Slide Slam E8
Selective disruption of sentence comprehension by transcranial alternating current stimulation over the left inferior frontal cortex
Shinri Ohta1, Wakana Oishi1; 1Kyushu University, Fukuoka, Japan
Previous neuroimaging studies have demonstrated that the cortical activity of different frequency bands tracked the construction of hierarchical structures, such as syllables, phrases, and sentences. Using transcranial alternating stimulation (tACS), which can modulate the specific frequency band of the cortical activity non-invasively, we examined whether the modulation of the cortical activity that corresponded to sentence structure construction changed sentence comprehension. We recruited 15 right-handed native speakers of Japanese (8 males, 21.9±0.8 years), who had no history of neurological or psychiatric diseases. The same participants were tested for both tACS and sham sessions. We used 96 Japanese sentences and 96 word strings (total 192 stimuli). Each sentence consisted of three noun phrases and one verb (e.g., Shimbun-no kisha-ga machigai-o naoshita, A newspaper reporter corrected a mistake), immediately followed by a question consisted of a subject and a verb (kisha-ga naoshita?, Did the reporter correct it?). Each word list consisted of four nouns or verbs (aratta tataita naoshita okutta, washed hit corrected sent). In the present experiment, we used a sentence comprehension task and a short-term memory task. In the sentence comprehension task, the participants judged whether the meaning of the sentence matched with the question, while in the short-term memory task, they judged the order of words. We used a double-blinded sham-controlled design. Stimulation was delivered using DC-Stimulator Plus (NeuroConn GmbH, Germany). The two electrodes were placed over F7 and Fp2 according to the International 10-20 EEG system, which were right above the left IFG and the right forehead, respectively. For the tACS, stimulation was given for 20 minutes (±2 mA, 0.5 Hz, 5 cm * 7 cm saline-soaked sponge electrodes, >10 kΩ). Sham stimulation, which controlled for the placebo effect, ramped up to ±2 mA over 10 s, remained at that level for 30 s, and ramped back down over 10 s. The participants felt the initial ramp-up event in the sham session, which is the most noticeable, without receiving an effective stimulation. Before and after the tACS and sham sessions, the participants performed the sentence comprehension and short-term memory tasks. The participants showed high accuracies (>90%) and short reaction times to comprehension questions (RTs, <1000ms). To consider the random variabilities of participants and stimuli, we analyzed the RTs by using a linear mixed-effect model (lme4 and lmerTest packages on R). We found that the model with the effects of Stimulation (baseline vs. tACS vs. Sham) and Task (sentence vs. memory) was significantly better than the simpler model without the effect of Stimulation (𝝌2(4)=18, p=0.001), suggesting the effect of tACS was significant. Moreover, the sham stimulation over the left IFG significantly decreased the RTs of the sentence comprehension task (t(2510)=−3.7, p=0.0002), indicating the learning effect. In contrast, the tACS over the left IFG did not show such effect (t(2505)=−1.7, p=0.08). In the present study, we demonstrated that the tACS over the left IFG interrupted the sentence comprehension task but not the short-term memory task, suggesting the causal relationship between the left IFG activation and sentence structure constructions.