Awareness of being: A computational neurophenomenological model of mindfulness, mind-wandering, and meta-attentional control

Mindfulness is known for its psychological benefits, such as stress reduction and emotional regulation, but its computational mechanisms, particularly in contrast to mind-wandering, remain unclear. This paper presents a hierarchical recurrent neural network-based agent model grounded in the free-energy principle to explore these processes within the framework of allostasis. The model integrates interoceptive, proprioceptive, and exteroceptive predictive processing, minimizing variational free energy throughout past and future contexts. Simulations of resting state showed that inferred (mentally simulated) homeostatic context modulated attentional shifts between a "being mode" (focused on present interoceptive perception) and a "doing mode" (involving proprioceptive and exteroceptive sensorimotor loops). We also identified a meta-attentional parameter that controls these shifts by influencing cognitive attitudes toward internally generated beliefs across network modules, spanning from past to future. By manipulating this parameter, we replicated mindfulness states, where attention remained focused on the present interoceptive state despite minimal sensory prediction errors. This study offers insights into the computational dynamics of mindfulness and mind-wandering, laying the groundwork for future explorations of consciousness and pure awareness—a state of consciousness devoid of content.