Performance Reassessment of a Real‐time Seizure‐detection Algorithm on Long ECoG Series
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
PURPOSE: Automated seizure detection and blockage requires highly sensitive and specific algorithms. This study reassessed the performance of an algorithm by using a more extensive database than that of a previous study and its suitability for safety/efficacy closed-loop studies to block seizures in humans. METHODS: Up to eight electrocorticography (EcoG) channels from 15 subjects were analyzed off-line. Visual and computerized analyses of the data were performed by different (blinded) investigators. Independent visual analysis also was performed for clinical seizures and for detections identified only by the algorithm. The following were computed: FP rate, number of FNs, latency to automated detection, warning rate for clinical onset and warning times, seizure duration/intensity, and interrater agreement. Adaptations to improve performance were performed when indicated. RESULTS: Fourteen subjects met inclusion criteria. Generic algorithm "relative sensitivity" for clinical seizures was 100%; two undetected subclinical seizures and two unclassified seizures were captured after adaptation. FPs/day were zero in seven and fewer than one in an additional three subjects. Adaptations for four subjects with greater than 1 FP/day (7.7-66.6/day) reduced the rate to 0 in one subject and to fewer than five FP/day (1.7-4.2/day) in the remainder. Generic latency to automated detection was <5 s in eight of 13 subjects, and in 12 of 13 after adaptation. Detections provided warning of clinical onset in three of four subjects in whom it always followed electrographic onset, and in four of four after adaptation. Interrater agreement was low for FPs and EDs. CONCLUSIONS: The generic algorithm demonstrated high sensitivity, specificity, and speed, characteristics further enhanced by adaptation. This algorithm is well suited for seizure detection/warning and use in safety/efficacy closed-loop therapy studies.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it