Objective: To prospectively identify the predictors of motor learning gains in a heterogeneous cohort of people with Parkinson’s disease (PD).

Background: Retaining improvements of motor skills after training can be challenging in PD. So far, prospective studies examining the impact of motor and non-motor symptoms on learning fine motor skills are lacking.

Method: This cohort study included 97 PD patients without dementia (HY I-IV; 47% females) who trained the Swipe-Slide Pattern (SSP) task (2 weeks). The task involved swiping predefined patterns as fast and accurately as possible. The primary outcome (SSP-Time) was assessed at baseline, after two weeks of training and four weeks follow-up. Learning gains reflected the change from baseline to follow-up (higher scores = better). Using the PROBAST guidelines, two prediction models were developed: 1) including predominantly motor-related factors, i.e., age, sex, levodopa equivalent daily dosage, disease severity (MDS-UPDRS-III), freezing of gait status, a mobile phone task (MPT) and the dexterity questionnaire (DEXTQ-24); and 2) including non-motor symptoms, i.e., global cognition (MoCA), depression (GDSS), anxiety (PAS), apathy (SAS) and sleep quality (PSQI). As per PROBAST rules, baseline SSP-Time was not included, yet correlation analyses investigated its association with learning gains.

Results: Both models significantly predicted overall learning gains (motor: F(7,89)=7.4, p<0.001, R²=0.369; non-motor: F(5,91)=9.8, p<0.001, R²=0.315). The first model showed that worse disease severity (β=0.402, 95%CI [15.6 101.5]) and worse MPT performance (β=0.270, 95%CI [18.4 261.6]) predicted greater learning gains. The non-motor model illustrated that worse global cognition (β=-0.464, 95%CI [-466.0 -95.0]) and greater anxiety (β=0.199, 95%CI [7.2 108.1]) predicted better outcomes. Lastly, a strong negative correlation between learning and the baseline SSP-Time was found (ρ=-0.744, p<0.001).

Conclusion: As noted previously, worse baseline task performance is indicative of more “room for improvement” after training. In line, we found that preserved motor learning capacity is determined by worse disease severity, daily life Mobile Phone proficiency, cognition and anxiety. Hence, our results provide robust evidence that both motor and non-motor features need to be considered when developing personalized learning programs in PD.

« Back to 2025 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/clinical-predictors-of-home-based-motor-learning-of-touchscreen-skills-in-parkinsons-disease/