Learning to fear danger is essential for survival. However, overactive, relapsing fear behavior in the absence of danger is a hallmark of disabling anxiety disorders that affect millions of people. Its suppression is thus of great interest, but the necessary brain components remain incompletely identified. We studied fear suppression through a procedure in which, after acquiring fear of aversive events (fear learning), subjects were exposed to fear-eliciting cues without aversive events (safety

more »

... learning), leading to suppression of fear behavior (fear extinction). Here we show that inappropriate, learning-resistant fear behavior results from disruption of brain components not previously implicated in this disorder: hypothalamic melanin-concentrating hormone–expressing neurons (MNs). Using real-time recordings of MNs across fear learning and extinction, we provide evidence that fear-inducing aversive events elevate MN activity. We find that optogenetic disruption of this MN activity profoundly impairs safety learning, abnormally slowing down fear extinction and exacerbating fear relapse. Importantly, we demonstrate that the MN disruption impairs neither fear learning nor related sensory responses, indicating that MNs differentially control safety and fear learning. Thus, we identify a neural substrate for inhibition of excessive fear behavior.