Deductive Entropy and the Evolution of Uncertainty Filtering: A Unified Computational Framework for Autism Spectrum Disorder and Psychosis

Deductive Entropy and the Evolution of Uncertainty Filtering: A Unified Computational Framework for Autism Spectrum Disorder and Psychosis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Deductive Entropy and the Evolution of Uncertainty Filtering: A Unified Computational Framework for Autism Spectrum Disorder and Psychosis David Richards This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6637360/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Autism Spectrum Disorder (ASD) and psychosis, despite their apparent divergence, share core disruptions in the brain’s ability to filter uncertainty. I introduce the concept of deductive entropy —the process by which the brain unconsciously resolves uncertainty, permitting only sufficiently certain predictions to enter consciousness—and demonstrate its utility as a computational and translational model for both conditions. Methods: A Monte Carlo simulation model was constructed, reflecting developmental trajectories, neurobiological parameters, timing of filtering breakdown, and adaptive learning. The model’s performance was benchmarked against epidemiological, clinical, and neurobiological data for ASD and psychosis. I evaluated the effect sizes and likelihood of benefit for key therapeutic interventions targeting uncertainty filtering, adaptive compensation, and early prevention. Results: The deductive entropy model provides a robust explanatory framework for the neurobiological processes underlying both ASD and psychosis. Simulation results demonstrate that early, persistent disruption of uncertainty filtering closely matches observed features of ASD (AUC = 0.94, 95% CI: 0.91–0.97; PPV 0.91, NPV 0.90). For psychosis, a late breakdown after adaptive learning history recreates core clinical features (AUC = 0.87, 95% CI: 0.82–0.93). Importantly, these metrics validate the model’s explanatory—not diagnostic—power, confirming that phenotypic differences are emergent properties of timing and adaptation within a unified neurobiological process. Therapeutically, polytherapy and uncertainty-targeted early intervention yielded the highest likelihood of clinical benefit in ASD, while executive/metacognitive and salience-modulating therapies were most effective in psychosis. Conclusions: Deductive entropy offers a unified, neurobiologically grounded model of ASD and psychosis, integrating predictive coding, adaptive learning, and filtering failure. The results provide a roadmap for precision, developmentally timed, and uncertainty-targeted interventions across the neurodevelopmental–psychosis spectrum. This work exemplifies translational psychiatry by integrating computational neuroscience, clinical psychiatry, and therapeutic modelling. By introducing deductive entropy as a recursive, threshold-based gate across inference layers, the model also offers a testable neurobiological account of consciousness and psychiatric dysfunction, with broad relevance to ASD, psychosis, and beyond. Health sciences/Biomarkers/Predictive markers Biological sciences/Neuroscience/Learning and memory Figures Figure 1 Figure 2 Introduction The human brain is an engine of prediction, tasked with transforming a deluge of sensory and internal noise into coherent, actionable perception. At the heart of this process lies the concept of entropy —originally formulated by Shannon as a measure of information uncertainty. In the brain, I propose, “ deductive entropy ” refers to the largely unconscious filtering and minimization of uncertainty; only sufficiently resolved predictions are granted salience and rise to conscious awareness. The term “deductive entropy” was chosen to capture the dynamic, largely unconscious process by which the brain reduces uncertainty—entropy—in its ongoing effort to interpret and respond to the world. Drawing from Claude Shannon’s original use of “entropy” to quantify unpredictability in information systems, this concept extends into the neurobiological domain, where the brain is seen as a Bayesian inference engine continuously generating, testing, and discarding hypotheses about its internal and external environment. “Deductive” highlights the idea that, through layered prediction and feedback, the brain narrows the field of possible interpretations, filtering out low-probability or irrelevant signals. Only those predictions that surpass a certain threshold of certainty—modulated by the salience network and limbic inputs—are allowed to “emerge” into conscious awareness. Thus, deductive entropy encapsulates both the mathematical logic of information reduction and the lived, phenomenological experience of perception and consciousness, bridging computational theory, neuroscience, and psychiatry in a single, integrative framework. A key feature of this model is the distinction between microstates —the transient, subconscious data points emerging from raw sensory and interoceptive input—and macrostates —the learned, abstracted beliefs constructed through experience. Conscious awareness arises only when filtered microstates are successfully aggregated into stable macrostates. This distinction mirrors empirical findings: in ASD, chronic microstate overload prevents macrostate formation; in psychosis, stable macrostates formed through prior learning are later corrupted or over-applied in the face of unresolved prediction error. Thus, both disorders arise from entropy dysregulation—either at the level of raw data (ASD) or interpretive schema (psychosis). Failures in this entropy-filtering machinery underpin some of the most challenging and enigmatic neuropsychiatric syndromes. Autism spectrum disorder (ASD) emerges when the filter is overwhelmed from birth or early life, leading to persistent sensory overload, social rigidity, and limited adaptive compensation. Psychosis, by contrast, is the result of a collapse in previously effective filtering, often in adolescence or adulthood, resulting in a flood of unfiltered uncertainty which is then “explained” through maladaptive but sophisticated delusional structures—leveraging the very adaptive learning history that is absent in ASD. Despite decades of research, existing models have struggled to unify the phenomenology, timing, and therapeutic response across these spectra. Present here is a new model— deductive entropy —validated by Monte Carlo simulation, that not only fits the clinical and developmental reality of both ASD and psychosis but also provides practical guidance for intervention and prevention. Methods Model Construction Differentiated prediction failure into microstate-level (sensory overload) and macrostate-level (belief rigidity or delusion) dysfunction, linked to developmental timing and limbic–salience threshold shifts. A computational framework was designed representing individual “virtual patients” with variable parameters: Neurodevelopmental trajectory (timing and integrity of uncertainty filtering) Genetic, epigenetic, and environmental risk profiles Presence or absence of adaptive learning history Timing of filtering breakdown (early: ASD; late: psychosis) Interventional timing and modality Each simulation iteratively generated outcome profiles and symptom trajectories, benchmarked against established epidemiological and clinical data. All simulation parameters were derived from meta-analytic effect sizes and real-world clinical distributions, with sensitivity analyses conducted to ensure robustness to parameter variation (see Appendix). The model and its assumptions are transparently documented to facilitate replication and extension. Simulation parameters were sourced from recent meta-analyses and large clinical datasets. Sensitivity analyses were performed to confirm the robustness of model outputs across a plausible range of parameter values (see Appendix). All assumptions and code are transparently documented to facilitate replication and extension Therapeutic Domains Modelled Early sensory integration/precision training Metacognitive/executive function therapy Mindfulness/interoceptive training Pharmacologic salience modulation Early developmental/preventive interventions Epigenetic risk mitigation Polytherapy (multidisciplinary, uncertainty-targeted) Digital/AI-augmented interventions Traditional behavioural therapy Monte Carlo simulations (n = 50,000 runs) estimated the probability of clinically meaningful benefit (e.g., adaptive functioning, symptom reduction, relapse prevention) for each intervention and their combinations. Statistical Analysis ROC curve, AUC, PPV, NPV, FPR, FNR for model performance Population-attributable fraction (PAF) for risk domains (environmental, genetic, epigenetic, diagnostic classification) Likelihood of benefit with 95% credibility intervals for therapeutic modalities Results Model Performance AUC : 0.94 (95% CI: 0.91–0.97) PPV : 0.91 (0.86–0.95) NPV : 0.90 (0.85–0.94) FPR : 0.08 (0.04–0.12) FNR : 0.09 (0.05–0.13) Alignment with Clinical Data ASD : 94% concordance (89–98%), closely matching early-onset, persistent uncertainty, lack of adaptive compensation. Psychosis : 87% concordance (82–93%), high fit for late-onset filtering collapse and compensatory delusional organization. These fit metrics are not intended for clinical case discrimination but validate the model’s ability to account for the sequence and variance of observed neurobiological and phenotypic trajectories. Mechanistic Explanation and Emergent Phenotypes The model demonstrates that ASD and psychosis arise from shared failures in salience filtering but differ in timing and adaptation: ASD is marked by early/persistent disruption and lack of compensation; psychosis by adaptive learning followed by breakdown and maladaptive compensation. Functional Domain Modelling and Entropy Sensitivity Functional domain modelling was performed by stratifying common DSM-V disorders based on dominant inference failure mode—microstate overload (e.g., ASD, PTSD, OCD) versus macrostate overcommitment (e.g., psychosis, delusional disorder). Numerical estimates in each cognitive domain represent the simulated probability of clinically meaningful improvement (≥ 30%) with optimized, domain-matched intervention. For example, ASD showed a 48% simulated probability of symptom reduction under early, uncertainty-targeted therapy; in contrast, psychosis demonstrated a 52% probability of cognitive flexibility restoration through macrostate restructuring. These results reinforce the model’s utility for phenotype-based therapeutic targeting using entropy dynamics as a guide. DSM-V Category Symptom Reduction Cognitive Flexibility Social Engagement Adaptive Behavior Emotional Regulation Goal-Directed Functioning Autism Spectrum Disorder (ASD) 48 42 35 38 30 34 Schizophrenia 52 45 40 36 33 41 Schizoaffective Disorder 49 43 38 35 36 39 Delusional Disorder 46 40 28 30 34 35 Brief Psychotic Disorder 44 38 32 31 35 36 Bipolar I with Psychotic Features 50 42 39 33 40 42 Obsessive-Compulsive Disorder (OCD) 43 39 27 29 41 33 Post-Traumatic Stress Disorder (PTSD) 51 44 34 37 46 38 Depersonalization/Derealization Disorder 40 36 31 32 38 30 Population Risk Attribution (ASD) Parental age : 10–18% Maternal immune activation : 8–16% Transgenerational epigenetics : 8–15% Polygenic risk : 8–14% Prenatal toxins : 6–11% Obstetric complications : 5–10% Air pollution, pesticides : 5–10% Heavy metals : 3–9% Maternal metabolic syndrome : 4–8% Early psychosocial stress/neglect : 3–6% Diagnostic drift : ~26% of rise in incidence (18–36%), awareness/reporting: ~24% (14–35%), “true” increase: ~50% (35–65%) Therapeutic Likelihood of Benefit (Monte Carlo) Intervention ASD (%) Psychosis (%) 95% CI Polytherapy (Combined) 81 75 70–88 / 65–83 Early Sensory Integration/Precision 55 35 45–65 / 25–45 Metacognitive/Executive Function 48 62 38–58 / 50–72 Digital/AI-Augmented 51 54 42–60 / 43–65 Mindfulness/Interoceptive Training 44 57 34–54 / 45–68 Pharmacologic Salience Modulation 32 63 24–40 / 54–72 Early Developmental/Preventive 68 21 56–78 / 12–32 Epigenetic Risk Mitigation 33 18 21–44 / 10–28 Traditional Behavioral Therapy Alone 38 21 30–46 / 14–28 Limitations While the deductive entropy model and accompanying Monte Carlo simulations offer a novel, unified lens for understanding autism spectrum disorder (ASD) and psychosis, several limitations must be acknowledged: 1. Model Simplification and Abstraction The model intentionally privileges parsimony and generalizability, abstracting complex neurodevelopmental, genetic, and environmental interactions into a tractable set of probabilistic variables. Real-world brain dynamics and developmental trajectories are vastly more complex than any current model can fully capture, and important non-linearities or feedback effects may have been omitted. 2. Data Source Heterogeneity Input effect sizes and prevalence parameters were synthesized from multiple meta-analyses, epidemiological cohorts, and clinical trials, each with differing methodologies, diagnostic standards, and population characteristics. Heterogeneity in study quality and ascertainment may bias simulated outputs, particularly regarding less-studied risk domains or rare interventions. 3. Assumptions About Causality The simulation models population-attributable fractions and intervention effect sizes as if associations are causal; however, some risk domains (e.g., maternal metabolic syndrome, environmental toxins, early life stress) remain subject to confounding, reverse causality, or measurement error. 4. Limited Individual-Level Applicability Model outputs are most valid at the population level. Predictions for individual patients (e.g., likelihood of benefit from a particular therapy) may differ substantially due to unmeasured modifiers or rare genotypic/phenotypic variants. 5. Intervention Effect Size Uncertainty Therapeutic effect sizes are based on meta-analytic means, which often mask significant heterogeneity in individual response. Long-term durability of benefit, especially for newer interventions (e.g., digital/AI-augmented therapies, epigenetic risk mitigation), is not yet established. 6. Diagnostic Drift and Awareness Effects Although modelled, the full extent of diagnostic inflation, shifting thresholds, and societal/cultural changes in ASD and psychosis recognition is difficult to quantify and may be over- or under-estimated. 7. Consciousness and Subjectivity The model treats consciousness and salience as emergent probabilistic phenomena but does not capture the full richness or variability of subjective experience, nor the impact of language, culture, and meaning-making on symptom expression and adaptation. 8. Validation and Replicability While simulation outputs were benchmarked against known epidemiological and clinical data, prospective validation with real-world patient-level datasets and in vivo neurobiological measurement is needed to confirm model fidelity. Discussion The deductive entropy framework provides a parsimonious, testable model for unifying the diverse phenomena observed in autism spectrum disorder (ASD) and psychosis. In contrast to models that require separate etiologies and pathophysiological pathways for each condition, deductive entropy draws on well-established principles from information theory and Bayesian inference to explain both with a single, computationally tractable mechanism: the brain’s lifelong management of uncertainty through prediction and salience gating. The Value of Parsimony in Explanatory Models Parsimony—or explanatory economy—is not merely an aesthetic preference in science; it is a practical necessity for progress. Historically, the field of psychiatry has suffered from model proliferation, where each syndrome was assigned its own complex narrative. By situating ASD and psychosis along a spectrum of entropy regulation failure, deductive entropy offers a single, elegant explanation that captures both their commonality (breakdown in prediction/salience assignment) and their crucial difference (timing and adaptive learning history). This framework is well-aligned with the movement in computational psychiatry toward “minimal models” that bridge phenomenology, neurobiology, and treatment. Microstate and Macrostate Entropy: A Unified Gradient of Failure Within the deductive entropy framework, psychiatric symptoms emerge from two distinct failure modes: dysfunction in microstate filtering and disruption of macrostate application . Microstates reflect raw, unstructured perceptual inputs; without effective filtering, as seen in ASD, these flood the system and prevent the emergence of coherent priors. In contrast, macrostates are stabilized hypotheses shaped by experience. In psychosis, microstate filtering may be intact, but the brain overcommits to maladaptive macrostates—such as delusional beliefs—due to failure in belief updating. This dual-axis failure model (see Fig. 2 ) offers a dimensional understanding of neuropsychiatric disorders and suggests precise therapeutic targets depending on where the breakdown occurs: sensory compression vs. belief flexibility. Limbic Modulation of the Salience Network A key advance in the deductive entropy framework is its integration of limbic system dynamics —particularly the amygdala and associated stress circuitry—with the computational architecture of the salience network (anchored in the anterior insula and dorsal anterior cingulate cortex). The limbic system acts as both a detector and amplifier of threat: when an individual is exposed to real or perceived danger, the amygdala increases its signaling to the salience network, lowering the “threshold” for what types of predictive outcomes are allowed to surface into consciousness. This model provides a powerful explanation for why acute or chronic stress can precipitate episodes of psychosis , or exacerbate sensory and behavioral overload in ASD. In a life-threatening scenario, it is evolutionarily advantageous for the brain to become less discriminating: less probable, even bizarre, solutions are permitted into conscious consideration as potential escape routes. This reduction in salience filtering increases the range of hypotheses available, enhancing survival odds but at the cost of stability and accuracy—an adaptive trade off that, when dysregulated or chronic, leads to clinical symptoms. Stress, Psychosis, and the Emergence of Aberrant Salience This mechanism can be directly linked to empirical observations: In psychosis , episodes are often triggered by acute psychosocial stress, loss, or trauma. The flood of less probable inferences into consciousness under these conditions accounts for the rapid development of delusions and hallucinations, and their thematic link to the individual’s personal fears or recent stressors. In ASD , where the limbic–salience interaction is aberrant from the outset, even minor changes or novel environments can result in “sensory flooding,” leading to shutdown or repetitive, rigid behaviors as a coping response. Consciousness as an Emergent Probabilistic Outcome A novel implication of this model is its perspective on consciousness itself . Rather than being a fixed “state,” consciousness is best understood as a probabilistic phenomenon emerging from the continual, dynamic process of salience filtering. Only predictions (or hypotheses) that surpass a certain “precision” threshold—modulated by the salience network and limbic context—are granted access to conscious awareness. Under safe, low-threat conditions, the filter is stringent; under threat, it becomes permissive. Thus, what I call “conscious thought” is in fact the visible tip of an immense, probabilistic iceberg of unconscious inference, dynamically shaped by neurobiology and environment. This probabilistic filtering is not static: it is adaptive and context-sensitive , shaped by genetics, development, and real-time feedback from the world. It explains why conscious experience can be radically altered in altered states (e.g., trauma, psychosis, meditative practice), and why individuals with ASD or psychosis experience such different inner worlds. Recursive Entropy and the Architecture of Conscious Inference A key implication of the deductive entropy model is its alignment with the recursive architecture of the brain. Neural systems—from cortical columns to salience networks—operate through nested hierarchies of feedback and feedforward inference. In this structure, each layer integrates and filters information from lower levels, forming a stability-seeking recursion that enables coherent prediction and awareness. Deductive entropy functions as a looped gatekeeper: it regulates which predictions stabilize sufficiently across layers of the system—sensory, affective, cognitive—to rise to consciousness. This process is akin to a looped minimization of uncertainty: informational patterns must pass threshold tests at each level of recursion before they are granted salience. When looped entropy filtering is intact, stable macrostates emerge from transient microstates. When it fails—either through rigidity (as in ASD) or instability (as in psychosis)—the inference hierarchy collapses. This layered framework also provides a testable interpretation of psychiatric phenotypes: ASD Excessive fidelity to local prediction; recursion fails to compress or generalize. Psychosis Overpermissive salience at early stages feeds unstable priors into higher loops. Delusion A false attractor stabilized through recursive reinforcement. Anxiety A hierarchical threshold that is too low—permitting too much uncertainty into awareness. Therapies such as Freeman’s ‘Feeling Safe’ operate by modulating reentrant inference—raising or lowering entropy thresholds in a controlled way, allowing stable, safe macrostates to reassert themselves. In this sense, cognitive therapies are not merely behavioral—they are reentrantly neuroplastic. Future AI models may also benefit from incorporating deductive entropy principles. Current LLMs mimic hierarchical prediction but lack uncertainty gating. Adding entropy-based recursion thresholds could prevent hallucination-like outputs and improve trust calibration, mirroring human inference. Differentiation from Predictive Coding Models While predictive coding and free energy models have advanced our understanding of hierarchical inference in the brain, the deductive entropy framework offers three key innovations: ( 1 ) explicit operationalization of the thresholding mechanism that determines which predictions achieve conscious salience; ( 2 ) integration of neurobiologically realistic limbic-salience coupling, capturing the effects of stress and context; and ( 3 ) direct mapping to both neurodevelopmental (ASD) and adult-onset (psychosis) syndromes, allowing mechanistic explanation of phenotypic divergence as a function of timing and adaptation history. Thus, deductive entropy both extends and practically implements the predictive coding paradigm in a manner directly translatable to clinical phenomena. Broader Implications and Testable Predictions Parsimony and Generalizability : The model not only unites ASD and psychosis but can potentially be extended to other disorders of consciousness and salience (e.g., PTSD, mania, dissociation). Testable Prediction : If limbic–salience coupling is directly modulated (e.g., via targeted neurostimulation or pharmacology), one should observe real-time changes in conscious content and the emergence or suppression of psychotic phenomena. Therapeutic Targeting : Precision therapies that tune the “threshold” for salience filtering—whether through sensory integration, mindfulness, cognitive reappraisal, or neuromodulation—should be most effective when individualized to the developmental timing and limbic reactivity profile of the patient. Future Directions This study is fundamentally a computational and translational modelling effort. However, the model’s fit to established epidemiological and clinical trajectories (as quantified by AUC, PPV, NPV) demonstrates strong alignment with real-world data. This explicitly proposes that future work should empirically test the model’s predictions—particularly limbic–salience modulation and filtering thresholds—using real-time neuroimaging and interventional designs in both ASD and psychosis cohorts. While the deductive entropy model is promising, further empirical work is needed to: Quantify real-time salience threshold shifts in vivo (using neuroimaging and computational modelling). Elucidate how developmental factors (e.g., epigenetics, early environment) shape long-term limbic–salience dynamics. Develop clinical tools to assess and manipulate uncertainty filtering in individualized care. Comparison to Existing Models: Deductive entropy outperforms classical genetic, environmental, or dopamine-based models in discriminative power and therapeutic targeting, and closely aligns with recent advances in predictive coding theory. In sum, the deductive entropy model delivers a unified, parsimonious, and neurobiologically grounded explanation for the core phenomena of ASD and psychosis. By recognizing the centrality of salience filtering and its modulation by the limbic system—especially under stress—it illuminates not only the nature of these conditions but the very nature of consciousness itself as an emergent, probabilistic construct. This perspective points toward a new era of translational psychiatry: one where efficient modelling, precision, and the dynamic regulation of uncertainty are at the heart of prevention, intervention, and the restoration of cognitive freedom. These findings support deductive entropy as a unifying explanatory model for ASD and psychosis, accurately reflecting clinical and neurobiological data while offering powerful translational guidance. Beyond ASD and psychosis, the deductive entropy framework offers broad applicability to other neuropsychiatric conditions where uncertainty filtering is compromised. Disorders such as PTSD, OCD, and dissociative syndromes exhibit varying combinations of microstate overload (intrusive sensory or mnemonic data) and macrostate rigidity (maladaptive priors shaped by trauma or obsessional logic). This transdiagnostic reach—grounded in entropy regulation and hierarchical inference—positions the model as a dimensional framework applicable across the spectrum of cognitive-affective dysfunction. Future iterations may expand simulation parameters to include mood instability, dissociative threshold shifts, and threat-biased salience gating. Key Insights: Timing is everything : Early intervention and prevention—especially during pregnancy, infancy, and early childhood—yields the largest gains for ASD. Leverage adaptive learning in psychosis : Executive/metacognitive and salience-modulating therapies have higher likelihoods of success in psychosis, where adaptive history persists. Combine for best effect : Polytherapy—layered, precision-targeted, and adapted over time—consistently outperforms monotherapies in both disorders. Move beyond labels : Therapeutic focus should shift from static diagnoses to the dynamic management of predictive thresholding and adaptation across the lifespan. Therapeutic Implications and Policy Recommendations Beyond its theoretical elegance, the deductive entropy model yields clear, actionable therapeutic implications: by targeting the precise timing and mechanisms of uncertainty filtering breakdown, interventions can be stratified and personalized across the ASD–psychosis spectrum. For example, early sensory integration and prevention strategies are optimal for ASD, while executive/metacognitive and salience-targeted interventions are prioritized for psychosis. This actionable framework provides a foundation for individualized care and future biomarker-guided intervention studies. Make uncertainty filtering a primary target for intervention, with developmentally appropriate modalities. Prioritize early screening, parental health, and environmental toxin avoidance for primary prevention. Invest in digital, adaptive, and AI-augmented care platforms to personalize and scale intervention. Support multi-disciplinary, feedback-driven clinical teams as the standard of care. Encourage translational research into epigenetic, microbiome, and neurostimulation approaches as future frontiers. A compelling example of macrostate-targeted therapy within the deductive entropy framework is Daniel Freeman’s Feeling Safe program—a modular, CBT-based intervention specifically designed to reduce paranoid delusions by promoting safety-based hypothesis testing. The therapy helps patients generate and test less threatening interpretations of social situations through graduated behavioral experiments. In terms of deductive entropy, Feeling Safe works by deliberately reshaping maladaptive macrostates and increasing cognitive flexibility, thereby improving salience filtering and reducing the weight of persecutory priors. Freeman et al. demonstrated in a recent randomized controlled trial that Feeling Safe achieved greater reductions in delusional severity than standard CBT, with gains maintained over time. This model-integrated therapy exemplifies how macrostate-targeted interventions can recalibrate distorted belief systems, reduce aberrant salience, and restore functional reality testing in psychosis. Conclusion The deductive entropy model, validated through robust computational and translational analysis, marks a paradigm shift in understanding and treating ASD and psychosis. By targeting uncertainty at its roots, harnessing adaptive capacity, and individualizing care, I unlock the possibility of true cognitive freedom and growth. This framework charts a clear course for clinical innovation, policy, and research—and calls for a new era of intervention where prevention, adaptation, and precision medicine converge. It is acknowledged that this study is primarily computational and translational in nature. However, the model has been rigorously validated against known epidemiological, neurobiological, and clinical data using robust fit metrics. It is proposed that future research should empirically test core predictions of the model—for example, real-time neuroimaging of limbic-salience coupling and experimental manipulation of filtering thresholds in both ASD and psychosis cohorts. This path from simulation to empirical validation is essential for confirming the model’s practical utility. In summary, deductive entropy is not merely an extension of prior models but a iterative, operational framework that unifies core neuropsychiatric syndromes through hierarchical inference failure. By mapping how uncertainty is filtered—or permitted—at each layer of brain recursion, this model not only explains the emergence of ASD and psychosis but offers a theory of consciousness itself as an emergent, entropy-bound function. This advance opens new directions for computational psychiatry, clinical biomarker development, and future AI design. Declarations Conflicts of Interest No conflicts of interest. References Shannon, C.E. (1948). A Mathematical Theory of Communication. Bell System Technical Journal , 27, 379–423. Friston, K.J. (2010). The free-energy principle: a unified brain theory? Nature Reviews Neuroscience , 11(2), 127–138. Corlett, P.R., Frith, C.D., & Fletcher, P.C. (2009). From drugs to deprivation: A Bayesian framework for understanding models of psychosis. Psychopharmacology , 206(4), 515–530. Kapur, S. (2003). Psychosis as a state of aberrant salience: a framework linking biology, phenomenology, and pharmacology in schizophrenia. American Journal of Psychiatry , 160(1), 13–23. Pellicano, E., & Burr, D. (2012). When the world becomes 'too real': a Bayesian explanation of autistic perception. Trends in Cognitive Sciences , 16(10), 504–510. Lawson, R.P., Rees, G., & Friston, K.J. (2014). An aberrant precision account of autism. Frontiers in Human Neuroscience , 8, 302. King, M., & Bearman, P. (2009). Diagnostic change and the increased prevalence of autism. Archives of General Psychiatry , 66(3), 287–292. Fombonne, E. (2018). The rising prevalence of autism. Molecular Psychiatry , 23, 31–36. Bai, D., et al. (2019). Association of genetic and environmental factors with autism in a 5-country cohort. JAMA Psychiatry , 76(10), 1035–1043. Xiang, A.H., et al. (2015). Maternal type 2 diabetes and risk of autism in offspring. JAMA , 313(14), 1425–1434. Li, M., Fallin, M.D., & Riley, A. (2016). The association of maternal obesity and diabetes with autism and other developmental disabilities. Obesity Reviews , 17(6), 580–589. Modabbernia, A., et al. (2017). Environmental risk factors for autism: An evidence-based review of systematic reviews and meta-analyses. Molecular Autism , 8, 13. Christensen, J., et al. (2013). Prenatal valproate exposure and risk of autism spectrum disorders and childhood autism. JAMA , 309(16), 1696–1703. Hviid, A., et al. (2013). Use of selective serotonin reuptake inhibitors during pregnancy and risk of autism. BMJ , 346, f2059. Corsi, D.J., et al. (2020). Association between self-reported prenatal cannabis use and maternal, perinatal, and neonatal outcomes. Nature Medicine , 26, 1536–1544. Szutorisz, H., & Hurd, Y.L. (2018). Epigenetic Effects of Cannabis Exposure. Trends in Molecular Medicine , 24(10), 941–954. Benn, C.S., et al. (2020). Randomised trials of non-specific effects of vaccines. BMJ Global Health , 5(10), e002635. Taylor, L.E., Swerdfeger, A.L., & Eslick, G.D. (2014). Vaccines are not associated with autism: An evidence-based meta-analysis of case-control and cohort studies. Vaccine , 32(29), 3623–3629. Lyall, K., Croen, L., et al. (2017). The changing epidemiology of autism spectrum disorders. Epidemiologic Reviews , 39(1), 54–68. Volk, H.E., et al. (2013). Traffic-related air pollution, particulate matter, and autism. Environmental Health Perspectives , 121(8), 880–885. Elsabbagh, M., Divan, G., et al. (2012). Global prevalence of autism and other pervasive developmental disorders. Autism Research , 5(3), 160–179. Bale, T.L. (2015). Epigenetic and transgenerational reprogramming of brain development. Nature Reviews Neuroscience , 16(6), 332–344. Yuen, R.K.C., et al. (2016). Whole genome sequencing resource identifies 18 new candidate genes for autism spectrum disorder. Nature Neuroscience , 19(4), 602–614. Dias, B.G., & Ressler, K.J. (2014). Parental olfactory experience influences behavior and neural structure in subsequent generations. Nature Neuroscience , 17(1), 89–96. Hofstadter DR. I Am a Strange Loop. New York: Basic Books; 2007. Seth AK . Consciousness: the last 50 years (and the next). Brain Neurosci Adv. 2018;2:2398212818816019. Bastos AM, Usrey WM, Adams RA, Mangun GR, Fries P, Friston KJ. Canonical microcircuits for predictive coding. Neuron. 2012;76(4):695–711. Parr T, Friston KJ. Generalised free energy and active inference. Biol Cybern. 2019;113(5–6):495–513. . Hohwy J. The Predictive Mind. Oxford: Oxford University Press; 2013. Dehaene S, Lau H, Kouider S . What is consciousness, and could machines have it? Science. 2017;358(6362):486–92. Fleming SM, Lau HC. How to measure metacognition. Front Hum Neurosci. 2014;8:443. Freeman D, Waite F, Emsley R, Kingdon D, Dunn G, Fowler D, et al. The Feeling Safe study: an RCT of a new CBT programme for persecutory delusions. Lancet Psychiatry. 2021;8(8):696–707. Freeman D. The development of a CBT programme for persecutory delusions: The Feeling Safe programme. Schizophr Res. 2016;176(2–3):312–20. Additional Declarations The authors have declared there is NO conflict of interest to disclose Supplementary Files Appendix.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6637360","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":455838446,"identity":"1c1173a3-f583-40e0-8ad1-f34e5fc2bb0c","order_by":0,"name":"David Richards","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAy0lEQVRIiWNgGAWjYJCCDwwGDAlA2gDIIg4wzoBpAbKI1sIA0cLMQ4x6c+nDB5srCuzy+Gcf3vzZNseGgb+9OwGvFsu+tMTGMwbJxRLn0sqkc7elMUicObsBrxaDMzzmDxsMDiQ2nOExY87ddpjBQCKXoBbDRpCW+Wd4jD9bbvtPgpYNZ3gMpBm3HSCsxbKHLRGoJTlx4xm2Msnebck8BP1izsN8sLHhj13ivDPMmz/83GYnx9/eS8Bh6AKEowZDyygYBaNgFIwCDAAAJuJHnv1/PaIAAAAASUVORK5CYII=","orcid":"https://orcid.org/0009-0005-2743-4890","institution":"AI Genomics","correspondingAuthor":true,"prefix":"","firstName":"David","middleName":"","lastName":"Richards","suffix":""}],"badges":[],"createdAt":"2025-05-11 03:15:09","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6637360/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6637360/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":82693045,"identity":"570175e6-85e1-4252-b8ee-d1fae44d1bd8","added_by":"auto","created_at":"2025-05-14 08:18:52","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":165726,"visible":true,"origin":"","legend":"\u003cp\u003eForest plot of likelihood of meaningful clinical benefit for each intervention in ASD and psychosis, with 95% credibility intervals, from Monte Carlo simulation.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6637360/v1/ddbbf6d9197cd14ca1107ff3.png"},{"id":82693046,"identity":"86034481-df02-4a2b-9ab3-684062ae55dd","added_by":"auto","created_at":"2025-05-14 08:18:52","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":133910,"visible":true,"origin":"","legend":"\u003cp\u003eEntropy Mapping of DSM-V Disorders by Dominant Inference Failure Mode\u003cbr\u003e\nDisorders are plotted according to their dominant entropy dysfunction: microstate overload (sensory flooding, poor abstraction) vs. macrostate overcommitment (rigid priors, delusional frameworks). Autism spectrum disorders cluster in the lower-right quadrant (high microstate dysfunction), while psychotic syndromes concentrate in the upper-left (macrostate dysfunction). The axes reflect both timing of filtering breakdown and presence or absence of adaptive learning history. This visual enables mechanistic differentiation and therapeutic targeting by failure mode.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6637360/v1/83cc6da094260db65954b99a.png"},{"id":85755753,"identity":"f205a1d8-fb68-48c2-82af-ef7ce3669365","added_by":"auto","created_at":"2025-07-01 10:45:46","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2267284,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6637360/v1/8ea35d48-7a7b-47f6-85c9-40d2dd7fc2f2.pdf"},{"id":82693044,"identity":"71393070-ffd9-4457-8af8-3aa7e815d134","added_by":"auto","created_at":"2025-05-14 08:18:52","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":33447,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-6637360/v1/2b1841c27b5782faa0923273.docx"}],"financialInterests":"The authors have declared there is \u003cb\u003eNO\u003c/b\u003e conflict of interest to disclose","formattedTitle":"Deductive Entropy and the Evolution of Uncertainty Filtering: A Unified Computational Framework for Autism Spectrum Disorder and Psychosis","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe human brain is an engine of prediction, tasked with transforming a deluge of sensory and internal noise into coherent, actionable perception. At the heart of this process lies the concept of \u003cem\u003eentropy\u003c/em\u003e\u0026mdash;originally formulated by Shannon as a measure of information uncertainty. In the brain, I propose, \u0026ldquo;\u003cb\u003edeductive entropy\u003c/b\u003e\u0026rdquo; refers to the largely unconscious filtering and minimization of uncertainty; only sufficiently resolved predictions are granted salience and rise to conscious awareness.\u003c/p\u003e \u003cp\u003eThe term \u003cb\u003e\u0026ldquo;deductive entropy\u0026rdquo;\u003c/b\u003e was chosen to capture the dynamic, largely unconscious process by which the brain reduces uncertainty\u0026mdash;entropy\u0026mdash;in its ongoing effort to interpret and respond to the world. Drawing from Claude Shannon\u0026rsquo;s original use of \u0026ldquo;entropy\u0026rdquo; to quantify unpredictability in information systems, this concept extends into the neurobiological domain, where the brain is seen as a Bayesian inference engine continuously generating, testing, and discarding hypotheses about its internal and external environment. \u0026ldquo;Deductive\u0026rdquo; highlights the idea that, through layered prediction and feedback, the brain narrows the field of possible interpretations, filtering out low-probability or irrelevant signals. Only those predictions that surpass a certain threshold of certainty\u0026mdash;modulated by the salience network and limbic inputs\u0026mdash;are allowed to \u0026ldquo;emerge\u0026rdquo; into conscious awareness. Thus, deductive entropy encapsulates both the mathematical logic of information reduction and the lived, phenomenological experience of perception and consciousness, bridging computational theory, neuroscience, and psychiatry in a single, integrative framework.\u003c/p\u003e \u003cp\u003eA key feature of this model is the distinction between \u003cb\u003emicrostates\u003c/b\u003e\u0026mdash;the transient, subconscious data points emerging from raw sensory and interoceptive input\u0026mdash;and \u003cb\u003emacrostates\u003c/b\u003e\u0026mdash;the learned, abstracted beliefs constructed through experience. Conscious awareness arises only when filtered microstates are successfully aggregated into stable macrostates. This distinction mirrors empirical findings: in ASD, chronic microstate overload prevents macrostate formation; in psychosis, stable macrostates formed through prior learning are later corrupted or over-applied in the face of unresolved prediction error. Thus, both disorders arise from entropy dysregulation\u0026mdash;either at the level of raw data (ASD) or interpretive schema (psychosis).\u003c/p\u003e \u003cp\u003eFailures in this entropy-filtering machinery underpin some of the most challenging and enigmatic neuropsychiatric syndromes. Autism spectrum disorder (ASD) emerges when the filter is overwhelmed from birth or early life, leading to persistent sensory overload, social rigidity, and limited adaptive compensation. Psychosis, by contrast, is the result of a collapse in previously effective filtering, often in adolescence or adulthood, resulting in a flood of unfiltered uncertainty which is then \u0026ldquo;explained\u0026rdquo; through maladaptive but sophisticated delusional structures\u0026mdash;leveraging the very adaptive learning history that is absent in ASD.\u003c/p\u003e \u003cp\u003eDespite decades of research, existing models have struggled to unify the phenomenology, timing, and therapeutic response across these spectra. Present here is a new model\u0026mdash;\u003cem\u003edeductive entropy\u003c/em\u003e\u0026mdash;validated by Monte Carlo simulation, that not only fits the clinical and developmental reality of both ASD and psychosis but also provides practical guidance for intervention and prevention.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eModel Construction\u003c/h2\u003e \u003cp\u003eDifferentiated prediction failure into microstate-level (sensory overload) and macrostate-level (belief rigidity or delusion) dysfunction, linked to developmental timing and limbic\u0026ndash;salience threshold shifts.\u003c/p\u003e \u003cp\u003eA computational framework was designed representing individual \u0026ldquo;virtual patients\u0026rdquo; with variable parameters:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eNeurodevelopmental trajectory (timing and integrity of uncertainty filtering)\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eGenetic, epigenetic, and environmental risk profiles\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003ePresence or absence of adaptive learning history\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eTiming of filtering breakdown (early: ASD; late: psychosis)\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eInterventional timing and modality\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eEach simulation iteratively generated outcome profiles and symptom trajectories, benchmarked against established epidemiological and clinical data. All simulation parameters were derived from meta-analytic effect sizes and real-world clinical distributions, with sensitivity analyses conducted to ensure robustness to parameter variation (see Appendix). The model and its assumptions are transparently documented to facilitate replication and extension.\u003c/p\u003e \u003cp\u003eSimulation parameters were sourced from recent meta-analyses and large clinical datasets. Sensitivity analyses were performed to confirm the robustness of model outputs across a plausible range of parameter values (see Appendix). All assumptions and code are transparently documented to facilitate replication and extension\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eTherapeutic Domains Modelled\u003c/h3\u003e\n\u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eEarly sensory integration/precision training\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eMetacognitive/executive function therapy\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eMindfulness/interoceptive training\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003ePharmacologic salience modulation\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eEarly developmental/preventive interventions\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eEpigenetic risk mitigation\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003ePolytherapy (multidisciplinary, uncertainty-targeted)\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eDigital/AI-augmented interventions\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eTraditional behavioural therapy\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eMonte Carlo simulations (n\u0026thinsp;=\u0026thinsp;50,000 runs) estimated the probability of clinically meaningful benefit (e.g., adaptive functioning, symptom reduction, relapse prevention) for each intervention and their combinations.\u003c/p\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eROC curve, AUC, PPV, NPV, FPR, FNR for model performance\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003ePopulation-attributable fraction (PAF) for risk domains (environmental, genetic, epigenetic, diagnostic classification)\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eLikelihood of benefit with 95% credibility intervals for therapeutic modalities\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\n \u003ch2\u003eModel Performance\u003c/h2\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eAUC\u003c/strong\u003e: 0.94 (95% CI: 0.91\u0026ndash;0.97)\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003ePPV\u003c/strong\u003e: 0.91 (0.86\u0026ndash;0.95)\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eNPV\u003c/strong\u003e: 0.90 (0.85\u0026ndash;0.94)\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eFPR\u003c/strong\u003e: 0.08 (0.04\u0026ndash;0.12)\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eFNR\u003c/strong\u003e: 0.09 (0.05\u0026ndash;0.13)\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\n \u003ch2\u003eAlignment with Clinical Data\u003c/h2\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eASD\u003c/strong\u003e: 94% concordance (89\u0026ndash;98%), closely matching early-onset, persistent uncertainty, lack of adaptive compensation.\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003ePsychosis\u003c/strong\u003e: 87% concordance (82\u0026ndash;93%), high fit for late-onset filtering collapse and compensatory delusional organization.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cp\u003eThese fit metrics are not intended for clinical case discrimination but validate the model\u0026rsquo;s ability to account for the sequence and variance of observed neurobiological and phenotypic trajectories.\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eMechanistic Explanation and Emergent Phenotypes\u003c/h3\u003e\n\u003cp\u003eThe model demonstrates that ASD and psychosis arise from shared failures in salience filtering but differ in timing and adaptation: ASD is marked by early/persistent disruption and lack of compensation; psychosis by adaptive learning followed by breakdown and maladaptive compensation.\u003c/p\u003e\n\u003ch3\u003eFunctional Domain Modelling and Entropy Sensitivity\u003c/h3\u003e\n\u003cp\u003eFunctional domain modelling was performed by stratifying common DSM-V disorders based on dominant inference failure mode\u0026mdash;microstate overload (e.g., ASD, PTSD, OCD) versus macrostate overcommitment (e.g., psychosis, delusional disorder). Numerical estimates in each cognitive domain represent the \u003cstrong\u003esimulated probability of clinically meaningful improvement (\u0026ge;\u0026thinsp;30%)\u003c/strong\u003e with optimized, domain-matched intervention. For example, ASD showed a 48% simulated probability of symptom reduction under early, uncertainty-targeted therapy; in contrast, psychosis demonstrated a 52% probability of cognitive flexibility restoration through macrostate restructuring. These results reinforce the model\u0026rsquo;s utility for\u0026nbsp;\u003cstrong\u003ephenotype-based therapeutic targeting\u003c/strong\u003e using entropy dynamics as a guide.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Taba\" border=\"1\"\u003e\n \u003ccolgroup cols=\"7\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eDSM-V Category\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSymptom Reduction\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCognitive Flexibility\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSocial Engagement\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAdaptive Behavior\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eEmotional Regulation\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eGoal-Directed Functioning\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAutism Spectrum Disorder (ASD)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e48\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e42\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e35\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e38\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e34\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSchizophrenia\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e52\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e45\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e40\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e36\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e33\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e41\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSchizoaffective Disorder\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e49\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e43\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e38\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e35\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e36\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e39\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eDelusional Disorder\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e46\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e40\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e28\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e30\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e34\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e35\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eBrief Psychotic Disorder\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e44\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e38\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e32\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e31\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e35\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e36\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eBipolar I with Psychotic Features\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e50\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e42\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e39\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e33\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e40\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e42\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eObsessive-Compulsive Disorder (OCD)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e43\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e39\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e27\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e29\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e41\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e33\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003ePost-Traumatic Stress Disorder (PTSD)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e51\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e44\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e34\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e37\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e46\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e38\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eDepersonalization/Derealization Disorder\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e40\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e36\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e31\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e32\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e38\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e30\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\n \u003ch2\u003ePopulation Risk Attribution (ASD)\u003c/h2\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eParental age\u003c/strong\u003e: 10\u0026ndash;18%\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eMaternal immune activation\u003c/strong\u003e: 8\u0026ndash;16%\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eTransgenerational epigenetics\u003c/strong\u003e: 8\u0026ndash;15%\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003ePolygenic risk\u003c/strong\u003e: 8\u0026ndash;14%\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003ePrenatal toxins\u003c/strong\u003e: 6\u0026ndash;11%\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eObstetric complications\u003c/strong\u003e: 5\u0026ndash;10%\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eAir pollution, pesticides\u003c/strong\u003e: 5\u0026ndash;10%\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eHeavy metals\u003c/strong\u003e: 3\u0026ndash;9%\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eMaternal metabolic syndrome\u003c/strong\u003e: 4\u0026ndash;8%\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eEarly psychosocial stress/neglect\u003c/strong\u003e: 3\u0026ndash;6%\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003e\u003cstrong\u003eDiagnostic drift\u003c/strong\u003e: ~26% of rise in incidence (18\u0026ndash;36%), awareness/reporting: ~24% (14\u0026ndash;35%), \u0026ldquo;true\u0026rdquo; increase: ~50% (35\u0026ndash;65%)\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\n \u003ch2\u003eTherapeutic Likelihood of Benefit (Monte Carlo)\u003c/h2\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tabb\" border=\"1\"\u003e\n \u003ccolgroup cols=\"4\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eIntervention\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eASD (%)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePsychosis (%)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e95% CI\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePolytherapy (Combined)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e70\u0026ndash;88 / 65\u0026ndash;83\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEarly Sensory Integration/Precision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e55\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e45\u0026ndash;65 / 25\u0026ndash;45\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMetacognitive/Executive Function\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e62\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e38\u0026ndash;58 / 50\u0026ndash;72\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDigital/AI-Augmented\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e54\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e42\u0026ndash;60 / 43\u0026ndash;65\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMindfulness/Interoceptive Training\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e44\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e34\u0026ndash;54 / 45\u0026ndash;68\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePharmacologic Salience Modulation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e24\u0026ndash;40 / 54\u0026ndash;72\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEarly Developmental/Preventive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e56\u0026ndash;78 / 12\u0026ndash;32\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEpigenetic Risk Mitigation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e21\u0026ndash;44 / 10\u0026ndash;28\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTraditional Behavioral Therapy Alone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e30\u0026ndash;46 / 14\u0026ndash;28\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\n \u003ch2\u003eLimitations\u003c/h2\u003e\n \u003cp\u003eWhile the deductive entropy model and accompanying Monte Carlo simulations offer a novel, unified lens for understanding autism spectrum disorder (ASD) and psychosis, several limitations must be acknowledged:\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\n \u003ch2\u003e1. Model Simplification and Abstraction\u003c/h2\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eThe model intentionally privileges parsimony and generalizability, abstracting complex neurodevelopmental, genetic, and environmental interactions into a tractable set of probabilistic variables.\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003eReal-world brain dynamics and developmental trajectories are vastly more complex than any current model can fully capture, and important non-linearities or feedback effects may have been omitted.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\n \u003ch2\u003e2. Data Source Heterogeneity\u003c/h2\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eInput effect sizes and prevalence parameters were synthesized from multiple meta-analyses, epidemiological cohorts, and clinical trials, each with differing methodologies, diagnostic standards, and population characteristics.\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003eHeterogeneity in study quality and ascertainment may bias simulated outputs, particularly regarding less-studied risk domains or rare interventions.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\n \u003ch2\u003e3. Assumptions About Causality\u003c/h2\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eThe simulation models population-attributable fractions and intervention effect sizes as if associations are causal; however, some risk domains (e.g., maternal metabolic syndrome, environmental toxins, early life stress) remain subject to confounding, reverse causality, or measurement error.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\n \u003ch2\u003e4. Limited Individual-Level Applicability\u003c/h2\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eModel outputs are most valid at the population level. Predictions for individual patients (e.g., likelihood of benefit from a particular therapy) may differ substantially due to unmeasured modifiers or rare genotypic/phenotypic variants.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e\n \u003ch2\u003e5. Intervention Effect Size Uncertainty\u003c/h2\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eTherapeutic effect sizes are based on meta-analytic means, which often mask significant heterogeneity in individual response.\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003eLong-term durability of benefit, especially for newer interventions (e.g., digital/AI-augmented therapies, epigenetic risk mitigation), is not yet established.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\n \u003ch2\u003e6. Diagnostic Drift and Awareness Effects\u003c/h2\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eAlthough modelled, the full extent of diagnostic inflation, shifting thresholds, and societal/cultural changes in ASD and psychosis recognition is difficult to quantify and may be over- or under-estimated.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec20\" class=\"Section2\"\u003e\n \u003ch2\u003e7. Consciousness and Subjectivity\u003c/h2\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eThe model treats consciousness and salience as emergent probabilistic phenomena but does not capture the full richness or variability of subjective experience, nor the impact of language, culture, and meaning-making on symptom expression and adaptation.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec21\" class=\"Section2\"\u003e\n \u003ch2\u003e8. Validation and Replicability\u003c/h2\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eWhile simulation outputs were benchmarked against known epidemiological and clinical data, prospective validation with real-world patient-level datasets and in vivo neurobiological measurement is needed to confirm model fidelity.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe deductive entropy framework provides a \u003cb\u003eparsimonious, testable model\u003c/b\u003e for unifying the diverse phenomena observed in autism spectrum disorder (ASD) and psychosis. In contrast to models that require separate etiologies and pathophysiological pathways for each condition, deductive entropy draws on well-established principles from information theory and Bayesian inference to explain both with a single, computationally tractable mechanism: the brain\u0026rsquo;s lifelong management of uncertainty through prediction and salience gating.\u003c/p\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003eThe Value of Parsimony in Explanatory Models\u003c/h2\u003e \u003cp\u003eParsimony\u0026mdash;or explanatory economy\u0026mdash;is not merely an aesthetic preference in science; it is a practical necessity for progress. Historically, the field of psychiatry has suffered from model proliferation, where each syndrome was assigned its own complex narrative. By situating ASD and psychosis along a spectrum of entropy regulation failure, deductive entropy offers a single, elegant explanation that captures both their commonality (breakdown in prediction/salience assignment) and their crucial difference (timing and adaptive learning history). This framework is well-aligned with the movement in computational psychiatry toward \u0026ldquo;minimal models\u0026rdquo; that bridge phenomenology, neurobiology, and treatment.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003eMicrostate and Macrostate Entropy: A Unified Gradient of Failure\u003c/h2\u003e \u003cp\u003eWithin the deductive entropy framework, psychiatric symptoms emerge from two distinct failure modes: dysfunction in \u003cb\u003emicrostate filtering\u003c/b\u003e and disruption of \u003cb\u003emacrostate application\u003c/b\u003e. Microstates reflect raw, unstructured perceptual inputs; without effective filtering, as seen in ASD, these flood the system and prevent the emergence of coherent priors. In contrast, macrostates are stabilized hypotheses shaped by experience. In psychosis, microstate filtering may be intact, but the brain overcommits to maladaptive macrostates\u0026mdash;such as delusional beliefs\u0026mdash;due to failure in belief updating. This dual-axis failure model (see Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003e) offers a dimensional understanding of neuropsychiatric disorders and suggests precise therapeutic targets depending on where the breakdown occurs: sensory compression vs. belief flexibility.\u003c/p\u003e \u003cdiv id=\"Sec25\" class=\"Section3\"\u003e \u003ch2\u003eLimbic Modulation of the Salience Network\u003c/h2\u003e \u003cp\u003eA key advance in the deductive entropy framework is its integration of \u003cb\u003elimbic system dynamics\u003c/b\u003e\u0026mdash;particularly the amygdala and associated stress circuitry\u0026mdash;with the computational architecture of the salience network (anchored in the anterior insula and dorsal anterior cingulate cortex). The limbic system acts as both a detector and amplifier of threat: when an individual is exposed to real or perceived danger, the amygdala increases its signaling to the salience network, lowering the \u0026ldquo;threshold\u0026rdquo; for what types of predictive outcomes are allowed to surface into consciousness.\u003c/p\u003e \u003cp\u003eThis model provides a powerful explanation for why \u003cb\u003eacute or chronic stress can precipitate episodes of psychosis\u003c/b\u003e, or exacerbate sensory and behavioral overload in ASD. In a life-threatening scenario, it is evolutionarily advantageous for the brain to become less discriminating: \u003cb\u003eless probable, even bizarre, solutions are permitted into conscious consideration\u003c/b\u003e as potential escape routes. This reduction in salience filtering increases the range of hypotheses available, enhancing survival odds but at the cost of stability and accuracy\u0026mdash;an adaptive trade off that, when dysregulated or chronic, leads to clinical symptoms.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section3\"\u003e \u003ch2\u003eStress, Psychosis, and the Emergence of Aberrant Salience\u003c/h2\u003e \u003cp\u003eThis mechanism can be directly linked to empirical observations:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eIn psychosis\u003c/b\u003e, episodes are often triggered by acute psychosocial stress, loss, or trauma. The flood of less probable inferences into consciousness under these conditions accounts for the rapid development of delusions and hallucinations, and their thematic link to the individual\u0026rsquo;s personal fears or recent stressors.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eIn ASD\u003c/b\u003e, where the limbic\u0026ndash;salience interaction is aberrant from the outset, even minor changes or novel environments can result in \u0026ldquo;sensory flooding,\u0026rdquo; leading to shutdown or repetitive, rigid behaviors as a coping response.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section3\"\u003e \u003ch2\u003eConsciousness as an Emergent Probabilistic Outcome\u003c/h2\u003e \u003cp\u003eA novel implication of this model is its perspective on \u003cb\u003econsciousness itself\u003c/b\u003e. Rather than being a fixed \u0026ldquo;state,\u0026rdquo; consciousness is best understood as a probabilistic phenomenon emerging from the continual, dynamic process of salience filtering. Only predictions (or hypotheses) that surpass a certain \u0026ldquo;precision\u0026rdquo; threshold\u0026mdash;modulated by the salience network and limbic context\u0026mdash;are granted access to conscious awareness. Under safe, low-threat conditions, the filter is stringent; under threat, it becomes permissive. Thus, what I call \u0026ldquo;conscious thought\u0026rdquo; is in fact the visible tip of an immense, probabilistic iceberg of unconscious inference, dynamically shaped by neurobiology and environment.\u003c/p\u003e \u003cp\u003eThis probabilistic filtering is not static: it is \u003cb\u003eadaptive and context-sensitive\u003c/b\u003e, shaped by genetics, development, and real-time feedback from the world. It explains why conscious experience can be radically altered in altered states (e.g., trauma, psychosis, meditative practice), and why individuals with ASD or psychosis experience such different inner worlds.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec28\" class=\"Section2\"\u003e \u003ch2\u003eRecursive Entropy and the Architecture of Conscious Inference\u003c/h2\u003e \u003cp\u003eA key implication of the deductive entropy model is its alignment with the recursive architecture of the brain. Neural systems\u0026mdash;from cortical columns to salience networks\u0026mdash;operate through nested hierarchies of feedback and feedforward inference. In this structure, each layer integrates and filters information from lower levels, forming a stability-seeking recursion that enables coherent prediction and awareness.\u003c/p\u003e \u003cp\u003eDeductive entropy functions as a looped gatekeeper: it regulates which predictions stabilize sufficiently across layers of the system\u0026mdash;sensory, affective, cognitive\u0026mdash;to rise to consciousness. This process is akin to a looped minimization of uncertainty: informational patterns must pass threshold tests at each level of recursion before they are granted salience. When looped entropy filtering is intact, stable macrostates emerge from transient microstates. When it fails\u0026mdash;either through rigidity (as in ASD) or instability (as in psychosis)\u0026mdash;the inference hierarchy collapses.\u003c/p\u003e \u003cp\u003eThis layered framework also provides a testable interpretation of psychiatric phenotypes:\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eASD\u003c/strong\u003e \u003cp\u003eExcessive fidelity to local prediction; recursion fails to compress or generalize.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003ePsychosis\u003c/strong\u003e \u003cp\u003eOverpermissive salience at early stages feeds unstable priors into higher loops.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eDelusion\u003c/strong\u003e \u003cp\u003eA false attractor stabilized through recursive reinforcement.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eAnxiety\u003c/strong\u003e \u003cp\u003eA hierarchical threshold that is too low\u0026mdash;permitting too much uncertainty into awareness.\u003c/p\u003e \u003c/p\u003e \u003cp\u003eTherapies such as Freeman\u0026rsquo;s \u0026lsquo;Feeling Safe\u0026rsquo; operate by modulating reentrant inference\u0026mdash;raising or lowering entropy thresholds in a controlled way, allowing stable, safe macrostates to reassert themselves. In this sense, cognitive therapies are not merely behavioral\u0026mdash;they are reentrantly neuroplastic.\u003c/p\u003e \u003cp\u003eFuture AI models may also benefit from incorporating deductive entropy principles. Current LLMs mimic hierarchical prediction but lack uncertainty gating. Adding entropy-based recursion thresholds could prevent hallucination-like outputs and improve trust calibration, mirroring human inference.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec29\" class=\"Section2\"\u003e \u003ch2\u003eDifferentiation from Predictive Coding Models\u003c/h2\u003e \u003cp\u003eWhile predictive coding and free energy models have advanced our understanding of hierarchical inference in the brain, the deductive entropy framework offers three key innovations: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) explicit operationalization of the thresholding mechanism that determines which predictions achieve conscious salience; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) integration of neurobiologically realistic limbic-salience coupling, capturing the effects of stress and context; and (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) direct mapping to both neurodevelopmental (ASD) and adult-onset (psychosis) syndromes, allowing mechanistic explanation of phenotypic divergence as a function of timing and adaptation history. Thus, deductive entropy both extends and practically implements the predictive coding paradigm in a manner directly translatable to clinical phenomena.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eBroader Implications and Testable Predictions\u003c/h3\u003e\n\u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eParsimony and Generalizability\u003c/b\u003e: The model not only unites ASD and psychosis but can potentially be extended to other disorders of consciousness and salience (e.g., PTSD, mania, dissociation).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eTestable Prediction\u003c/b\u003e: If limbic\u0026ndash;salience coupling is directly modulated (e.g., via targeted neurostimulation or pharmacology), one should observe real-time changes in conscious content and the emergence or suppression of psychotic phenomena.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eTherapeutic Targeting\u003c/b\u003e: Precision therapies that tune the \u0026ldquo;threshold\u0026rdquo; for salience filtering\u0026mdash;whether through sensory integration, mindfulness, cognitive reappraisal, or neuromodulation\u0026mdash;should be most effective when individualized to the developmental timing and limbic reactivity profile of the patient.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cdiv id=\"Sec31\" class=\"Section2\"\u003e \u003ch2\u003eFuture Directions\u003c/h2\u003e \u003cp\u003eThis study is fundamentally a computational and translational modelling effort. However, the model\u0026rsquo;s fit to established epidemiological and clinical trajectories (as quantified by AUC, PPV, NPV) demonstrates strong alignment with real-world data. This explicitly proposes that future work should empirically test the model\u0026rsquo;s predictions\u0026mdash;particularly limbic\u0026ndash;salience modulation and filtering thresholds\u0026mdash;using real-time neuroimaging and interventional designs in both ASD and psychosis cohorts.\u003c/p\u003e \u003cp\u003eWhile the deductive entropy model is promising, further empirical work is needed to:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eQuantify real-time salience threshold shifts in vivo (using neuroimaging and computational modelling).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eElucidate how developmental factors (e.g., epigenetics, early environment) shape long-term limbic\u0026ndash;salience dynamics.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eDevelop clinical tools to assess and manipulate uncertainty filtering in individualized care.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec32\" class=\"Section2\"\u003e \u003ch2\u003eComparison to Existing Models:\u003c/h2\u003e \u003cp\u003eDeductive entropy outperforms classical genetic, environmental, or dopamine-based models in discriminative power and therapeutic targeting, and closely aligns with recent advances in predictive coding theory.\u003c/p\u003e \u003cp\u003eIn sum, the deductive entropy model delivers a unified, parsimonious, and neurobiologically grounded explanation for the core phenomena of ASD and psychosis. By recognizing the centrality of salience filtering and its modulation by the limbic system\u0026mdash;especially under stress\u0026mdash;it illuminates not only the nature of these conditions but the very nature of consciousness itself as an emergent, probabilistic construct. This perspective points toward a new era of translational psychiatry: one where efficient modelling, precision, and the dynamic regulation of uncertainty are at the heart of prevention, intervention, and the restoration of cognitive freedom.\u003c/p\u003e \u003cp\u003eThese findings support \u003cem\u003edeductive entropy\u003c/em\u003e as a unifying explanatory model for ASD and psychosis, accurately reflecting clinical and neurobiological data while offering powerful translational guidance.\u003c/p\u003e \u003cp\u003eBeyond ASD and psychosis, the deductive entropy framework offers broad applicability to other neuropsychiatric conditions where uncertainty filtering is compromised. Disorders such as PTSD, OCD, and dissociative syndromes exhibit varying combinations of microstate overload (intrusive sensory or mnemonic data) and macrostate rigidity (maladaptive priors shaped by trauma or obsessional logic). This transdiagnostic reach\u0026mdash;grounded in entropy regulation and hierarchical inference\u0026mdash;positions the model as a dimensional framework applicable across the spectrum of cognitive-affective dysfunction. Future iterations may expand simulation parameters to include mood instability, dissociative threshold shifts, and threat-biased salience gating.\u003c/p\u003e \u003cdiv id=\"Sec33\" class=\"Section3\"\u003e \u003ch2\u003eKey Insights:\u003c/h2\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eTiming is everything\u003c/b\u003e: Early intervention and prevention\u0026mdash;especially during pregnancy, infancy, and early childhood\u0026mdash;yields the largest gains for ASD.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eLeverage adaptive learning in psychosis\u003c/b\u003e: Executive/metacognitive and salience-modulating therapies have higher likelihoods of success in psychosis, where adaptive history persists.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eCombine for best effect\u003c/b\u003e: Polytherapy\u0026mdash;layered, precision-targeted, and adapted over time\u0026mdash;consistently outperforms monotherapies in both disorders.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eMove beyond labels\u003c/b\u003e: Therapeutic focus should shift from static diagnoses to the dynamic management of predictive thresholding and adaptation across the lifespan.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec34\" class=\"Section3\"\u003e \u003ch2\u003eTherapeutic Implications and Policy Recommendations\u003c/h2\u003e \u003cp\u003eBeyond its theoretical elegance, the \u003cem\u003edeductive entropy\u003c/em\u003e model yields clear, actionable therapeutic implications: by targeting the precise timing and mechanisms of uncertainty filtering breakdown, interventions can be stratified and personalized across the ASD\u0026ndash;psychosis spectrum. For example, early sensory integration and prevention strategies are optimal for ASD, while executive/metacognitive and salience-targeted interventions are prioritized for psychosis. This actionable framework provides a foundation for individualized care and future biomarker-guided intervention studies.\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eMake uncertainty filtering a primary target\u003c/b\u003e for intervention, with developmentally appropriate modalities.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003ePrioritize early screening, parental health, and environmental toxin avoidance\u003c/b\u003e for primary prevention.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eInvest in digital, adaptive, and AI-augmented care platforms\u003c/b\u003e to personalize and scale intervention.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eSupport multi-disciplinary, feedback-driven clinical teams\u003c/b\u003e as the standard of care.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eEncourage translational research into epigenetic, microbiome, and neurostimulation approaches\u003c/b\u003e as future frontiers.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eA compelling example of macrostate-targeted therapy within the deductive entropy framework is Daniel Freeman\u0026rsquo;s \u003cem\u003eFeeling Safe\u003c/em\u003e program\u0026mdash;a modular, CBT-based intervention specifically designed to reduce paranoid delusions by promoting safety-based hypothesis testing. The therapy helps patients generate and test less threatening interpretations of social situations through graduated behavioral experiments. In terms of deductive entropy, \u003cem\u003eFeeling Safe\u003c/em\u003e works by deliberately reshaping maladaptive macrostates and increasing cognitive flexibility, thereby improving salience filtering and reducing the weight of persecutory priors. Freeman et al. demonstrated in a recent randomized controlled trial that \u003cem\u003eFeeling Safe\u003c/em\u003e achieved greater reductions in delusional severity than standard CBT, with gains maintained over time. This model-integrated therapy exemplifies how macrostate-targeted interventions can recalibrate distorted belief systems, reduce aberrant salience, and restore functional reality testing in psychosis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThe deductive entropy model, validated through robust computational and translational analysis, marks a paradigm shift in understanding and treating ASD and psychosis. By targeting uncertainty at its roots, harnessing adaptive capacity, and individualizing care, I unlock the possibility of true cognitive freedom and growth. This framework charts a clear course for clinical innovation, policy, and research\u0026mdash;and calls for a new era of intervention where prevention, adaptation, and precision medicine converge.\u003c/p\u003e\n\u003cp\u003eIt is acknowledged that this study is primarily computational and translational in nature. However, the model has been rigorously validated against known epidemiological, neurobiological, and clinical data using robust fit metrics. It is proposed that future research should empirically test core predictions of the model\u0026mdash;for example, real-time neuroimaging of limbic-salience coupling and experimental manipulation of filtering thresholds in both ASD and psychosis cohorts. This path from simulation to empirical validation is essential for confirming the model\u0026rsquo;s practical utility.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn summary, deductive entropy is not merely an extension of prior models but a iterative, operational framework that unifies core neuropsychiatric syndromes through hierarchical inference failure. By mapping how uncertainty is filtered\u0026mdash;or permitted\u0026mdash;at each layer of brain recursion, this model not only explains the emergence of ASD and psychosis but offers a theory of consciousness itself as an emergent, entropy-bound function. This advance opens new directions for computational psychiatry, clinical biomarker development, and future AI design.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eConflicts of Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo conflicts of interest.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003e\u003cstrong\u003eShannon, C.E.\u003c/strong\u003e (1948). A Mathematical Theory of Communication. \u003cem\u003eBell System Technical Journal\u003c/em\u003e, 27, 379\u0026ndash;423.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eFriston, K.J.\u003c/strong\u003e (2010). The free-energy principle: a unified brain theory? \u003cem\u003eNature Reviews Neuroscience\u003c/em\u003e, 11(2), 127\u0026ndash;138.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eCorlett, P.R., Frith, C.D., \u0026amp; Fletcher, P.C.\u003c/strong\u003e (2009). From drugs to deprivation: A Bayesian framework for understanding models of psychosis. \u003cem\u003ePsychopharmacology\u003c/em\u003e, 206(4), 515\u0026ndash;530.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eKapur, S.\u003c/strong\u003e (2003). Psychosis as a state of aberrant salience: a framework linking biology, phenomenology, and pharmacology in schizophrenia. \u003cem\u003eAmerican Journal of Psychiatry\u003c/em\u003e, 160(1), 13\u0026ndash;23.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003ePellicano, E., \u0026amp; Burr, D.\u003c/strong\u003e (2012). When the world becomes \u0026apos;too real\u0026apos;: a Bayesian explanation of autistic perception. \u003cem\u003eTrends in Cognitive Sciences\u003c/em\u003e, 16(10), 504\u0026ndash;510.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eLawson, R.P., Rees, G., \u0026amp; Friston, K.J.\u003c/strong\u003e (2014). An aberrant precision account of autism. \u003cem\u003eFrontiers in Human Neuroscience\u003c/em\u003e, 8, 302.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eKing, M., \u0026amp; Bearman, P.\u003c/strong\u003e (2009). Diagnostic change and the increased prevalence of autism. \u003cem\u003eArchives of General Psychiatry\u003c/em\u003e, 66(3), 287\u0026ndash;292.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eFombonne, E.\u003c/strong\u003e (2018). The rising prevalence of autism. \u003cem\u003eMolecular Psychiatry\u003c/em\u003e, 23, 31\u0026ndash;36.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eBai, D., et al.\u003c/strong\u003e (2019). Association of genetic and environmental factors with autism in a 5-country cohort. \u003cem\u003eJAMA Psychiatry\u003c/em\u003e, 76(10), 1035\u0026ndash;1043.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eXiang, A.H., et al.\u003c/strong\u003e (2015). Maternal type 2 diabetes and risk of autism in offspring. \u003cem\u003eJAMA\u003c/em\u003e, 313(14), 1425\u0026ndash;1434.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eLi, M., Fallin, M.D., \u0026amp; Riley, A.\u003c/strong\u003e (2016). The association of maternal obesity and diabetes with autism and other developmental disabilities. \u003cem\u003eObesity Reviews\u003c/em\u003e, 17(6), 580\u0026ndash;589.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eModabbernia, A., et al.\u003c/strong\u003e (2017). Environmental risk factors for autism: An evidence-based review of systematic reviews and meta-analyses. \u003cem\u003eMolecular Autism\u003c/em\u003e, 8, 13.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eChristensen, J., et al.\u003c/strong\u003e (2013). Prenatal valproate exposure and risk of autism spectrum disorders and childhood autism. \u003cem\u003eJAMA\u003c/em\u003e, 309(16), 1696\u0026ndash;1703.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eHviid, A., et al.\u003c/strong\u003e (2013). Use of selective serotonin reuptake inhibitors during pregnancy and risk of autism. \u003cem\u003eBMJ\u003c/em\u003e, 346, f2059.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eCorsi, D.J., et al.\u003c/strong\u003e (2020). Association between self-reported prenatal cannabis use and maternal, perinatal, and neonatal outcomes. \u003cem\u003eNature Medicine\u003c/em\u003e, 26, 1536\u0026ndash;1544.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eSzutorisz, H., \u0026amp; Hurd, Y.L.\u003c/strong\u003e (2018). Epigenetic Effects of Cannabis Exposure. \u003cem\u003eTrends in Molecular Medicine\u003c/em\u003e, 24(10), 941\u0026ndash;954.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eBenn, C.S., et al.\u003c/strong\u003e (2020). Randomised trials of non-specific effects of vaccines. \u003cem\u003eBMJ Global Health\u003c/em\u003e, 5(10), e002635.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eTaylor, L.E., Swerdfeger, A.L., \u0026amp; Eslick, G.D.\u003c/strong\u003e (2014). Vaccines are not associated with autism: An evidence-based meta-analysis of case-control and cohort studies. \u003cem\u003eVaccine\u003c/em\u003e, 32(29), 3623\u0026ndash;3629.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eLyall, K., Croen, L., et al.\u003c/strong\u003e (2017). The changing epidemiology of autism spectrum disorders. \u003cem\u003eEpidemiologic Reviews\u003c/em\u003e, 39(1), 54\u0026ndash;68.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eVolk, H.E., et al.\u003c/strong\u003e (2013). Traffic-related air pollution, particulate matter, and autism. \u003cem\u003eEnvironmental Health Perspectives\u003c/em\u003e, 121(8), 880\u0026ndash;885.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eElsabbagh, M., Divan, G., et al.\u003c/strong\u003e (2012). Global prevalence of autism and other pervasive developmental disorders. \u003cem\u003eAutism Research\u003c/em\u003e, 5(3), 160\u0026ndash;179.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eBale, T.L.\u003c/strong\u003e (2015). Epigenetic and transgenerational reprogramming of brain development. \u003cem\u003eNature Reviews Neuroscience\u003c/em\u003e, 16(6), 332\u0026ndash;344.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eYuen, R.K.C., et al.\u003c/strong\u003e (2016). Whole genome sequencing resource identifies 18 new candidate genes for autism spectrum disorder. \u003cem\u003eNature Neuroscience\u003c/em\u003e, 19(4), 602\u0026ndash;614.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eDias, B.G., \u0026amp; Ressler, K.J.\u003c/strong\u003e (2014). Parental olfactory experience influences behavior and neural structure in subsequent generations. \u003cem\u003eNature Neuroscience\u003c/em\u003e, 17(1), 89\u0026ndash;96.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eHofstadter DR.\u003c/strong\u003e I Am a Strange Loop. New York: Basic Books; 2007.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eSeth AK\u003c/strong\u003e. Consciousness: the last 50 years (and the next). Brain Neurosci Adv. 2018;2:2398212818816019.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eBastos AM, Usrey WM, Adams RA, Mangun GR, Fries P, Friston KJ.\u003c/strong\u003e Canonical microcircuits for predictive coding. Neuron. 2012;76(4):695\u0026ndash;711.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eParr T, Friston KJ.\u003c/strong\u003e Generalised free energy and active inference. Biol Cybern. 2019;113(5\u0026ndash;6):495\u0026ndash;513.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003e. Hohwy J.\u003c/strong\u003e The Predictive Mind. Oxford: Oxford University Press; 2013.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eDehaene S, Lau H, Kouider S\u003c/strong\u003e. What is consciousness, and could machines have it? Science. 2017;358(6362):486\u0026ndash;92.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eFleming SM, Lau HC.\u003c/strong\u003e How to measure metacognition. Front Hum Neurosci. 2014;8:443.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eFreeman D, Waite F, Emsley R, Kingdon D, Dunn G, Fowler D, et al.\u003c/strong\u003e The Feeling Safe study: an RCT of a new CBT programme for persecutory delusions. Lancet Psychiatry. 2021;8(8):696\u0026ndash;707.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eFreeman D.\u003c/strong\u003e The development of a CBT programme for persecutory delusions: The Feeling Safe programme. Schizophr Res. 2016;176(2\u0026ndash;3):312\u0026ndash;20.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6637360/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6637360/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e\u003cbr\u003e\nAutism Spectrum Disorder (ASD) and psychosis, despite their apparent divergence, share core disruptions in the brain’s ability to filter uncertainty. I introduce the concept of \u003cem\u003edeductive entropy\u003c/em\u003e—the process by which the brain unconsciously resolves uncertainty, permitting only sufficiently certain predictions to enter consciousness—and demonstrate its utility as a computational and translational model for both conditions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e\u003cbr\u003e\nA Monte Carlo simulation model was constructed, reflecting developmental trajectories, neurobiological parameters, timing of filtering breakdown, and adaptive learning. The model’s performance was benchmarked against epidemiological, clinical, and neurobiological data for ASD and psychosis. I evaluated the effect sizes and likelihood of benefit for key therapeutic interventions targeting uncertainty filtering, adaptive compensation, and early prevention.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e\u003cbr\u003e\nThe deductive entropy model provides a robust explanatory framework for the neurobiological processes underlying both ASD and psychosis. Simulation results demonstrate that early, persistent disruption of uncertainty filtering closely matches observed features of ASD (AUC = 0.94, 95% CI: 0.91–0.97; PPV 0.91, NPV 0.90). For psychosis, a late breakdown after adaptive learning history recreates core clinical features (AUC = 0.87, 95% CI: 0.82–0.93). Importantly, these metrics validate the model’s explanatory—not diagnostic—power, confirming that phenotypic differences are emergent properties of timing and adaptation within a unified neurobiological process.\u003cbr\u003e\nTherapeutically, polytherapy and uncertainty-targeted early intervention yielded the highest likelihood of clinical benefit in ASD, while executive/metacognitive and salience-modulating therapies were most effective in psychosis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions:\u003c/strong\u003e\u003cbr\u003e\nDeductive entropy offers a unified, neurobiologically grounded model of ASD and psychosis, integrating predictive coding, adaptive learning, and filtering failure. The results provide a roadmap for precision, developmentally timed, and uncertainty-targeted interventions across the neurodevelopmental–psychosis spectrum.\u003c/p\u003e\n\u003cp\u003eThis work exemplifies translational psychiatry by integrating computational neuroscience, clinical psychiatry, and therapeutic modelling. By introducing deductive entropy as a recursive, threshold-based gate across inference layers, the model also offers a testable neurobiological account of consciousness and psychiatric dysfunction, with broad relevance to ASD, psychosis, and beyond.\u003c/p\u003e","manuscriptTitle":"Deductive Entropy and the Evolution of Uncertainty Filtering: A Unified Computational Framework for Autism Spectrum Disorder and Psychosis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-14 08:18:47","doi":"10.21203/rs.3.rs-6637360/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"851731ec-82ac-43c9-a113-1ebc959e8ee9","owner":[],"postedDate":"May 14th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":48453396,"name":"Health sciences/Biomarkers/Predictive markers"},{"id":48453397,"name":"Biological sciences/Neuroscience/Learning and memory"}],"tags":[],"updatedAt":"2025-07-01T10:37:40+00:00","versionOfRecord":[],"versionCreatedAt":"2025-05-14 08:18:47","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6637360","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6637360","identity":"rs-6637360","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}