Module 7

Computational Psychiatry

Bayesian-brain failure modes map onto psychiatric symptom dimensions. Aberrant precision on priors is proposed to underlie hallucinations and delusions in schizophrenia; elevated sensory precision underlies autistic hypersensitivity; and altered reward prediction errors underlie depression and addiction. Computational psychiatry formalises these mappings as testable parameter shifts.

1. Schizophrenia — Weak Priors

Adams 2013 proposed that schizophrenia involves attenuated precision on high-level priors. The consequence: ambiguous sensory data are over-weighted, producing hallucinations (endogenous sensory generation) and delusions (aberrant inference about causes). The diagnostic hollow-mask illusion — normally perceived as convex despite contrary depth cues — is resistable in schizophrenia because the convex-face prior is down-weighted.

2. Autism HIPPEA

Pellicano & Burr 2012 proposed HIPPEA (High, Inflexible Precision of Prediction Errors in Autism): elevated sensory-channel precision. The consequence: reduced generalisation, sensory overload, and a preference for structured predictable environments. Evidence includes psychophysical tasks (Palmer 2017) and pupillometry (Lawson 2017). HIPPEA integrates symptom diversity (sensory, social, perceptual) under a single computational profile.

Simulation: Precision Profiles

Python

script.py37 lines

import numpy as np, matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt

# Adams 2013 hollow-mask illusion simulation:
# aberrant precision on priors -> schizophrenia resistant to illusion
# normal vs high-precision prior processing

observations = np.linspace(-3, 3, 200)

def posterior(obs, prior_mean, prior_prec, like_prec):
    post_prec = prior_prec + like_prec
    post_mean = (prior_prec*prior_mean + like_prec*obs) / post_prec
    return post_mean

prior_mean = 0.0  # prior favors convex face
typical = [posterior(o, prior_mean, 1.0, 1.0) for o in observations]
schiz_low_prior = [posterior(o, prior_mean, 0.3, 1.0) for o in observations]
autism_high_like = [posterior(o, prior_mean, 1.0, 3.0) for o in observations]

fig, ax = plt.subplots(figsize=(11, 5), facecolor='#0a0a1a')
ax.set_facecolor('#111827'); ax.tick_params(colors='#cbd5e1')
for s in ax.spines.values(): s.set_color('#334155')
ax.plot(observations, observations, color='#94a3b8', ls=':', label='No prior (=obs)')
ax.plot(observations, typical, color='#a78bfa', lw=2.5, label='Typical')
ax.plot(observations, schiz_low_prior, color='#dc2626', lw=2.5, label='Schizophrenia (weak prior)')
ax.plot(observations, autism_high_like, color='#34d399', lw=2.5, label='Autism HIPPEA (high likelihood prec)')
ax.set_xlabel('Sensory observation (ambiguous)', color='#cbd5e1')
ax.set_ylabel('Posterior percept', color='#cbd5e1')
ax.set_title('Precision-weighted inference: clinical profiles',
             color='#c4b5fd', fontweight='bold')
ax.legend(facecolor='#1e293b', edgecolor='#334155', labelcolor='#cbd5e1')
ax.grid(True, color='#334155', alpha=0.3)
plt.tight_layout()
plt.savefig('output.png', dpi=120, bbox_inches='tight', facecolor='#0a0a1a')
print('Schizophrenia: attenuated priors -> over-weighting of ambiguous sensory data -> hallucinations')
print('Autism HIPPEA: excessive sensory precision -> reduced generalisation / sensory hypersensitivity')

Click Run to execute the Python code

Code will be executed with Python 3 on the server

3. Depression, Addiction, Hallucinogens

Depression is increasingly modelled as suppressed reward prediction errors and rigid, over-confident priors about failure. Addiction involves excess reward prediction error on drug-associated cues plus impaired inhibition. Psychedelics (psilocybin, LSD) are proposed to destabilise high-level priors via 5-HT2A activation, temporarily enabling “re-inference” and explaining both mystical experiences and therapeutic effects in treatment-resistant depression (Carhart-Harris 2019 REBUS model).

Key References

• Adams, R. A., Stephan, K. E., Brown, H. R., Frith, C. D. & Friston, K. J. (2013). “The computational anatomy of psychosis.” Front. Psychiatry, 4, 47.

• Pellicano, E. & Burr, D. (2012). “When the world becomes ‘too real’: a Bayesian explanation of autistic perception.” Trends Cogn. Sci., 16, 504–510.

• Carhart-Harris, R. L. & Friston, K. J. (2019). “REBUS and the anarchic brain: toward a unified model of the brain action of psychedelics.” Pharmacol. Rev., 71, 316–344.

• Stephan, K. E. et al. (2016). “Computational neuroimaging strategies for single-patient predictions.” Neuroimage, 145, 180–199.

Share:X Reddit LinkedIn

← Module 6 Module 8: Applications →