Changes in the magnitude and direction of physiological measures (EMG, EEG, temperature, etc.) are not strongly related to the reduction of clinical symptoms in biofeedback therapy. Previously, nonspecified perceptual, cognitive, and emotional factors related to threat perception (Wickramasekera, 1979, 1988, 1998) may account for the bulk of the variance in the reduction of clinical symptoms. The mean magnitude of these previously nonspecified or placebo factors is closer to 70% when both the therapist and patient believe in the efficacy of the therapy. This powerful placebo effect is hypothesized to be an elicited conditioned response (Wickramasekera, 1977a, 1977c, 1980, 1985) based on the memory of prior healings. These memories of healing are more resistant to extinction if originally acquired on a partial rather than continuous reinforcement schedule. High and low hypnotic ability in interaction with threat perception (negative affect) is hypothesized to contribute to both the production and reduction of clinical symptoms. High and low hypnotic ability respectively are hypothesized to be related to dysregulation of the sympathetic and parasympathetic arms of the autonomic nervous system. Biofeedback is hypothesized to be most effective for reducing clinical symptoms in people of low to moderate hypnotic ability. For people high in trait hypnotic ability, training in self-hypnosis or other instructional procedures (e.g., autogenic training, progressive muscle relaxation, mediation, CBT, etc.) will produce the most rapid reduction in clinical symptoms.