Abstracts

Rationale: A question in English most frequently starts with an interrogative word (e.g., WHAT). In Japanese, conversely, a question with the same meaning can begin or end with an interrogative, and native Japanese speakers commonly use both types of sentences. Our presurgical epilepsy evaluation includes intracranial measurement of high-gamma activity during an auditory sentence comprehension task. By asking patients questions beginning or ending with an interrogative, we determined the effect of word order on the spatiotemporal dynamics of high-gamma modulations (65-95 Hz) elicited by sentence comprehension. Methods: The study included 17 native Japanese speakers with drug-resistant focal epilepsy (age: 15-54 years) who underwent extraoperative ECoG recording in two tertiary epilepsy centers in Japan. Each patient was instructed to overtly answer each of the 96 auditory sentence questions of 1.8 s. Thereby, 48 trials consisted of [interrogative], [adverb or object], and [verb], whereas the other 48 consisted of (adverb or object), (verb), and (interrogative). The word frequency and total numbers of syllables within sentences were controlled between the two trial groups. By employing time-frequency analysis on ECoG signals, we compared the spatiotemporal dynamics of high-gamma modulations between the two trial groups. Results: The response time did not significantly differ between the two trial groups. The spatiotemporal dynamics of high-gamma modulations were grossly similar between the groups. High-gamma augmentation initially involved the bilateral superior-temporal gyri during sentence listening, subsequently involved the widespread temporal-frontal-parietal regions in the left hemisphere, and eventually involved the bilateral pre- and post-central gyri during overt responses. Left premotor-prefrontal high-gamma augmentation took place earlier during sentence listening in the trials ending with an interrogative than in those beginning with. Conversely, high-gamma augmentation following sentence offset involved larger extents of the left temporal-parietal regions in the trials beginning with an interrogative than in those ending with. Conclusions: The effect of word order in sentence questions on the large-scale neural dynamics during sentence comprehension may be modest. This observation is consistent with the notion that our brain mostly processes the meaning of an entire sentence question as a chunk rather than treating each word strictly in a piece-by-piece manner. The observed difference in high-gamma profiles between the trial groups has raised a hypothesis that a sentence beginning with a non-interrogative, concrete word may more rapidly elicit a top-down prediction process in the prefrontal-premotor region. Funding: Intramural Research Grant (28-4: Clinical Research for Diagnostic and Therapeutic Innovations in Developmental Disorders) from Neurological and Psychiatric Disorders.JSPS KAKENHI Grant Number 19K09494Grant-in-Aid for Scientific Research on Innovative Areas [17H05936, 18H05058] from MEXTNIH grant NS64033.