Article
fMRI resting state connectivity between language areas as defined by direct electrocortical stimulation in low grade glioma patients
Search Medline for
Authors
Published: | June 9, 2017 |
---|
Outline
Text
Objectives: The aims of this study were (1) to examine the extend of resting state functional connectivity between language regions of interest (ROI) as identified by intraoperative direct electrocortical stimulation (DES) and to compare these findings with negative control regions; (2) to determine whether Resting State Connectivity (RSC) can discriminate between DES-positive language regions and nearby DES-negative regions.
Methods: We conducted a retrospective study of 10 low-grade glioma patients with language positive DES regions on at least two separate cortical lobes. Language regions were identified by intraoperative DES. We compared seed to seed analysis of RSC between and within groups of positive and negative ROI. The latter regions were randomly chosen, both at close distance (10 mm, same gyrus) and farther away (> 10 mm, different gyrus).
Results: Median connectivity within the group of positive language ROIs (n=31) was significantly higher than within either the group of close- or far-negative controls, 0.332 [0.206-0.391] vs. 0.203 [0.049-0.291] and 0.215 [0.093-0.301] respectively (p<0.05). The median connectivity between positive language ROIs did not significantly differ from the median connectivity between the close-negative and positive language seeds; 0.332 [0.206-0.391] vs. 0.282 [0.156-0.375] (p>0.44), but it did differ from the median connectivity between the far-negative and positive language seeds; 0.332 [0.206-0.391] vs. 0.091 [0.028-0.230] (p<0.001).
Conclusion: On a group level, median RSC correlates with DES for mapping of surgically relevant language areas in patients with LGG. However, RSC does not seem able do discriminate between positive or negative language ROIs on a surgical relevant resolution without a-priori knowledge. Surgical application of resting state protocols for identification of language function should therefore be addressed with caution. However, the results do suggest a shared neuronal and functional basis between both techniques.