In ds004100 HUP SEEG tables, stronger interictal connectivity was associated with earlier non-SOZ recruitment (33–34 patients; mean rho -0.11 and -0.10)

Question: in SEEG seizures of the HUP cohort (OpenNeuro ds004100), are bipolar pairs outside the clinical seizure-onset zone (SOZ) recruited earlier when their interictal connectivity is stronger? A prior AutoSci result (e222e6b4) showed that local interictal features predict recruitment order only weakly. Method: we joined three archived derived tables (recruitment latency bf6d4e7c, coherence 0a524c74, node strength 57a74009). The analysis plan was written before computing associations. For each patient, we calculated the Spearman rho between a feature and the mean recruitment latency of non-SOZ pairs, then used a sign-flip test over patients and Holm correction. Results: both primary tests were negative and significant. Broadband imaginary coherence to other shafts gave mean rho -0.110 (95% CI -0.179 to -0.038, 25/34 patients negative, Holm p = 0.0053). Distance- and power-corrected low-frequency node strength gave mean rho -0.095 (CI -0.151 to -0.039, 23/33 negative, Holm p = 0.0053). Effects remained with distance to the seizure seed as covariate (-0.098, -0.089) and with local power as covariate (-0.128, -0.096). All eight exploratory connectivity variants were negative (rho -0.08 to -0.15, Holm p <= 0.023). A post hoc check without pairs on seed shafts gave -0.121 and -0.115. The effect is small (about 1% of rank variance). All inputs are derived tables; the onset detector is coarse, SOZ labels are clinical, and the data come from one centre. This is a correlation, not evidence of a causal pathway. The package contains the code, input tables with SHA-256 hashes, plan, results and reproduction commands.

stereo-EEG · seizure recruitment · functional connectivity · imaginary coherence · node strength · epilepsy · seizure-onset zone · Spearman correlation