Neural network for confirmation of coil location in transcranial magnetic stimulation by motor evoked potential and force

We proposed an artificial neural network to confirm the location of stimulating coil in transcranial magnetic stimulation (TMS) from data of finger force and electromyography (EMG). In experiments, finger forces of the right index finger and motor evoked potentials (MEPs) of the muscles involved in the generation of the finger forces were measured when the primary motor cortex in the left hemisphere of the cerebrum was stimulated by TMS. The measured finger forces and MEPs varied trial by trial although the stimulating coil and subject's head were fixed. The neural network learned the mapping from the MEPs and/or the finger forces to the coil location. After sufficient learning, the neural network was able to classify unlearned MEPs and finger forces into corresponding coil locations.