BME PhD Dissertation Defense - Lia Bonacci


BME PhD Dissertation Defense - Lia Bonacci


Title: “Towards Better Communication at the Cocktail Party: Neural Correlates of Spatial Attention in Complex Auditory Scenes”

Committee: Barbara Shinn-Cunningham, PhD – BU BME (Advisor) H. Steven Colburn, PhD – BU BME (Chair) Nancy Kopell, PhD – BU Math Michael Qin, PhD – Office of Naval Research’s Robotic, Autonomic, and Autonomy Program Inyong Choi, PhD – The University of Iowa’s Department of Communication Sciences and Disorders

Abstract: The ability to direct and maintain attention is essential for communication in noisy environments. Often times, knowing where to attend is helpful for this purpose, allowing selection of targets in one location and suppression of distractors elsewhere. This ability is effortless for most, yet the way in which the brain performs this selection is still not well understood. Furthermore, while many can successfully converse when surrounded by irrelevant speech and noise, hearing-impaired individuals struggle to do so, even with current assistive technologies. In order to address these issues, this work aims to characterize neural correlates of spatial attention under a variety of circumstances using electroencephalography (EEG), which could be used to guide processing of acoustic signals. First, we characterize EEG correlates of visual spatial attention to dynamic stimuli within scenes of different spatial complexities. Second, we characterize these same correlates during an auditory spatial attention task and explore how pitch differences between competing auditory streams may affect the use of spatial features during sustained attention. Finally, we compare the identified correlates between normal-hearing and hearing-impaired listeners in order to determine if a weakened ability to spatially attend affects the observed neural signatures. Thus, the work presented in this dissertation serves to shed light on the brain's ability to direct and maintain spatial attention in a variety of contexts. By characterizing EEG signatures within these contexts, we also provide considerations for future technology that uses non-invasive neural measures to enhance communication abilities of hearing-impaired listeners in complex auditory scenes.


11:00am on Wednesday, March 6th 2019


610 Commonwealth Ave (CILSE), room 809


ENG Home, BME Home