Abstract
The trends in movement-related functional activity measurement for brain-computer interface (BCI) are mostly associated with the central lobe of the brain. This consideration may be a faulty approach for the paralyzed patient. This limitation demands an alternative approach for movement-related BCI. For the first time, we propose the prefrontal hemodynamics for implementing movement-related BCI. This paper aims to model the activation pattern and the classification performances of the prefrontal hemodynamics regarding the movement-related events. Utilizing functional near-infrared spectroscopy (fNIRS) the changes in the concentration of the oxidized hemoglobin and deoxidized hemoglobin regarding voluntary and imagery movements are acquired. With necessary preprocessing, the fNIRS signals are statistically analyzed to localize the most significant activated regions regarding the applied stimuli. The experiment shows that movement-related events have a strong correlation with the prefrontal hemodynamics. The patterns of the movement-related hemodynamics are modeled by polynomial regression and used to classify the voluntary and imagery events based on the maximum similarity approach. The resulting classification accuracies are found promising that proves the effectiveness of the prefrontal fNIRS signal to be effective in movement-related brain functionality analysis.
Highlights
The neural activations regarding the movement-related events are one of the most important research areas for implementing the practical brain-computer interface (BCI)
We have demonstrated the conventional time-domain feature extraction and classification strategy to classify the signals by four different classification algorithms: Artificial Neural Network (ANN), Support Vector Machines (SVM), Linear Discriminant Analysis (LDA), and k-Nearest Neighbor
We found that the classification accuracies of the proposed and conventional methods are almost similar
Summary
The neural activations regarding the movement-related events are one of the most important research areas for implementing the practical brain-computer interface (BCI). Depending on the electrical activities, brain function measuring modality is electroencephalography (EEG) and magnetoencephalography (MEG). MEG is not widely used for functional brain imaging due to its high degree of noise sensitivity [2]. In the field of BCI, it has a very slight scope of operation because of its very high cost, motion sensitivity, and heavyweight. On this contrary, Functional Near-Infrared Spectroscopy (fNIRS) is another technique for functional brain imaging that provides very good spatial resolution (~11.5cm), moderate temporal resolution (up to 100Hz), portability to use, high value of signal to noise ratio (SNR), VOLUME XX, 2017
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
