Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Fronto-temporal language network highly selective for sign language relative to action semantics, regardless of iconicity
Poster Session B, Friday, October 25, 10:00 - 11:30 am, Great Hall 3 and 4
Akshi Akshi1, Miriam Hauptman1, Matthew Sampson1, Qi Cheng2, Marina Bedny1; 1Johns Hopkins University, 2University of Washington
Studies of sign languages provide insight into the role of modality (i.e., auditory vs. visuo-manual) in shaping the neurobiology of language. In speakers of auditory languages, left fronto-temporal language networks are highly selective for language over non-linguistic tasks, including action semantics and working memory (Fedorenko et al., 2020). Like spoken languages, sign languages activate fronto-temporal networks (e.g., Neville et al., 1998). Responses to sign languages are partially distinct from responses to pantomime and gesture in group analyses (MacSweeney, et al., 2004, Emmorey, et al., 2011). The present functional magnetic resonance imaging (fMRI) study uses individual subject analyses to test the selectivity of the language network for sign language over action semantics and visual working memory in deaf signers of American Sign Language (ASL), a visuo-manual language. To test whether iconicity affects language selectivity, we compared non-classifier and partially iconic classifier constructions, matched in meaning and word-order. We use handling classifiers where handshapes convey how objects are handled, thus resembling pantomimic actions. Congenitally profoundly deaf participants (N=9) viewed ASL transitive and di-transitive sentences featuring six verbs (Give, Take, Open, Close, Pick-up, Put-down), with various agents and objects (e.g., The man opened the bottle. The woman gave the book to the man.) in these constructions: Subject-Verb-Object with Non-Classifier verb-form (SVO NCL), Object-Subject-Verb with Non-Classifier verb-form (OSV NCL), and Object-Subject-Verb with Classifier verb-form (OSV CL). In control conditions, participants watched action videos (action control) matched semantically to the ASL sentences (e.g., video of a man opening a bottle) and meaningless videos that preserved the presence of motion, human bodies and faces (low-level control). Participants also completed a spatial working memory task, in which they had to remember four (easy) or eight (hard) locations within a 3x4 grid (Fedorenko et al., 2011). We localized bilateral prefrontal and lateral temporal language regions in each subject using leave-one-run-out analysis (top 5% vertices ASL>Control). Inferior frontal and lateral temporal areas exhibited highly selective responses to language relative to both action and low-level control conditions (ps<.005). Responses to classifier and non-classifier constructions and for SVO and OSV word orders were equally language-selective (ps>.05). Next, we localized fronto-parietal working-memory regions in individual participants (working memory hard>easy, top 5% vertices). These regions did not respond preferentially to language compared to action control or low-level control (ps>.2). In whole cortex maps, ASL sentences activated a bilateral fronto-temporal network which was slightly left-biased on average, whereas action videos activated posterior superior temporal and occipito-temporal areas, predominantly in the right hemisphere. The working memory task activated fronto-parietal areas previously observed in hearing people. In deaf signers, the fronto-temporal language network is highly selective for visuo-manual language relative to action observation and working memory. Even when the meanings of the observed actions and sentences are matched, fronto-temporal networks exhibit a robust preference for sentences. Selectivity for visuo-manual language in this network is invariant to iconicity, i.e., classifier vs. non-classifier constructions. Language dissociates from action semantics and working memory networks irrespective of language modality.
Topic Areas: Signed Language and Gesture, Meaning: Lexical Semantics