Discovery of the new biometric response

     The main objective of the prototype development was to demonstrate the efficacy of the mechanical and optical design that placed a metal barrier between the LEDs and the phototdetector. Additionally, I wished to explore ways of overcoming the pulse oximetry motion artifact problem that results in missed SpO2 data and false alarms. Earlier prototype studies involved use of LEDs ranging from 500 nm to 660 nm; thus the graph labeling of "Amber" for the red values. Peak and trough sampling of the signal was performed by timing from an electromechanically sensed pulse at the radial artery. Two time interval values were set in the LabVIEW program using a graph cursor on full waveform graphs. Each heart cycle produced a peak and trough intensity value of red light and a peak and trough intensity value of infrared light. Calculation of SpO2 was performed with these data using the formula published in multiple pulse oximeter US patents.

     The sensor design and data calculation process appeared to be successful based on the prototype sensor producing SpO2 values that closely aligned with simultaneous values from a commercial pulse oximeter during hypoxia challenge tests performed by the subject briefly breathing nitrogen gas. Curiosity then led to also graphing the raw peak and trough values in a waveform graph. The above is a "screen-copy" of the graph generated during one of these studies. The upper trend, with trough and peak infrared (810 nm) values colored red and pink, respectively, rises as the hypoxic stress increases. Conversely, the lower trend, with trough and peak red (660 nm) values colored yellow and orange, respectively, decreases as hypoxic stress increases. The striking divergence of the non-pulsatile, average (DC) red and infrared data values far exceeded the pulsatile trough-peak (AC) variations and was unexpected. The graph also shows multiple tandem distortions that were purposely induced during the study by varying the pressure of the subject's thumb on the sensor's optical face to simulate sensor motion.

     Research of published literature did not find prior reference to the DC signal divergence. My first thought was that the divergence, although not likely to be useful in calculating blood oxygen, could at least indicate whether hypoxic stress was occurring when SpO2 calculations were distorted by the sensor motion. The initial pulse oximeter development work began in early 1999. The above image was recorded during prototype testing in January 2000, but the source of the DC divergence could not be resolved at that time. No further development work was performed until 2014 when we formed the company and rebuilt the prototype to further investigate the DC signals. The availability of miniaturized Bluetooth communications technology in 2014 also enabled creation of wearable devices. For additional details on the patented sensor technology, please refer to the "Patent Portfolio" page on this website.