Anterior insula and mid-cingulate cortex differentially regulate anxiety and fear brain–body responses

Abstract

Heightened reactivity to predictable and unpredictable threat is a hallmark of anxiety, yet causal evidence for the roles of deep interoceptive and salience-network hubs remains limited. Using low-intensity focused ultrasound (LIFU), we transiently modulated the ventral anterior insula (vAI) and anterior mid-cingulate cortex (aMCC) in healthy adults performing the No–Predictable–Unpredictable (NPU) threat task. We quantified anxiety-potentiated responses (APR) and fear-potentiated responses (FPR) along with brain–body markers including EMG startle responses, EEG event-related potentials, electrodermal responses and heart rate (HR). Mixed-effects models revealed a significant LIFU × Task interaction such that LIFU to the vAI reduced APR symptom reports and LIFU to the aMCC reduced FPR symptom reports. Physiological changes largely paralleled behavior: vAI modulation reduced APR-linked startle, EEG response, and HR, while aMCC effects were confined to FPR (EEG and HR reductions). Brain-Body-Behavior analysis revealed that LIFU to the vAI increased the coupling of both HR and EEG with subjective reports such that changes in HR and the EEG evoked potential predicted symptom ratings but only during the APR task. In addition, baseline HR/HRV correlated with anxiety across individuals but did not moderate these neuromodulation effects. These findings provide causal, site-specific evidence that the vAI preferentially supports anticipatory anxiety under uncertainty, whereas the aMCC contributes to brain–body mobilization for predictable threat. Targeted neuromodulation of the vAI increases brain-body-behavior coupling to reduce physiological response that translates to reduced symptoms. Neuromodulation targeting the vAI may enable precision interventions for anxiety-related disorders. Significance Statement Anxiety arises from dysregulated interactions between the brain and body, but causal evidence for the cortical hubs governing these loops has been lacking. Using low-intensity focused ultrasound (LIFU), we transiently modulated the ventral anterior insula (vAI) and anterior mid-cingulate cortex (aMCC) during a threat uncertainty task. vAI stimulation reduced anticipatory anxiety symptoms under uncertainty and strengthened heart rate and EEG coupling with subjective reports, consistent with aberrant interoceptive prediction processing. aMCC stimulation selectively dampened cue-evoked fear without enhancing brain–body alignment, consistent with a role in defensive mobilization. These dissociable effects across physiology, behavior, and report provide causal evidence for interoceptive–autonomic integration and highlight circuit-specific neuromodulation as a potential therapy for anxiety.

Abstract via europepmc.

Flags from extraction

• exposures[0].unspecified_domain.pressure_kpa — Peak negative pressure ~900 kPa and 27 W/cm2 Isppa described as 'outside the head' for all participants; medium not named, so placed in unspecified_domain. No in situ values reported (k-Wave simulations done but values not given in the text).
• exposures[1].unspecified_domain.pressure_kpa — Same free-field-like value applied to the aMCC transducer (H-104) as stated for all participants.
• exposures[1].target.terms — Paper's methods call the target 'right dorsal anterior cingulate cortex (aMCC)'; elsewhere anterior mid-cingulate cortex.
• sham_type — Active sham: a high-impedance material inserted into the gel attenuated the ultrasound before the skull; coded as inactive_transducer.
• n_sessions_per_subject — Four visits, of which three (vAI, aMCC, sham) were LIFU/sham sessions; the first visit was imaging only.