Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Speech Perceptual Adaptation in Dyslexia: Behavioral and Neural Responses
Poster Session C, Thursday, October 1, 10:45 am - 12:45 pm, Wangari Maathai
Natalia López Rojas1, Daniel del Olmo1, Òscar Cervera1, Gaël Cordero1,2, Begoña Díaz1; 1Universitat Internacional de Catalunya (UIC), 2Basque Center on Cognition, Brain & Language (BCBL)
Speech perceptual adaptation allows listeners to adjust speech perception to variable input across speakers. Dyslexia, a reading disability linked to a phonological deficit, is associated with difficulties in identifying phoneme invariants, which may hamper adapting to speakers’ idiosyncratic productions. The present study investigated this issue by comparing 16 adults with dyslexia and 16 matched controls in their ability to flexibly adapt speech perception. Participants completed a phoneme categorization task using synthesized vowels, where the endpoint vowels /a/ and /e/ were recorded by a female speaker. They also completed three oddball paradigms before and after a distributional learning (DL) phase. During DL exposure, continuum tokens were presented in a bimodal distribution centered on each participant’s most ambiguous stimulus to induce a shift toward categorizing it as /e/. Three oddball paradigms were used to assess the Mismatch Negativity (MMN), an electroencephalographic (EEG) index of auditory discrimination. In the Trained Voice condition, the endpoint /a/ was frequently presented (standard) and the participant-specific ambiguous token was infrequently presented (deviant). In the Unfamiliar Voice condition, the same stimuli as in the Trained Voice condition were used, but with relevant speaker characteristics modified to change voice identity. In the Auditory Control condition, two complex tones were presented with the five formant frequencies of the two stimuli from the Trained Voice condition. Before DL, dyslexics showed different categorization patterns compared to controls, suggesting a shifted category boundary. After DL, both groups responded similarly, with the category boundary shifting in the expected direction. Regarding the EEG data, deviant stimuli elicited larger negativities than standard stimuli, consistent with the MMN response. Greater MMNs were observed in the Auditory Control and Trained Voice conditions compared to the Unfamiliar Voice condition. A significant Group × Condition interaction indicated different patterns of condition sensitivity across groups. In controls, the MMN was larger in the Auditory Control condition than in both speech conditions. In the dyslexia group, the MMN was greater in the Trained Voice than in the Unfamiliar Voice condition and neither differed from the Auditory Control. The behavioral findings show that individuals with dyslexia can flexibly adjust speech perception similarly to controls. Initial group differences disappeared after DL, suggesting that distributional learning may help people with dyslexia adjust speech perception to speaker characteristics. Regarding the EEG data, diminished discrimination responses were present in the dyslexia group. Contrary to the behavioral data, the MMN amplitude for the trained stimuli did not change after learning in either group. These results suggest a dissociation between behavioral adaptation and automatic neural discrimination of speech, suggesting that short-term perceptual learning may influence phoneme categorization without impacting phoneme discrimination processes indexed by the MMN.
Topic Areas: Speech Perception, Disorders: Developmental