Classifying human motor imagery abilities from heart rate variability analysis: a preliminary study

This study investigates the assessment of motor imagery (MI) ability in humans through the analysis of heartbeat dynamics. Previous studies have demonstrated that MI processes strongly influence the autonomic nervous system (ANS) activity and, consequently, this reflects on the dynamics of ANS correlates such as the Heart Rate Variability (HRV). Here, we propose to extract a set of linear and nonlinear features from the HRV signals to characterize good and bad imagers. The feature set was used as input of a pattern recognition system based on the support vector machine in order to automatically recognize good and bad imagers using only cardiovascular information. To this aim, we designed an experiment where twenty volunteers performed visual and kinaesthetic imagery tasks. Results showed an accuracy of classification between good and bad imagers over 74%.