Two experiments were conducted to determine whether subjects can learn to control the amplitudes of their contingent negative variation (CNV) through biofeedback training. In Experiment 1, the effect of feedback training on the subjects' discriminability of CNV amplitudes was examined using a simple reaction time task paradigm. Sixteen subjects were asked to try to discriminate their own CNV amplitude levels by pressing a key. Feedback information on discrimination performance was given after the S2 presentation. It was found that biofeedback training led to a significant improvement in discrimination performance. In Experiment 2, ten subjects were given immediate feedback about the amplitudes of CNV observed between S1 and S2. Subjects were instructed to either increase or decrease their CNV amplitude with the aid of immediate feedback information. Results indicated that subjects could learn to regulate their CNV amplitudes in the required directions. It was also found that subjects were able to extend the effects of feedback training to the post-feedback period when feedback signals were no longer presented.