roles of the cerebellum and striatum in sub-second and supra-second timing: Support for an initiation, continuation, adjustment, and termination (ICAT) model of temporal processing. Neurosci Biobehav Rev. 2016;71:739–55.
Avanzino L, Bove M, Pelosin E, Ogliastro C, Lagravinese G, Martino D. The cerebellum predicts the temporal consequences of observed motor acts. PLoS One. 2015;10:e0116607.
Trillenberg P, Verleger R, Teetzmann A, Wascher E, Wessel K. On the role of the cerebellum in exploiting temporal contingencies:; evidence from response times and preparatory EEG potentials in patients with cerebellar atrophy. Neuropsychologia. 2004;42:754–63.
Bareš M, Lungu O, Liu T, Waechter T, Gomez CM, Ashe J. Impaired predictive motor timing in patients with cerebellar disorders. Exp Brain Res. 2007;180:355–65.
Bareš M, Lungu OV, Husarova I, Gescheidt T. Predictive motor timing performance dissociates between early diseases of the cerebellum and Parkinson's disease. Cerebellum. 2010;9:124–35.
Breska A, Ivry RB. Double dissociation of single-interval and rhythmic temporal prediction in cerebellar degeneration and Parkinson's disease. Proc Natl Acad Sci U S A. 2018;115:12283–8.
Breska A, Ivry RB. The human cerebellum is essential for modulating perceptual sensitivity based on temporal expectations. Elife. 2021;10:e66743.
Ivry RB, Richardson TC. Temporal control and coordination: the multiple timer model. Brain Cogn. 2002;48:117–32.
Ohmae S, Tanaka M. Two different mechanisms for the detection of stimulus omission. Sci Rep. 2016;6:20615.
Takeya R, Kameda M, Patel AD, Tanaka M. Predictive and tempo-flexible synchronization to a visual metronome in monkeys. Sci Rep. 2017;7:6127.
Kunimatsu J, Suzuki TW, Ohmae S, Tanaka M. Different contributions of preparatory activity in the basal ganglia and cerebellum for self-timing. Elife. 2018;7:e35676.
Tanaka M, Kunimatsu J, Suzuki TW, Kameda M, Ohmae S, Uematsu A, et al. Roles of the Cerebellum in Motor Preparation and Prediction of Timing. Neuroscience. 2021;462:220–34.
Ohmae S, Uematsu A, Tanaka M. Temporally specific sensory signals for the detection of stimulus omission in the primate deep cerebellar nuclei. J Neurosci. 2013;33:15432–41.
Uematsu A, Ohmae S, Tanaka M. Facilitation of temporal prediction by electrical stimulation to the primate cerebellar nuclei. Neuroscience. 2017;346: 190–6.
Kameda M, Ohmae S, Tanaka M. Entrained neuronal activity to periodic visual stimuli in the primate striatum compared with the cerebellum. Elife. 2019;8:e48702.
Kunimatsu J, Tanaka M. Striatal dopamine modulates timing of self-initiated saccades. Neuroscience. 2016;337:131–42.
Mates J, Muller U, Radil T, Poppel E. Temporal integration in sensorimotor synchronization. J Cogn Neurosci. 1994;6:332–40.
Repp BH. Sensorimotor synchronization: a review of the tapping literature. Psychon Bull Rev. 2005;12:969–92.
Repp BH, Su YH. Sensorimotor synchronization: a review of recent research (2006–2012). Psychon Bull Rev. 2013;20:403–52.
Matsuda S, Matsumoto H, Furubayashi T, Hanajima R, Tsuji S, Ugawa Y, et al. The 3-second rule in hereditary pure cerebellar ataxia: a synchronized tapping study. PLoS One. 2015;10:e0118592.
Tokushige SI, Terao Y, Matsuda S, Furubayashi T, Sasaki T, Inomata-Terada S, et al. Does the Clock Tick Slower or Faster in Parkinson's Disease? - Insights Gained From the Synchronized Tapping Task. Front Psychol. 2018;9:1178.
Pöppel E. A hierarchical model of temporal perception. Trends Cogn Sci. 1997;1:56–61.
Penhune VB, Zattore RJ, Evans AC. Cerebellar contributions to motor timing: a PET study of auditory and visual rhythm reproduction. J Cogn Neurosci. 1998;10:752–65.
Bijsterbosch JD, Lee KH, Hunter MD, Tsoi DT, Lankappa S, Wilkinson ID, et al. The role of the cerebellum in sub- and supraliminal error correction during sensorimotor synchronization: evidence from fMRI and TMS. J Cogn Neurosci. 2011;23:1100–12.
Pecenka N, Engel A, Keller PE. Neural correlates of auditory temporal predictions during sensorimotor synchronization. Front Hum Neurosci. 2013;7:380.
Molinari M, Leggio MG, Thaut MH. The cerebellum and neural networks for rhythmic sensorimotor synchronization in the human brain. Cerebellum. 2007;6:18–23.
Schlerf JE, Spencer RM, Zelaznik HN, Ivry RB. Timing of rhythmic movements in patients with cerebellar degeneration. Cerebellum. 2007;6:221–31.
Lungu OV, Bares M, Liu T, Gomez CM, Cechova I, Ashe J. Trial-to-trial Adaptation: Parsing out the Roles of Cerebellum and BG in Predictive Motor Timing. J Cogn Neurosci. 2016;28:920–34.
Schmitz-Hubsch T, du Montcel ST, Baliko L, Berciano J, Boesch S, Depondt C, et al. Scale for the assessment and rating of ataxia: development of a new clinical scale. Neurology. 2006;66:1717–20.
Storey E, Tuck K, Hester R, Hughes A, Churchyard A. Inter-rater reliability of the International Cooperative Ataxia Rating Scale (ICARS). Mov Disord. 2004;19:190–2.
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.
Aschersleben G, Prinz W. Synchronizing actions with events: the role of sensory information. Percept Psychophys. 1995;57:305–17.
Teki S, Grube M, Kumar S, Griffiths TD. Distinct neural substrates of duration-based and beat-based auditory timing. J Neurosci. 2011;31:3805–12.
Vorberg D, Hambuch R. Timing of two-handed rhythmic performance. Ann N Y Acad Sci. 1984;423:390–406.
Schulze HH, Vorberg D. Linear phase correction models for synchronization: parameter identification and estimation of parameters. Brain Cogn. 2002;48:80–97.
Mates J. A model of synchronization of motor acts to a stimulus sequence. I. Timing and error corrections. Biol Cybern. 1994;70:463–73.
Mates J. A model of synchronization of motor acts to a stimulus sequence. II. Stability analysis, error estimation and simulations. Biol Cybern. 1994;70:475–84.
Diedrichsen J, Ivry RB, Pressing J. (2003). Cerebellar and basal ganglia contributions to interval timing. in Meck WH, (Ed.), Functional and neural mechanisms of interval timing. CRC Press/Routledge/Taylor & Francis Group; 2003. 457 – 83.
Teki S, Grube M, Griffiths TD. A unified model of time perception accounts for duration-based and beat-based timing mechanisms. Front Integr Neurosci. 2012;5:90.
Wohlschläger A, Koch R, (Eds.), Synchronization error: An error in time perception. Lisse (The Netherlands): Swets & Zeitlinger; 2000.
Ivry RB, Keele SW, Diener HC. Dissociation of the lateral and medial cerebellum in movement timing and movement execution. Exp Brain Res. 1988;73:167–80.
Tada M, Nishizawa M, Onodera O. Redefining cerebellar ataxia in degenerative ataxias: lessons from recent research on cerebellar systems. J Neurol Neurosurg Psychiatry. 2015;86:922–8.
Sakakibara S, Aiba I, Saito Y, Inukai A, Ishikawa K, Mizusawa H. (2014). Clinical features and MRI findings in spinocerebellar ataxia type 31 (SCA31) comparing with spinocerebellar ataxia type 6 (SCA6). [Article in Japanese] Rinsho Shinkeigaku. 2014;54:473-9.
Aschersleben G. Temporal control of movements in sensorimotor synchronization. Brain Cogn. 2002;48:66–79.
Paillard J. Research on simultaneity in the execution of reactive movements. C R Seances Soc Biol Fil. 1948;142:935–7.
Fraisse P. Les synchronizations sensori-motrices aux rythmes. in Requin J, (Ed.), Anticipation et Comportement. Paris (France): Centre National de la Recherche Scientifique; 1980. 233 – 57.
O'Boyle DJ, Freeman JS, Cody FW. The accuracy and precision of timing of self-paced, repetitive movements in subjects with Parkinson's disease. Brain. 1996;119:51–70.
Broersen R, Onuki Y, Abdelgabar AR, Owens CB, Picard S, Willems J, et al. Impaired spatio-temporal predictive motor timing associated with spinocerebellar ataxia type 6. PLoS One. 2016;11:e0162042.
Schulz JB, Borkert J, Wolf S, Schmitz-Hubsch T, Rakowicz M, Mariotti C, et al. Visualization, quantification and correlation of brain atrophy with clinical symptoms in spinocerebellar ataxia types 1, 3 and 6. Neuroimage. 2010;49:158–68.
Kim JM, Lee JY, Kim HJ, Kim JS, Kim YK, Park SS, et al. The wide clinical spectrum and nigrostriatal dopaminergic damage in spinocerebellar ataxia type 6. J Neurol Neurosurg Psychiatry. 2010;81:529–532.
Xie T, Appelbaum D, Bernard J, Padmanaban M, Pu Y, Gomez C. Evaluation of parkinsonism and striatal dopamine transporter loss in patients with spinocerebellar ataxia type 6. J Neurol. 2016;263:2302–7.
Yoshida K, Asakawa M, Suzuki-Kouyama E, Tabata K, Shintaku M, Ikeda S, et al. Distinctive features of degenerating Purkinje cells in spinocerebellar ataxia type 31. Neuropathology. 2014;34:261–7.
Liu T, Xu D, Ashe J, Bushara K. Specificity of inferior olive response to stimulus timing. J Neurophysiol. 2008;100:1557–61.
Müller K, Schmitz F, Schnitzler A, Freund HJ, Aschersleben G, Prinz W. Neuromagnetic correlates of sensorimotor synchronization. J Cogn Neurosci. 2000;12:546–55.
Müller K, Aschersleben G, Schmitz F, Schnitzler A, Freund HJ, Prinz W. Inter- versus intramodal integration in sensorimotor synchronization: a combined behavioral and magnetoencephalographic study. Exp Brain Res. 2008;185:309–18.
Pollok B, Muller K, Aschersleben G, Schmitz F, Schnitzler A, Prinz W. Cortical activations associated with auditorily paced finger tapping. Neuroreport. 2003;14:247–50.
Pollok B, Muller K, Aschersleben G, Schnitzler A, Prinz W. The role of the primary somatosensory cortex in an auditorily paced finger tapping task. Exp Brain Res. 2004;156:111–7.
Pollok B, Gross J, Muller K, Aschersleben G, Schnitzler A. The cerebral oscillatory network associated with auditorily paced finger movements. Neuroimage. 2005;24:646–55.
Miall C. The Storage of Time Intervals Using Oscillating Neurons. Neural Comput. 1989;1:359–71.
Matell MS, Meck WH. Neuropsychological mechanisms of interval timing behavior. Bioessays. 2000;22:94–103.

No competing interests reported.

SuppFigure1.tif
Supplementary Figure 1. Mean + 2 standard deviations (SDs) of synchronizing tapping error (STE) as a function of inter-stimulus interval (ISI). These values indicate the upper limit of the STE distribution. Among all the disease types, these values tended to increase along with the ISI. These values were largest in multiple system atrophy (MSA) and smallest in spinocerebellar ataxia (SCA)31 groups.
SuppFigure2.tif
Supplementary Figure 2. Time courses of synchronizing tapping error (STE) at inter-stimulus interval (ISI) 3600 ms across trials in different subject groups, plotted as described in Fig.4. There was no significant correlation between the tap counts and STE for all subject groups.

Download PDF

Version 1

posted

You are reading this latest preprint version

Roles of the cerebellum and basal ganglia in temporal integration: insights gained from the synchronized tapping task

Status:

Version 1

Abstract

Figures

Introduction

Patients And Methods

Results

Discussion

Correction mechanisms for synchronized tapping

Roles of the cerebellum and basal ganglia in synchronized tapping

Integration of cerebellar and basal ganglia systems for temporal integration

Superior tapping performance in SCA31

Limitations

Conclusions

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1