Two Fears, Two Brains: Why "Fear of Fear" and "Fear of the Unknown" Drive Avoidance Through Separate Circuits

Avoidance is the engine of every anxiety disorder. It is also the clinician's central problem: it relieves distress in the moment, prevents corrective learning, and is the behaviour exposure therapy must reverse. Yet "avoidance" has long been treated as a single construct. This Molecular Psychiatry study from Hannah Berg, Shmuel Lissek and colleagues at the University of Minnesota (with collaborators at UT Austin, UCSD and Harvard) takes a sharper instrument to it, asking whether two of the best-validated transdiagnostic risk traits route to avoidance through the same brain machinery — or different machinery entirely.

The design is a transdiagnostic one, deliberately mixing anxiety, trauma- and stressor-related, and obsessive-compulsive disorders under the umbrella of anxiety-related disorders (ARDs). Participants completed self-reports of anxiety sensitivity and intolerance of uncertainty, then performed an avoidance task during fMRI. A threat cue was paired with shock; safety cues and generalisation stimuli of graded resemblance to the threat let the researchers separate two cognitive phases — threat reactivity (the immediate appraisal of a cue) and mental simulation (the deliberative window before committing to avoid or approach).

The dissociation is clean. During threat reactivity, anterior insula and dorsomedial prefrontal cortex tracked how threat-relevant each stimulus was — expected, and reassuringly consistent with the salience literature. The novel finding is moderation: both ARD status and, independently, trait anxiety sensitivity strengthened the relationship between anterior-insula threat reactivity and maladaptive avoidance. In other words, "fear of fear" does not simply raise the volume of interoceptive alarm; it changes how strongly that alarm dictates behaviour.

Intolerance of uncertainty told a different story, and only in the simulation phase. Using multi-voxel pattern analysis to decode upcoming avoidance from neural activity, the team found that in high-IU individuals, avoidance behaviour was less concordant with patterns in motor cortex and intraparietal sulcus. The "fear of the unknown" appears to loosen the link between what the planning system represents and what the person does — a signature of indecision and over-deliberation rather than of an over-loud threat signal.

For the practitioner, the value is conceptual precision. The study does not hand us a biomarker to order, but it does formalise something many clinicians sense: that the patient who avoids because their body screams danger is not the same as the patient who avoids because they cannot tolerate not knowing. The first is an interoceptive, insula-driven problem — the natural target of interoceptive exposure and anxiety-sensitivity-focused work. The second is a deliberative, uncertainty-driven problem — better served by uncertainty-tolerance training and behavioural experiments that build comfort with ambiguous outcomes. Treating both with a generic "do the exposure" risks aiming at the wrong mechanism.

A Double Dissociation, Not a Severity Gradient

It would be easy to read these traits as two ways of being "more anxious". The neural data argue otherwise: the two traits engage different regions, in different task phases, with different computational signatures. That is the hallmark of distinct mechanisms, not points on one continuum — and it is why a single transdiagnostic label can obscure clinically meaningful heterogeneity.

From Self-Report to Mechanism

Anxiety sensitivity and intolerance of uncertainty have decades of questionnaire validation behind them. What this work adds is mechanistic grounding: it shows these are not merely correlated self-descriptions but markers of separable neural processes. That matters for the field's push toward mechanism-matched intervention, where the trait a patient scores high on would point to a specific therapeutic lever.