Interoceptive cardiac signals selectively enhance fear memories

Fear is coupled to states of physiological arousal. We tested how learning and memory of threat, i.e. conditioned fear, is influenced by interoceptive signals. Forty healthy individuals were exposed to two threat (CS+, paired with electrocutaneous shocks) and two safety (CS-) stimuli, specifically time-locked to either cardiac ventricular systole (when arterial baroreceptors signal cardiovascular arousal to brainstem), or diastole (when these afferent signals are quiescent). Threat learning was indexed objectively using skin conductance responses (SCRs). During acquisition of threat contingencies, cardiac effects dominated: Stimuli (both CS+ and CS-) presented at systole evoked greater SCR responses, relative to stimuli (both CS+ and CS-) presented at diastole. This difference was amplified in more anxious individuals. Learning of conditioned fear was established by the end of the acquisition phase, which was followed by an extinction phase when unpaired CSs were presented at either the same or switched cardiac contingencies. One day later, electrocutaneous shocks triggered the reinstatement of fear responses. Subsequent presentation of stimuli previously encoded at systole evoked higher SCRs. Moreover, only those participants for whom stimuli had the same cardiac-contingency over both acquisition and extinction phases retained conditioned fear memory (i.e. CS + > CS-). Our findings reveal two important cardiac afferent effects on threat learning and memory: 1) Cardiac signals bias processing towards threat. 2) Cardiac signals are a context for fear memory; altering this context can disrupt the memory. These observations suggest how threat reactivity may be reinforced and maintained by both acute and enduring states of cardiac arousal.