32 patients with intracerebral tumours underwent nTMS and fMRI prior to surgery. The cortical hotspots (HS) and Centres of Gravity (CoG) of M1 of the hand, the foot and the tongue representation were determined for each localizer technique and controlled by DCS. Euclidean distances (ED) between nTMS/fMRI and the respective DCS HS/CoG’s were computed in 3D. Moreover, nTMS, fMRI and DCS maps were projected onto the cortical surface following normalization to a standard brain. After smoothing, overlap volumes between both nTMS and fMRI with DCS functional maps (voxels) were calculated.
CoG’s of the nTMS maps were located significantly closer to the respective DCS CoG’s as compared to fMRI (p < 0.001), mainly due to differences in the congruency of the hand (nTMS-DCS/fMRI-DCS: 9.7 ± 6.0/17.3 ± 9.6 mm) and the foot (nTMS-DCS/fMRI-DCS: 16.2 ± 20/21.7 ± 17 mm) representation. However, comparison of hotspots did not reveal statistically significant differences. Regarding the overlap of entire functional maps, nTMS showed a higher agreement with DCS as compared to fMRI (p < 0.01), mostly attributable to the somatotopic hand representation (nTMS-DCS/fMRI-DCS: 75 ± 17%/57 ± 34%; p < 0.01).
Comparison of both M1 CoG’s and volume overlaps suggest that nTMS agrees better with the gold standard of intraoperative DCS as compared to fMRI, especially regarding the hand and the foot representation and, thus, is recommendable for presurgical M1 mapping. However, both methods should be regarded as complementary, particularly in respect of the tongue representation which is less easy to depict by nTMS.