Analysing Wireless Eeg Based Functional Connectivity Measures With Respect To Change In Environmental Factors

In this paper we present a systematic exploration to formulate a predictive model of the human cognitive process with the changing environmental conditions at workplace. We select six different environmental conditions with small change in temperature/ventilation representative of realistic work environment having manual control. EEG data were acquired through 19-channel wireless system from three participants and CO2, Temperature, Relative humidity were recorded throughout the six conditions. The EEG data was pre-processed using an artifact reduction algorithm and 129 neurophysiological features were extracted from functional connectivity measures using complex network analysis. The environmental data were processed to generate 15 time/frequency domain features. Five best features selected through a ranking algorithm for all the variables across the six conditions were processed to formulate a model (environmental parameters as predictors) using retrospective 10-fold cross-validation in conjunction with multiple linear regression. The model was prospectively evaluated over 10 runs on a test set to predict the EEG variable across the six conditions and parameters corresponding to the run producing least root mean square error were reported. Our exploration shows that the condition having no modulation of the ambient environmental parameters reflects the optimum condition for predicting the EEG features using the examined environmental parameters.