On medication

Note: N = number of participants; CD = standard deviation of clock component estimated with the Wing-Kristofferson model; MD = standard deviation of motor component estimated with the Wing-Krist-offerson model; Sig. = significance (* indicates that the difference between the experimental and control group is statistically reliable); HY = Hohn and Yahr scale.

responses and a motor system that implements these responses. Based on a small set of assumptions, namely, that the component sources are independent of each other and that feedback mechanisms do not play a role in producing the next tap, estimates of the two components can be obtained from the autocovariance function of the intertap intervals. The assumptions of the model have received substantial empirical support (Ivry and Hazeltine, 1995; Wing, 1980).

Ivry and Keele (1989) provided the first large-scale study in which the performance of patients with different neurological disorders was assessed with the Wing-Kristofferson continuation task. Patients with lesions of the cerebellum exhibited significant increases in variability on this task. The increase was especially marked in the estimate of the clock component, and in a more detailed analysis of a subgroup of patients, the clock deficit was found to be associated with damage to the lateral regions of the neocerebellum (Ivry et al., 1988). In contrast, lesions centered in the medial cerebellum were associated with increases in the estimate of motor implementation noise. This double dissociation is in agreement with the anatomical projections of the output pathways of the cerebellum: the lateral regions primarily ascend to motor, premotor, and prefrontal regions of the cerebral cortex, while the medial regions innervate brain stem and spinal cord regions of the descending motor pathways. The finding of a clock deficit in patients with neocerebellar lesions has been replicated (Franz et al., 1996; Ivry et al., in press).

The results from experiments involving PD patients yield a more ambiguous picture. Two studies have reported that medicated PD patients perform similarly to age-matched controls in terms of overall variability during the continuation phase (Duchek et al., 1994; Ivry and Keele, 1989), whereas one study has shown a deficit in performance similar to that observed with cerebellar patients (Harrington et al., 1998a). An earlier study (Pastor et al., 1992) had also pointed to a deficit in nonmedicated PD patients. However, each participant completed only a single trial per interval, rendering estimates of the variability components suspect. O'Boyle et al. (1996) found significant increases in estimates of clock and motor implementation for patients tested off L-dopa medication. On the other hand, Ivry and Keele (1989) reported no change in performance as a function of medication level.

Greater convergence is found in studies involving PD patients with unilateral symptoms either taking L-dopa medication (O'Boyle et al., 1996) or tested in a drug-naive state (Ivry and Keele, 1989; Wing et al., 1984). In these cases, the patients showed consistent increases in the estimate of clock variability when tapping with the impaired hand; the motor estimate was not significantly changed. For example, Keele and Ivry (1987) tracked one patient over a 2-week period during which he began L-dopa therapy. His performance showed a marked improvement over the test sessions with the decrease in variability solely associated with a reduction in the estimate of clock variability.

In sum, the neuropsychological studies suggest that disorders of the cerebellum and basal ganglia can result in increased variability during repetitive tapping. The increase may be isolated to either of the two component processes proposed in the Wing-Kristofferson model. The source of the increase following cerebellar pathology appears to be dependent on lesion location within the cerebellum. In PD, the deficit is less consistently observed, but when present, is almost always restricted to the clock component. Ivry and Hazeltine (1995) point out that the term clock is misleading in the Wing-Kristofferson model since this estimate refers to all sources of variability other than that associated with motor implementation. As such, they propose that the term clock variability should be replaced by central variability, encompassing all aspects of motor planning and preparation that occur prior to movement initiation. In this view, the cerebellum and basal ganglia might both add to central variability, but in distinct ways (Ivry and Keele, 1989).

Four studies have used functional imaging to investigate neural regions involved in the production of rhythmic movements. We defer our review of two of these to

Section 19.3.2. In the third study (Penhune et al., 1998), participants reproduced a sequence of isochronous intervals with the right hand, with the target pace indicated by either an auditory or visual metronome. Compared to just listening to the metronome, increased activation in the auditory condition was observed in the left globus pallidus and right anterior cerebellum (lobules V and VI). In the visual condition, activation was observed in the left lateral cerebellum (VIIa) in addition to the anterior cerebellar site. No basal ganglia activation was found in the visual condition. Another set of comparisons involved conditions in which participants produced either novel or well-practiced rhythmic patterns. Prominent cerebellar activation was found for the novel conditions, spanning left and right cerebellar hemispheres (VIIa and VIIb) and anterior and posterior vermal areas (III and IV, VIIIa and VIIIb). This pattern was similar for both auditory and visual stimulus conditions. The basal ganglia also showed a bilateral increase in activation during novel rhythms, but again, this increase was limited to the auditory condition. Kawashima et al. (2000) recently reported congruent results, in which they compared visually triggered and memory-timed finger taps. Again, this comparison was significant for the anterior cerebellum, but not for the basal ganglia. These results point to a consistent involvement of the cerebellum in the processing of temporal information. In contrast, the results are less consistent for the basal ganglia and are restricted to one sensory domain.

19.2.2 The Perception of Timed Events

Table 19.2 summarizes the performance of cerebellar and PD patients on time perception tasks. The table is restricted to studies in which participants judged the duration of a stimulus; we excluded papers in which the temporal judgment was based on gap detection or simultaneity (Artieda et al., 1992). Temporal acuity on duration discrimination tasks is conventionally assessed by varying the difference between a standard and a comparison stimulus. The participant's responses are fitted with a logistic distribution function to estimate the point of subjective equality (PSE) and variability, typically quantified as correct performance on 75% of the trials (corresponding to ±1 SD of the logistic distribution). The table also shows the variability data in a normalized form in which the SD has been divided by the PSE. Interestingly, the results indicate a considerable range in the Weber fractions across the studies, probably reflecting the different methodologies used to make the threshold estimates.

Deficits for cerebellar patients were reported by Ivry and colleagues in three studies (Casini and Ivry, 1999; Ivry and Keele, 1989; Mangels et al., 1998). Note that many of the same patients were tested in the 1998 and 1999 experiments and thus do not constitute independent samples. Nichelli et al. (1996) also reported elevated thresholds for intervals ranging from 100 to 600 msec. However, their group of patients with cerebellar degeneration performed comparable to controls for intervals ranging from 100 to 325 msec. As with the tapping data, the PD results are less consistent. Ivry and Keele (1989) observed no increase in the difference threshold in a group of medicated PD patients. In contrast, Harrington et al. (1998a) observed a significant impairment in a different group of medicated PD patients with intervals of 300 and 600 msec.

0 0

Post a comment