Slide Slam F5
ERP and time-frequency correlates of phonological and temporal deviants in dyslexic readers
Alexandra Emmendorfer1, Sonja Kotz1, Bernadette Jansma1, Milene Bonte1; 1Maastricht University
Developmental dyslexia is a reading deficit of neurobiological origin, associated with persistent difficulties in fluent word recognition and impaired spelling abilities. Further deficits in speech processing are reported, with dyslexic individuals showing decreased sensitivity to sublexical features. In typical development, the brain exploits phonotactic and temporal regularities in speech through statistical learning, to achieve accurate phonological representations. These representations are crucial in early word learning, and may further contribute to reading development. The passive oddball paradigm offers an opportunity to study implicit processing of these features. The ERP component of interest is the mismatch negativity (MMN), a marker of experience-dependent change detection. Prior research has shown a sensitivity of the MMN to variations in phonotactic probability (Bonte et al., 2005) but not syllable stress (Emmendorfer et al., 2020) in Dutch speakers. This sensitivity is reduced in adults (Noordenbos et al., 2013) and children (Bonte et al., 2007) with dyslexia. In the time-frequency domain, auditory change detection is associated with increased theta inter-trial coherence (ITC) in deviant trials (Fuentemilla et al., 2008). Prior research has reported a further enhancement of theta ITC in children with dyslexia (Halliday et al., 2014). We used EEG to investigate the sensitivity to phonotactic and temporal regularities in speech in adults (aged 18-30) with and without dyslexia in a passive, multi-feature oddball paradigm. We manipulated phonological and temporal regularities in Dutch pseudowords varying in phonotactic probability (formal deviant) and syllable stress (temporal deviant). Each stimulus was presented as a formal deviant, temporal deviant, or standard. This allowed comparing identical stimuli across conditions. To understand the mechanisms underlying formal and temporal change detection, we perform both ERP and time-frequency analyses. ERP analyses revealed that both formal and temporal deviants elicit an MMN. For formal deviants, phonotactic probability modulates MMN latency, with high phonotactic probability deviants eliciting an earlier MMN compared to low phonotactic probability deviants, indicating facilitated change detection. These responses were overall delayed in dyslexic readers. We did not observe modulations of MMN in temporal deviants. In the time-frequency domain, we observed the typical increased theta ITC in formal deviants, with a trend towards enhanced ITC in dyslexic individuals. The theta ITC was modulated by syllable stress: Formal deviants with second syllable stress elicited greater theta phase-locking. Temporal deviants elicited decreased delta/theta ITC. The current results indicate an overall increase in theta ITC and slower ERP responses to formal deviants in dyslexic readers, but no group differences related to sublexical regularity. They support prior evidence of MMN sensitivity to phonotactic probability. The fact that we do not report group differences for phonotactic probability may be explained by comparably high reading levels of dyslexic university students. The dissociable oscillatory patterns observed for formal and temporal deviants are striking. As the acoustic markers of syllable stress are typically associated with increased theta ITC, the current pattern of oscillatory results may suggest that temporal deviants disrupt the tracking of the regular speech rhythm within a stimulus sequence, while change detection in formal deviants is associated with sensory memory processes.