Wider fear generalization seen in the brain at clinical high risk for psychosis

People at clinical high risk for psychosis show a subtly broader pattern of fear generalization in the brain, according to a new study published in Translational Psychiatry. The finding comes from a functional magnetic resonance imaging study of fear conditioning and generalization, and it suggests that a basic associative memory process may be linked to psychosis risk as well as to transdiagnostic vulnerability for psychiatric illness.

The study addressed a long-standing question: whether altered associative learning observed in psychiatric conditions reflects ongoing symptoms or an underlying risk factor shared across disorders. To probe this, the researchers recruited 72 participants with subclinical psychopathology. Of these, 23 met criteria for the Clinical High-Risk for Psychosis syndrome (CHR+) and 49 did not (CHR-).

During the experiment, participants viewed face stimuli while lying in an MRI scanner. Some faces were paired with a mild electrical shock (the CS+) and others were not (the CS-). Crucially, the paradigm also included eight face “morphs” — intermediate stages between the CS+ and CS- — that were tailored to each participant’s perceptual discrimination ability, allowing the researchers to measure how fear responses spread across stimuli that resembled the shock-paired face.

After the scan session, participants rated the likelihood that each stimulus had previously been followed by a shock. These memory ratings, together with fear generalization-related neural responses, were then analyzed using both categorical and individual-level psychometric modeling approaches. The neural analysis was conducted in a subset of the sample comprising 16 CHR+ and 40 CHR- participants.

The expected patterns of fear generalization appeared in the anterior insula and the superior frontal gyrus, and in the memory ratings, confirming that the paradigm worked as intended across the groups.

The psychometric modeling revealed a key difference: the CHR+ group showed a significantly greater threshold in responses of the left anterior insula, representing a wider fear generalization curve, compared with the CHR- group. In other words, in people at high risk for psychosis, neural signals of fear extended more broadly across stimuli resembling the threat-associated face.

Across the whole sample, regardless of CHR status, symptoms of anxiety were associated with a wider fear generalization gradient. This transdiagnostic link suggests that the breadth of fear generalization may track vulnerability to psychiatric illness more generally, rather than being specific to psychosis risk alone.

The authors conclude that specific features of fear generalization may be linked both to CHR status and to transdiagnostic risk for psychiatric illness. Because the study examined people with subclinical psychopathology rather than established diagnoses, the findings point toward altered basic learning processes as a possible marker of risk, while leaving open the question of how such changes relate to the later emergence of clinical symptoms.

Subject of Research: Fear generalization and its neural correlates in individuals at clinical high risk for psychosis

Article Title: Neurobehavioral changes in fear generalization in clinical high risk for psychosis

Article References: Clauss, J. A., Babadi, B., Tuominen, L., Deng, W., Leathem, L., & Holt, D. J. (2026). Neurobehavioral changes in fear generalization in clinical high risk for psychosis. Translational Psychiatry. https://doi.org/10.1038/s41398-026-04494-5

Image Credits: AI Generated

DOI: 10.1038/s41398-026-04494-5

Keywords: fear generalization, psychosis risk, clinical high risk, fMRI, anterior insula, fear conditioning, anxiety, associative learning, Translational Psychiatry, psychometric modeling, Neurobehavioral, changes