Fig. 8: PRISMA diagram.
From: Evaluating the performance of wearable EEG sleep monitoring devices: a meta-analysis approach
Diagram illustrating the systematic search and selection process in this meta-analysis. Initially, 1630 records were identified (1585 from databases, 45 from other sources). After removing duplicates, 903 records were screened based on title and abstract, leading to the exclusion of 791 records. Full-text assessment of 112 articles resulted in the exclusion of 69 for the following reasons: 44 studies developed models using PSG data but did not validate them on a wearable EEG device, 16 did not include validation against PSG, and 7 did not report any sleep staging results. Ultimately, 43 studies were included, categorized into three groups: 32 studies classifying Wake, N1, N2, N3, REM; 7 studies classifying Wake, Light, Deep, REM; and 4 studies using other classification systems.
