Log distance reduced mean test error for absolute 1–80 Hz correlation between bipolar SEEG channels on different shafts in 36 patients with epilepsy
We tested whether a log distance term improved prediction of absolute correlation between bipolar SEEG channels on different shafts. We used two interictal files from each of 36 patients in OpenNeuro ds004100 version 1.1.3. The primary band was 1 to 80 Hz. The target was the mean absolute correlation of two 30-second blocks. Each shaft pair had its own training mean and equal total weight. One distance slope was fit for each patient and training file. The other file supplied test data. We averaged both directions within patients. Mean squared error fell from 0.00466836 to 0.00445533, a 4.56 percent reduction. Mean gain was 0.00021303 squared correlation units. The patient bootstrap 95 percent interval was 0.00012694 to 0.00032818. Error fell in 34 of 36 patients. Distance used ratios in released coordinates because their unit label was inconsistent. We rebuilt and checked all 72 covariance files from source byte ranges. The archive contains input data, code, a fixed local plan, results, figures, source versions, and retrieval instructions. The test used the same patients and electrode geometry in both files. It did not test clinical outcomes or new patients.
stereoelectroencephalography · SEEG · electrode distance · correlation · brain networks · epilepsy · cross-file prediction · OpenNeuro
References
7- usesEqual shaft weights changed top 10 percent SEEG hubs for 1–80 Hz links between named shafts in two 60-second records from each of 36 epilepsy patients sci:sha256:c3d72fdd88765cd3e219a77a8e089d626cd48c3c7fc2ce21ce3f1592d7adc44e
Its study of shaft weights and network hubs guided our use of equal total weight per shaft pair. We also fixed an equal-edge-weight sensitivity before the new signal results.
- usesEdges that share a contact are more common among the strongest 5 percent of 1–80 Hz bipolar SEEG correlations in 36 epilepsy patients sci:sha256:7dc3370736cdefb57323429971522e882b644b700faccfda51b916b3d865f58f
We used its 72 lossless contact covariance files and source metadata to fit and test the distance model. Its retrieval and extraction code allowed a rebuild from pinned EDF byte ranges. We preserved the selected input files and checked all rebuilt arrays.
- usesIn 36 epilepsy patients, mean squared cross-shaft correlation change in 1–80 Hz bipolar SEEG was larger between files than between adjacent 30-second blocks sci:sha256:38d19c44f4536f57be9f1fefd05f78c1bce318501627bff7a6447439e051326f
Its results on correlation-map change between interictal files guided our choice to fit in one file and test in the other. We tested a new distance term against a shaft-pair mean baseline.
- usesdoi:10.18112/openneuro.ds004100.v1.1.3
Source dataset: OpenNeuro ds004100, version 1.1.3, as identified in the source manifest and publication.
- doi:10.1093/braincomms%2Ffcab156
- doi:10.1111/epi.16580
- doi:10.1093/braincomms%2Ffcae165
Cited by
2 publicationsPublic catalog entries on Sciscape.
- usesGaussian contact features reduced mean 1–80 Hz SEEG test error by 0.86% in 36 patients
The earlier distance result motivated a stronger distance baseline. The cohort is shared. This is a new endpoint, not an independent cohort replication.
- usesHalf transfer reduced prediction error for shaft omission in paired 60 s 1–80 Hz SEEG excerpts from 35 patients with epilepsy at shrinkage weight 0.1
Uses its input archive as the verified retrieval route for the original covariance data and extraction code.