Distance controls remove DTI–CCEP node-degree coupling in 8 SEEG patients

We reassessed the association between DTI structural connectivity degree (SCD) and CCEP effective connectivity degree (ECD) per electrode contact in OpenNeuro ds005448. The source study reported a positive pooled association after adjustment for node proximity (13 patients: 5 ECoG, 8 SEEG; ECD coefficient 0.1256, p = 0.0054). We reproduced the pooled model (ECD coefficient 0.1253, SE 0.045, t = 2.78) and found a similar association in the 8 SEEG patients using the same linear proximity control (coefficient 0.126, 95% CI 0.065 to 0.187, t = 4.03). We then controlled for contact geometry using edge probabilities within 5 mm distance bins. With distance-expected SCD and ECD included, the SEEG ECD coefficient was 0.021 (95% CI -0.048 to 0.090, t = 0.59); the distance-expected SCD coefficient was 0.97 (t = 10.6). The median per-patient correlation of distance-residual degrees was 0.07 (Wilcoxon p = 0.38), and the combined permutation test gave p = 0.20. A distance-bin shuffle also reproduced the observed mean proximity-adjusted correlation (0.147 null mean; 0.149 observed mean). In these data, contact geometry explains the node-level degree coupling, and DTI degree did not identify CCEP hub contacts beyond geometry. A small true effect remains possible (SEEG residual-model 95% CI -0.034 to 0.101). The analysis used the authors’ derivative networks and did not re-detect N1 responses or rerun tractography. It included 8 SEEG patients with 44–101 nodes each; networks were binary and symmetric, and the geometry analysis used Euclidean distance and 5 mm bins.

DTI · CCEP · node degree · contact geometry · structural connectivity · effective connectivity · stereo-EEG · OpenNeuro