Psychotherapy & Pathology

In one sentence: Mental disorders aren’t separate diseases with separate causes — they’re the same system stuck in different ways. Each disorder maps to a specific “misconfiguration” of the mind’s architecture, and therapy means adjusting specific settings.

Theory sources: BM (pathology taxonomy, biases, radicalization), AGI_F (failure modes, self-repair), NM (reconsolidation, hub displacement, bias mechanisms), EMT (immune system of personality)

One Architecture, Many Disorders

BMC models a healthy mind as a system in balance: emotional drives and learned beliefs working together, with moderate flexibility and enough energy. Each mental disorder is a specific departure from this balance:

Comorbidity is predictable: Disorders that are “close neighbors” in this parameter space co-occur more frequently. ADHD + depression (both involve focus instability), OCD + autism (both involve filter over-sensitivity).

Three Therapeutic Levers

BMC identifies three mechanistically distinct ways to intervene. Each targets a different part of the architecture:

Lever 1: The 6-Hour Rewriting Window (Reconsolidation)

When a memory is recalled with genuine surprise (something that contradicts what you expected), it enters a ~6-hour window where it can be updated:

Critical: Recall without surprise does not open the window (Pedreira et al., 2004). This explains:

• Why simple repetition doesn’t change deep beliefs
• Why exposure therapy works only when the expected catastrophe doesn’t occur (surprise!)
• Why cognitive restructuring must involve genuine shock, not just logical argument

PTSD application: The fear memory is recalled in a safe context (surprise: the expected danger is absent). Within 6 hours: extinction training can update the fear memory itself, rather than just creating a competing “it’s safe” memory (Schiller et al., 2010).

Lever 2: Hub Displacement (Shifting Core Beliefs)

Core pathological beliefs (“I’m worthless,” “the world is dangerous,” “I can’t cope”) are hubs — the most connected nodes in the belief network. They resist direct confrontation because everything is wired through them.

The therapeutic strategy: don’t attack the hub directly. Instead, build a competing hub that gradually absorbs connections:

Counter-intuitive prediction: Peripheral beliefs change first. Therapy should target the edges before approaching the core — like peeling an onion from the outside.

Lever 3: Filter Recalibration (The Mind’s Immune System)

The mind has an “immune system” (the I-layer) that decides which new information to accept and which to reject. It can be miscalibrated in both directions:

The filter operates at three levels:

Cognitive Biases: Useful Features, Not Random Errors

The ~200 documented cognitive biases emerge from 6 mechanisms that serve useful purposes in the healthy mind:

Therapeutic implication: You can’t debias generally. Training someone to overcome anchoring (a WM-limits bias) does not improve their resistance to confirmation bias (a hub-inertia bias). Each mechanism requires a different intervention.

The Bias → Pathology Spectrum

Depression: A Detailed Example

To show how the model works in practice, here’s depression analyzed through BMC:

What’s happening:

1. Chronic stress keeps the fear/grief systems activated → working memory captured
2. Reduced desk space → fewer ideas processed per cycle → less learning progress
3. No progress on anything → curiosity gaps become aversive instead of motivating
4. Stuck gaps + no progress = rumination (looping on unsolvable problems)
5. Rumination drains energy → further shutdown → positive feedback loop

How to intervene at each point:

Radicalization: How Extreme Beliefs Form

Radicalization follows a predictable 5-phase trajectory that BMC maps precisely:

Why adolescents are vulnerable: The prefrontal cortex (the filter) isn’t fully developed until ~25. Young people have weak filtering + peak curiosity + emotional instability — the perfect storm.

Deradicalization prediction: Reactivating dormant belief clusters (exposure to diverse perspectives) is more effective than direct counter-argument (which is blocked by hub inertia).

Testable Predictions

Formalization

For readers interested in the mathematical treatment:

Healthy range parameters:

• Balance (G/M ratio): 0.8–2.5
• Stability ($\sigma_{SW}$): near 1.0 (critical regime)
• Modularity ($Q$): 0.2–0.4
• Energy ($E_{available}$): above depletion threshold

Radicalization index:

Hub displacement dynamics:

Peripheral-first prediction:

Where $C_E(i)$ is eigenvector centrality — the more central a belief, the harder it is to change.

Hub inertia (resistance to change):

Full formal treatment: BM Parts V–VI, IX, NM Parts V–VI, IX, AGI_F Part VI.

Next: AI Safety translates these same failure modes and safeguards into the design of artificial minds.