Larger, but not better, implicit motor adaptation ability inherent in medicated Parkinson’s disease patients: a smart-device-based study

Generating appropriate motor commands is an essential brain function. To achieve proper motor control in diverse situations, predicting future states of the environment and body and modifying the prediction are indispensable. The internal model is a promising hypothesis about brain function for generating and modifying the prediction. Although several findings support the involvement of the cerebellum in the internal model, recent results support the influence of other related brain regions on the internal model. A representative example is the motor adaptation ability in Parkinson’s disease (PD) patients. Although this ability provides some hints about how dopamine deficits affect the internal model, previous findings are inconsistent; some reported a deficit in the motor adaptation ability in PD patients, but others reported that the motor adaptation ability of PD patients is comparable to that of healthy controls. A possible factor causing this inconsistency is the difference in task settings, which yield different cognitive strategies in each study. Here, we demonstrate a larger, but not better, motor adaptation ability in PD patients than healthy controls while reducing the involvement of cognitive strategies and concentrating on implicit motor adaptation abilities. This study utilizes a smart-device-based experiment that enables motor adaptation experiments anytime and anywhere with less cognitive strategy involvement. The PD patients showed a significant response to insensible environmental changes, but the response was not necessarily suitable for adapting to the changes. Our findings support compensatory or paretic cerebellar functions in PD patients from the perspective of motor adaptation.


Introduction
individuals because the task switch can be involved in a gradual perturbation with a large amplitude.
Here, we investigated the motor adaptation ability of PD patients while decreasing the influence of 69 the task switch as much as possible. To reduce the influence of the task switch, we relied on a gradually 70 applied perturbation whose existence was noticeable by 1 out of the 82 participants in our previous study 71 [11]; the visuomotor rotation changed by one degree in each trial, and the maximum value of the rotation 72 was 15 degrees. Based on a previous study, task switches are less involved in gradually applied 15 degree 73 visuomotor rotation than in gradually applied 45 degree visuomotor rotation [9]. We demonstrate that 74 the PD patients showed a larger, but not better, motor adaptation ability than elderly individuals and 75 young individuals, rather than a compatible or impaired ability. 76 In addition to reducing the influence of the task switch, we also decreased the burden to participate Fifty-four subjects participated in the current study; their ages and sexes are summarized in Table 1 The PD patients were outpatients satisfying the following inclusion criteria. The elderly individuals 97 were inpatients with broken lower-limb bones that did not disturb performance in our study. The young 98 individuals were volunteers. All of the participants were naive to the purpose of the study.

99
The PD patients were clinically evaluated based on the Unified Parkinson's Disease Rating Scale The cursor position on the tablet display (d x , d y ) that was controlled by participants was determined as

120
The a x and a y were sampled at 200 Hz. The cursor position in the accelerometer coordinate system was 121 transformed into the position on the tablet display (d x , d y ) by multiplying by the rotation matrix R; The required task was to tilt the held tablet device appropriately. Corresponding to the tilting motion, 129 the cursor displayed on the tablet moved (a yellow circle with a 4.5 mm radius on the Nexus 9). The 130 participants were instructed to move the cursor toward the visually instructed target (a purple circle with 131 a 4.5 mm radius on the Nexus 9) also on the display in a straightforward manner within two seconds 132 ( Fig. 1A). At the beginning of each trial, the subjects needed to tilt the tablet to set the cursor at the 133 initial position in the center of the tablet screen (a blue circle with a 9.0 mm radius on the Nexus 9) The subjects participated in 20 practice trials and 80 learning trials. In the first 20 practice trials, the target position was pseudorandomly set to either 60, 75, 90, 105, or 120 degrees without any visuomotor rotation (90 degrees indicated the 12 o'clock position on the tablet display). In the following 80 learning 142 trials, the target position was fixed at 90 degrees. The learning trials were divided into two parts. In 143 the first 40 learning trials, the subjects experienced gradually increasing and vanishing clockwise (CW) 144 perturbation. In the latter 40 trials, the subjects underwent counterclockwise (CCW) perturbation.

145
Half of the participants experienced the CW perturbation first, and the other half of the participants 146 experienced the CCW perturbation first. No subjects were aware of the existence of the perturbation.

147
The experiment typically took less than 30 minutes.   The averaged learning effects of the ith subject, , and the phase ϕ i . The temporal 174 delay was calculated by temporally sliding the fragments of imize the squared error from the fragments of the perturbation sequence p = (p 9 , ..., p 32 ). The squared 176 error between the learning effects and the perturbation is hereafter referred to as the task error. We chose were chosen to be (p 9 , ..., p 32 ).

182
After determining ∆ * i as ∆ * i = arg min ∆i ( 1 quantify the trajectories, we calculated the trajectory error as the sum of the lateral deviations within 192 each trial:

218
The learning effects showed between-trial variation depending on the between-trial varying perturba-219 tion (Fig. 1C). The shaded area in Fig. 1C denoted the trial numbers (trials 12-35) when the learning 220 effects of PD patients were significantly different from 0 (t-test p<0.01 [corrected]). We compared the 221 learning effects averaged across the trials denoted by the shaded area in each subject (Fig. 1D) larger than the those of the elderly individuals and those of the young individuals in these measures.

229
In contrast to the learning effects, there was no group effect (F(2,51)=1.35, p=0.27) and no significant 230 difference among the three groups (p>0.28) in the RMSE (Fig. 1E), the error between the learning effects 231 and the perturbation (a detailed definition of this metric is in the Methods section). These results indicate 232 that the PD patients showed more substantial learning effects than the elderly individuals and the young 233 individuals but a comparable ability to minimize the RMSE.

234
To further study the learning effects, we decomposed the learning effects into three components: the 235 amplitude A, to quantify the magnitude of the response to the perturbation; the phase ϕ, to quantify the 236 similarity between the learning curves and the perturbation; and the temporal delay ∆, to quantify the 237 temporal sensitivity of the response to the perturbation.

238
For the amplitude measurement ( Fig. 2A), there was a significant group effect (F(2,51) = 6.76,  larger amplitudes compared to the other two groups ( Fig. 2A). In addition, a slightly faster response 262 delay in the PD patients contributed to the large learning effects (Fig. 2C). 263 We further considered other factors that may affect the larger substantial learning effects in PD expect some correlation between these metrics. In contrast to this assumption, there was no significant 276 correlation between the magnitude and the normalized movement times (Fig. 3B, r=0.15, p=0.27).

277
Additionally, there was no correlation between the phase and the normalized movement time (r=0.10, 278 p=0.46), between the lag and the normalized movement time (r=-0.15, p=0.28). These results indicate 279 that the movement time was not a significant factor affecting the learning effects.

280
Following previous studies that have reported that PD patients showed a large amount of feedback  We further investigated the relationship between the clinical scores and the learning effects (Fig. 4).

291
There was no significant correlation between all the recorded attributes and the clinical scores (i.e., age,