Electrocorticography (ECoG) is an alternative approach to less invasive brain-machine interfaces (BMI). We developed a method to decode multiple muscle activities and trajectory of arm movements from ECoG measurements in our previous study. Although wrist step-tracking movement is one of the most frequent movements performed in everyday life, as our best knowledge, the prediction of multiple targets during wrist step-tracking movement has not been investigated.
The aim of this study is to classify eight-targets using ECoG signals for the prosthesis control. Each ECoG signal was filtered by different bandpass filters for sensorimotor rhythms, normalized by the standard z-score, and smoothed by a Gaussian filter. We used sparse logistic regression to find the best weights between frequency bands of ECoG and targets. We obtained averages of 10-fold cross validation 33.55 ± 9.40 and 34.38 ± 12.93 percent for two section data against 12.5 % chance level.