Dhatri Devaraju, Ph.D.
Lecturer/Researcher, Department of Communication Sciences and Disorders
University of Wisconsin-Madison
Leveraging electroencephalography and machine learning in studying neural mechanisms of attention control and working memory
Effective speech perception requires sustained attention and constant updating and manipulation of speech information stored in the working memory. Electroencephalography (EEG) can be an effective non-invasive tool that can be used to assess how these processes interplay at different timescales in the brain, which can be a window into understanding disordered attentional control in individuals with communication disorders. The talk will focus on two studies assessing neural mechanisms underlying attentional control in individuals who stutter. The first study specifically evaluates inhibitory control in children who stutter using a child-friendly visual go/no-go task. Although there were no significant differences in children who stutter compared to their peers in terms of spatio-temporal analyses of their event-related potentials, different patterns of frontal beta rhythms were evident between the groups. The second study aimed at evaluating the effects of sustained attention and working memory on phonological processing using an auditory n-back task in adults who stutter. Adults who stutter showed altered neural representations in terms of spatio-temporal distributions compared to their peers. Further, I will be talking about the current advances that we are making in leveraging machine learning algorithms to classify syllables based on EEG responses in an n-back task under increased attention and working memory demands.