Comodulation masking release with delayed signal onsets

The detectability of a 1250‐Hz, 50‐ms signal was assessed in a comodulation masking release (CMR) setting. One 50‐Hz noise band (the “masker”) was centered at 1250 Hz; other noise bands were centered at 850, 1050, 1450, and 1650 Hz (the “cue” bands). The masker and/or cue bands were gated prior to the onset of the signal by an amount (“fringe”) that was varied across blocks of trials. The noise bands and the signal always ended together. The temporal envelopes of the noise bands were all correlated, all correlated except for the masker band, or all uncorrelated. When all of the noise bands were gated synchronously, the CMR grew from 1 dB in the burst condition to 7 dB for a 450‐ms fringe, due to a greater improvement in detectability in the correlated condition compared to the uncorrelated and all‐uncorrelated conditions. When the cue bands were gated before the masker, the average CMR was larger (4–6 dB) than when the masker was gated before the cue bands (2–4 dB). These differences in improvement with increasing fringe duration may be attributable to differences in neural adaptation in the correlated and uncorrelated conditions. [Work supported by NINCDS Grant NS15895.]