Author(s): NIETSCHMANN, P., FRANKLIN, S., PAPADOPOULOU, T., FRANKLIN, D.W., Institution: TECHNICAL UNIVERSITY OF MUNICH, Country: GERMANY, Abstract-ID: 2163

INTRODUCTION:
Humans exhibit remarkable abilities in learning motor skills. While motor performance continues to improve with practice, even after years of learning¹, most studies only examine the learning of novel dynamics over several hundred trials² ³. That is, most of our understanding of the parameters and characteristics of motor learning rely on studies that examine only the early adaptation processes. While some studies of complex task adaptation have examined thousands of trials, the adaptation to a single velocity-dependent force field is usually considered to be complete within a few hundred movements² ³. Here we investigate motor adaptation to a single novel force field with over seven thousand trials and multiple weeks of training, to uncover the process of human motor learning.
METHODS:
Ten right-handed participants performed training sessions once every week for eight consecutive weeks, with a follow-up session 15 weeks later to assess retention. Participants performed forward reaching movements with their right hand using a robotic manipulandum in a velocity dependent curl force field. Throughout the sessions we probed the adaptation process by measuring the decay, spontaneous recovery, retention, angular generalization, and transfer of the motor memory to the opposite limb.
RESULTS:
Participants adapted quickly to the dynamics, showing a fast reduction in kinematic error and a rapid increase in predictive force compensation within the first session. Interestingly, we find continuous adaptation over the subsequent sessions. Moreover, there is strong retention of the predictive forces between sessions, increasing gradually until the fifth session. We observe a strong increase in spontaneous recovery between the first and seventh session, indicating a stronger motor memory, but little change in the decay rates over the training weeks. Despite extensive training, we found no evidence of transfer to the left hand. However, while generalization to adjacent reaching directions was stable across training weeks, we did find a slight narrowing of the Gaussian-like angular generalization with practice, suggesting a slight fine tuning of the spatial properties of the motor memory. Finally, even after 15 weeks without practice, participants immediately exhibited similar levels of predictive adaptation to the dynamics.
CONCLUSION:
Our results show that learning continuously progressed over time, creating a stable motor memory that is retained for prolonged periods without task exposure. Notably, while predictive adaptation developed over time, other aspects, such as the decay rate, remained relatively unaffected during our training period. Our research highlights motor memory formation and durability and could inform training strategies aimed at developing motor skills or advancing rehabilitation.

¹Crossman, Ergonomics. 1959, 2:153-166.
²Franklin et al., J Neurophysiol. 2012, 108:467-478.
³Shadmehr and Mussa-Ivaldi, J Neurosci. 1994 14:3208-3224.