Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Cerebellar Contribution in Learning-Related and Error-Related Activity during Artificial Grammar Learning
Poster Session D, Saturday, September 13, 5:00 - 6:30 pm, Field House
Zhuoya Liu1, Manson Cheuk-Man Fong1, Matthew King-Hang Ma1, William Shiyuan Wang1; 1Research Centre for Language, Cognition, and Neuroscience, The Hong Kong Polytechnic University
During rule acquisition automation, the posterior cerebellum shows reduced learning-related activity, reflecting its plasticity (Balsters & Ramnani, 2011, J. Neurosci.). However, changes in error-related cognitive activity remain less understood. In the cerebrum, the activation was suppressed by the automation of familiar syntactic knowledge but enhanced by repetition of unfamiliar knowledge (Weber et al., 2016, J. Neurosci.). As Crus I/II of the posterior lobe is associated with advanced cognitive processing, we examined if these regions support the learning- and error-related activities differently. In this study, we used longitudinal fMRI to investigate the cerebellum’s role in artificial grammar learning. In accordance with Balsters and Ramnani (2011), we distinguished learning-related (i.e., correct processing) and error-related (i.e., incorrect processing) activities based on whether participants successfully applied an internal model of grammar knowledge. This study has two objectives: (1) To determine if Crus I/II were engaged in correct and incorrect processing differently, we compared these activities in the early and late stages and correlated activation differences with participants’ reaction times (RT) to account for the individual differences. (2) To assess whether Crus I/II exhibit different plasticity in correct and incorrect processing, we compared changes in the two activities between stages and examined the interaction between activity type and stage. Seventeen Cantonese-speaking participants (mean age = 20.9, SD = 2.51) completed one fMRI session before (pre-fMRI) and one after training (post-fMRI), following the first behavioral learning session. Inside the scanner, participants judged the grammaticality of the sentence within 3.2 s. The pre- and post-fMRI sessions shared the same paradigms but used different stimuli. Behaviourally, participants demonstrated an overall improvement in grammar processing following training, with significantly higher d-prime scores in the post-fMRI than pre-fMRI, t(16) = 4.69, p < .001. However, there was no significant difference in RT, t(16) = 0.89, p = .384. Regarding the difference between correct vs. incorrect judgment, the bilateral Crus I/II activities were higher for correct than incorrect grammar judgment in both sessions (pre and post). However, while this contrast was not associated with RT in the pre-fMRI session, a significantly negative correlation was revealed in the left Crus I in the post-fMRI session. These findings suggested that while the bilateral Crus I/II were relatively disengaged in incorrect grammar judgment, the contrast between correct vs. incorrect judgment was especially prominent in the left Crus I for participants with rapid grammaticality judgment. Regarding plasticity (pre–post differences), the activity in bilateral Crus I/II decreased from early to late stage for correct judgment, which is consistent with Balsters and Ramnani (2011); however, no significant change was observed in incorrect judgment activity. There was no interaction effect between activity type and stage. Overall, these findings suggest that the Crus I/II play a greater role in learning-related activities during grammar processing, with potential hemispheric differences that link to individual differences. This research was supported by Faculty Reserve awarded to M.C-M.F. by the Faculty of Humanties, HKPolyU, and a postgraduate studentship awarded to Z. L.
Topic Areas: Language Development/Acquisition, Syntax and Combinatorial Semantics