M1 neural decoding of finger movements using a priori neural activities before movements

One of the efforts on brain-machine interfaces (BMI) is reducing the system delay to perform sophisticated movements. Motor commands in brain are produced in advance before its movement in central nervous system. In the field of electroencephalogram (EEG), this issue has been studied widely as readiness potential (RP) to overcome time critical problem. Neural responses prior to movement in M1, primary motor cortex have been observed. In this paper, we present the neural decoding of finger or wrist movements using neural activity data collected from 0.5 sec. to 1.0 sec. when the movement onset happened at 1.0 sec. With novel temporal window, 30 randomly selected neurons could perform competent decoding accuracy as high as 95%.