Presentation Information
[P3-20]Temporal Jitter in Music Reveals Robust Early Stream Formation and Enhanced Attentional Selection via Attention Recruitment
*Shu Sakamoto1,2, Emily Wood1,2, Harris Miller1, Ellia Baines1, Kevin Yang1, Lily Eshraghi1, Laurel J. Trainor1,2 (1. Department of Psychology, Neuroscience, and Behavior, McMaster University (Canada), 2. McMaster Institute of Music and the Mind (Canada))
Keywords:
Auditory Scene Analysis,Temporal Response Functions,Jitter,EEG,Attention Decoding
Auditory scene analysis involves organizing sounds into perceptual streams. Our prior work indicates early, automatic stream formation for music is more robust than for speech. The present study investigated if temporal regularity of stimuli, a key bottom-up feature that differentiates music from speech, affects this early automatic musical stream formation. Participants (N=15) listened to two simultaneous custom-composed polyphonic piano pieces, spatialized via head-related transfer functions. The degree of note onset jitter within pieces was varied. Tasks were to either detected targets in one stream (segregation task) or both streams (integration). We recorded 128-channel electroencephalography (EEG) and used multivariate temporal response functions (mTRFs) to reconstruct the spectral flux of stimuli, comparing two representational models: a Separated model representing independent neural processing of streams and a Combined model representing unified stream processing. Results replicated our previous findings of early stream segregation where the Separated model outperformed the Combined model at an EEG-to-stimulus lag of 62.5–85.9 ms in both tasks. Crucially, no significant interaction occurred between the Separated versus Combined representational models and jitter level in either task, suggesting note onset regularity did not modulate early, automatic stream formation. However, a significant main effect of jitter was observed, suggesting general neural encoding was enhanced for stimuli with higher jitter in both tasks. Furthermore, for the segregation task, higher jitter also enhanced attentional selection of the attended stream, evident even at early processing latencies (39.1-117.2 ms). This suggests greater temporal irregularity, which is cognitively demanding, recruits greater top-down attention when segregating streams. In conclusion, while early, automatic musical stream formation was robust to note onset regularity, increased temporal irregularity (higher jitter) recruited greater processing resources, enhancing general neural encoding and aiding attentional selection in a complex auditory scene.