Inter-muscular coherence features to classify upper limb simple tasks

The application of Hybrid Brain-Computer Interfaces (BCI) for post-stroke hand motor rehabilitation requires the investigation of new electromyographic (EMG) features, potentially able to identify pathological synergies to be discouraged. Inter-muscular coherence (IMC) is gaining attention as a descriptor of the mechanisms behind abnormal motor control in stroke patients. With the ultimate goal to exploit IMC features to control BCIs, this work aims at (a) characterizing finger extension and grasping tasks by IMC features, (b) assessing IMC feature performance in classifying different conditions. Classification results (accuracy equal to 0.81 ± 0.19) pave the way for IMC feature application in hybrid BCI control.