Although many patients with panic disorder and agoraphobia respond well to treatment, for others the response is unsatisfactory. Investigating why some respond well and some do not is the only way that we can ever hope to improve treatment for nonresponders. Brain imaging studies have identified many key neuronal circuits in panic disorder. A large co-operative study in this issue of the Journal, conducted by Lueken et al. (1) and carried out across several German psychiatric centers, has now monitored how activity in these circuit changes are involved in panic attacks with agoraphobia during treatment. The authors’ efforts to develop a multisite consortium with matched scanners and clinical programs made their study possible. Studies from this group and others have suggested that a network involving the coupling of the prefrontal cortex and amygdala plays a role in panic attacks with agoraphobia, as does hippocampal activity. In comparing responders to a cognitive-behavioral program with nonresponders, Lueken et al. were able to show that “patients who did not respond to treatment exhibited enhanced activation in the pregenual anterior cingulate cortex, the amygdala, and the hippocampus during safety signal processing compared with responders” (1). Nonresponding patients activated a circuit that suggested danger in response to a safety signal, so instead of getting the messages that “things are okay,” they continued to register mild danger signals. Thus, for these patients, the cognitive exposure appears to partially reinforce rather than diminish the feeling of danger, through a misinterpretation of the safety signal. Looking closely at the data, there is a continuum across these responses rather than two discrete groups of clear responders and nonresponders. This is also the case clinically, since the 50% dividing line for response or nonresponse to cognitive therapy did not lead to two distinct groups. Rather, the patients exhibited a continuum of therapeutic responses. Thus, there are some very good and very poor responders at the two extremes, where these differences are most apparent, but there are also many patients in the middle who receive some benefit and some normalization of neuronal activity. Clinically, this suggests, as the authors note, that cognitive-behavioral therapy may be appropriate for a majority of patients, since even most nonresponders showed changes in the right direction following therapy.