| The listener's response process is modeled as a temporal window through which the incoming stimulus waveform is viewed with the window positioned so as to maximize signal-to-noise ratio. To test this model, listeners were asked to detect a .5 ms pulse in a noise background during a 1500 ms observation interval. Using the probe-signal method (G. Z. Greenberg and W. D. Larkin, J. Acoust. Soc. Am., 44, 1513-1523, 1968) in a yes/no task; a primary signal which had a fixed temporal position was presented on 40%, a probe signal (variable temporal position) on 10%, and noise alone on 50% of the trials. Listeners' performance, which was measured as hit rate, was maximal at those positions near the primary. In separate experiments, this effect was shown to be independent of the position of the primary.; In companion experiments not utilizing the probe-signal method, it was determined that signals positioned at any point within the observation interval were equally detectable if presented in isolation. However, if a signal could occur with equal probability at any of seven positions (temporal uncertainty), then performance worsened. This decrease can be accounted for by a temporal window which changes from a narrow configuration yielding a large signal-to-noise ratio in the first instance to a broadened state producing a reduced signal-to-noise ratio in the second case. |
