Psychiatry Companion Paper¶
Jean-Paul Niko · February 2026
\begin{center} {\LARGE\bfseries\color{sectionblue} Intelligence Filters, Affective Geometry, and [3pt] Cognitive Thermodynamics: [3pt] A Mathematical Framework for Psychiatric [3pt] Diagnosis and Treatment Response}
{\large Jean-Paul Niko} [4pt] {} [2pt] {\smallniko@triptomean.com} [12pt] {\small February 2026} \end{center}
Abstract
Contemporary psychiatric diagnosis relies on categorical systems (DSM-5, ICD-11) that poorly capture the dimensional, overlapping, and dynamic nature of mental disorders. We present a mathematical framework---drawn from the "Intelligence as Geometry" (RTSG) program---that models psychiatric conditions as characteristic perturbations of a multi-dimensional cognitive architecture. Each individual carries an variable-dimensional intelligence vector (n=12 for humans) \(\bI \in \\mathbb{R}^{n(e)}\) representing capacities across symbolic, spatial, linguistic, social, mnemonic, auditory, kinesthetic, and evaluative domains. Psychiatric conditions are formalized as filter operators \(\bF\) that distort this vector in disorder-specific patterns: depression attenuates evaluative and mnemonic components while amplifying narrow symbolic loops (rumination); PTSD hyperactivates evaluative gain (hypervigilance) while distorting mnemonic temporal indexing; mania corresponds to high-velocity, high-curvature trajectories through affective space. The framework introduces cognitive thermodynamics---entropy, temperature, free energy, and phase transitions for mental states---providing a mathematical account of treatment resistance (deep energy basins), temporary worsening before improvement (thermodynamic necessity), and the impossibility of "removing" trauma (filter irreversibility). The compatibility matrix \(\bK\) predicts treatment interaction effects and optimal sequencing. We derive 15 testable clinical hypotheses, each specified with the measurement modality (fMRI, EEG, EMA, psychometric) required for validation. The framework is dimensional, quantitative, transdiagnostic, and aligned with the Research Domain Criteria (RDoC) initiative.
% ═══════════════════════════════════════════════════════════════════════════════
Introduction: Beyond Categorical Diagnosis¶
% ═══════════════════════════════════════════════════════════════════════════════
This paper presents a mathematical framework for psychiatry derived from the Intelligence as Geometry (RTSG) program. The core insight is that psychiatric conditions are not categories to be assigned but geometric distortions of a multi-dimensional cognitive profile that can be characterized, measured, and predicted with algebraic precision.
The limitations of categorical diagnosis are well documented. The DSM-5 treats disorders as discrete kinds, yet comorbidity rates routinely exceed 50%, suggesting that the boundaries between categories are artifacts of the classification system rather than features of the underlying pathology [caspi2014,kotov2017]. The Research Domain Criteria (RDoC) initiative [insel2010] called for dimensional approaches grounded in neuroscience, but has struggled to provide the mathematical formalism that would make "dimensional" more than a slogan.
Our framework provides exactly this formalism. We model each individual's cognitive architecture as an variable-dimensional intelligence vector \(\bI \in \\mathbb{R}^{n(e)}\), where each component represents capacity in a distinct cognitive domain. Psychiatric conditions are then formalized as filter operators---linear or nonlinear maps \(\bF: \\mathbb{R}^{n(e)} \to \\mathbb{R}^{n(e)}\)---that distort the intelligence vector in disorder-specific patterns. This yields:
[nosep] - Dimensional diagnosis: An 8D profile, not a binary label. - Comorbidity explained: Disorders with overlapping filter signatures naturally co-occur---high comorbidity is predicted by the framework. - Treatment interaction: The compatibility matrix \(\bK\) predicts which treatment combinations amplify or interfere. - Treatment resistance: Formalized as the depth of a free-energy basin, with quantitative implications for intervention design. - Trauma irreversibility: A theorem, not a metaphor---filters compose but do not uncommit.
We develop each of these in turn, deriving 15 testable clinical hypotheses along the way. The paper is self-contained but draws on the full RTSG mathematical apparatus detailed in the master treatise [niko2026].
% ═══════════════════════════════════════════════════════════════════════════════
The Intelligence Vector as Clinical Cognitive Profile¶
% ═══════════════════════════════════════════════════════════════════════════════
Intelligence Vector
\tA
An individual's cognitive profile is an variable-dimensional vector: \begin{keyeq} [ \bI = (I_{\mathrm{symb}},\; I_{\mathrm{spat}},\; I_{\mathrm{ling}},\; I_{\mathrm{soc}},\; I_{\mathrm{mnem}},\; I_{\mathrm{aud}},\; I_{\mathrm{kin}},\; I_{\mathrm{eval}}) \in \mathbb{R}^{n(e)} ] \end{keyeq} Each component \(I_t\) represents capacity in a cognitive type, measured in standardized units ("cogs") calibrated via pairwise ELO tournaments on type-specific tasks.
The eight types correspond to distinct but interacting cognitive systems:
\begin{center} \small
[Table — see PDF for formatted version]
\end{center}
Critically, these types are not independent. They interact through the compatibility matrix \(\bK\).
Compatibility Matrix
\tA
The symmetric matrix \(\bK \in \R^{8 \times 8}\) encodes pairwise interactions between intelligence types: \begin{keyeq} [ K_{st} \begin{cases}
1 & \text{types \(s\) and \(t\) amplify each other (positive coupling)}
= 1 & \text{independent}
< 1 & \text{types \(s\) and \(t\) interfere (inhibitory coupling)} \end{cases} ] \end{keyeq} The diagonal entries \(K_{tt} = 1\). Off-diagonal entries are estimated from cognitive training transfer studies, neuroimaging functional connectivity, and psychometric cross-loadings.
For clinical purposes, a patient's profile is not the raw vector \(\bI_{\mathrm{raw}}\) but the effective vector after all filters have been applied: \(\bI_{\mathrm{eff}} = \bF(\bI_{\mathrm{raw}})\). The filter formalism makes this precise.
% ═══════════════════════════════════════════════════════════════════════════════
Filter Operators: A Mathematical Language for Psychopathology¶
% ═══════════════════════════════════════════════════════════════════════════════
The central claim of this paper is that every psychiatric condition can be characterized by a filter signature---a specific pattern of attenuation, amplification, and distortion applied to the intelligence vector.
Cognitive Filter
\tA
A cognitive filter is a map \(\bF: \\mathbb{R}^{n(e)} \to \\mathbb{R}^{n(e)}\) that transforms the intelligence vector. In the linear regime, \(\bF\) is an \(8 \times 8\) matrix: \begin{keyeq} [ \bI_{\mathrm{eff}} = \bF \cdot \bI_{\mathrm{raw}}, \qquad F_{st} \in [0, \infty) ] \end{keyeq} A diagonal filter \(\bF = \diag(f_1, \ldots, f_8)\) attenuates (\(f_t < 1\)) or amplifies (\(f_t > 1\)) each type independently. Off-diagonal entries model cross-type distortion.
The RTSG framework identifies five species of filter, operating at different timescales:
\begin{center} \small
[Table — see PDF for formatted version]
\end{center}
Filter Composition
\tA
Filters compose by matrix multiplication: \begin{keyeq} [ \bI_{\mathrm{eff}} = F_{\mathrm{attn}} \circ \Fstate \circ \Fcult \circ \Fdev \circ \Fceil(\bI_{\mathrm{raw}}) ] \end{keyeq} The composition is generally non-commutative: the order in which filters are applied matters. In particular, \(\Fdev \circ \Fceil \neq \Fceil \circ \Fdev\) whenever developmental experience interacts with substrate constraints.
Filter Irreversibility
\tA
For any lossy filter \(\bF\) (i.e., \(\det(\bF) < 1\)), there exists no inverse filter \(\bF^{-1}\) such that \(\bF^{-1} \circ \bF = \mathrm{Id}\) on the full type space. Information destroyed by a filter cannot be recovered.
This is the mathematical statement of a clinical reality: you cannot "un-trauma" a brain. Trauma therapy does not remove the traumatic filter---it adds compensatory filters that partially restore the effective intelligence vector through a different pathway. The composition \(\bF_{\mathrm{therapy}} \circ \bF_{\mathrm{trauma}}\) can approximate normative function, but the underlying trajectory through filter space is permanently altered.
% ═══════════════════════════════════════════════════════════════════════════════
Disorder Filter Signatures¶
% ═══════════════════════════════════════════════════════════════════════════════
Each psychiatric condition corresponds to a characteristic filter signature---a specific pattern of attenuations, amplifications, and cross-type distortions. We present six canonical signatures, each derived from the RTSG filter formalism and mapped to clinical phenomenology.
Major Depressive Disorder¶
Depression Filter
\tB
The depressive state filter \(\Fstate^{\mathrm{dep}}\) is characterized by: \begin{keyeq}
\end{keyeq} The amplification of \(I_{\mathrm{symb}}\) is narrow-band: it enhances repetitive symbolic loops while suppressing novel symbolic exploration. Formally, \(f_{\mathrm{symb}}\) acts as a high-gain, low-bandwidth filter---amplifying a small subspace of symbolic processing (self-referential negative cognition) while attenuating the rest.
Post-Traumatic Stress Disorder¶
PTSD Filter
\tB
The trauma filter \(\Ftrauma^{\mathrm{PTSD}}\) combines a state filter with a permanent developmental modification: \begin{keyeq}
\end{keyeq} The mnemonic distortion is not a simple attenuation. Rather, the trauma filter introduces off-diagonal coupling: \(F_{\mathrm{mnem},\mathrm{eval}} \gg 0\), meaning that mnemonic retrieval becomes yoked to evaluative threat assessment. Memories are retrieved not by temporal context but by threat-relevance.
Bipolar Disorder: Mania¶
Manic Trajectory
\tB
Mania is not a static filter but a trajectory through the psychophysiological state space \(\bPsi(t)\) characterized by: \begin{keyeq}
\end{keyeq} where \(\kappa\) denotes the curvature of the trajectory through PAD affective space and \(r_{\mathrm{basin}}\) is the radius of the attractor basin around the current state.
The filter signature during mania includes:
Obsessive-Compulsive Disorder¶
OCD Filter
\tB
The OCD filter is characterized by a tight symbolic loop with failed evaluative termination: \begin{keyeq}
\end{keyeq} The distinguishing feature from depression is that the OCD loop is anxiogenic (driven by threat-evaluation) rather than ruminative (driven by self-referential negative cognition). The evaluative component is not attenuated but disconnected---\(K_{\mathrm{eval},\mathrm{symb}}\) drops, meaning evaluative judgment cannot terminate the symbolic loop even though \(I_{\mathrm{eval}}\) itself remains functional.
ADHD¶
ADHD Filter
\tB
ADHD is characterized by a hypervisor spectral shift---too many eigenvalues above the attention threshold: \begin{keyeq}
\end{keyeq} A flat eigenvalue spectrum means that many cognitive types simultaneously compete for conscious presentation. Nothing achieves clear dominance in the attention simplex, producing the phenomenology of distractibility, rapid topic-switching, and difficulty sustaining single-type focus.
Autism Spectrum¶
Autism Filter
\tB
Autism involves a developmental filter \(\Fdev^{\mathrm{ASD}}\) that reshapes the \(\bK\) matrix: \begin{keyeq}
\end{keyeq} This is not a deficit model: the autistic \(\bK\) matrix produces different synergy patterns. The reduced social coupling means that social information provides less amplification to other types, while the enhanced symbolic-spatial coupling produces the characteristic strengths in pattern recognition, systematization, and detail-oriented processing [baron-cohen2009].
% ═══════════════════════════════════════════════════════════════════════════════
The Affective Geometry of Mental States¶
% ═══════════════════════════════════════════════════════════════════════════════
Beyond the intelligence vector, each individual occupies a position in affective space---a geometric model of emotional state that is richer and more clinically informative than single-scale measures like the PHQ-9 or GAD-7.
Emotional State Vector
\tB
The emotional state at time \(t\) is a vector \(\be(t) \in \R^3\) in the PAD space of Mehrabian and Russell [mehrabian1974]: \begin{keyeq} [ \be(t) = \bigl(v(t),\; a(t),\; d(t)\bigr) ] \end{keyeq} where \(v \in [-1,1]\) is valence (pleasure/displeasure), \(a \in [0,1]\) is arousal (activation intensity), and \(d \in [-1,1]\) is dominance (sense of control). The norm \(\|\be\| = \sqrt{v^2 + a^2 + d^2}\) measures emotional intensity.
Psychophysiological State
\tB
The full psychophysiological state combines the intelligence vector with the emotional state: \begin{keyeq} [ \bPsi(t) = \bigl(\bI_{\mathrm{eff}}(t),\; \be(t)\bigr) \in \R^{11} ] \end{keyeq} A psychiatric trajectory is a curve \(\bPsi: [0,T] \to \R^{11}\) through this combined space.
This geometric framing yields several clinical constructs:
[nosep] - Velocity \(\|\dot{\bPsi}\|\): Rate of state change. High in mania, low in depression. - Curvature \(\kappa\): Trajectory unpredictability. High in rapid cycling, low in stable states. - Basin depth: The energy barrier that must be overcome to exit the current state. Deep basins correspond to treatment-resistant states. - Basin radius: The range of perturbations the state can absorb without transitioning. Large in mania (expansive), small in anxiety (brittle).
% ═══════════════════════════════════════════════════════════════════════════════
Affective Modulation of Cognitive Performance¶
% ═══════════════════════════════════════════════════════════════════════════════
Emotion does not merely accompany cognition---it modulates the effective intelligence vector. The affective modulation operator formalizes how emotional state alters cognitive performance.
Affective Modulation Operator
\tB
The modulation operator \(M(\be): \\mathbb{R}^{n(e)} \to \\mathbb{R}^{n(e)}\) transforms the intelligence vector as a function of emotional state: \begin{keyeq} [ \bI_{\mathrm{modulated}} = M(\be) \cdot \bI_{\mathrm{eff}} ] \end{keyeq} In the linear regime, \(M(\be) = \mathrm{Id} + \sum_{j} e_j \cdot \bG_j\), where \(\bG_j\) are gain matrices for each PAD dimension and \(e_j \in \{v, a, d\}\).
The modulation produces three clinically significant regimes:
[nosep] - Facilitation (\(M_{tt} > 1\)): Moderate arousal enhances performance on the relevant type. The Yerkes--Dodson law is a consequence: \(M_{tt}(a)\) is an inverted-U function of arousal.
-
Disruption (\(M_{tt} < 1\)): Extreme emotional states (high arousal with low dominance, i.e., panic) attenuate cognitive function across all types except evaluative threat detection. This is the phenomenology of anxiety: \(I_{\mathrm{eval}}\) is amplified while everything else is suppressed.
-
Dominance asymmetry: Low dominance (\(d < 0\)) selectively attenuates \(I_{\mathrm{ling}}\) and \(I_{\mathrm{soc}}\) (the "can't speak" and "can't connect" effects of feeling powerless), while high dominance amplifies \(I_{\mathrm{eval}}\) and \(I_{\mathrm{symb}}\) (the "decisive clarity" of agency).
% ═══════════════════════════════════════════════════════════════════════════════
Cognitive Thermodynamics of Mental Illness¶
% ═══════════════════════════════════════════════════════════════════════════════
The RTSG framework introduces thermodynamic concepts---entropy, temperature, free energy, and phase transitions---to the clinical domain, providing a mathematical account of treatment dynamics.
Cognitive Temperature and Entropy¶
Cognitive Temperature
\tA
The cognitive temperature \(\Tcog \geq 0\) is the variance of attention fluctuations around equilibrium: \begin{keyeq} [ \Tcog = \frac{1}{n-1}\sum_{t=1}^{n} \lambda_t^*\, \bigl\langle(\delta\lambda_t)^2\bigr\rangle ] \end{keyeq} where \(\lambda^*\) is the equilibrium attention allocation and \(\delta\lambda_t = \lambda_t - \lambda_t^*\).
Clinical interpretation: At \(\Tcog \approx 0\), the patient is locked into a rigid attentional pattern (obsessive focus, catatonia, perseveration). At high \(\Tcog\), attention wanders without structure (delirium, manic distractibility, dissociation). Healthy function operates in a moderate range where the system can both sustain focus and flexibly redirect.
Cognitive Free Energy and Treatment Landscapes¶
Cognitive Free Energy
\tB
The cognitive free energy combines internal energy with entropic cost: \begin{keyeq} [ \Fcog = U_{\mathrm{cog}} - \Tcog \cdot S_{\mathrm{cog}} ] \end{keyeq} where \(U_{\mathrm{cog}}\) is the cognitive energy (cost of maintaining the current allocation) and \(S_{\mathrm{cog}} = -\sum_t \lambda_t \ln \lambda_t\) is the attention entropy.
Mental states correspond to minima of the free energy landscape \(\Fcog(\lambda)\). This yields a powerful clinical vocabulary:
[nosep] - A stable disorder (chronic depression, persistent OCD) is a deep local minimum---the free energy basin is deep enough that the system cannot escape via thermal fluctuations alone. - Treatment resistance = basin depth exceeds available activation energy. The deeper the basin, the more intensive the intervention required to escape it. - Relapse = the system falls back into its original basin after insufficient perturbation. The disorder basin still exists; treatment must either fill it in (structural change) or move the system far enough away that the basin is no longer the closest minimum. - Temporary worsening before improvement is a thermodynamic necessity: escaping a deep basin requires climbing the energy barrier, which transiently increases \(\Fcog\). The well-documented clinical observation of initial symptom worsening during treatment (e.g., increased anxiety during exposure therapy, activation syndrome with SSRIs) has a mathematical explanation.
Second Law of Cognitive Thermodynamics
\tB
For an isolated cognitive system, the attention entropy \(S_{\mathrm{cog}}\) does not decrease over time: [ \frac{dS_{\mathrm{cog}}}{dt} \geq 0 ] Applied to treatment: any intervention that reduces entropy (increases order, focuses attention) requires energy input from outside the system---therapeutic effort, pharmacological modulation, environmental structuring, or social support.
Phase Transitions¶
Cognitive Phase Transition
\tB
A phase transition occurs when continuous variation of a control parameter (medication dose, stress level, sleep deprivation) causes a discontinuous change in the equilibrium attention allocation \(\lambda^*\): \begin{keyeq} [ \exists\, \theta_c : \quad \lim_{\theta \to \theta_c^-} \lambda^(\theta) \neq \lim_{\theta \to \theta_c^+} \lambda^(\theta) ] \end{keyeq}
% ═══════════════════════════════════════════════════════════════════════════════
Treatment Interaction and Sequencing via the \(\bK\) Matrix¶
% ═══════════════════════════════════════════════════════════════════════════════
The compatibility matrix \(\bK\) does not merely describe cognitive type interactions---it predicts which treatment combinations will amplify or interfere with each other.
Treatment Interaction Prediction
\tB
Let treatment \(A\) target intelligence type \(s\) and treatment \(B\) target type \(t\). If \(K_{st} > 1\), the treatments are synergistic: applying both produces greater improvement than the sum of individual effects. If \(K_{st} < 1\), the treatments interfere: the combination is subadditive.
This generates specific, testable predictions for treatment combinations:
\begin{center} \small
[Table — see PDF for formatted version]
\end{center}
% ═══════════════════════════════════════════════════════════════════════════════
Comorbidity as Filter Overlap¶
% ═══════════════════════════════════════════════════════════════════════════════
The filter framework provides a natural explanation for the high rates of psychiatric comorbidity that embarrass categorical systems.
Comorbidity Prediction
\tB
Two disorders \(D_1\) and \(D_2\) with filter signatures \(\bF_1\) and \(\bF_2\) will show high comorbidity if and only if their filter signatures share common components: \begin{keyeq} [ \text{Comorbidity}(D_1, D_2) \propto \frac{|\bF_1 \circ \bF_2 - \bF_1| + |\bF_1 \circ \bF_2 - \bF_2|}{|\bF_1 - \mathrm{Id}| + |\bF_2 - \mathrm{Id}|} ] \end{keyeq} When \(\bF_1\) and \(\bF_2\) perturb overlapping components, the composition \(\bF_1 \circ \bF_2\) is close to both individual filters---each disorder "prepares the ground" for the other.
Examples: [nosep] - Depression + Anxiety: Both attenuate \(I_{\mathrm{eval}}\) (depression: attenuation of positive evaluation; anxiety: amplification of threat evaluation). The shared evaluative perturbation means each condition exacerbates the other. - PTSD + Depression: Trauma filter attenuates \(I_{\mathrm{soc}}\) and distorts \(I_{\mathrm{mnem}}\); depression filter attenuates \(I_{\mathrm{eval}}\) and \(I_{\mathrm{mnem}}\). The shared mnemonic distortion is the comorbidity mechanism. - ADHD + Anxiety: ADHD flattens the eigenvalue spectrum; anxiety amplifies \(I_{\mathrm{eval}}\). These are orthogonal perturbations---the comorbidity rate should be lower than depression-anxiety. This is consistent with epidemiological data.
% ═══════════════════════════════════════════════════════════════════════════════
Connection to Existing Frameworks¶
% ═══════════════════════════════════════════════════════════════════════════════
The RTSG psychiatric framework connects to several major research programs:
Research Domain Criteria (RDoC). The RDoC initiative [insel2010] proposed dimensional, biologically grounded constructs to replace categorical diagnosis. Our eight intelligence types map onto RDoC domains: \(I_{\mathrm{eval}}\) corresponds to Positive and Negative Valence Systems, \(I_{\mathrm{soc}}\) to Systems for Social Processes, \(I_{\mathrm{mnem}}\) to Cognitive Systems (memory), and the attention simplex to Arousal/Regulatory Systems. The framework adds what RDoC lacks: algebraic structure (the \(\bK\) matrix) and dynamic formalism (cognitive thermodynamics).
HiTOP. The Hierarchical Taxonomy of Psychopathology [kotov2017] organizes psychopathology dimensionally into spectra and subfactors. The filter formalism provides a generative model: each HiTOP spectrum corresponds to a principal component of the filter perturbation matrix \(\bF - \mathrm{Id}\). The Internalizing spectrum reflects evaluative-mnemonic filter perturbations; the Externalizing spectrum reflects evaluative-kinesthetic perturbations with disinhibition.
Network Theory. Borsboom's network approach [borsboom2017] models disorders as self-reinforcing symptom networks. The \(\bK\) matrix is the formal counterpart: it specifies the coupling strengths between cognitive dimensions. Symptoms in Borsboom's networks correspond to attenuated or amplified intelligence types, and the self-reinforcing loops correspond to positive-feedback cycles in the filter composition (e.g., \(I_{\mathrm{eval}} \downarrow \to I_{\mathrm{soc}} \downarrow\) via \(K_{\mathrm{eval},\mathrm{soc}} > 1 \to I_{\mathrm{eval}} \downarrow\) further via social isolation removing evaluative support).
Computational Psychiatry. Friston's free energy principle [friston2014] models psychopathology as aberrant predictive processing. Our cognitive free energy \(\Fcog\) is structurally analogous but operates on the intelligence vector rather than sensory prediction. The frameworks are compatible: \(\Fcog\) governs the macro-level attention landscape, while Friston's variational free energy governs micro-level perceptual inference.
Constructed Emotion Theory. Barrett's theory [barrett2017] that emotions are constructed from interoceptive predictions is compatible with our affective modulation operator. The PAD emotional state vector \(\be(t)\) is the low-dimensional summary of Barrett's interoceptive prediction state, and the modulation operator \(M(\be)\) formalizes how this constructed emotional state alters cognitive performance.
% ═══════════════════════════════════════════════════════════════════════════════
Summary of Testable Predictions¶
% ═══════════════════════════════════════════════════════════════════════════════
We collect the testable predictions distributed throughout the paper:
% ═══════════════════════════════════════════════════════════════════════════════
Limitations and Scope¶
% ═══════════════════════════════════════════════════════════════════════════════
This framework is a research tool, not a diagnostic instrument. Several important limitations must be noted:
Parameter estimation. The filter signatures presented here (e.g., \(f_{\mathrm{eval}} \approx 0.3\) for depression) are theoretically derived estimates. Empirical calibration requires systematic cognitive profiling studies across clinical populations. The framework predicts the pattern of perturbation; the exact values are an open research program.
Linearity assumption. The filter formalism uses linear operators in first approximation. Actual psychiatric perturbations likely involve nonlinear interactions (e.g., threshold effects where \(I_{\mathrm{eval}}\) below a critical value triggers a qualitatively different regime). The linear model is a tractable starting point, not a claim about the true dynamics.
Pairwise \(\bK\) matrix. The compatibility matrix captures pairwise interactions between intelligence types. Higher-order interactions (three-way synergies or inhibitions) may be clinically significant but are not modeled in the current framework.
Individual variation. Different patients with the same diagnosis may have different filter signatures. The canonical signatures presented here are population-level prototypes; personalized psychiatry would estimate individual filter profiles from comprehensive cognitive assessment.
No clinical validation yet. The 15 predictions are empirically testable but have not yet been tested within this framework. The framework should be evaluated by its predictive success, not adopted as clinical truth before validation.
% ═══════════════════════════════════════════════════════════════════════════════
Conclusion¶
% ═══════════════════════════════════════════════════════════════════════════════
We have presented a mathematical framework for psychiatry that models mental disorders as characteristic filter perturbations of a multi-dimensional cognitive architecture. The framework provides:
[nosep] - Dimensional diagnosis: 8D cognitive profiles replace binary labels. - Filter signatures: Each disorder is characterized by a specific pattern of attenuation, amplification, and cross-type distortion. - Treatment prediction: The \(\bK\) matrix predicts treatment synergies and interferences. - Dynamic formalism: Cognitive thermodynamics provides a mathematical account of treatment resistance, relapse, and temporary worsening. - Comorbidity explanation: High comorbidity is a prediction of overlapping filter signatures, not an embarrassment for the classification system. - Filter irreversibility: The mathematical impossibility of "removing" a lossy filter provides formal foundations for trauma-informed care.
The framework is aligned with dimensional approaches (RDoC, HiTOP), compatible with computational psychiatry (Friston's free energy), and generates 15 testable predictions that can be evaluated with existing neuroimaging, psychometric, and ecological momentary assessment technologies. It offers not a finished theory but a mathematical scaffold on which a quantitative, transdiagnostic psychiatry can be built.
This paper is a companion extraction from the master treatise "Intelligence as Geometry" (Niko, 2026). The full mathematical development, including proofs, additional applications, and the complete filter formalism, is available in the parent document.
% ═══════════════════════════════════════════════════════════════════════════════
\color{sectionblue¶
Three-Space Psychiatry}
The three-space ontology reframes psychiatric pathology as disruptions of the instantiation process or the filter cascade.
Id filter dysregulation. PTSD represents chronic Id filter activation: the pre-filter remains engaged long after the survival threat has passed, anchoring consciousness to a narrow band of \(t_\mathbb{R}\) and suppressing \(t_{\mathrm{lat}}\) navigation. The "flashback" is a failure of the Id to release: consciousness is trapped in a past-\(t_\mathbb{R}\) moment, unable to navigate laterally. Treatment targets Id recalibration.
Psychosis as filter cascade inversion. In psychotic states, the normal filter ordering (Id \(\to\) Ceiling \(\to\) Dev \(\to\) Cult \(\to\) State \(\to\) Attn) may invert: attention filters override cultural and developmental filters, producing instantiations inconsistent with shared \(\PS\). Hallucinations are private instantiations that fail to converge with the community's shared instantiation (\(\CSp_{\mathrm{shared}}\)).
Dissociation as hypervisor failure. Dissociative disorders correspond to disruption of the hypervisor fixed point (Proposition XIII.7 of the monograph). When the contraction ratio \(\kappa\) of the actualization operator approaches or exceeds 1, the hypervisor becomes unstable, and the self-interpreting ground state fragments. Depersonalization is the phenomenological signature of a hypervisor approaching instability; dissociative identity disorder represents multiple competing near-fixed-points.
References¶
See PDF for full bibliography.¶
v2 Integration: GNEP Failure Modes & Lyapunov Diagnostics (TMP-20260217)¶
Psychiatric conditions as GNEP failure modes:
| Condition | GNEP Failure | Lyapunov Signature |
|---|---|---|
| ADHD-spectrum | Hypervisor overload | λ near 0, high variance |
| OCD-spectrum | Assembly locked | λ << 0, rigid attractor |
| Dissociation / DID | Assembly fragmentation | λ bifurcates across sub-agents |
| Antisocial pathology | Id constraint violation | \(\lambda < 0\) (GL ground state), wrong attractor |
| Mania / psychosis-spectrum | Drive D excess | \(\lambda > 0\) (GL above critical temperature), unstable |
Lyapunov diagnostic principle: - \(\lambda < 0\) (GL ground state): stable attractor (health or pathological fixation) - λ = 0: bifurcation (acute crisis, insight moment, medication onset) - \(\lambda > 0\) (GL above critical temperature): unstable/chaotic (acute psychosis, mania, dissociation)
The Schopenhauer-Nietzsche Transition maps onto therapeutic trajectory: undirected distress (σdW) → directed recovery (μdt + σdW) → stable integration (\(\lambda < 0\) (GL ground state)).