Quality analysis of heart rate derived from functional near-infrared spectroscopy in stress assessment

Mahya Mirbagheri a, Naser Hakimi a, Elias Ebrahimzadeh a,b,c,d, S. Kamaledin Setarehdan a,*

a Control and Intelligent Processing Center of Excellence, School of Electrical and Computer Engineering, College of Engineering, University of Tehran, Tehran, Iran
b Seaman Family MR Research Centre, University of Calgary, Calgary, Alberta, Canada
c Department of Clinical Neurosciences, University of Calgary, Calgary, Alberta, Canada
d Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada

Background: Functional Near-Infrared Spectroscopy (fNIRS) is a non-invasive technique for studying brain hemodynamics. Since brain hemodynamics also involves components from the heart rate (HR), it is possible to extract the HR signal from the fNIRS (EHR). By removing global systemic changes from the EHR, we hypothesized that the EHR includes physiological information beyond HR, which is the functional response of the brain.

Method: To evaluate this hypothesis, the mental states of 10 healthy subjects were assessed under a stressful condition induced using the Montreal Imaging Stress Task (MIST). EHR and Reference HR (RHR) signals were then extracted from fNIRS and ECG, respectively. Next, a method based on Independent Component Analysis (ICA) was applied in order to discriminate the independent sources in the EHR signal. Thereafter, the least correlated components with RHR were reconstructed, which is the so-called target signal of EHR.

Results: Similarity measurement of EHR with fNIRS signals revealed a significant correlation between them. To show the effectiveness of the existing information in EHR signals, they were employed for mental stress assessment by using the SVM classifier, leading to an accuracy of 97.3 � ۰.۹%, which was 13%, 30.8% and 1.8% greater than those of the fNIRS, RHR, and EHR signals, respectively.

Conclusions: We showed not only that the EHR signal is a brain-related response, but also that it contains useful information for mental stress level measurement. Moreover, it is concluded that the EHR signal includes information from both brain and heart, which is hence useful for applications in which brain and heart functions are altered.