gms | German Medical Science

Objective: Accurate and reliable localization of brain regions associated with language processing is important to improve planning prior to neurosurgical interventions and to reduce the risk of postoperative neurological deficits. This is especially important for patients who suffer from a tumor in close vicinity of those areas. Due to the disease itself associated with aphasia or other language related impairment, it is often difficult for patients compared to healthy subjects to work on a standard task for the localization of the language areas, such as a semantic decision task (SDT) or a word generation task (WGT). Therefore, using fMRI techniques without active participation, we assessed a novel passive listening paradigm and compared it to SDT and WGT in a cohort of healthy subjects.

Methods: 23 healthy subjects participated in a test-retest fMRI study (two sessions) with three fMRI language paradigms: WGT, SDT and a new passive listening task, in the latter sentences were presented interrupted by blocks of white noise. Language specific BOLD activity was identified for all paradigms on individual and group level. Furthermore, to determine reliability of fMRI activation patterns intraclass-correlation coefficients (ICC) were assessed.

Results: The passive listening task resulted in reliable activation of a bilateral fronto-temporal language-related cluster including Broca’s and Wernicke’s area at single subject and group level (Session1: T=23.23; p<0.001; Session 02: T=21.31; p<0.001). SDT showed as expected activation of left frontal cluster encompassing Broca’s area (Session1: T=10.16; p<0.001, Session 2: T=8.90; p<0.001), and WGT further activated its right-hemispheric homologue (Session1: left T=12.83; p<0.001, right T=8.49; p<0.001; Session 2: left T=12.47; p<0.001, right T=7.88; p<0.001). Therefore, across all three tasks Broca’s area could be localized consistently. No significant difference in activation patterns was found across both sessions, supporting high test-retest reliability of cluster location (at p<0.05, FWE corrected), also at single subject level with ICCs of 0.54±0.21 and 0.52±0.23 (passive), 0.68±0.11 (SDT) and 0.49±0.26 (WGT).

Conclusion: The passive listening paradigm is suitable to reliably activate language areas, it is easily applicable even without active participation of the patient, and is therefore a helpful alternative to standard active paradigms for preoperative non-invasive localization of language areas.