Real-time brain activation detection by FPGA implemented Kalman filter

This paper presents an embedded system for real-time multi-channel brain activity detection by implementing the Kalman filter (KF) core on a field-programmable gate array (FPGA). The KF with a model driven approach is implemented on an FPGA, for the first time as per our knowledge. The model driven based brain activation model and its parameters' estimation methodology by KF is depicted from Aqil et al., 2012 (Detection of event-related hemodynamic response to neuroactivation by dynamic modeling of brain activity). The multiple instantiations of the KF core along with the coding of necessary compatibilities amongst the KF cores and with a single communication core, allow the parallel processing of multiple measurement channels. Through a serial universal asynchronous receiver/transmitter core, the fNIRS data is communicated to the system where it is being concurrently processed in 32-bit single precision IEEE754 format. The proposed fNIRS-KF embedded system is verified by an fNIRS dataset in real-time.