Neural-signal electroencephalogram (EEG) methods to improve human-building interaction under different indoor air quality

Abstract

In this study, neural-signal electroencephalogram (EEG) methods to improve human-building interaction under different indoor air quality conditions were investigated. Experiment was conducted to study correlations between EEG frequency bands and subjective perception as well as task performance. Machine learning-based EEG pattern recognition methods as feedback mechanisms were also investigated. Results showed that EEG theta band (4–8 Hz) correlated with subjective perceptions, and EEG alpha band (8–13 Hz) correlated with task performance. These EEG indices could be utilized as more objective metrics in addition to questionnaire and task-based metrics. For the machine learning-based EEG pattern recognition methods, the linear discriminant analysis (LDA) and support vector machine (SVM) classifiers can classify mental states under different indoor air quality conditions with high accuracy. In general, the EEG theta and alpha bands as more objective indices and the machine learning-based EEG pattern recognition methods as real-time feedback mechanisms have good potential to improve the human-building interaction.