Machine Learning to Estimate the Amount of Training to Learn a Motor Skill

Machine Learning (ML) has been widely and successfully employed in different fields to estimate information from datasets. However, the necessary time to learn a motor task or to rehabilitate is mainly determined by the professional experience of medical doctor, physiotherapist and so on. Thus, this work introduces a software to measure the performance of subjects on a experiment performing a tracing task, which requires motor learning, and uses ML algorithms on the dataset acquired during this experiment. The task is divided into 1 session that has 3 blocks and each block is composed of 10 trials whereas each trial is one word. Furthermore, ML algorithms - namely k-nearest neighbours, decision tree, support vector machines and multilayer-perceptron neural network - are applied on the collected data from the experiment to estimate which block the subject currently is. The results demonstrated that there was motor learning, as well as that is possible to apply classification models to predict the block of the subject with decision tree achieving statistically significant (p-value < 0.01) best predictions. The proposed approach may be useful for health professionals when estimating the amount of training a patient requires to learn a motor task or rehabilitate.