An Intelligent Model to Predict the Impact on Health Due to Commuting to Work on a Regularbasis

A regular commute to work can become a source of chronic stress, which is reflected in physicaland emotional responses. We sought to investigate the physiological effects of commuting in GreaterLondon as measured by electroencephalography (EEG) and blood pressure (BP) and to develop aneurophysiological model which predicts higher stress levels using machine learning. Participantswere 45 healthy volunteers (18 females), with a mean age of 32 years. Mode of commute werebus (n=8), driving (n=6), cycling (n=7), train (n= 9), tube (n=13) and both bus and train (n=2).Participants wore non-invasive wearable biosensor technology to measure EEG, and BP during theirmorning commute, each day consecutively for 5 days. Positive andNegative Affect Schedule (PANAS)ratings were acquired before and after each commute as a measure of their state. Correlation analysiswas applied to find the significant features associated with stress as measured by a reduction in thepositive ratings in the PANAS. Machine learning was applied to select the best algorithmic structurethat predicted high-stress levels. Following the commutes, BP and EEG beta waves were significantlyincreased and the positive PANAS rating decreased from 34.73 to 28.60. Combining EEG beta wavesand BP predicted systolic BP were higher when the EEG Beta low power exceeded Alpha low powerpost commute then pre commute using Random Forest at 83% accuracy, K-Nearest Neighbour at 80%accuracy, Support Vector Machine at 80% accuracy, and Naïve Bayes at 73% accuracy.