To the Editor: We thank Isaac Marks, M.D., F.R.C.Psych., for his interesting comment. He pointed out that exercise itself induces prolonged exposure to some situations that are often avoided by patients suffering from panic disorder with agoraphobia. Although we tried to avoid specific cognitive or exposure techniques, we agree that is impossible to perform outdoor running without having exposure at the same time. Some of our patients had marked initial difficulties in coming to our running group once a week and could only run when a friend or relative accompanied them. In such cases, our study design did not allow for encouraging the patients to run on their own in order to increase the intensity of exposure. Yet, we are not able to separate the therapeutic effects of motor activity from the beneficial effects of exposure. To do so, it would be necessary to have one group of patients exercising at home (e.g., by using bicycle ergometry) and to compare this group to patients treated by standard exposure techniques and—if possible—to a third group with combined exercise and exposure treatment. However, mere motor activity at home would confront patients with internal stimuli such as sweating and palpitations that might lead to interoceptive conditioning. Dr. Marks emphasizes that such a mechanism might also contribute to the beneficial effects of exercise. We mentioned in our Discussion that interoceptive conditioning has indeed been used in cognitive behavioral approaches to help patients reattribute certain somatic cues to nonpathological vegetative functions. Again, for methodological reasons, we did not discuss these experiences and cognitions with the patients from our study, in an attempt to restrict the brief talking sessions to general support only. However, we observed that the experience of being able to run 3 or 4 miles does not remain without influence on dysfunctional cognitions, especially those related to somatic concerns. A more detailed analysis of Dr. Bandelow’s Panic and Agoraphobia Scale subscales revealed that the most prominent effect of exercise was related to a marked decrease of somatic concerns (59.5% mean change from baseline).