Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Are deaf readers sensitive to semantic violations under high orthographic overlap in the parafovea?
Poster Session F, Friday, October 2, 2:45 - 4:45 pm, Wangari Maathai
This poster is part of the Sandbox Series.
Zed Sehyr1, Samantha Schwarz2, Sara Milligan2, Elizabeth Schotter2; 1Chapman University, 2University of South Florida
During sentence reading, information extracted from upcoming words in parafoveal vision can facilitate subsequent foveal word recognition. Event related potential (ERP) studies using rapid serial visual presentation (RSVP) with flankers have shown that detection of orthographic violations in the parafovea modulates processing demands during subsequent foveal semantic integration, eliciting an N400 (pN400) effect (Caliskan et al., 2023; Laszlo & Federmeier, 2009). Recently, Nestor et al. (2025) demonstrated that in hearing readers, parafoveal preview primarily supports orthographic, rather than semantic processing with downstream facilitation of foveal processing contingent on parafoveal detection of orthographic mismatch. Skilled deaf readers exhibit a wider perceptual span and efficient eye movements than hearing readers (Bélanger et al., 2012), suggesting that they may extract parafoveal information differently. In this study, we investigated whether deaf readers differ from hearing readers in their sensitivity to semantic violations under high orthographic overlap during parafoveal perception, and how parafoveal availability influences subsequent foveal processing strategies. Using experimental design based on Nestor et al., participants read highly constraining sentences using an RSVP with flankers paradigm combined with a parafoveal masking manipulation. Sentence final target words were either (1) expected, (2) anomalous orthographic neighbors (e.g., bank for bark), or (3) anomalous orthographic non-neighbors (e.g., wine for bark). Targets were either visible or masked when presented parafoveally and were always visible when presented foveally. Parafoveal visibility was manipulated to isolate the contribution of parafoveal preview to subsequent foveal processing. ERP analyses focused on the N400 time window from three hundred to five hundred fifty milliseconds time locked to parafoveal and foveal word onset, as well as the Late Positive Component from five hundred fifty to eight hundred milliseconds. Preliminary visual inspection of data from 23 hearing readers and 15 deaf readers reveal distinct ERP profiles across groups. Consistent with prior findings, hearing readers showed a robust pN400 for anomalous non-neighbors, however, no effect for the orthographic neighbors. In the fovea, N400 amplitude was reduced for all both anomaly types when words were visible parafoveally, however, N400 was reduced only for the expected and anomalous neighbor when parafoveal preview was masked. This result only partially replicates the prior study by Nestor et al. The deaf participants did not show pN400 effects for either anomaly type during parafoveal viewing. A subset of eight skilled deaf readers showed a subtle N400 effect to anomalous non-neighbors when words were parafoveally masked, but not when they were visible, suggesting that deaf readers may utilize parafoveal information in ways not indexed by a typical N400 response. Ongoing data collection aims to: a) increase sample power to detect small to medium size effects, b) better match groups on reading and language proficiency and c) to examine how individual differences in reading skill relate to parafoveal processing. The study will clarify whether deaf readers rely on an alternative predictive or orthographic processing strategies during sentence comprehension.
Topic Areas: Reading, Meaning: Lexical Semantics