Presentation

M. Blake Rafferty¹, Eun Jin Paek¹, Tim Saltuklaroglu¹, Kevin Reilly¹, Steffanie Barber¹, Devin M. Casenhiser¹; ¹University of Tennessee Health Science Center

Tracking language in naturalistic speech: an EEG investigation of the mechanisms underlying aphasic comprehension Individuals with aphasia (IWA) often exhibit marked difficulties in sentence comprehension. Mechanistic accounts of aphasic comprehension have proposed a diverse array of possible causes, such as: 1. intermittent sentence parsing disruptions (Caplan, Michaud, and Huffman, 2015), 2. reduced processing speed (Burkhardt, Piñango, and Wong, 2003), 3. reduced memory resources (Dronkers, et al., 2000), and 4. increased susceptibility to interference/noise (Dickey, Choy, and Thompson, 2007; Gibson et al., 2013). The present study investigates these possible mechanisms by assessing differences in cortical responses for IWA and a group of healthy controls as they passively listen to a set of naturalistic sentences that vary in their length and complexity. We specifically intend to assess how each suggested mechanism may alter the cortical “tracking” of abstract linguistic units (e.g., words and syntactic phrases), which has been suggested as a possible mechanism for internally representing abstract linguistic information. Tracking will be quantified via Mutual Information (MI) between EEG responses and the acoustic envelope of the stimulus materials, filtered at the rate of occurrence for words and syntactic phrases. Critically, because MI is a single measure of statistical dependency between two continuous signals, it may be used as a valid measure of cortical response for relatively few experimental trials. This is specifically relevant to studies of IWA because it allows for a drastically reduced amount of time for recording, consequently reducing the overall task-demands and fatigue that IWA may experience during studies with many trials. To determine whether IWA exhibit altered cortical tracking, MI at word and phrase rates will be compared. Following this, possible mechanisms for aphasic comprehension will also be operationalized and compared between groups. Intermittent deficits in IWA (1) will be quantified as the overall variance in MI that each subject exhibits for items of interest—for example, to words at the word rate and to items that close syntactic phrases at the phrase rate. To assess reduced processing speed (2), we plan to quantify the temporal lags that contain the highest amount of MI at each item of interest (i.e., words/phrase closures) for each subject. To assess reduced memory resources (3), we plan to assess how MI relates to sentence length in each group, and to evaluate susceptibility to interference in IWA (4), we intend to assess how MI relates to syntactic/lexical ambiguity. Following between group comparisons, each factor will be entered into a regression analysis investigating the ways that each relates to severity and aphasia type. These results may provide novel insights into cortical function/disfunction in IWA during online comprehension and may serve as useful tools for evaluating the success of interventions.

Topic Areas: Disorders: Acquired,