A division of labor between power and phase coherence in encoding attention to stimulus streams

Alessandro Tavano, David Poeppel

Research output: Contribution to journalArticle

Abstract

Both time-based (when) and feature-based (what) aspects of attention facilitate behavior, so it is natural to hypothesize additive effects. We tested this conjecture by recording response behavior and electroencephalographic (EEG) data to auditory pitch changes, embedded at different time lags in a continuous sound stream. Participants reacted more rapidly to larger rather than smaller feature change magnitudes (deviancy), as well as to changes appearing after longer rather than shorter waiting times (hazard rate of response times). However, the feature and time dimensions of attention separately contributed to response speed, with no significant interaction. Notably, phase coherence at low frequencies (delta and theta bands, 1–7 Hz) predominantly reflected attention capture by feature changes, while oscillatory power at higher frequency bands, alpha (8–12 Hz) and beta (13–25 Hz) reflected the orienting of attention in time. Power and phase coherence predicted different portions of response speed variance, suggesting a division of labor in encoding sensory attention in complex auditory scenes.

Original languageEnglish (US)
Pages (from-to)146-156
Number of pages11
JournalNeuroImage
Volume193
DOIs
StatePublished - Jun 2019

    Fingerprint

Keywords

  • Attention
  • Deviancy
  • Oscillation
  • Phase coherence
  • Reaction time

ASJC Scopus subject areas

  • Neurology
  • Cognitive Neuroscience

Cite this