Reading focus
Syllabus: What Is a Mental Disorder?
Compare a proposed interaction among symptoms with a model that summarizes how they vary together. Identify the causal claim that the network account adds.
Why you’re reading it. End by separating the organization of psychopathology from an explanation of how a difficulty might persist.
A question to consider. Sketch a proposed cycle linking poor sleep, fatigue and withdrawal. What repeated observation or intervention could test one arrow, and why would a correlation among the three not establish the whole cycle?
Why this reading
Borsboom develops a theory in which interactions among symptoms can sustain a disordered state. The contribution is explanatory: it asks whether symptoms participate in maintaining problems rather than merely reflecting an underlying common cause. Read the paper for the causal commitments and the evidence that would be required to test them.
Before you read
The theoretical contrast is with an account in which an underlying disorder causes its symptoms. A statistical latent variable need not automatically have that causal interpretation. Three useful terms are connectivity, the strength of relations in the proposed symptom system; external field, influences outside that system; and hysteresis, persistence after a triggering influence recedes.
Read for
- Compare the network account with a common-cause interpretation: what mechanisms does each propose?
- Examine the role of bridge symptoms, external influences, and persistence after a trigger ends.
- Distinguish a causal theory of symptom interaction from a statistical network of associations.
- Ask what observations or interventions would discriminate between explanations.
Consider afterwards
- If there is no common cause, what does Wakefield's "dysfunction" refer to? Can the harmful-dysfunction analysis be restated in network terms?
- A network model and a latent-variable model can fit the same cross-sectional data. What kind of evidence would tell them apart?
Editorial context
A cross-sectional association network does not establish the within-person causal dynamics described by the theory. Bringmann et al. (2022) discuss temporal measurement, node selection, and the distinction between within-person and between-person effects. HiTOP offers a descriptive organization of covariation, so it need not be an incompatible causal theory. Evaluate each framework at the level of the claim being made.