Brain-Computer Interface Based Triggering Functional Electrical Stimulation Device for Foot Drop Rehabilitation
Salma Mahgoub Hussein Ahmed, Magdi B. M. Amien , Yousif Elhadi Elsideag Ahmed
كلية الهندسة والتكنولوجيا-جامعة الجزيرة · السودان
Functional Electrical Stimulation (FES) is considered to be the most effective treatment for the foot drop patients with severe hemiplegia. The FES technique uses smearing harmless levels of electric current to activate the impaired or disabled neuromuscular system in order to accomplish the lost function. The combination of the two technologies of Brain Computer Interface (BCI) with the FES has been created as an enhancement to functional electrical stimulation therapy technique in which activation of the stimulation was achieved by identifying changes in the Electroencephalogram (EEG) oscillatory activity indicating the attempt to move. The objective of this study was to generate better rehabilitation BCI based method for foot drop patients due to stroke through integrating BCI and FES technologies. Graz dataset ‘A’ has been used, processed and analyzed using MATLAB Mathworks and the open-source toolbox BioSig to load and open the signals after the appropriate preprocessing. Then, EEG data was classified into either “idling” or “dorsiflexion” state using Linear Discriminant Analysis (LDA) according to their selected features based on the spectral information using the Power Spectrum Density (PSD) as a frequency domain analysis method. The LDA classifier had been applied to maximize the distance between the two classes means and minimize the inter class variance, while PSD has been implemented to present at which frequency range variations were strong. The performance of the developed method recorded an efficient computational time, and approved great performance by achieving 79%, 99.7%, 87.9%, 98.9%, 89.4%, for sensitivity, specificity, F1 factor, precision, and the area under the curve respectively, with 95.6%, classification accuracy. For further future works, the study recommended building a low-cost electric circuit of FES device to import the output of the BCI triggering signal as a control signal, and comparing BCI to flex sensor triggering signal according to FES factors such as waveform, amplitude, pulse duration, frequency, ramp up/down, on/off time, and polarity.