Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
Context-driven word production suppresses high-gamma-band activity in the middle temporal cortex
Poster A3 in Poster Session A, Tuesday, October 24, 10:15 am - 12:00 pm CEST, Espace Vieux-Port
Andrey Zyryanov1,2, Semen Kudryavtsev1, Irina Makarova3, Anastasia Matuschak3, Elizaveta Gordeyeva4, Nikita Utyashev4, Alexandra Balatskaya4, Madina Odeniyazova4, Nikita Pedyash4, Alexey Dimertsev4, Andrey Zuev4, Oleg Bronov4, Tommaso Fedele1, Olga Dragoy1,5, Svetlana Malyutina1; 1HSE University, Moscow, Russia, 2University of Tübingen, Germany, 3HSE University, Nizhniy Novgorod, Russia, 4National Medical and Surgical Center named after N. I. Pirogov, Moscow, Russia, 5Institute of Linguistics, Russian Academy of Sciences, Moscow, Russia
Long-standing evidence that semantically related distractors impede object naming (semantic interference) suggests that multiple lexical-semantic representations coactivate during word production. The neural dynamics underlying this coactivation, however, remain obscure. While semantic interference consistently modulates neural activity in the left posterior temporal cortex, it does so in inconsistent directions ranging from suppression (Dirani & Pylkkänen, 2021; Piai et al., 2014; de Zubicaray et al., 2013) to enhancement (Gauvin et al., 2021). Furthermore, it remains unclear how the findings from naming with distractors translate into more naturalistic, context-driven word production. We investigated the spatiotemporal dynamics of context-driven word production using stereotactic EEG recordings in 10 subjects with refractory epilepsy while they performed a sentence completion task. Subjects listened to a sentence with the final word omitted and responded with a single context-appropriate noun. The sentences varied in response agreement (mean 0.48, range 0.07–0.98 in 141 healthy subjects). Lower agreement predicted longer response latencies in correct trials (t = –13.9, p < 0.001), reflecting semantic interference. Among a total of 720 artifact-free bipolar channels, the task robustly modulated signal amplitude in the high-gamma band (HGA, 70–150 Hz) in 81 channels (as measured using a time-resolved linear regression model of evoked responses, test-set correlation between true and predicted HGA 0.10–0.48, all p < 0.001). Hierarchical clustering of these channels based on average HGA in sentence- and response-locked time windows separated them into three spatiotemporal clusters. Two clusters (34 and 17 channels) predominantly spanned bilateral superior temporal, left inferior frontal, and bilateral precentral gyri. They showed well-established auditory and motor HGA enhancement time-locked to sentence comprehension and to response articulation, respectively. The third cluster (30 channels) was located mainly in bilateral middle temporal gyri (MTG) and showed a suppression of HGA around sentence offset. This suppression showed consistent single-trial temporal dynamics: It built up throughout sentence comprehension, ceased immediately before response onset, and was followed by HGA enhancement at the start of articulation in some channels. Our findings suggest that suppressed MTG activity previously reported by neuroimaging studies of semantic interference in naming also subserves context-driven word production. These findings are best explained by the models of word production where lexical representations interact via inhibitory connections over the course of word selection (e.g., Howard, 2006).
Topic Areas: Language Production, Control, Selection, and Executive Processes