Errors or Adaptations? A Critical Review of Predictive Processing in Psychiatry

Predictive processing (PP) accounts often characterize post-traumatic stress disorder (PTSD) and depression as maladaptive and epistemically distorting, due to wide divergences between brain generated top-down models and bottom-up sensory inputs. This review questions that characterization. First, trauma survivors with PTSD or depression may lessen risk by overestimating threats, with hypervigilance sustaining a desirable gap between anticipated problems and harms that would otherwise occur. Second, PP psychiatric frameworks prescribe how trauma survivors ought to assess matters yet give limited justification, introducing tacit normative assumptions into neurocomputational models. This repeats in certain PP assessments of schizophrenia and psychosis, which presuppose Western concepts of self as normative neurocognitive ideals. Third, PP accounts claim that cognition evolved primarily for action, not veridical representation; but their notion of prediction error can tacitly invoke the concept of veridical representation. Fourth, PP defenders have asserted that depressive slow-downs follow from maladaptive brain-based regulatory models. However, physiological problems may instead make activity strenuous, such that slowing down is adaptive. This position advanced in this review is that atypical mental outlooks need not be epistemically distorted, and that mismatches between anticipatory models and outcomes—when they occur—can sometimes index adaptive success rather than failure.