Converting Spectral Evoked-to-Background Ratio intoTime-Domain Signal-to-Noise Ratio - Validation for HighFrequency Oscillations

Abstract

N -Interval Fourier Transform Analysis (N -FTA) allows for spectral separation of an evoked target signal from uncorrelated background activity. It computes the frequency-dependent evoked-to-background ratio (EBR). The developed method allows for conversion of the spectral EBR into expected values for improvement of signal-to-noise ratio with progressing sweep count. Our study presents the mathematical basis for this conversion along with a validation for simulated and recorded data. The major findings are: O_LIThree factors enter the calculus of the expected SNR: the ratio of durations of the single sweep cycle and the evoked response window, the mean EBR in the spectral target band, and the sweep count. By conversion of all factors to dB, the expected SNR is defined by their sum. C_LIO_LIThe two fundamental theories governing the improvement of SNR with increasing sweep count, the law of large numbers and the uncertainty principle of signal processing, deliver identical results. C_LIO_LIConversion of EBR to expected SNR was successfully validated by simulated and recorded data and can be applied to all types of evoked data. C_LI Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=112 SRC="FIGDIR/small/662283v1_ufig1.gif" ALT="Figure 1"> View larger version (18K): org.highwire.dtl.DTLVardef@6f34d6org.highwire.dtl.DTLVardef@f5d607org.highwire.dtl.DTLVardef@1b6a18eorg.highwire.dtl.DTLVardef@1288e35_HPS_FORMAT_FIGEXP M_FIG C_FIG