Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Investigating representational overlap in sublexical structure between fingerspelling and text
Poster Session D, Thursday, October 1, 4:30 - 6:30 pm, Wangari Maathai
This poster is part of the Sandbox Series.
Brennan Terhune-Cotter1, Samuel Evans1, Cathy Price1, Meaghan Spedden1, Patrick Rosenburg1, Mairéad MacSweeney1; 1University College London
Fingerspelling in British Sign Language (BSL) is an orthographic code that explicitly visualizes English text on the hands. Better fingerspelling skill is connected to better reading outcomes in deaf children (Allen, 2015), and may constitute an alternative pathway to skilled reading for this population. Fingerspelling primes English text, suggesting shared orthographic representations (Lee et al., 2025; Sehyr et al., 2023). Neural evidence likewise indicates overlap between fingerspelling and text: both engage bilateral ventral occipitotemporal (vOT) and posterior temporal cortex. An additional finding is that orthographic tuning in vOT is left-lateralized in hearing speakers but bilateral in deaf signers (Glezer et al., 2025). This population difference may stem from the integration of orthographic and manual processing during fingerspelling; an interpretation that is supported by evidence that vOT is also responsive to non-linguistic hand stimuli in hearing non-signers (e.g. Peelen and Downing, 2007). While these findings indicate fingerspelling and text share neural resources, we do not know whether common neural responses within regions such as vOT reflect shared sublexical representations or co-located but distinct neural populations. The current study investigates this question. The second research question will examine whether there is an impact of grain size on processing of fingerspelling and text. Fluent fingerspelling contains coarticulation, in which adjacent letters are combined into fluid movement ‘envelopes’ that abstract over individual letters in a word. Efficient comprehension of fluent fingerspelling may require the activation of coarse-grained movement representations which are not strongly linked to individual letter sequences. This is conceptually analogous to coarse-grained processing of written words which is achieved by approximating relative letter positions over words (i.e., open bigrams) (Grainger and Ziegler, 2011). We hypothesize that, due to its reliance on movement, coarse-grained processing of fingerspelling engages visual phonological processes, rather than orthographic processes typically located in vOT. We will scan twenty-five deaf BSL signers with 3T fMRI while they read written and watch fingerspelled nonwords. To assess the impact of modality on orthographic processing (RQ1), we will use searchlight representational similarity analysis (RSA) to test for a modality effect in the neural representations of individual words within vOT. We predict that modality-specific representations will occur in posterior vOT, with cross-modal representations emerging in more anterior regions. To compare activation for fine- versus coarse-grained representations (RQ2), half of the items will be presented as single-letter sequences, therefore removing coarse structure from both the written words and fingerspelled words. We predict that, in contrast to predicted vOT activation for letter-by-letter structure (in both modalities) and coarse text structure (open bigrams), coarse fingerspelling structure will be primarily processed in posterior temporal cortex, which has been implicated in the processing of biological communicative movements in non-signers (Peelen and Downing, 2007). These data will further our understanding of how fingerspelling may provide a ‘bridge’ to reading through the activation of shared sublexical structure.
Topic Areas: Signed Language and Gesture, Reading