A Systematic Review and Meta-Analysis Evaluating the Clinical Impact and Accuracy of Artificial Intelligence in EEG for the Early Detection of Nonconvulsive Seizures
Bibliographic record
Abstract
Artificial intelligence-integrated electroencephalography (AI-EEG) has demonstrated promise in the early detection of nonconvulsive status epilepticus (NCSE), particularly in emergency and intensive care settings with limited access to trained EEG technologists. This review includes 20 studies, of which 12 were incorporated into a meta-analysis assessing the diagnostic accuracy of AI-EEG. The pooled sensitivity reached 95%, with a specificity of 83%. However, when the pretest probability of NCSE is 40%, false positives may occur in approximately one in seven patients. Commercial AI-EEG platforms have shown a reduction in unnecessary antiepileptic drug (AED) administration compared to clinical judgment alone. Four prospective cohort studies reported a 26% relative risk reduction (RR -0.26; 95% CI -0.50 to -0.02; p = .03) in unnecessary AED use. Additionally, AI-EEG shortened the median time to EEG acquisition in resource-limited settings-from 4.5 hours (IQR 3.2-6.8) to 2.1 hours (IQR 1.5-3.4). A sub-analysis from an industry-sponsored trial suggested potential benefits of AI-EEG in reducing morbidity and ICU length of stay, though evidence remains insufficient for definitive conclusions. Despite these advantages, rapid-deployment AI-EEG systems face challenges: lack of video integration makes it difficult to distinguish seizures from artifacts or behavioral events, and limited electrode coverage may miss central brain activity. Moreover, AI algorithms tend to overread sharp and spike activities compared to human interpretation. Further investigator-initiated studies are needed to evaluate the diagnostic yield of AI-EEG beyond its simplified setup, assess its true impact on patient outcomes, and determine its feasibility for large-scale clinical implementation. .
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 | 0.010 | 0.033 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.015 | 0.030 |
| Bibliometrics | 0.004 | 0.006 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 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".