Show summary
Auditory perception involves separating streams of sounds in our surroundings that originate from different sources and have varying temporal structures. The ability to selectively focus on one amongst multiple sound streams, as well as the ability to extract regularities to anticipate forthcoming sounds, are therefore essential. These abilities are often referred to as selective attention and prediction, respectively. Selectively attending to and predicting specific information involve processing multiple dimensions of sensory information, such as its content (what), location (where), and timing (when). So far, the research fields of attention and sensory prediction have been mainly concerned with the ‘what’- and ‘where’-dimensions of information processing. The temporal aspects of sensory prediction and attention are therefore far less understood, despite the obvious importance of timing in perceiving rapidly changing environments. Accordingly, the present work aimed to investigate the specific roles of attention and prediction during perception of temporal structure, as well as their potential interaction.
The first goal was to identify electrophysiological signatures of temporal attention during perception of sound sequences. It was found (Paper I) that modulation of neural activity indexing voluntary attention to sound, aligned more strongly to sound timing in attended compared to ignored rhythmic sound streams. The second goal was to identify electrophysiological and behavioral correlates of temporal prediction. Previous studies demonstrated how the brain responds more strongly to surprising sounds (i.e., deviant pitch) as their probability of occurring decreases (i.e., statistical learning). The second study (Paper II) showed that temporal predictability similarly affected early perceptual responses, as these were diminished when sound timing was more predictable (i.e., less frequent rhythm violations). Additionally, the third study (Paper III) found that presenting sounds at statistically surprising intervals improved processing of coinciding visual targets, indexed by faster responses. Finally, the third goal was to address the independence of- as well as potential interaction between attention and prediction. The reported electrophysiological and behavioral effects of temporal predictability (Papers II and III) occurred regardless of whether the sounds were task-relevant, in line with assumptions that prediction is an inherent feature of perceptual processing. However, the attention-allocation effects reported in Paper I appeared to be enhanced when the auditory rhythmic stimuli were more predictable, suggesting interaction.
In summary, it was demonstrated how to conceptualize and study the independent roles of attention and prediction during perception of temporal structure. Furthermore, this permitted investigating potential interaction between attentional and predictive processing. It was shown how attention to varying temporal structures in rapidly presented sound stimuli, is reflected in ongoing neural dynamics. Furthermore, we found similar effects of temporal predictability to those reported for prediction of sound features (e.g., pitch). Specifically, the expectedness of particular intervals between sounds was affected by the probability of their occurrence (i.e., statistical learning). Finally, predictive processing related to sound timing appeared to operate regardless of attention (Papers II and III). However, the potential interaction demonstrated in Paper I highlights that while predictive processing might operate automatically, this does not rule out the possibility of an interplay between mechanisms. In conclusion, the results might have important implications for expanding our knowledge regarding how these mechanisms are implemented in the brain in future studies (e.g., identifying crucial brain structures and specific neural mechanisms). Furthermore, deficient temporal processing is a characteristic of certain disorders (e.g., Parkinson’s disease), making this research also clinically relevant. Accordingly, findings of the present work encourage future investigation of attention and prediction as distinct cognitive mechanisms.
