Soft Boundary-based Neurofeedback Training procedure: A Method to Control EEG Signal Features during Neurofeedback Training Using Fuzzy Similarity Measures

Abstract

Most conventional neurofeedback training methods, often lead the subject toward increasing or decreasing the training features, which may result into side effects for the subject. In order to solve this problem, a training method which uses neurofeedback to learn how to control EEG features during training, was presented. This method defines a preliminary cluster and a target cluster. During training, the fuzzy similarity of training features of subjects’ EEGs towards the target cluster center is calculated and if it is greater than a threshold, subject’s score is increased. A scoring index (the number of points given to the subject for each success) is defined, whose value is determined according to subject’s performance. The results show that excessive increase or decrease of training features of subject’s brain signals will result in a decrease in scores received by the subject and lead the subject toward controlling feature of brain signals within target cluster’s range.