Validation of a New AI Seizure Detection Algorithm for Pre-term Neonates, Term Neonates and Infants
Bibliographic record
Abstract
Abstract Seizures are frequent in the first year of life and they are associated with increased risk of long-term neurological disability. Accurate diagnosis of seizures in neonates and infants requires continuous electroencephalogram (EEG) monitoring with expert interpretation. While rapid detection of seizures allows timely management and may influence neurodevelopmental outcomes, it may not be feasible in settings without round-the-clock expert interpretation. We evaluated the real-world performance of Clarity (Ceribell Inc.), a novel algorithm for automated seizure detection from EEG of preterm neonates, term neonates, and infants in a validation dataset of patients of postnatal age 0-1 years and gestational age 22-42 weeks who underwent continuous EEG monitoring at three hospitals in the United States. We quantified algorithm performance using the following metrics: (1) area under the curve for classification of 10-second EEG segments as seizure or no seizure; (2) sensitivity and false positives in detecting seizure episodes; and (3) sensitivity, specificity and negative/positive predictive value in identifying recordings where seizure duration reached or exceeded predetermined thresholds. The validation dataset consisted of 756 EEG recordings with a cumulative duration of 5230 hours (median: 2.94 hours, interquartile interval: 1.46-12.62 hours) from 167 preterm neonates, 323 term neonates, and 266 infants. A total of 568 seizures occurred in 75 recordings, with a cumulative seizure duration of 15.13 hours. The algorithm achieved an area under the curve of 0.96. The algorithm was correct 99-100% of the time when ruling out seizures in a recording. It correctly detected 75-83% of seizure episodes and identified 89-94% of recordings containing seizures. Performance was equivalent between preterm neonates, term neonates, and infants and across the three participating hospitals. Overall, the Clarity algorithm demonstrated high performance in detecting and ruling out seizures from EEG data in preterm neonates, term neonates, and infants across three hospitals. This demonstrates real-world feasibility and high reliability of automatic seizure detection with Clarity for continuous EEG monitoring in the first year of life.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.008 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".