Subject-specific Functional ROIs Enhance Reliability in Language FMRI

Functional MRI can be used to identify individual language-sensitive brain regions in the setting of presurgical diagnostics to improve functional postoperative outcome. In this study, a proven language task was adapted into German and tested with regard to its effectiveness, robustness and reliability in a time frame appropriate for the clinical setting. In addition, two different analysis approaches were compared to address the problem of arbitrary statistical thresholds commonly used in the clinical routine to derive contrast maps.

On two different days, 24 healthy volunteers were examined in a 3T MRI, whereby the task was run twice in each session. The fMRI included two conditions in a block design, reading of sentences and reading of pronounceable nonword lists. We quantified brain activity by using subject-specific, functionally defined ROIs on the one hand and standardized, anatomically defined ROIs on the other. We then tested, whether the two different analyses indicated robust activation of language-sensitive brain regions, and whether effect sizes were reliable across sessions.

Subject-specific functional ROIs as well as anatomical ROIs led to significant positive effect sizes in the major language sensitive regions of the left hemisphere. However, subject-specific functional ROIs resulted in significantly larger effect sizes and a higher reliability in comparison to anatomical ROIs.

The choice of analysis method has a significant impact on the result. For paradigms with short measurement times and little signal change as common in clinical routine, it is highly recommended to use the subject-specific functional ROIs approach.