Modelling a BCI system to estimate FES stimulation intensity for individual stroke survivors in foot drop cases

This work provides a model to estimate stimulation level required for hemiplegic patient to achieve foot dorsiflexion using BCI approach. The model's input is EEG from motor cortex of hemiplegic patient. The model's output is estimation of stimulation strength. The delta mean and alpha peak frequencies are found to be suitable as EEG features to establish the relationship between EEG parameters and stimulus strength for this model. This model is self-adaptive in nature. As therapeutic procedure continues, the model uses a feedback loop, so that it can fine tune the stimulus strength to optimum value for individual patients. The model is prepared by using analysis of EEG feature from all the frequency bands and stimulus strength as observed clinically in 15 stroke patients while achieving dorsiflexion.