Room: Exhibit Hall | Forum 8
Purpose: Language task-based fMRI (tb-fMRI) has been employed as a clinical routine for pre-operative language mapping. Standard analyzing method can reliably localize major language centers such as Brocaâ€™s Area (BA) and Wernickeâ€™s Area (WA), however, do not elucidate how they functionally interact with each other to form comprehension. Our main objective was to investigate the language network of healthy and diseased subjects from tb-fMRI data by employing graph theoretical techniques.
Methods: Nine healthy right-handed adults performed a verbal fluency fMRI task using verb generation in response to auditory nouns. One patient with low-grade glioma with no language impairment (suspect reorganization) performed a category and letter fMRI task respectively before and after the surgery. We developed an innovative pipeline to analyze/model the brain under both conditions. First, we extracted time series from each voxel of the fMRI images to fit with the task model. The ones fit well (p < 0.001) will be kept as the network nodes. Then, these nodes were grouped together as modules according to their anatomical locations. The intra and inter-modular links was inferred by employing statistical inference method.
Results: Our results show a robust common language network across healthy individuals with quantified modular link strengths, which unveiled the crucial role of left ventral part of the Pre-Motor Area (Pre-MA) played in language. This has not been addressed much in the current neuroscience literature. In contrast to the intact brain, where BA and WA connect to each other by a direct link, the diseased brain possibly underwent reorganization presents an indirect link through pre-Supplementary Motor Area (pre-SMA).
Conclusion: This new approach allows one to understand the functional connectivity between language-related areas activated during a specific task, should lead to useful clinical applications, for example, to provide information about the individualized organization of language networks to guide neurosurgery.
IM/TH- Image Analysis (Single modality or Multi-modality): Computer-aided decision support systems (detection, diagnosis, risk prediction, staging, treatment response assessment/monitoring, prognosis prediction)