Grade of False Feedback and Belief in Biofeedback

Abstract

One of the critical issues in biofeedback training is to assess the skill or extent of belief that the trainee has acquired up to the present stage, as a result of the training. It is a conjecture, but seems probable that the observation of the behavior of the trainee is useful, when he is given a feedback, which is not a faithful representation of what has actually occurred. When the trainee has a firm belief as to what the correct action is to induce the desired event, he will not be affected by a disturbing feedback. While, when the trainee is skeptical of how he should act, he will be disturbed greatly. Of course, this kind of test should not be continued for long since the state of the trainee may change, and there may also be an objection form ethical viewpoint in cheating the trainee. In this study, the situation is simplified to examine the possibility of the above idea. The probabilistic model is constructed, where the major parameters are the probability of the trainee to act correctly, and the probability of the system to feedback honestly or dishonestly. As an example of a task, which is not very easy, the examinee is instructed to watch the flexor Capri ulnaris EMG signal, as well as the feedback signal, to maintain some 10 to 20 % of the maximum voluntary contraction. It is shown, by comparing the result of experiment and the result obtained from the model, that the extent of the belief of the trainee is low at the early stage of experiment, but is gradually improved with the progress of the training, which partially supports the conjecture.