Canalization and plasticity in psychopathology

Carhart-Harris and colleagues revive the biological concept of canalization to propose a parsimonious, transdiagnostic model of psychopathology. They contrast canalization — the development of stable, low-variance phenotypes — with forms of plasticity that increase phenotypic variability. Central to their framing is precision (inverse variance): canalization corresponds to increased precision and reduced sensitivity to new evidence, whereas a distinct form of plasticity they introduce, TEMP (Temperature or Entropy Mediated Plasticity), corresponds to decreased precision, increased variance, and higher learning rates. This distinction is used to reinterpret how entrenched cognitive and behavioural styles can develop and persist after adversity, distress or dysphoria. The paper sets out to do three things. First, it develops a theoretical model (CANAL) that treats the general psychopathology or 'p' factor as arising from entrenched canalized phenotypes. Second, it links this model to existing computational frameworks (notably the free-energy principle) and a specific mechanistic account of psychedelic action (the REBUS model), arguing that transient increases in TEMP can enable a simulated-annealing-like reconfiguration of maladaptive attractors. Third, the authors map this framework onto existing basic and clinical neuroscience findings, propose candidate biomarkers and experimental paradigms to test the model, and discuss clinical implications — including how psychedelic therapy plus supportive psychotherapy might counteract canalization.

This is a theoretical, integrative article rather than an empirical study. The investigators synthesise evidence and concepts from developmental biology (Waddington's canalization), complexity science, dynamical systems and the free-energy principle, together with preclinical and human neuroscience and clinical findings on psychedelics and other rapid-acting interventions. From these sources they formalise two constructs — canalization (close to Hebbian, associative consolidation) and TEMP (an entropy- or 'temperature'-related early plasticity) — and map them onto a free-energy or energy-landscape metaphor in which precision corresponds to curvature and depth of attractor basins. Rather than reporting new data, the authors draw on published animal and human studies to link stages of the proposed annealing process to measurable markers. They recommend experimental approaches including high-temporal-resolution electrophysiology (EEG) to index early increases in neural entropy, and MRI-based pre/post scans to detect downstream functional and anatomical reweighting. Proposed neurophysiological measures include Lempel–Ziv complexity and a newer metric termed CSER (Complexity of State-Space Entropy Rate) applied to cortical recordings, while neuroimaging outcomes include metrics such as brain network modularity, white-matter diffusivity and synaptic-density markers. Behavioural assays suggested for operationalising canalization include measures of cognitive flexibility and associative learning, and the authors advocate cross-species paradigms and longitudinal designs to capture dynamics rather than cross-sectional snapshots.

The paper does not present original empirical results; instead it assembles prior findings that the authors interpret as consistent with the CANAL and TEMP/REBUS frameworks. Key empirical observations cited include: evidence that classic serotonergic psychedelics (5-HT2A agonists) increase measures of neural entropy in humans, with one multimodal study reporting increased brain entropy under high-dose psilocybin and a change in prefrontal-to-subcortex diffusivity one month later, and the acute entropy increase predicting longer-term improvements in psychological well-being. Psychedelics and 5-HT2A agonism are reported to promote expression of plasticity-related genes, such as BDNF, and to increase cortical spine formation in vitro and in vivo, particularly in cortex, findings the authors link to heterosynaptic plasticity. Additional cited evidence includes studies showing that LSD accelerates learning rates in animals and increases learning rate in humans, acute increases in sensitivity to music and suggestion under psychedelics, sub-acute gains in cognitive flexibility, and longer-term psychological changes (for example in personality or nature relatedness) after psychedelic experiences. The dissociative anesthetic ketamine is noted to reverse stress-induced dendritic spine atrophy in animal models and to show synaptic-density changes in humans measured one day after a single dose. Conversely, chronic stress and past adversity are associated with dendritic atrophy, reduced cortical thickness, reduced grey matter volume and markers of accelerated brain ageing, which the authors treat as neurobiological substrates for canalization. At the network level, the authors report findings that psilocybin therapy for depression was associated with decreased brain network modularity (interpreted as increased global integration) that correlated with symptom improvement in two independent samples; similar decreases in modularity have been reported after electroconvulsive therapy, although the literature contains conflicting results and some large-sample studies report different modularity patterns. The extracted text does not consistently provide sample sizes, effect sizes or confidence intervals for these cited studies, so quantitative anchoring in this summary is limited to directional and mechanistic characterisations as presented by the authors. The authors also synthesise clinical and translational considerations drawn from the literature: potential therapeutic benefits of combining an acute TEMP-like entropic phase with post-acute psychotherapy, and potential risks whereby increased plasticity could be hijacked to entrench maladaptive patterns (a metaplastic or iatrogenic canalization). They identify candidate contraindications for psychedelic therapy as speculative but include early psychosis, schizophrenia, certain personality disorders, bipolar I disorder and younger age as situations warranting caution. Proposed empirical tests include EEG/CSER or Lempel–Ziv indices predicting downstream markers of heterosynaptic plasticity across species, replications of post-treatment modularity changes, and behavioural assays of cognitive flexibility.

Carhart-Harris and colleagues interpret their synthesis to propose that entrenched canalization — excessive precision of priors that narrows an individual's phenotypic state-space — is a core component of psychopathology. They argue that such canalized phenotypes often develop as defensive responses to adversity, distress or dysphoria and endure because they minimise uncertainty at the cost of adaptability. The authors position their CANAL model within the free-energy framework and align it with the REBUS account of psychedelic action: psychedelics, via 5-HT2A agonism, produce an acute reduction in precision (an entropic increase or TEMP), which they liken to simulated annealing. This early phase, they propose, enables belief states to escape entrenched attractors and permits downstream heterosynaptic reweighting that can rebalance the system when followed by supportive psychotherapy. They relate the model to a wide range of psychopathological phenomena — internalizing and externalizing spectra, addictions, OCD, eating disorders, functional somatic syndromes and thought disorders — arguing that many common symptoms can be framed as canalized coping strategies that were once adaptive or momentarily effective. The authors interpret evidence of psychedelic-induced entropy, increased synaptogenesis, changes in modularity, and rapid symptom improvements across diagnoses as broadly consistent with their annealing account, while emphasising that these links remain provisional and require rigorous testing. The paper acknowledges important limitations: the model is extrapolative and speculative, empirical gaps exist across species and measurement modalities, and the literature includes conflicting neuroimaging findings (for example regarding modularity). The authors note methodological challenges that have hampered biomarker discovery in psychiatry — cross-sectional designs, heterogeneous symptoms, poor longitudinal sampling and high inter- and intra-subject variability — and recommend longitudinal, provocation and cross-species paradigms to test their hypotheses. They also discuss ethical and clinical implications: psychedelic-induced increases in TEMP might be therapeutic when combined with high-quality psychological support, but may precipitate decompensation in vulnerable individuals or be misused to shape beliefs inappropriately. Finally, the authors call for empirical validation of candidate biomarkers (EEG entropy metrics, modularity, synaptic-density measures), behavioural operationalisations of canalization, and research into individual differences (including possible polygenic susceptibility) that influence propensity for TEMP and canalization.

The paper concludes by restating the central proposal: a parsimonious model that conceptualises entrenched canalization — acquired in response to adversity and distress — as a principal component of psychopathology. The authors suggest that interventions which transiently increase TEMP, notably serotonergic psychedelics administered with gently guiding psychological support, may counter over-potentiated canalization by enabling an annealing-like reconfiguration of neural and psychological state-spaces. They emphasise that this account is a hypothesis-generating framework and invite empirical tests of its core predictions and biomarkers.