Quantify the detection rate of covert consciousness in behaviorally unresponsive patients, compare yields across paradigms and settings, and evaluate prognostic associations.

Covert consciousness is a condition of preserved awareness after critical injury or illness that evades routine bedside detection. Reported rates of covert consciousness vary widely across studies, modalities, and settings.

Systematic searches of PubMed, Web of Science, and Scopus (January 2006–September 31, 2025). Inclusion: Studies of disorders of consciousness using pre-specified positivity rules for active or passive paradigms. Primary meta-analysis used each study’s main paradigm; additional paradigms were reserved for sensitivity analyses. Random-effects GLMMs estimated pooled detection overall and by paradigm/modality; pre-specified meta-regression tested moderators (analytic stringency, sessions per patient, task class, accuracy).

94 studies were identified (n=2,652). The pooled detection of covert consciousness was 32.8% (95% CI 28.7–37.3; I²=76.6%). Cognitive-motor dissociation (active task) was detected at 31.5% (95% CI 26.0–37.6; k=61). Covert cortical processing (passive paradigm) detected at 35.0% (95% CI 27.9–42.8; k=25). By modality, EEG (including brain-computer interfaces) exceeded fMRI (active task: 36.1%, 95% CI 28.1-45, k=37 vs 22.1%, 15.6–30.4, k=16). Detection rate was lower in acute settings in comparison to chronic (23.4%, 95% CI 16.6–31.9; k=14 vs 28.7–39.5; k=65). In meta-regression, greater analytic stringency (i.e. more conservative positivity rules) was associated with lower detection (β=−0.514, p<0.001), while more EEG sessions per patient predicted higher detection (β=0.858, p<0.001). In BCI studies, higher accuracy thresholds were linked to lower pooled detection (β=−1.119, p=0.023). On follow-up, covertly conscious patients had larger CRS-R gains (Hedges g=1.40, 95% CI 0.44–2.37; k=8).

Approximately one-third of behaviorally unresponsive patients demonstrate evidence of covert consciousness. Yield is higher with EEG, in mixed/chronic settings, and lower under stricter thresholds. Positivity is associated with improved functional gains. Results support standardized, multimodal, serial assessment with explicit thresholds for interpretation.