Awareness during electromyographic biofeedback: of signal or process?

During relaxation training, awareness of trial-to-trial changes in frontalis-muscle tension levels was assessed with and without auditory electromyographic (EMG) biofeedback. Immediately after each 128-sec training trial, the subject was required to guess whether muscle tension indexed by EMG activity increased ("Up") or decreased ("Down") relative to the immediately preceding trial. The probability of correct guessing, P(c), improved as the absolute difference in EMG increased between trials only when biofeedback was presented. For subjects not receiving biofeedback, P(c) remained low even when the absolute difference between trials was large. Subjects in each condition employed a strategy to guess "Down" more often consistent with the expectation that they were being trained to relax. The "Down" set strategy was shown to be separable from the informational basis of P(c) provided by biofeedback. This procedure can be employed to evaluate central assumptions of biofeedback relating to posttraining awareness of changes in muscle tension and the relationship between awareness and control of muscle tension.