Surface somatosensory evoked potential detection by FPGA based multi-adaptive filter

Surface somatosensory evoked potentials (SSEP) collected from conscious subjects usually presents poor signal-to-noise ratio (SNR), requiring several hundreds ensembles averaging to provide a meaningful waveform. A FPGA based adaptive filtering is proposed to perform fast and accurate SSEP extraction by fixed-point adaptive noise canceller (ANC). In 6 normal subjects and 1 neurological abnormal patient, the latency and the peak-to-peak amplitude in SSEP by FPGA based ANC technique were compared with that measured by ensemble averaging. Using 100 trials ANC processed SSEP was sufficient to extract a waveform in equivalent to that extracted by 1000 trials ensemble averaging. The use of fixed-point ANC based on FPGA proved to shorten SSEP measurement time and provide varying information underlying SSEP.