gms | German Medical Science

Objective: Non-invasive assessment of hemispheric language dominance by functional MRI (fMRI) is of great clinical interest e.g. for planning of epilepsy or brain tumour surgery in potentially language-eloquent localization, but still remains challenging. Clustered sparse acquisition can improve language mapping by avoiding acoustic contamination by scanner noise as well as by minimizing task-related head movement artifacts, offering however a relatively short scanning time compared to conventional (i.e., single volume) sparse acquisition. We, here, set out to compare the potential of clustered versus standard sparse fMRI acquisition to detect hemispheric language dominance using different regions of interest (ROI).

Methods: 18 healthy, right-handed native speakers of German (m=4) underwent fMRI (Philips Ingenia 3T) during a picture naming task. Sparse acquisition acquired one full volume of the brain every 12s resulting in a silent period of 9.2s (TE: 30ms, FA: 90°, 36 slices, voxel size: 3mm isotropic, 14 min per session). Clustered acquisition acquired three volumes each 12s (TR: 1.2s, 8.4s silence, 7min per session). Here, two slices were acquired simultaneously using the “Simultaneous Multislice Imaging” function of the scanner. Data were analysed using SPM12/Matlab. JuBrain Anatomy Toolbox was used to define the ROIs taken into concern, i.e., inferior frontal gyrus (STG), superior temporal gyrus (STG) and angular gyrus (ANG). AveLI was used to compute laterality indices.

Results: Both protocols revealed activation in the IFG, the ANG and the STG. The extent of activation was comparable between sparse and clustered acquisition. Spatial agreement as indicated by dice similarity was 54.6%. Highest AveLIs were found for clustered sparse acquisition when using the IFG or the IFG+ANG as ROIs (61% of subjects left-lateralized; median AveLI IFG = 0.27 / IFG+ANG = 0.17).

Conclusion: Clustered acquisition offers the advantages of sparse acquisition with halved examination time and shows substantial agreement regarding the task-related activations. Moreover, clustered acquisition seems to be comparatively well suited to detect hemispheric language dominance, especially when considering the IFG or a combination of the IFG and the ANG as ROIs.