The Interoceptive–Metacognitive Architecture of Fatigue: Partial Mediation and Resilience Buffering

The problem here to be studied is whether interoceptive predictive processing and metacognitive appraisal jointly explain persistent fatigue and everyday impairment. Contemporary theory suggests that sustained interoceptive prediction error undermines allostatic self-efficacy, with fatigue operating as a metacognitive signal that expected costs outweigh benefits.

Methods

We conducted a preregistered, cross-sectional study in a UA/EU higher-education cohort (N = 300, age 18–35). Participants completed MAIA-2 (interoceptive sensibility), MCQ-30 (metacognitive beliefs), BIS/BAS (approach– avoidance), FSS-9 (fatigue), IPF (psychosocial functioning; coded higher = greater impairment), and ARS-30 (academic resilience). An ESEM measurement model (with multi-group invariance by sex and region) preceded a structural model estimating direct paths to IPF and an indirect path via FSS; ARS-30 was tested as a moderator. Robust maximum likelihood was used with FIML for missing data; indirect effects employed bootstrap CIs.

Results

A three-factor ESEM (Interoceptive Sensibility, Metacognitive Beliefs, Approach–Avoidance) fit well and showed metric—plus partial scalar— invariance (sex, UA/EU). In structural analyses, metacognitive threat/uncontrollability was associated with higher fatigue and greater impairment, whereas interoceptive sensibility showed protective associations (lower fatigue, better functioning). Fatigue carried a substantive portion of metacognitive influence on impairment (partial mediation), with residual direct effects intact (complementary mediation). Approach–avoidance tendencies contributed modestly. Academic resilience attenuated the metacognitive threat → fatigue pathway. Core paths were invariant across sex and region.

Conclusions

Findings are consonant with predictive-processing and cost– benefit control accounts: persistent expectation–evidence mismatch amplifies perceived costs and degrades confidence in bodily control, with fatigue as the gatekeeper to down-regulated action. The work establishes a measurement-tight platform for longitudinal and post-exertional paradigms incorporating behavioural interoceptive accuracy/insight to quantify prediction error directly.