Abstract
Experiments involving the detection of a tone presented in noise form the basis of many models of psychophysics that are used to estimate frequency resolution and cochlear compression for listeners with normal and impaired hearing. Despite the wide-spread application of tone-in-noise experiments, the cues used by listeners to detect the tone are a matter of debate. Here we adopt a molecular psychophysics approach – decision variable correlation (DVC) – to evaluate the extent to which trial-by-trial responses are consistent with candidate decision variables, including stimulus energy or the temporal envelope. We measured detection thresholds in normal-hearing adults for a 1000-Hz tone presented in one-third octave noise centered on the tone frequency. The tone and noise were gated simultaneously and thresholds were measured for durations of 10, 20, 50, and 100ms. We limited the reliability of energy-based cues by incorporating a roving-level paradigm. Similarly, the use of short tones (e.g., 10 ms) limited the reliability of envelope-based cues. Our results reveal that listeners adopted a strategy that emphasized energy-based cues for short-duration tones and envelope-based cues for long-duration tones. This finding suggests that models of human psychophysics may benefit from incorporating a decision device that adjusts cue weights based on stimulus duration.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have