In vivo investigation of ear canal pulse oximetry during hypothermia

Fig. 1. Block diagram showing the measurement setup used during the cold exposure test. PPG, ECG, and temperature signals were acquired using the setup

Fig. 2. A 3D sketch of ear canal PPG sensor, photograph of an assembled ear canal PPG sensor placed inside the right ear of a volunteer and a photograph of the finger sensor

Fig. 3. The raw infrared PPG signals acquired from the ear canal (green trace) and the finger (black trace) of a volunteer

Fig. 4. The power spectrum of the infrared raw PPG signals acquired from the ear canal and the finger of four healthy volunteers, highlighting the respiratory and cardiac components. The power of the respiratory component in the EC PPG signals is significantly higher in all subjects

Fig. 5. The infrared AC PPG signals acquired from the ear canal and the finger, and the simultaneously acquired ECG signal from one of the volunteers

Fig. 6. Infrared AC PPG signals acquired from (a) the ear canal, and (b) the finger of a volunteer and the simultaneously acquired (c) skin and (d) core temperature signal for the entire duration of the study. The Y-axis on the left shows the amplitude of the PPG signals while the Y-axis on the right shows the temperature. The spike in finger PPG signal at 11th min is a movement artefact

Fig. 7. NPA of a the finger and b the ear canal PPG signals acquired from 15 volunteers during all three stages of the experiment. Each box shows the mean NPA measured across a 2 min period in all the volunteers. The red line in each box shows the median value of the data, the shows the mean, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$+$$\end{document}+ shows the outliers

Fig. 8. Box and whiskers plots demonstrating the change in mean SpO\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$_2$$\end{document}2 measured for every two minutes of the study in a the finger and b the ear canal. SpO\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$_2$$\end{document}2 estimated from the finger have dropped significantly towards the end of the study when compared to the ear canal. The red line in each box shows the median value of the data, the shows the mean, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$+$$\end{document}+ shows the outliers

Fig. 9. The red finger PPG signal (green trace) and the ECG signal (blue trace) acquired from a volunteer towards the end of the cold exposure, along with the peaks detected (red trace) by the peak detection algorithm. The amplitude of the red finger PPG signals was sometimes so weak that the peak detection algorithm could not distinguish between the heart beat peaks and the noise peaks

Fig. 10. Red PPG signals acquired from the fingers of the volunteers in whom accurate estimation of SpO\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$_2$$\end{document}2 was not possible during the cold exposure. The PPG signal is indistinguishable from noise in these cases