Soft boundary-based neurofeedback training based on fuzzy similarity measures: A method for learning how to control EEG Signal features during neurofeedback training

Abstract

BackgroundMost commonly used neurofeedback training (NFT) methods are able to assist subjects towards an increase/decrease in EEG features. So, it is possible that the enhancement/inhabitation in a subject’s EEG features exceed normal limits if the process of changes in brain activity in the subject is very successful. This issue may also bring about a reduction in the effectiveness of NFT. New methodA soft boundary-based NFT method was proposed for learning how to control the EEG features during training. According to this method, an initial group was defined within which the training features of subjects’ EEG signals were placed prior to training and a target group was considered referring to what the features of the EEG signals should be shifted towards during training. In the course of training, the fuzzy similarity of EEG features of subject towards the target group center was measured and the subject’s score was increased if their fuzzy similarity was higher than a threshold. Within this method, an adaptive scoring index (the scores assigned to subjects for each achievement) was defined whose value was determined according to brain activity of the subject. ResultsIncrease/decrease in large amounts in the training features of subject’s EEG could lead to a descending trend in the scores received using the proposed method. Comparison with existing methodsThe proposed method may assist subjects to control their EEG signal features within the target group range. ConclusionThe proposed method may be able to prevent the side effects of neurofeedback.