Cortical thickness, schizophrenia, and causality in psychiatry: when the trace is mistaken for the cause

A commentary published in Nature Mental Health by Professor Andrea Raballo, Full Professor of Psychiatry and Director of the REMEDI Lab at USI, together with Michele Poletti (IRCCS-AUSL Reggio Emilia) and Antonio Preti (University of Turin), challenges a prevalent notion in neuroscience: that specific brain features "determine" or "protect" an individual from schizophrenia.

The reality, the authors argue, is far more complex, and this complexity has direct consequences for how we understand—and communicate—mental illness.

The problem with causality in psychiatry

In recent years, large-scale genetic studies have produced fascinating results: certain brain characteristics appear to cause psychiatric disorders or, alternatively, shield us from them. A recent study published in the same journal, Nature Mental Health, utilised an advanced technique—Mendelian randomisation—to show that greater cortical thickness reduces the risk of schizophrenia. The figures are robust. The method is rigorous. However, the interpretation warrants a moment of reflection. This is the starting point of the Matters Arising article authored by Professor Andrea Raballo, Michele Poletti, and Antonio Preti, recently published in Nature Mental Health.

Schizophrenia does not appear; it develops

The central issue is that schizophrenia does not suddenly emerge in adulthood as if a switch were flipped. It develops over time, often over years or decades, through progressive stages: subtle deviations in childhood development, silent alterations of subjectivity in adolescence, and increasingly structured symptoms leading up to clinical onset. "The cortical thickness we measure in the adult brain is not the origin of this process. It is the sedimentary record." Treating that measurement as a direct cause risks freezing a deeply dynamic process into a single snapshot. The brain is not a fixed system; it is the cumulative result of interactions between genes, environment, relations, and personal history, continuing at least until puberty and beyond.

Why this matters beyond the laboratory

This is not merely a technical matter. When science communicates that a brain feature "causes" a psychiatric illness, it implicitly conveys a specific idea: that the disorder is written into a person's biology in a fixed and predetermined way, even before that person emerges as a subject. This narrative—even when stemming from rigorous, well-intentioned research—can fuel stigma. It can influence how patients interpret their own history, how families assign meaning to suffering, and how society conceives of psychiatric vulnerability. The contribution from USI does not deny the value of genetic and neurobiological studies. On the contrary, it recognises the power of these tools while calling for them to be paired with reflection on the conceptual quality of what they measure. A model built on an ill-defined object will always inherit the same error, regardless of how sophisticated the method may be.

From causal inference to developmental inference

The proposal is to change the question. Not simply: "Does X cause schizophrenia?" but rather: "How does vulnerability become structural over time, through biology and lived experience?" This perspective sits at the heart of the mission of the REMEDI Lab (REthinking MEntal health through Clinical and Data Intelligence) at USI: to integrate advanced data analysis tools—predictive models, AI, and computational linguistics—with clinical intelligence capable of preserving the complexity of psychiatric phenomena. It is not enough to predict better; we must truly understand what we are modelling.