Jones2021 - So What, Implications of Network Theory for Research and Practice

There have been large shifts in the conceptualisation of psychopathology over the last 100 years. From psychoanalysis to behaviorism to brain circuitry.

The medical model assumes that psychological problems are caused by disease entities that exist independently of their signs and symptoms. “Treat the disease, not the symptoms” assumes that there is a disease independent of the symptoms to treat.

Underlying disease entities that give rise to specific symptom patterns.

Modern psychiatric research is also dominated by the assumption that psychological problems are caused by diseases or aberrations in the brain: lesions, abnormalities in structure and function, chemical imbalances, aberrant communication between networks, etc.

Network theory of psychopathology

Signs and symptoms of a disorder co-occur because of a network of causal relations among the symptoms themselves.

Symptoms may be initially activated through outside forces (e.g., a traumatic event) but the symptoms cohere as a syndrome that persists over time due to mutually reinforcing relationships among the symptoms (page 5)

Analysis of psychopathology networks

Network nodes

Network edges

How network theory relates to assumptions in the medical model of psychopathology

Assumption 1: Mental illlnesses exist independently of their signs and symptoms

Network theory argues that signs and symptoms are the mental illness. Most things that have been called symptoms are not indicators of underlying diseases, but are Network nodes in a causal web.

a pattern of pervasive worry is not caused by GAD, the pattern of pervasive worry is GAD. (page 7)

Mental illness is a state that the network can be in.

Assumption 2: Classification of psychological problems should follow the medical model

The medical model works well when:

  • Specific causes reliablly lead to the same consequences
  • Some consequences are unique to a given cause
  • Different causes are independent of one another
  • Consequences don’t feed back into the causes

Why is this classification system lackluster in psychopathology?

  1. Diagnostic boundaries are extremely fuzzy, since many symptoms appear in multiple disorders.
  2. We also have Equifinality and Equipotentiality.
  3. Symptoms have wide-reaching causal effects (social anxiety -> problematic alcohol use -> feelings of worthlessness, loneliness, depression)

the fuzziness between diagnostic categories arises because mental disorders are overlapping communities of causally interacting components, not discrete disease entities (page 10)

Assumption 3: Psychological problems are caused by diseases or aberrations in the brain

Three arguments against this view by Borsboom2018 - Brain disorders Not really Why network structures

  1. If all symptoms depend on a common latent variable (brain aberration), then we can describe the disorder as having a common cause. However, if there are many interacting causes such as in a network model, then the reductionist view is blocked.
  2. Many symptoms are rational in the context of another (hand washing is rational if I really believe my hand has been contaminated), so the content of thought is needed to explain the full symptom pattern and a neurobiological cause is implausible.
  3. Symptom networks rely on cultural and historical contexts

Given the lack of progress in identifying reliable brain markers of most mental problems over the past century, we believe there is good reason for us to devote greate energy to alternative accounts, such as network theory. (page 14)

They think that network theory should be integrated with genetic and neurobiological factors.

Implications of network theory for research

Avoid reification

Diagnostic labels should not fall into the trap of Reification and be treated as established disease entities.

Shift the level of analysis

Focus on individual psychological problems and the relationships among them rather than disorders. (Eiko Fried’s research on depression)

Move beyond symptoms alone

The causal feedback loops generating mental disorders will not only consist of symptoms that are highly distressing, highly visible, or specific to the disorder. For example genetic factors or trauma should be incorporated.

Specify formal theories

Mathematical or computational models, to formally test them in empirical studies.

Take causality seriously and use experiments

Use randomized experiments to investigate the causal interactions, this is particularly underused in network approach literature so far.