Prefrontal contributions to the stability and variability of thought and conscious experience

This work is a conceptual review and theoretical synthesis rather than a primary empirical study. The investigators integrate two conceptual frameworks (DFT and REBUS) with existing neuroscientific evidence about prefrontal subregions and large-scale brain networks to generate hypotheses about how PFC nodes contribute to stability and variability in thought and conscious experience. The paper draws on multiple lines of evidence reported in the literature, including functional neuroimaging, lesion studies, transcranial magnetic stimulation, intracranial recordings, neurochemical measures, and clinical observations from conditions such as depression, schizophrenia, anorexia nervosa and alexithymia. It also integrates findings from experimental pharmacology of serotonergic psychedelics (e.g., psilocybin, LSD, DMT) and from studies examining network-level brain dynamics (functional connectivity, hierarchical information flow, measures related to entropy). The extracted text does not present an explicit methods section describing literature search strategies, inclusion criteria, databases searched, or dates covered. Instead, the paper appears to be a narrative, theory-driven review that organises empirical findings around the three central concepts adopted from DFT and REBUS: deliberate constraints, automatic constraints, and top-down beliefs.

The authors map empirical findings onto the three central mechanisms (deliberate constraints, automatic constraints, top-down beliefs) and discuss how specific prefrontal subregions and networks contribute to stability or variability in mental dynamics. Deliberate constraints: Regions of the frontoparietal and cingulo-opercular control networks are linked to deliberate, top-down constraint of thought. The dorsolateral PFC (DLPFC) is associated with maintaining and implementing rules, biasing other brain regions to favour goal-consistent contents, supporting prospective memory and resisting short-term temptations; disruptions to DLPFC (lesion or inhibitory TMS) impair prospective memory and self-control. The rostrolateral PFC (RLPFC) represents multiple rule–outcome associations and supports cognitive branching, planning and strategies that may organise thought over longer timescales; RLPFC engagement has been observed in creative tasks that require abstract, supraordinate instructions. The dorsal anterior cingulate cortex (dACC) is implicated in performance monitoring, valuation of actions and online adjustment of behaviour (a proposed basis for ‘‘tenacity’’), aligning outward actions and internally directed thought with goals. Together, these control nodes can increase stability by constraining content and transitions, while also enabling brief, targeted variability (for example during creative idea generation) through controlled shifts in strategy. Automatic constraints: Multiple networks and PFC nodes contribute to automatic, non-volitional constraints. The default network core (DN CORE), centred on the anterior medial PFC (aMPFC), is tied to abstract self-referential schemas and mnemonic elaboration; aMPFC recruitment is linked to self-related processing and to stereotyped, schematic streams of thought such as rumination, which manifest as rigid, low-variability mental states. The medial temporal lobe subcomponent of the default network (DN MTL), especially hippocampal regions, appears to generate spontaneous, perceptually detailed and associative thought contents; hippocampal activity often precedes reports of spontaneous thoughts, supporting transitions via associative cueing and thereby increasing variability when constraints are low. In contrast, the ventromedial PFC / medial orbitofrontal cortex (mOFC) contributes value-based elaboration and schema-guided construction of thought, biasing hippocampal generation toward highly valued, schema-consistent sequences and thus enhancing stability. The third default subcomponent (DN SUB3), with a dorsomedial PFC (dMPFC) node, underlies abstract, constructive simulation and ‘‘gist’’-oriented processing (social cognition and other high-construal tasks). dMPFC recruitment can bind mnemonic contents via abstracted meaning, promoting stability; however, altered dMPFC function is implicated in schizophrenia, where hyperactivity or misbinding of self/other representations may increase associative variability and contribute to hyper-associative symptoms. Salience and ventral attention networks: Prefrontal nodes in these networks detect salient internal or external stimuli and can either stabilise experience around a dominant signal or increase variability by shifting attention among multiple salient inputs. The rostral ACC (rACC) supports viscero-sensory appraisal and integrates interoceptive signals with autobiographical self-knowledge; its recruitment can lead to persistent focus on bodily states (increasing stability) or, when interpretability is compromised (e.g., alexithymia), to unpredictable viscero-sensory-driven variability. The right inferior frontal gyrus (IFG) detects spatially localised salient external stimuli, triggers attention shifts and can constrain motor responses; in dynamic external environments it contributes to frequent reorientations (increasing variability), whereas in stable external contexts its recruitment can reduce attentional lapses and thereby decrease experiential variability. Top-down beliefs and psychedelics: Using the REBUS account, the authors describe beliefs (priors) as top-down constraints that normally stabilise lower-level processing. Psychedelics are proposed to relax these priors (reduce precision-weighting), increasing bottom-up information flow and neural entropy and thus expanding experiential variability. Empirical observations consistent with this account include decreased functional connectivity within DN CORE (reduced aMPFC–posterior cingulate coupling) following psilocybin and LSD, correlations between reductions in aMPFC activation/blood flow and subjective intensity of effects, and altered glutamate concentrations in aMPFC linked to anxious ego-dissolution. The rACC shows early connectivity changes under psilocybin, suggesting bodily self-related beliefs may relax before higher-level autobiographical beliefs; psychedelics also reduce the functional integrity and modular segregation of control networks and flatten hierarchical gradients, consistent with reduced capacity for goal-related beliefs to constrain dynamics. Overall, the review emphasises that relaxation of mental, bodily and goal-related priors under psychedelics corresponds with increased bottom-up signalling and greater variability in thought and conscious experience. Context dependence: Across sections, the authors underscore that the same neural constraints can yield increased stability or increased variability depending on context—for example, persistent salient interoceptive signals can stabilise experience, whereas variable viscero-sensory input or disrupted interpretability can increase variability. Clinical conditions (rumination in depression, altered dMPFC function in schizophrenia, heightened dACC–retrosplenial coupling in anorexia nervosa, alexithymia-related rACC alterations) are used to illustrate how region-specific patterns map onto stable or variable dynamics.

The authors interpret their synthesis as evidence that integrating the DFT and REBUS frameworks with regionally specific neuroscience offers a useful way to understand how prefrontal subregions bias the stability versus variability of thought and conscious experience. They argue that deliberate control systems (frontoparietal and cingulo-opercular networks) generally stabilise cognition by imposing rules and goal-aligned constraints, whereas default network and salience-related PFC nodes can either stabilise or destabilise experience depending on whether they impose schematic/value-based constraints or enable associative, hippocampal-driven generation. Psychedelics, via 5-HT2A agonism, are interpreted within REBUS as relaxing high-level priors (self-, bodily- and goal-related beliefs), producing increased bottom-up influence, higher neural entropy and richer, more variable conscious states. Compared with earlier research that emphasised either cognitive control or spontaneous generation in isolation, the present synthesis highlights their interaction and the importance of dynamics (transitions and temporal stability) rather than static descriptions of content. The authors stress that neural-level constraints do not map straightforwardly onto phenomenological stability: the effect of a given region on experienced variability depends critically on environmental and internal context, on timescale, and on interactions with other networks. Key limitations and uncertainties acknowledged include the complexity of mapping network-level findings onto subregional specialisations and the challenges inherent in combining multi-scale explanations (regional specificity versus whole-brain dynamics). The paper also notes that much of the psychedelic literature relies on whole-brain measures that can obscure subregional effects, and that empirical causal claims about region-specific roles remain to be tested more directly. Implications and future directions proposed by the authors include: (1) empirical work to clarify how neural constraints relate contextually to conscious-state stability versus variability; (2) investigations of functional and anatomical gradients across adjacent prefrontal regions (for example along ventral–dorsal axes of the ACC); (3) wider adoption of fine-grained, regionally specific analyses in neuroscience and psychedelic research (seed-based structural, functional and neurochemical approaches alongside global measures); (4) studies linking evolutionary cortical expansion and serotonin 2A receptor distribution to hierarchical gradients and cognitive function; and (5) applying refined subregional predictions to clinical assessment and interventions (examples mentioned include early-detection behavioural tasks, models of psychedelic-assisted therapy and proposals to explore psychedelics for disorders of consciousness). The authors conclude that progress will require multi-level approaches that respect both network dynamics and regional specialisation, and that doing so has potential clinical relevance.