A biophysical framework for conscious state transitions based on intrinsic neuronal properties: the dynamic contextual element memory (DCEM) model

Unlike artificial systems that execute externally specified commands, the brain constructs contextual representations to guide behaviors that satisfy innate and learned desires generated by the Affective Valuation System (AVS). However, the neural mechanisms by which contextual representations are formed, selected, maintained, and replaced to generate conscious state transitions remain unknown. Here, we propose the Dynamic Contextual Element Memory (DCEM) model as a biophysical framework within the dorsolateral prefrontal cortex. The DCEM is conceived as a dynamic contextual memory consisting of numerous logical Contextual Elements (CEs), each implemented by multiple physical cortical columns. AVS signals initiate contextual exploration in the DCEM and guide contextual selection, maintenance, and transition. Recurrent interactions between the DCEM and the parietal and temporal association cortices generate γ-band contextual candidates from sensory and memory information. Physical columns of candidates selected by Contextual Element Decision Neurons (CEDeN) in the mediodorsal thalamus receive synchronized feedback and form a coherent β-band Contextual Element Loop (CE Loop) representing a single logical Contextual Element. Outputs from DCEM Layer 5 redistribute the selected context throughout the association cortices, allowing AVS to shape candidates. CE Loops that lose CEDeN support following re-evaluation of behavioral outcomes or new information fail to maintain synchrony and disappear. We therefore propose that conscious state transitions occur through the formation, maintenance, disappearance, and replacement of CE Loops. This model was naturally derived by extending the intrinsic biophysical properties of neurons identified in our studies to cortical networks while integrating accumulated findings from modern observational techniques, neuroanatomy, neurophysiology, and neuroscience.