You are viewing the SNL 2018 Archive Website. For the latest information, see the Current Website.

Poster D15, Friday, August 17, 4:45 – 6:30 pm, Room 2000AB

Impaired Incidental Phonetic Learning in People with Aphasia

Christopher Heffner1, David Saltzman1, Samantha Formica1, Emily Myers1;1University of Connecticut

Learning does not end after a stroke. People with aphasia (PWA) must learn or relearn many attributes of language to recover from their stroke, and it is likely that this learning process recruits multiple systems (Menke et al., 2009), including systems responsible for declarative and procedural learning. Although previous studies have found that PWA are slower to learn non-linguistic categories (Vallila-Rohter & Kiran, 2013, 2015), phonetic categories have not yet been investigated. In the present study, we investigate the acquisition of non-native phonetic categories in PWA, focusing on the influence of feedback on phonetic learning in PWA. Although nearly every proposal related to learning in PWA suggests that their procedural memory should be relatively spared, it is unclear if feedback encourages (Vallila-Rohter & Kiran, 2013b) or inhibits (Gabay, Dick, Zevin, & Holt, 2015) effective use of procedural learning. All PWA sustained left hemisphere damage, with variable lesion sites and behavioral consequences. Participants completed two tasks. In each task, they learned one of two phonetic continua—German fricatives and Arabic geminates—with assignment of continuum to task counterbalanced across participants. In the Incidental Task (based on Gabay et al., 2015), participants were told their job is to fight zombies by pressing a button corresponding to the location of a zombie on a screen. Although the participants were not told this before the experiment, the location of the zombie could be predicted based on the speech sounds they hear before the zombie appears. The pairings of zombies to sounds were scrambled during one block, with the degree of RT slowing serving as a measure of learning. In the Explicit Task, participants were explicitly told that they were going to learn to pair speech sounds with colored squares. After hearing a sound, they pressed a button corresponding to one of three colored squares. They were then given feedback on their responses: a checkmark for a correct response, and an X for an incorrect response. The index of learning for this experiment is the number of trials correct across the whole experiment. On the Incidental Task, there was little evidence that participants slowed down during the random block, suggesting that PWA could not exploit regularities in the input to learn during the Incidental Task. Yet on the Explicit Task, a majority of PWA showed evidence of learning, although none at the same rate as typical controls. The participants who successfully learned the contrast in the explicit condition tended to have spared frontal language areas, potentially indicating the importance of these areas for successful category learning; they also almost uniformly had lesions in temporal regions encompassing portions of Heschl’s gyrus, which may indicate that those regions are less important. The patterns for the non-learning participants were much more heterogeneous. This suggests that (1) PWA are impaired relative to controls on the acquisition of novel phonetic categories and (2) feedback may help promote successful exploitation of remaining learning systems on the part of PWA (Vallila-Rohter & Kiran, 2013).

Topic Area: Perception: Speech Perception and Audiovisual Integration