Task-based metacognition and insight in schizophrenia spectrum disorders: A systematic review and meta-analysis

Impaired insight is prevalent among individuals with psychosis spectrum disorders and is associated with treatment non-adherence, symptom relapse, and functional decline. Metacognition has been proposed as a mechanism underlying insight impairment, but it is defined and measured in different ways, ranging from experimental tasks assessing confidence in perceptual decisions to broader self-reflective capacities. Whether behavioral metacognitive tasks are associated with insight remains unclear. We therefore conducted a systematic review to examine the association between behavioral metacognitive task performance and insight in schizophrenia spectrum disorders. We searched the PsycINFO, PubMed, Embase, and Scopus databases in accordance with PRISMA guidelines. Studies were included if they assessed adults with psychosis spectrum disorders using a behavioral metacognitive task and a validated insight measure. A random-effects meta-analysis with subgroup analyses by task type was conducted on studies reporting extractable correlations. Seventeen studies were included; only 6 (35.3%) reported significant associations. Across the 12 studies eligible for pooling, the overall metacognition-insight correlation was small but significant (r = 0.18, 95% CI [0.05, 0.31]; I² = 60.2%). This effect varied substantially by task type: perceptual, probabilistic, metamemory, and social tasks showed no association with insight, whereas self-reflection paradigms (r = 0.46) and Meta-Wisconsin Card Sorting Test variants (r = 0.33) showed moderate associations. Task type accounted for approximately 89% of between-study heterogeneity. These results suggest that low-level, domain-specific metacognitive tasks do not capture processes relevant to insight, supporting hierarchical models of metacognition and the need for studies employing self-referential metacognitive tasks.