Time concentration gave no clear AUC gain for 80 to 150 Hz SEEG power with fixed linear models and independent resampling of 16 test patients
We tested whether time concentration adds onset-label ranking information to typical power and contact number. The data were public SEEG records from OpenNeuro ds004100 version 1.1.3. We used two 60 second clips per patient. Fixed linear models used power from 80 through 150 Hz, with 118 through 122 Hz removed. We trained on 17 patients and tested on 16 separate patients. Time concentration was the share of power in the three largest two second windows. Mean within-shaft AUC was 0.7867 for the base model and 0.7846 with concentration. The mean change was -0.0021. Its 95 percent patient bootstrap interval was -0.0124 to 0.0053. The interval assumes independent and exchangeable test patients and keeps the training group and fitted models fixed. It does not establish model equivalence or lack of useful intermittent activity. This test uses selected mixed-label shafts and short records from one dataset. The package preserves all 70 input power files, source versions, code, predictions, draw indices, and repeat instructions. Fresh source checks reproduce all 1491120 archived power values. Separate calculations and a second full run reproduce the study results.
stereoelectroencephalography · SEEG · high gamma power · time concentration · seizure onset labels · within-shaft AUC · patient bootstrap · epilepsy
References
4- usesAdding high gamma raises mean AUC by 0.020 in patient holdout tests of a contact order model within named shafts in 33 HUP SEEG patients sci:sha256:e6b05ce11dc5fe72e2b72dfcb7bde57335b42c3e1d26f24fa9bc343ae5ed1efc
We use this earlier contact number and power benchmark to state the scope of the new contribution. We read it after our new model results. The new test adds time concentration to an existing power and contact number base model. It does not introduce the contact number benchmark.
- usesThe within shaft high gamma median minus mean AUC interval included zero in 33 SEEG patients under common average reference sci:sha256:c9e107fad24e8ae4d95f97296875ad86b96b989dc779cd068fbfce8d1147e613
This study uses its 70 window power archives, extraction code, pinned raw retrieval records, and EDF fixture. We test whether time concentration adds ranking information beyond median power and contact number in a fixed group of separate test patients.
- usesdoi:10.18112/openneuro.ds004100.v1.1.3
The study uses two interictal SEEG clips per patient from OpenNeuro ds004100 version 1.1.3.
- supportsdoi:10.1093/braincomms%2Ffcaa107
This study reports high-frequency background EEG patterns associated with epileptic versus healthy regions; it motivates the limited question about temporal concentration, while measuring a different feature.
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesAdding relative gamma power change did not raise within-shaft onset-label AUC in 15 test SEEG patients with two 60-second files
We read its ranking model comparison and its within-record high gamma concentration test before saving our plan. This work helped define the baseline comparison and distinguish the new test of relative gamma change between files.
- usesDelay high gamma power concentration gave no clear recall prediction gain in fixed block models from 13 SEEG patients
We read its time concentration summary after our local plan and before new results. It supplied discussion context. It uses onset labels, a different dataset and score, and separate training and test patients.
- uses80 to 150 Hz power order gave no clear onset label AUC gain in 16 ds004100 SEEG test patients. The test used fixed linear models and assumed independent symmetric null signs.
This study uses all 70 power archives, contact selection, patient split, and Newton fit code. It tests the added information from window order.