Document Type

Article

Major

Physiology & Neurobiology

Mentor

Prof. Heather Read, Depts. of Psychological Sciences and Biomedical Engineering

Abstract

Prior studies have found alpha (8-12 Hz) and beta (15-30 Hz) oscillations measured with EEG both increase in power when people are performing speech-in-noise tasks. In theory, variation in speech-in-noise performance could reflect the ability to segregate and neurally encode background versus foreground sounds. Here, we aim to examine how alpha and beta oscillations play a role in ignoring background sounds versus attending to foreground speech sounds. We had thirty-four healthy young adults perform a speech-in-noise task while we recorded brain signals using 64-channel EEG. Subjects were instructed to ignore the randomly varied background “noise” sounds that onset at the beginning of each trial and attend to foreground digits. After listening to each digit sequence, subjects reported the digits heard. We analyzed the EEG data using custom MATLAB scripts developed by our lab, finding that speech-in-noise performance is easier when the background sound has high stationarity in acoustic features over time. Our results indicate a high involvement of alpha and beta oscillations in attending to foreground sounds amidst background noise, with the alpha oscillations increasing prior to foreground sound onset during background sound onset in order to suppress brain processing of the distracting background sound in preparation for focusing on the attended foreground digits. Interestingly, the increase of beta power prior to onset of the attended digit sequence supports the theory that beta oscillations engage to generate a working memory encoding of the ignored background sounds. Additionally, the more dynamically variable speech “Babble” background sound induced more beta oscillations, in theory reflecting more working memory processes and detection of temporal amplitude modulations over time for the speech “Babble” sounds as compared to the other sounds, such as “White Noise”. Given that the behavioral performance for correctly reporting the digit sequence was also lower for “Babble” background sounds, the higher beta power for “Babble” may index distractibility as well as a working memory representation of the “Babble” sound.

Comments

The author thanks the Holster Scholars Program, the University of Connecticut Honors Program, and the UConn Foundation for funding.

Download

Share

COinS