Diferencias entre la imaginación y ejecución motora durante los movimientos ponerse de pie y sentarse a través de modelos autorregresivos

Abstract

The analysis of motor imagination and execution (MI-ME) is one of the main research challenges in the field of brain-computer interfaces (BCI) based on electroencephalography (EEG). EEG signals play an important role in the learning, rehabilitation, and assistance of complex motor skills. However, this type of signal presents a highly non-stationary nature and noise. Electroencephalography signals proper of sensory-motor activity, corresponding to execution defects and motor imagination, were analyzed. The present research aims to determine the differences that exist between motor imagination and motor execution from a proof of concept, in the movements of sitting and standing. A comparative study was carried out between a method based on desynchronization / synchronization (ERDs) and a new method based on autoregressive moving average models with exogenous input (ARMAX). In this research, the normalized root mean square error (NRMSE), the mean square error (RMSE) and the mean absolute error (MAE) are used as evaluation metrics, in addition to the classification percentages. The results found show that it is possible to estimate and classify the movements of standing and sitting for the tasks of execution and motor imagination from channels C3, Cz and C4 with a high % of precision. It is expected, in the long term, to contribute to improve the quality of life of people with motor functional diversity.