Evidence for sensory-trace comparisons in spectral shape discrimination.

Area of Science:

• Auditory perception
• Psychoacoustics
• Signal detection theory

Background:

• Human auditory system's ability to detect subtle changes in complex sound spectra is crucial for speech and music perception.
• Understanding how spectral shape variations affect sound intensity discrimination is key to auditory modeling.
• Previous research suggests memory processes influence sound intensity perception, as described by Durlach and Braida's theory.

Purpose of the Study:

• To investigate how variations in the spectral shape of complex sounds (reference spectrum) impact observers' ability to discriminate intensity changes.
• To examine the role of different reference spectrum conditions (fixed, trial-to-trial, and presentation-to-presentation perturbation) on intensity discrimination.
• To test the applicability of Durlach and Braida's theory of auditory memory in explaining the observed discrimination patterns.

Main Methods:

• Three experimental conditions were used, varying the spectral shape of the reference sound: fixed, trial-to-trial differing, and presentation-to-presentation differing.
• A two-interval, two-alternative, forced-choice (2I-AFC) task was employed, with a signal of intensity increment to a 1000 Hz component.
• Key manipulations included the degree of amplitude perturbation in the reference spectrum and the interstimulus interval (ISI).

Main Results:

• Observers' ability to discriminate intensity changes varied significantly across the three reference spectrum conditions.
• The condition with reference spectra perturbed on every stimulus presentation showed evidence of comparisons based on 'sensory traces'.
• Conditions with fixed or trial-to-trial perturbed reference spectra were best explained by comparisons relative to the stimulus context.

Conclusions:

• The way reference spectra are varied critically influences the auditory system's comparison strategy for intensity discrimination.
• Durlach and Braida's theory provides a framework, suggesting 'sensory trace' comparisons occur when stimuli are highly variable.
• Contextual judgments, rather than sensory traces, dominate when reference spectra are stable or vary predictably within trials.