Elsevier

The Lancet Psychiatry

Volume 1, Issue 2, July 2014, Pages 148-158
The Lancet Psychiatry

Review
Computational psychiatry: the brain as a phantastic organ

https://doi.org/10.1016/S2215-0366(14)70275-5Get rights and content

Summary

In this Review, we discuss advances in computational neuroscience that relate to psychiatry. We review computational psychiatry in terms of the ambitions of investigators, emerging domains of application, and future work. Our focus is on theoretical formulations of brain function that put subjective beliefs and behaviour within formal (computational) frameworks—frameworks that can be grounded in neurophysiology down to the level of synaptic mechanisms. Understanding the principles that underlie the brain's functional architecture might be essential for an informed phenotyping of psychopathology in terms of its pathophysiological underpinnings. We focus on active (Bayesian) inference and predictive coding. Specifically, we show how basic principles of neuronal computation can be used to explain psychopathology, ranging from impoverished theory of mind in autism to abnormalities of smooth pursuit eye movements in schizophrenia.

Introduction

Computational psychiatry uses formal models of brain function to characterise the mechanisms of psychopathology, usually in a way that can be described in computational or mathematical terms. Computational psychiatry has arrived:1, 2, 3 the first international computational psychiatry meeting was held in 2013, and 2014 saw the inception of the Max Planck Society-University College London initiative on computational psychiatry and ageing research—and the first UK computational psychiatry course. Several computational psychiatry units are emerging worldwide.

In this Review, we aim to provide tangible examples of computational psychiatry and to explain how it motivates mechanistic research in systems neuroscience; research that is being, or will soon be, translated into clinical neuroscience. First, we consider the properties a computational formulation must possess to be useful in psychiatry. We focus on active inference or predictive coding as an example (panel). We emphasise the importance of active inference by showing how it contextualises other formal treatments. We conclude with some examples of how theoretical principles can unify apparently disparate aspects of psychiatric disorders. We will consider functional and dissociative symptoms, soft neurological signs in schizophrenia, interoceptive inference and autism, dysconnection models of delusional (false) beliefs, and formal models of interpersonal exchange. We chose these examples to show the breadth of psychopathology that can be understood in terms of one pathology; namely, false inference that can be ascribed to neuromodulatory failures at the synaptic level. This Review is prospective, in that most of the examples we consider relate to the promise of the future—much of the work that substantiates the points we make has yet to be undertaken.

Section snippets

The phantastic organ

Many formal or computational schemes could characterise psychopathology, ranging from parallel distributed processing or neural network theory and dynamical systems theory, to reinforcement learning and game theory. However, these theoretical frameworks do not address the central problem encountered in psychiatry—ie, the production of false beliefs. The problems that concern psychiatrists are, almost universally, abnormal beliefs and their behavioural sequelae (eg, dysmorphophobia, paranoid

Predictive coding and the Bayesian brain

Modern versions of Helmholtz's ideas are now among the most popular explanations for message passing in the brain and are usually portrayed in the setting of the Bayesian brain hypothesis as predictive coding.9, 10, 11, 12 Predictive coding is not a normative or descriptive scheme, it is a process theory with a biologically plausible basis—there is now much circumstantial anatomical and physiological evidence for predictive coding in the brain.12, 13, 14, 15, 16 In this scheme, neuronal

From the Bayesian brain to active inference

If the brain is a generative model of the world, then much of it must be occupied by modelling other people. In other words, individuals spend most of their time predicting the internal (proprioceptive) and external (exteroceptive) consequences of behaviour (both their own and that of others). To fully appreciate the bilateral nature of these predictions, inference can be considered in an embodied context. In this setting, perception can be understood as resolving exteroceptive prediction

Interoceptive inference

Recently, investigators described emotional processing in terms of predictive coding or inference about interoceptive or bodily states.38, 39 In active inference, motor reflexes are driven by proprioceptive prediction errors. Proprioceptive prediction errors compare primary afferents from stretch receptors with proprioceptive predictions that descend to α motor neurons in the spinal-cord and cranial nerve nuclei. This circuit effectively replaces descending motor commands with proprioceptive

Parallel-distributed processing, precision, and the dysconnection hypothesis

Part of the construct validity of active inference is that it leads to, and contexualises other formal approaches. For example, formal models of schizophrenia are often described in terms of neuronal disconnection. There are two versions of the disconnection hypothesis: the first is implied by Wernicke's sejunction hypothesis, which postulates an anatomical disruption or disconnection of association fibres;43 the second postulates abnormalities at the level of synaptic efficacy and plasticity,

Neuromodulation and false inference

We have introduced a computational framework for action and perception, with a special focus on the synaptic mechanisms that might underlie false inference in psychiatric disorders: in brief, the formal constraints implicit in predictive coding mandate modulatory gain control for ascending prediction errors. In an article in 2012, Edwards and colleagues79 illustrates how functional symptoms can be understood as false inference about the causes of abnormal sensations, movements, or their

Conclusion

In this Review, we have discussed how computational psychiatry can use formal models of perceptual inference and learning to provide a mechanistic and functional perspective on psychopathology and its underlying pathophysiology. We focused on inference as the overarching theoretical framework; largely because it can formalise perception and behaviour in terms of probabilistic beliefs. By assuming that the brain engages in some form of active inference, neuronal dynamics and message passing can

References (99)

  • AK Seth

    Interoceptive inference, emotion, and the embodied self

    Trends Cogn Sci

    (2013)
  • H Critchley et al.

    Will studies of macaque insula reveal the neural mechanisms of self-awareness?

    Neuron

    (2012)
  • J Panksepp et al.

    What is neuropsychoanalysis? Clinically relevant studies of the minded brain

    Trends Cogn Sci

    (2012)
  • G Winterer et al.

    Genes, dopamine and cortical signal-to-noise ratio in schizophrenia

    Trends Neurosci

    (2004)
  • M Laruelle

    Schizophrenia: from dopaminergic to glutamatergic interventions

    Curr Opin Pharmacol

    (2014)
  • JE Lisman et al.

    Circuit-based framework for understanding neurotransmitter and risk gene interactions in schizophrenia

    Trends Neurosci

    (2008)
  • J Lisman

    Excitation, inhibition, local oscillations, or large-scale loops: what causes the symptoms of schizophrenia?

    Curr Opin Neurobiol

    (2012)
  • KT Kishida et al.

    Imaging models of valuation during social interaction in humans

    Biol Psychiatry

    (2012)
  • AA Moustafa et al.

    A dopaminergic basis for working memory, learning and attentional shifting in Parkinsonism

    Neuropsychologia

    (2008)
  • QJ Huys et al.

    A Bayesian formulation of behavioral control

    Cognition

    (2009)
  • KT Kishida et al.

    Neuroeconomic approaches to mental disorders

    Neuron

    (2010)
  • M Moutoussis et al.

    Bayesian inferences about the self (and others): a review

    Conscious Cogn

    (2014)
  • A Peled

    Neuroscientific psychiatric diagnosis

    Med Hypotheses

    (2009)
  • E Pellicano et al.

    When the world becomes ‘too real’: a Bayesian explanation of autistic perception

    Trends Cogn Sci

    (2012)
  • MI Garrido et al.

    Repetition suppression and plasticity in the human brain

    Neuroimage

    (2009)
  • RJ Moran et al.

    An in vivo assay of synaptic function mediating human cognition

    Curr Biol

    (2011)
  • BB Averbeck et al.

    Probabilistic learning and inference in schizophrenia

    Schizophr Res

    (2011)
  • TV Maia et al.

    From reinforcement learning models to psychiatric and neurological disorders

    Nat Neurosci

    (2011)
  • H Helmholtz

    Concerning the perceptions in general. In: Treatise on physiological optics

    (1866)
  • P Dayan et al.

    The Helmholtz machine

    Neural Comput

    (1995)
  • TS Lee et al.

    Hierarchical Bayesian inference in the visual cortex

    J Opt Soc Am A Opt Image Sci Vis

    (2003)
  • J Hohwy

    The Predictive Mind

    (2013)
  • A Clark

    Whatever next? Predictive brains, situated agents, and the future of cognitive science

    Behav Brain Sci

    (2013)
  • MV Srinivasan et al.

    Predictive coding: a fresh view of inhibition in the retina

    Proc R Soc Lond B Biol Sci

    (1982)
  • RP Rao et al.

    Predictive coding in the visual cortex: a functional interpretation of some extra-classical receptive-field effects

    Nat Neurosci

    (1999)
  • K Friston

    Hierarchical models in the brain

    PLoS Comput Biol

    (2008)
  • D Mumford

    On the computational architecture of the neocortex. II. The role of cortico-cortical loops

    Biol Cybern

    (1992)
  • RA Adams et al.

    Predictions not commands: active inference in the motor system

    Brain Struct Funct

    (2013)
  • E Kandel

    The Age of insight: the quest to understand the unconscious in art, mind, and brain, from Vienna 1900 to the present

    (2012)
  • RL Gregory

    Perceptual illusions and brain models

    Proc R Soc Lond B Biol Sci

    (1968)
  • KJ Friston

    The fantastic organ

    Brain

    (2013)
  • CD Frith et al.

    Abnormalities in the awareness and control of action

    Philos Trans R Soc Lond B Biol Sci

    (2000)
  • JM Kilner et al.

    Predictive coding: an account of the mirror neuron system

    Cogn Process

    (2007)
  • H Brown et al.

    Active inference, sensory attenuation and illusions

    Cogn Process

    (2013)
  • H Feldman et al.

    Attention, uncertainty, and free-energy

    Front Hum Neurosci

    (2010)
  • A Clark

    The many faces of precision (Replies to commentaries on “Whatever next? Neural prediction, situated agents, and the future of cognitive science”)

    Front Psychol

    (2013)
  • RJ Moran et al.

    Free energy, precision and learning: the role of cholinergic neuromodulation

    J Neurosci

    (2013)
  • PS Goldman-Rakic et al.

    The anatomy of dopamine in monkey and human prefrontal cortex

    J Neural Transm Suppl

    (1992)
  • K Doya

    Modulators of decision making

    Nat Neurosci

    (2008)
  • Cited by (331)

    • To be or not to be: The active inference of suicide

      2024, Neuroscience and Biobehavioral Reviews
    • Computational Psychiatry and the Placebo Effect in Psychosis

      2023, Biological Psychiatry Global Open Science
    • The Placebo Effect in Psychosis: Why It Matters and How to Measure It

      2023, Biological Psychiatry Global Open Science
    View all citing articles on Scopus
    View full text