EEG-Based Sleep Stage Classification — Meta-Analysis Summary & Reference Implementation

This dataset accompanies a comparative meta-analysis of feature-extraction methods used in EEG-based sleep-stage classification. It provides the underlying data, summary tables and a reference implementation supporting the reported results. Contents: (1) a study-level decision sheet with per-study Newcastle-Ottawa Scale, Triad ABC and MIDO-Q (PROBAST+AI) quality scores and inclusion decisions; (2) per-paper strict and four-source per-stage F1 extractions; (3) publication-ready meta-analysis summary tables covering pooled effect sizes, heterogeneity, stage-level performance, diagnostic accuracy and publication-bias diagnostics; (4) computational pseudocode for the core algorithms (DerSimonian-Laird random-effects pooling, Egger regression, three-step screening); and (5) a Python reference implementation that reproduces the per-stage pooled F1 values. Methods: literature was retrieved from multiple academic databases (PubMed, Scopus, IEEE Xplore, OpenAlex and others), screened with a two-stage hybrid filter, and quality-assessed with a two-layer framework (Newcastle-Ottawa Scale >= 6 admission gate, then PROBAST+AI / Triad ABC / MIDO-Q). See README.md for full details. License: Creative Commons Attribution 4.0 International (CC BY 4.0).