Despite decades of research identifying neural correlates of consciousness, no consensus exists on which neural components are *necessary* for consciousness. We propose a four-loop architecture as the necessary neural substrate: (1) an ACC-mediodorsal thalamic loop detecting prediction errors and generating voluntary output drive; (2) a VTA-cortical dopaminergic loop sustaining the working memory substrate and computing prediction error magnitude; (3) a central lateral thalamocortical loop maintaining global integration; and (4) Layer 5 pyramidal recurrent loops gating sensory relay to global conscious access via apical-basal coincidence detection. We define consciousness as the unified operation of four functions: (1) detecting prediction errors and generating voluntary output drive; (2) sustaining the working memory substrate and computing prediction error magnitude; (3) maintaining global integration; and (4) gating sensory relay to global conscious access via apical-basal coincidence detection. The four loops are the necessary physical systems whose operation allows these functions to occur. The framework proposes that all four loops must synchronize within a 50-150ms window (varying with age and neural health); on this account, disrupting any one degrades rather than eliminating consciousness, while simultaneous multi-loop disruption eliminates consciousness reliably. No documented case exists of all four loop components simultaneously disrupted alongside preserved capacity for contingent, accurate reporting. The framework proposes an account of why anesthesia requires multiple pharmacological mechanisms, why partial disruption produces altered rather than absent consciousness, and why anesthetic agents targeting GABA-A, NMDA, and alpha-2 adrenergic receptors all reliably produce unconsciousness despite acting through distinct molecular mechanisms. It is also consistent with patient H.M. remaining fully conscious despite bilateral hippocampal ablation. The framework extends Thalamocortical Loop theory, corrects Global Neuronal Workspace theory's anatomical specification (consistent with preregistered adversarial collaboration evidence), and operationalizes Integrated Information Theory's integration criterion by replacing phi with a testable synchronization requirement, generating discriminating predictions that exceed what any prior single-mechanism theory produces.
Arthur Stewart (Sun,) studied this question.