In ds005169, all 51 one-sided visual hallucinations evoked by SEEG stimulation were contralateral, but only 1 of 112 stimulations evoked a complex hallucination

Question: OpenNeuro ds005169 (v1.0.0) has symptom labels for 43.2/50 Hz SEEG stimulations that caused visual effects. Do stimulations that cause complex hallucinations (shapes, people) need more current, or are they more anterior, than stimulations that cause elementary flashes? Method: we used only the metadata (events.tsv, electrodes.tsv in ScanRAS mm, SHA-256 of all 563 files). We wrote the plan (PLAN.md, hashed) before we examined all labels. Test: within-patient weighted difference with 20000 label permutations inside each patient, Holm correction. Result: of 112 events in 20 patients, 95 were elementary only, 1 was complex and 16 had neither flag. Thus the planned tests H1 and H2 could not be done. A hashed plan amendment then tested whether the labels agree with retinotopy. A1: all 51 one-sided events (14 patients, 100% in each patient) were in the field contralateral to the stimulated hemisphere (sign-flip p = 1.5e-4, Holm p = 4.5e-4; event binomial p = 8.9e-16). A2 (peripheral vs central, y) was not testable: no patient had both classes. A3 (upper vs lower quadrant, z): -6.7 mm in 2 patients, Holm p = 0.20. Secondary: white-matter vs grey-matter sites did not differ in current (-0.04 mA, 7 patients, p = 0.79). Median current was 0.5 mA. Limits: only stimulations with an effect are in the files; the current is the clinical value, not a threshold; ScanRAS is native space; clinicians who could know the stimulated side made the labels. Conclusion: the hemifield labels are valid for symptom-network maps, but this dataset cannot test the elementary-complex hierarchy.

stereo-EEG · visual hallucinations · cortical stimulation · retinotopy · OpenNeuro · visual field · clinical stimulation current