Effective Neurofeedback Training of Large Electroencephalogram Signals Using Serious Video Games

Neurofeedback can be utilized to treat various neuropsychiatric disorders in children. However, therapists primarily set threshold values for neurofeedback training. Thus, the training effect becomes subjective owing to the experience of the therapist. A clinically inexperienced therapist could set inappropriate thresholds, rendering the training ineffective. In this study, an effective neurofeedback system that includes signal processing of large amount of electroencephalogram (EEG) data and auto thresholding and provides various training contents was developed. The system uses a method that determines optimal threshold values, which are significant for an effective neurofeedback system. The success or failure of the activation and inhibition of specific EEG frequencies was determined based on these threshold values. The system determined an optimal threshold value to obtain the target success rate using a numerical optimization technique. The success or failure feedback for the reward and inhibit EEG frequencies was generated using auto thresholding. This feedback was sent to the training contents by the inter-process communication module to control the contents. Most training content was implemented as serious video games by using a commercial game engine. Success feedback on reward EEG frequency leads to game progress. By contrast, failure feedback on inhibiting EEG frequency hinders game progress. Consequently, the user gains the self-regulation ability to enhance the reward EEG frequency and suppress the inhibit EEG frequency. A pilot study involving five children with attention deficiency was conducted to demonstrate the effectiveness of the developed system. The results demonstrated that the childrent’s attention improved after neurofeedback training.