Assessing Behavioral and Neural Correlates of Change Detection in Spatialized Acoustic Scenes

Abstract

The ability to detect changes in complex auditory scenes is crucial for human survival, yet the neural mechanisms underlying this process remain elusive. This study investigates how the presence and location of sound sources impacts active auditory change detection as well as neural correlates of passive change detection. Stimuli were naturalistic temporal envelopes applied to synthesized broadband carriers designed to eliminate semantics and minimize contextual information while preserving naturalistic temporal envelopes and broadband spectra, presented in a spatial loudspeaker array. Behavioral change detection experiments tasked participants with detecting new sources added to spatialized and non-spatialized multi-source auditory scenes. In a passive listening experiment, participants were given a visual decoy task while neural data were collected via electroencephalography (EEG) during exposure to unattended spatialized scenes and added sources.

Our two behavioral experiments (N = 21 and 21) demonstrated that spatializing sounds facilitated change detection compared to non-spatialized presentation, but that performance declined with increasing number of sound sources and higher hearing thresholds at mid-high frequencies, exclusively in spatialized conditions. Slower reaction times were also observed when changes occurred from above or behind the listener, exacerbated by a higher number of sources. Two EEG experiments (N = 32 and 30), using the same stimuli, showed robust change-evoked responses. However, no significant differences were detected in our analysis as a function of spatial location of the appearing source. These findings provide fresh insights into the mechanisms of spatial auditory change detection, emphasizing the dynamic interplay of spatial cues, change location, and scene complexity.