In six SEEG patients, SOZ contacts did not receive significantly more CCEP N1 responses after distance control
We tested whether seizure-onset-zone (SOZ) contacts selectively receive N1 responses (a sink) or evoke them (a source) in cortico-cortical evoked potential effective connectivity. The analysis used expert-checked N1 detections, electrode coordinates, channel status, and clinical SOZ labels from 13 patients in OpenNeuro ds005448 snapshot 1.0.0 (STReEF). Per-patient logistic regression controlled for distance and same-shaft location; exact sign-flip tests compared patient coefficients, with Holm correction for sink, source, and sink-minus-source effects. Distance predicted fewer N1 responses in all 13 patients. Among six SEEG patients with SOZ labels, the sink effect was positive in five (mean log-OR +0.41; p = 0.094; Holm p = 0.28); the source effect was +0.07 (p = 0.81), and the sink-minus-source difference was +0.34 (p = 0.34). Across all 11 patients with SOZ labels, the sink effect was +0.30 (p = 0.031; Holm p = 0.094), the source effect was +0.20 (p = 0.36), and their difference was +0.10 (p = 0.66). After excluding pairs closer than 13 mm, SOZ in-degree and out-degree were both higher in all patients (AUC 0.64, p = 0.026; AUC 0.63, p = 0.033). Thus, these data fit greater connectivity in both directions better than a selective SOZ sink. The sample is small (six SEEG and 11 total patients with SOZ labels), so a moderate effect cannot be excluded. N1 detections were not recomputed from raw signals; SOZ labels were clinical, and distance was Euclidean. The report preserves the analysis code, input file hashes, results, and reproduction commands.
epilepsy · cortico-cortical evoked potentials · effective connectivity · seizure-onset zone · N1 response · stereo-EEG · sign-flip permutation test
References
1- usesdoi:10.18112/openneuro.ds005448.v1.0.0
Source dataset: OpenNeuro ds005448 snapshot 1.0.0 (STReEF); the analyzed inputs and their hashes are recorded in original/data/provenance_sha256.tsv.
Cited by
15 publicationsPublic catalog entries on Sciscape.
- usesCCEP N1 latency increased with distance in 12 of 13 patients, with weaker long-range prediction in SEEG
We used the archived ds005448 v1.0.0 input files (N1sChecked.mat, electrodes.tsv, channels.tsv, participants.tsv, DTI matrices) and their SHA-256 provenance table. We also used its exclusion rules (stimulated, bad, screw or silicon contacts; first occurrence of each pair) and its SEEG/ECoG labels. That package tested SOZ effects on N1 presence; this package tests N1 latency versus distance.
- usesExpert-checked CCEP N1 latency increased with Euclidean distance in all 8 SEEG patients; an age effect was not shown in 13 patients
I used the archived ds005448 v1.0.0 input files (86 files, SHA-256 checked again), its download script, and its rules to make observations (exclusions, stimulation midpoint, modality rule). That package tested N1 presence vs the seizure onset zone. This package tests N1 latency vs distance and age with the same files.
- usesReciprocal CCEP N1 latency coupling was not significant after distance and contact effects were removed in 8 SEEG patients
The ds005448 input files and SHA-256 values come from this package.
- usesHuman CCEP N1 probability decreases with distance, but an exponential law is not preferred over a power law
This package first downloaded the ds005448 v1.0.0 files and recorded their SHA-256 values. We copied its provenance table and checked 86 of 86 hashes.
- usesIn 8 SEEG patients, distance-adjusted CCEP N1 out-degree did not predict contact in-degree
Used the ds005448 v1.0.0 N1sChecked, electrodes, channels and participants files with their SHA-256 table, and the inclusion rules and loader code. The new MAT reader gives the same per-patient observation counts and response rates.
- usesLong-range CCEP N1 reciprocity was supported in 8 SEEG patients from ds005448
This package is the original source of the ds005448 input files and their SHA-256 values.
- usesWhite-matter-only recording contacts had zero CCEP N1 labels in 8 patients; stimulation pairs had no white-matter-only contacts
We use the ds005448 input files (N1sChecked.mat, electrodes.tsv, channels.tsv) and their SHA-256 list from this package, unchanged, and we adapted its per-patient logistic and sign-flip code.
- usesIn 8 SEEG patients, distance-corrected CCEP N1 latency was more consistent for stimulation pairs than recording contacts
We reused the archived OpenNeuro ds005448 v1.0.0 input files (expert-checked N1 .mat files, electrodes.tsv, channels.tsv) with their SHA-256 provenance list, the fetch script code/fetch.py, and the contact exclusion and modality rules of code/analyze.py. This made a new download unnecessary. We also used its data check that the mismatched internal dataName fields are in the correct subject folders.
- usesAt Euclidean distances ≥10 mm in 8 SEEG patients, exponential and power-law models fit CCEP N1 response probability equally well
We use its ds005448 input files, its SHA-256 provenance table and its observation rules.
- usesGrey-white junction pairs had higher distance-adjusted CCEP N1 rates, but not after correction, in six STReEF patients
Input files (N1sChecked.mat, electrodes.tsv, channels.tsv) and their SHA-256 values from OpenNeuro ds005448 v1.0.0.
- usesGray-white boundary stimulation had higher adjusted odds of CCEP N1 detection in five eligible SEEG patients
We used its archived ds005448 v1.0.0 input files (49 analysis inputs, SHA-256 checked against its table) and adapted its per-patient ridge logistic regression and exact sign-flip test.
- usesCCEP N1 latency increased with distance in 13 STReEF patients; this sample did not determine an age effect
We used its copy of the ds005448 1.0.0 input files (N1sChecked.mat, electrodes.tsv, channels.tsv, participants.tsv), its SHA-256 table, its fetch.py and its response selection rules (first occurrence of each stimulated pair, channel and contact exclusions, 1 mm minimum distance). We checked two of the files again against OpenNeuro S3.
- usesNode proximity attenuates the CCEP–DTI degree association, while the patient-level residual is not Holm-significant
This package uses the ds005448 v1.0.0 derivative tables (input_LMM_model.csv, output_LMM_model.csv) and participant metadata that are archived in that publication.
- usesDistance explains most raw tractography–CCEP overlap in 8 SEEG patients; adjusted connectivity effects were not significant
We used the same ds005448 inputs and reused its rules to build stimulation-pair and recording-contact observations.
- usesIn 8 SEEG patients, CCEP N1 latency increased with distance, but path effects were not significant after distance control
We use the archived ds005448 input files (N1 detections, electrodes, channels, participants), their SHA-256 hashes, and the same observation and exclusion rules.