Improved artifact rejection and isolation of compound action potentials by means of digital subtraction

When recording compound action potentials (CAPs) elicited by stimulating near the recording site, it may prove difficult to distinguish the CAP from the shock artifact because of their overlap in time. This problem is compounded when a pair of stimulation pulses is delivered because the CAP elicited by the test pulse (T pulse) may be partially superimposed on the artifact and response elicited by the conditioning pulse (C pulse) as well as on the T pulse artifact. Methods based on digital subtraction were used to address these problems. A record was obtained with the C-T interval adjusted to be slightly less than the absolute refractory period so that the T pulse would fail to elicit a CAP. A record consisting of a C pulse artifact and response was subtracted from this record to yield a 'pure' shock artifact. In principle, subtracting this 'artifact-only' record from records obtained with single pulses removes the shock artifact and yields a 'pure' response. An extension of this method was used to isolate T pulse responses from the C pulse artifact, C pulse response, and T pulse artifact. These methods proved effective in improving the isolation of the CAPs of interest from other features of the raw records. Limitations of the techniques and their complementarity with other methods of artifact reduction are discussed.