We present an estimate for the instantaneous discharge probability of a neurone, based on single-trial spike-train analysis. By detecting points where the neurone abruptly changes its firing rate and treating them specially, the method is able to achieve smooth estimates yet avoid the blurring of significant changes. This estimate of instantaneous discharge probability is then applied to the method of unitary event analysis. We show that the unitary event analysis as originally conceived is highly sensitive to firing-rate nonstationarities and covariations, but that it can be considerably improved if calculations of statistical significance use an instantaneous discharge probability instead of a firing-rate estimate based on averaging across multiple trials.

This content is only available as a PDF.
You do not currently have access to this content.