Poster E98, Monday, March 27, 2:30 – 4:30 pm, Pacific Concourse
fMRI Task Comparison for Pre-surgical Language Mapping in Neurosurgical Patients
Prashin Unadkat1, Luca Fumagalli1, Laura Rigolo1, Alexandra Golby1, Yanmei Tie1; 1Brigham and Women's Hospital, Harvard Medical School
Functional MRI (fMRI) based on blood-oxygen-level dependent (BOLD) signal is a brain mapping technique that has been used in pre-surgical planning of patients with brain lesions to understand the relationship between the eloquent cortex and brain lesion. Language function is particularly difficult to map due to the complex nature of language network. To compare the different language tasks used at our institution, we analyzed fMRI data from patients with brain tumors and epilepsy who underwent all three language paradigms: antonym generation (AG), sentence completion (SC), and auditory naming (AN). A total of 51 patients met these criteria and were included in this study. We calculated the Dice coefficient to evaluate the degree of overlap between the language maps for all three tasks using the same t-score that was utilized for clinical interpretation. The Dice coefficients (mean +/- SD) for the AG and SC, AG and AN, and SC and AN tasks were 0.291 +/- 0.132, 0.24 +/- 0.111, and 0.304 +/- 0.122 respectively. These results showed fair overlap between the t-maps for all three paradigms, suggesting that individual tasks activate different cortical areas for language function. Therefore, multiple language tasks maybe required to map all cortical areas related to language to increase the sensitivity of language fMRI for pre-surgical planning. Further research is needed to assess the effect of patients’ task performance and lesion (location, tumor grade and size) on the resulting language maps so that these factors can be taken into account when optimizing clinical language fMRI.
Topic Area: METHODS: Neuroimaging