Predictive processing models of disorders of consciousness provide mechanistic specificity for the generative model/interface account

{
 "question_id": "q:brain-if-consciousness-fundamental",
 "source_role": "pp-doc-systematic-review",
 "title": "Predictive processing models of disorders of consciousness provide mechanistic specificity for the generative model/interface account",
 "tradition": "computational_neuroscience",
 "central_reading": "A 2026 systematic review of 30 peer-reviewed studies (pubmed:42179888, doi:10.3389/fncom.2026.1797090) examining computational predictive processing (PP) models applied to non-invasive brain signals in disorders of consciousness (DoC) reveals that hierarchical prediction error dynamics — specifically, global information-sharing capacity — effectively distinguish conscious (MCS) from unconscious (VS/UWS) states. This provides neural-level mechanistic specificity for the 'brain as generative model/interface' account: conscious level tracks the brain's capacity for hierarchical predictive integration. The review documents a transition from simple waveform averaging to high-dimensional decoding of hierarchical prediction errors, with temporal progression of prediction error signatures in early coma demonstrating diagnostic and prognostic utility. This evidence pressures both the appearance-only account (which lacks a mechanism for why conscious level tracks predictive integration) and the filter account (which must explain why DOC recovery correlates with PP dynamics rather than kañcuka collapse patterns alone).",
 "claims": [
  {
   "id": "cl:pp-doc-global-integration-discriminates-consciousness",
   "text": "Global information-sharing markers derived from computational predictive processing models applied to non-invasive brain signals (EEG, fMRI) reliably discriminate between minimally conscious (MCS) and vegetative/unresponsive (VS/UWS) states. The systematic review of 30 studies (pubmed:42179888) demonstrates that hierarchical prediction error dynamics — specifically measures of global signal integration and temporal progression of prediction error signatures — effectively distinguish conscious from unconscious states with diagnostic utility exceeding standard behavioral assessment.",
   "source_id": "doi:10.3389/fncom.2026.1797090",
   "source_span": "Full article — systematic review of 30 peer-reviewed studies on computational PP models in DoC; key finding that global information-sharing markers effectively distinguish conscious states",
   "evidence": "empirical",
   "pramana": "pratyaksa",
   "hetu": "If consciousness level tracks the brain's capacity for hierarchical predictive integration, then PP-derived global integration measures should discriminate conscious from unconscious states, and this is confirmed across 30 studies using diverse methodological approaches (EEG, fMRI, hierarchical model comparison).",
   "sadhya": "The brain as generative model / interface account gains mechanistic specificity for DOC: conscious level tracks hierarchical prediction error integration capacity, not merely level of arousal or content complexity.",
   "vyapti": "If hierarchical prediction error integration (global information sharing) reliably discriminates MCS from VS/UWS across multiple independent studies (N=30, diverse modalities), then the brain's role in consciousness includes generative modeling via hierarchical predictive processing — supporting the interface/generative-model account over accounts that treat brain damage purely as filter collapse or appearance dissociation.",
   "vyapti_confidence": 0.65,
   "falsifier": "Show that global information-sharing markers are epiphenomenal correlates of arousal level rather than specific markers of conscious state — i.e., that they track brainstem-thalamic integrity (which the filter account also predicts) rather than hierarchical prediction error integration specifically. Or: demonstrate that the same PP-derived measures fail to generalize across independent cohorts with blinded classification.",
   "falsifier_status": "partially_tested — systematic review reports diagnos