Regular patterns stabilize auditory streams

Abstract

The auditory system continuously parses the acoustic environment into auditory objects, usually representing separate sound sources. Sound sources typically show characteristic emission patterns. These regular temporal sound patterns are possible cues for distinguishing sound sources. The present study was designed to test whether regular patterns are used as cues for source distinction and to specify the role that detecting these regularities may play in the process of auditory stream segregation. Participants were presented with tone sequences, and they were asked to continuously indicate whether they perceived the tones in terms of a single coherent sequence of sounds (integrated) or as two concurrent sound streams (segregated). Unknown to the participant, in some stimulus conditions, regular patterns were present in one or both putative streams. In all stimulus conditions, participants' perception switched back and forth between the two sound organizations. Importantly, regular patterns occurring in either one or both streams prolonged the mean duration of two-stream percepts, whereas the duration of one-stream percepts was unaffected. These results suggest that temporal regularities are utilized in auditory scene analysis. It appears that the role of this cue lies in stabilizing streams once they have been formed on the basis of simpler acoustic cues.

Multistability in auditory stream segregation: a predictive coding view

“…At this point, it is not clear how the human auditory system uses redundancy or how it deals with conflicting cues. Even less is known about the possible mechanisms mediating the effects of stimulus predictability on auditory stream segregation [48] and how these effects relate to grouping based on similarity cues [49]. The issue of how stimulus parameters govern the duration of the first perceptual phase in the auditory streaming paradigm requires further careful consideration, as it may provide important constraints on modelling the formation of auditory groups and how and when competition commences between the alternatives.…”

“…Both sets of tones contained small, random variations in frequency and intensity identical to Condition 1 reported in Bendixen et al [48]. Participants continuously indicated their perception, and their reports were analysed as described in Bendixen et al [48] and Denham et al [51]. Figure 2 shows the time course of the probability of reporting the segregated percept, averaged across 30 participants.…”

“…It seems reasonable to assume that predictability is used as a binding principle; yet, evidence has long been equivocal in this regard ( [47]; see discussion in Bregman [15]). However, a recent study based on the bi-stable nature of the auditory streaming paradigm [48] revealed an influence of the predictability of the sound sequence on grouping. Although grouping by prediction appeared to occur at a later stage than grouping based on perceptual similarity, detecting a predictive relationship between sounds substantially increased the stability of groups derived from similarity-based grouping.…”

“…Checking all possible combinations of perceptual proto-objects for their compatibility requires full synchronization among them, including aligning proto-objects with different temporal cycles, which is computationally highly demanding. Thus, with an increasing number of proto-objects (the case of multi-as opposed to bi-stability in perception; see these auditory [48,51] and visual [68] studies showing multi-stability), forming several organizations and establishing competition between them poses specific implementation problems for the hierarchical competition solution.…”

“…Here, we report data from one condition from the experiment in which, after 2 min, the frequency difference between the 'A' and 'B' sets of tones was suddenly increased from five semitones to seven semitones by lowering the 'A' tones by one semitone and elevating the 'B' tones by one semitone. Both sets of tones contained small, random variations in frequency and intensity identical to Condition 1 reported in Bendixen et al [48]. Participants continuously indicated their perception, and their reports were analysed as described in Bendixen et al [48] and Denham et al [51].…”