Spectral Separation of Evoked and Background Cortical Activity - Application to Short Latency Median Nerve Evoked Potentials

BackgroundSpectral analysis of repeatedly evoked potentials (EPs) is challenging since recordings contain a superposition of evoked signals and background activity. We developed a novel approach, N-interval Fourier Transform Analysis (N-FTA), which allows for reliable separation and accurate assessment of evoked and background spectral components. We applied this approach to spectral analysis of median nerve short latency EPs for identifying spectral bands of clinical relevance. MethodsWe performed right median nerve stimulation in two volunteers at 2.46 Hz and 3.95 Hz stimulation rate (600 and 1000 repetitions respectively). We applied N-FTA for splitting the periodically repeated evoked components from irregular background activity and investigated spectral components in the low, medium and high frequency (LF-, MF-, HF-) bands. We present a signal processing approach, which allows for accurately extracting diagnostically relevant features (signal morphology, latency and amplitude) using an 18 Hz to 240 Hz bandwidth. ResultsIn the LF-band < 10 Hz, evoked-to-background ratio (EBR) was below -15 dB due to the high level near DC background activity (1/f-activity, eye blinks,{theta} -band). Highest EBR was near -5 dB and obtained at a few tens of Hz. Relatively broad continuous segments of evoked activity were detected in the MF band. These frequencies of a few hundred Hz were linked to the signal segment between the N20 and P25 peaks in the time domain. High frequency oscillations (HFOs) near 600 Hz were approximately -25 dB below the background level and near the limit of the sensitivity of N-FTA. By subtracting stimulation artifacts and applying zero-phase filters it was possible to extract diagnostically relevant short latency EP features (signal morphology, latency and amplitude) only from the MF-band with a similar accuracy as a routinely used broad-band setting. HFOs displayed amplitudes of a few tenths of {micro}V. SummaryN-FTA allows for accurate, simultaneous spectral analysis of evoked and background activity in individual trials. The approach allowed for identifying target frequency bands of high evoked activity and for tailoring signal processing such, that morphology and latency can be obtained at a significantly reduced bandwidth. This should allow for development of more robust and faster recording techniques in the near future.