Search for a command to run...
We describe a class of parallel neural processing that operates below awareness, governs execution, automaticity, and is completely inaccessible to introspection. We propose a three-tier model, conscious processing (globally accessible, reportable experience implemented in four corticothalamic loops), subconscious processing (performance execution, inaccessible to introspection), and unconscious processing (no reportable experience and no directed voluntary execution, as in deep surgical anesthesia or coma). We identify four systems as the subconscious infrastructure for skilled execution and flow states, three active execution systems and one inhibitory gate whose suppression enables them, specifically the cerebellum (predictive forward modeling), basal ganglia (procedural habit execution), hippocampus (episodic encoding and execution gate), and locus coeruleus (brainstem; active sensory gating). During flow, the cerebellum, basal ganglia, and locus coeruleus are active while the hippocampus is suppressed; that suppression removes competing claims on behavioral control, enabling uninterrupted execution. Their collective state accounts for the full phenomenology of flow, including its sensory narrowing, automaticity and degraded episodic memory. The framework accounts for three phenomena that existing models do not fully account for: the learning-to-execution pipeline, in which skills consolidated through conscious practice in the cerebellum and basal ganglia, are executed without conscious supervision; performance interference, in which conscious reactivation during skilled execution degrades enhanced performance, because conscious monitoring interrupts systems that were executing without it; and flow states, in which the phenomenological reports of absorption, effortlessness, and loss of self-monitoring reflect the absence of conscious loop firing. The framework generates eleven falsifiable claims across six experimental tests distinguishing it from existing models of consciousness, expertise, and automaticity. These include an inverse relationship between conscious loop activity and subconscious system activity during flow, a double dissociation between conscious and subconscious contributions to learning versus execution, and a prediction that cerebellar and striatal activity during insight incubation predicts both insight occurrence and solution confidence at the moment of the aha report.