Patent Application
20220192549
In the human body, cardiac rhythm changes the blood volume passing through the arteries, which generates a pulsatile signal that can be optically measured using a light source and a detector; this optical sensing technique is known as photoplethysmography (PPG). Generally, the PPG signal is used for calculating heart rate by utilizing only one light source, and for measuring oxygen saturation (SpO2) by employing two light sources. Pulse oximeters measure SpO2 of blood by using PPG signals at two distinct wavelengths where light absorption in oxygenated and deoxygenated blood is different. PPG and oximetry can be performed in both transmission and reflection mode. Conventionally, transmission-mode pulse oximeter sensors composed of solid-state light-emitting diodes (LEDs) and photodiodes (PDs) are used to measure SpO2 at the extremities of the body where light can easily penetrate thin regions of tissue, such as the earlobes and the fingertips. However, this method of measuring SpO2 presents a few limitations—(i) Transmission-mode oximetry has limited sensing locations, and (ii) Solid-state LEDs and PDs do not conform well to the skin, therefore, reduce the signal-to-noise ratio (SNR).
Over the past few years, flexible and wearable sensors are getting significant attention in both academic research and industry due to their skin conformable form factors. Consequently, flexible optical sensors have been extensively studied for PPG and oximetry as they enhance SNR and provide design versatility. Sensor fabrication and sensing methodology remain a strong focus of recent reports. However, the reflectance oximeter sensor design, which is a crucial component of reflection-mode PPG and oximetry, is not well-reported in the literature. In addition, wearable reflection-mode PPG sensors and oximeters are prone to different types of noise, such as motion artifacts (MAs), thermal noise, and electromagnetic interference. Thermal noise and electromagnetic interference are each a high-frequency noise and can be eliminated through filtering. MA, however, is challenging to remove from the PPG signals. Adaptive filtering and comparing PPG signal to a reference accelerometer signal are popular techniques for reducing MAs. Furthermore, multi-channel PPG signals can also be utilized to extract heart rate and oxygenation information from channels that are less affected by MAs. The multi-channel PPG approach does not require additional hardware blocks or a reference signal.
In various embodiments herein, varied reflectance oximeter sensor designs with varying device geometry, light emitter and detector spacing are provided. In certain embodiments, an optical barrier between the emitter and the detector is provided to maximize sensor performance. Additionally, in certain embodiments, a printed, flexible, two-channel reflectance oximeter is provided, which is able to collect PPG signals using red and near-infrared (NIR) (or green) organic light-emitting diodes (OLEDs) and organic photodiodes (OPDs). Inverse-variance weighting and template matching algorithms are used, in certain aspects, to improve the detection of heart rate from the multi-channel PPG signals. The various embodiments are useful for and in wearable smart watches and wristbands.
According to an embodiment, a pulse oximeter device is provided that includes a first light emitting element configured to emit red light, a second light emitting element configured to emit green light or near infrared (NIR) light, and a sensor element configured to detect red and green light or detect red and NIR light. The sensor element, in an embodiment, has a substantially circular geometry, and wherein the first light emitting element and the second light emitting element each have a substantially arc-shaped geometry, and wherein the sensor element is positioned between the first light emitting element and the second light emitting element. In another embodiment, the first and second light emitting elements together have a substantially circular geometry, and wherein the sensor element comprises one or more substantially arc-shaped portions, and wherein the first and second light emitting elements are partially surrounded, or fully surrounded, by the sensor element. In another embodiment, the sensor element has a substantially rectangular geometry, and wherein the first light emitting element and the second light emitting element each have a substantially bracket-shaped geometry, and wherein the sensor element is positioned between the first light emitting element and the second light emitting element. In another embodiment, the first and second light emitting elements together have a substantially rectangular geometry, and wherein the sensor element comprises one or more substantially bracket-shaped portions, and wherein the first and second light emitting elements are partially surrounded, or fully surrounded, by the sensor element.
In certain aspects, the first light emitting element comprises a first light emitting diode (LED), wherein the second light emitting element comprises a second LED, and wherein the sensor element includes a photodetector.
In certain aspects, each of the first and second LEDs comprises an organic LED, and wherein the photodetector comprises an organic photodiode.
In certain aspects, the device further includes a flexible substrate, wherein the first light emitting element, the second light emitting element and the sensor element are formed on the flexible substrate. In certain aspects, the flexible substrate comprises polyethylene napthalate (PEN).
In certain aspects, the sensor element is configured to detect light transmitted through tissue containing blood. In certain aspects, the sensor element is configured to detect light reflected by tissue containing blood.
In certain aspects, the device further includes a signal processing element, such as one or more processors with associated memory, configured to receive and process the signals representing detected red and green light or detected red and NIR light output by the sensor element to produce signals that represent blood oxygenation content.
In certain aspects, the device further includes a light blocking feature positioned between the sensor and at least one of the first light emitting element or the second light emitting element.
In certain aspects, the device further includes a light blocking feature positioned between the sensor and both the first light emitting element and the second light emitting element. In certain aspects, the light blocking feature includes a black tape.
In certain aspects, a spacing between the sensor and each of the first light emitting element and the second light emitting element is between about 2 mm and about 6 mm.
In certain aspects, a wearable sensor device comprising a pulse oximeter device according to any of the described embodiments is provided.
In certain aspects, a method of making a pulse oximeter device according to any of the described embodiments is provided.
Some embodiments further include a non-transitory computer-readable storage medium storing program code including instructions that, when executed by a processor or processors, cause the one or more processors to perform one of the methods of processing the signals representing detected red and green light or detected red and NIR light output by the sensor element to produce signals that represent blood oxygenation content, as described herein. Non-exclusive examples of non-transitory computer-readable storage media include any medium that can store program code, such as a USB flash drive, a removable hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc.
Reference to the remaining portions of the specification, including the drawings and claims, will realize other features and advantages of the present invention. Further features and advantages of the present invention, as well as the structure and operation of various embodiments of the present invention, are described in detail below with respect to the accompanying drawings. In the drawings, like reference numbers indicate identical or functionally similar elements.
A schematic illustration of a two-channel (or two-pixel) wrist-worn reflectance PPG sensor, according to an embodiment, is shown in
Reflection-mode sensors typically require light emitters and detectors assembled on a substrate or a circuit board. LEDs (e.g., red and NIR LEDs or red and green LEDs) are placed on either side of the PD to assemble the sensor. The designs of commercially available optoelectronic sensors are limited in shape—typically rectangular, which do not provide much versatility to vary the sensor geometry. On the other hand, printed optoelectronics can be fabricated in various shapes and sizes. Herein, three different sensor geometry embodiments are discussed as shown in
PPG signal magnitudes vary appreciably based on the sensor placement locations on the wrist. Accordingly, three sensing locations were explored: (i) On top of the wrist, (ii) on top of the ulnar artery, and (iii) on top of the radial artery as shown in
A base polyethylene naphthalate (PEN) substrate is used to assemble the reflectance sensor. Inkjet-printed silver traces are used to route connections from the optoelectronics to the control electronics, which include an AFE and a microcontroller with a universal serial bus (USB) interface to a computer. The OLEDs and the OPD are printed on separate plastic substrates and then assembled on the PEN substrate with silver traces as shown in
Each OLED is then calibrated to the maximum current measured in a batch of devices, KOLED=maxISiPD(ISiPD). In the second step, the OPDs are calibrated by recording the OPD current while running the solid-state red LED at a fixed drive current. The red LED of the calibration platform is turned on and the OLEDs are turned off for calibrating the OPDs. The fabricated OPD detects light from the red LED and the measured photocurrent is recorded. Similar to the OLEDs, each OPD is then calibrated to the maximum OPD current measured in a batch of devices, KOPD=maxIOPD(IOPD). The obtained values are then used together with the measured PPG signal to calculate the calibrated signal magnitude. The calibration equation is given below, which is used to compare sensor performances for the three different geometries.
PPGcal=KOLED•KOPD•PPGmeas[mV]
After the calibration step, a fair comparison among the three different geometries can be performed. Additionally, another important design parameter, emitter-detector spacing, d, is evaluated.
The rectangular sensor includes OLEDs and an OPD that are all substantially square-shaped and which are placed side-by-side. Since the OLEDs do not surround the OPD from the top and the bottom, this scheme may be susceptible to ambient light, which could contribute to the noise of the measurement. Also, a significant amount of light coming out from the left edge of the red OLED and the right edge of the NIR OLED may not contribute to the measurement, and hence, gets lost. A perimeter light source that surrounds the OPD would be desired. The bracket design and the circular design schemes, where the light sources encompass the perimeter of the OPD, enhance measurement SNR. As shown in
For a direct comparison of the different geometries, the emitter-detector spacing was kept constant at 2 mm, and the device area of the OLEDs and the OPDs was kept the same for all three geometries as shown in
In reflectance PPG and oximetry, the light coming back from the arteries contributes to the signal, while the light scattered from the skin surface contributes to noise. Therefore, blocking the light scattered from the skin surface enhances SNR. In an embodiment, a light blocking feature is incorporated in the design, e.g., by utilizing an optical barrier between the OLEDs and OPD. In an embodiment, black tape may be used, e.g., cut into the shape that fits the area between the OLEDs and the OPD and applied therebetween to block scattered light.
Wearable PPG sensors are susceptible to thermal noise, electromagnetic interference, and MAs. While thermal noise and electromagnetic interference can be reduced with filtering, reducing MAs requires additional hardware and software processing. Adaptive filtering is a popular technique for addressing MAs in PPG signals. Another approach is to simultaneously record PPG and a reference signal such as an accelerometer signal and apply hybrid algorithms to determine heart rate (HR) and pulse oxygenation. Multi-channel PPG acquisition and processing can also be used to reduce MA by utilizing channels that are lightly influenced by MA. Multiple PPG channels add redundancy to the measurement for signal quality assessment, which is important for properly extracting HR and pulse oxygenation values. To process multi-channel data, two algorithms may be used: (1) Template matching (TM) with an ideal PPG signal, and (2) Inverse-variance weighting (IVW). The efficacy of both methods in acquiring high-quality PPG signal, and extracting HR are examined. The process flow of the template matching and inverse-variance weighting algorithms are shown in
Both TM and IVW algorithms are used to obtain a weighted PPG signal from multi-channel PPG. The equation for obtaining the weighted PPG is given in Eq. 2.
Here, PPGw is the weighted PPG from all channels, Wi is the weight for channel i determined by either of the two methods discussed in following subsections and PPGi is the PPG signal from channel i.
Template Matching (TM) with an Ideal PPG Signal
Template matching is a popular data processing techniques in biomedical signal processing. TM has been widely used in processing electroencephalography (EEG), electrocardiography (ECG), and PPG data. An ideal PPG template was used to determine the fidelity of the signal from each channel. The ideal template can be obtained from experimental data or by modeling. A small window is acquired from each channel after filtering. Then, troughs are detected to find the pulses in each window. Next, the correlation coefficient of each pulse with the ideal template is calculated. If the correlation coefficient is positive, the correlation coefficient is used as a weight to calculate the weighted average of the two signals. If the correlation coefficient for a pulse is negative, the pulse is ignored, i.e., the weight for that pulse in that channel is set to zero. Thus, using this method, the weight W's in Eq. 2 can be given by Eq. 3.
Here, ρi is the correlation coefficient between a pulse in channel i and the PPG template. HR and oxygenation values are determined from the weighted signal. The process flow for TM is presented in
In inverse-variance weighting algorithm, W, weight for channel i in Eq. 2 are assigned based on the standard deviation of heart rate variability in a specific time window. First, peaks and troughs of the signal from each window are determined. Then, the HR is calculated from the distances in between peaks or troughs. The channel with higher standard deviation in HR is assigned lower weight, because, in a small time-window of PPG signal, HR should not change too drastically. The weight assignment, in this case, is described by Eq. 4.
Here, σi is the standard deviation of HR in channel i. After assigning the weight, the weighted average of the signals (PPGw) is computed and the HR and other features are extracted from the signal. The process flow is presented in
To test the efficacy of the TM and IVW algorithms, a simulated dataset is used to determine HR variability over time. The simulated dataset is designed to represent HR variability while performing an exercise as shown in
The effect of noise on the accuracy of determining HR using TM and IVW algorithms is examined by adding noise of frequency below 5 Hz to one of the channels. This frequency range of noise is chosen because noise with a frequency above 5 Hz can be removed from the signal using a low-pass filter (LPF). The effect of SNR is shown in
In addition to the simulated dataset, three sets of PPG dual channel data reported by Zhang et al. (Zhang, Z., Pi, Z. & Liu, B. Troika: A general framework for heart rate monitoring using wrist-type photoplethysmographic signals during intensive physical exercise. IEEE Transactions on biomedical engineering 62, 522-531 (2015).) were used to test the efficacy of the TM and IVW algorithms. The results are summarized in Table 1, below. For the datasets, I and III the HR calculated using both methods are close to the ground truth HR, i.e., within about 2 b.p.m. However, in dataset II, both channels are more affected by MA, so the calculated HRs are further away from ground truth HR. For accurate detection of HR, at least one of the channels should be minimally affected by MA so that the PPG pulses are recognizable.
After validating the TM and IVW algorithms on the simulated and literature datasets, both methods were employed for processing the data collected by the printed multi-channel PPG sensor. The sensor is placed on the underside of the wrist, where Ch 1 collects data from the ulnar artery and Ch 2 collects data from the radial artery as shown in
By utilizing the versatility of printed electronics, optoelectronic sensors for PPG and oximetry may advantageously be fabricated in different shapes and sizes. Embodiments herein provide novel and non-conventional geometries such as bracket and circular designs to improve sensor performance. These sensor geometries demonstrated a clear improvement over the conventional rectangular sensor design. These sensor geometries also not only improved the PPG signal magnitudes but also decreased the overall sensor length and reduced power consumption. These sensor designs coupled with multi-channel redundancy can be incorporated into wrist-worn devices, making them extremely promising for wearable reflectance PPG and oximetry.
The OLEDs were printed on 125 μm thick indium tin oxide (ITO) patterned PEN substrates. The substrate was treated with (heptadecafluoro1,1,2,2-tetrahydrodecyl) (Gelest SIH5841.0) for 20 min under light vacuum (0.1-1 Torr) in order to make the surface hydrophobic. The substrate was then patterned by exposing the active area of the OLEDs with a plasma etcher to selectively etch off the hydrophobic layer. Then poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT:PSS) (Clevios AI4083, Heraeus), interlayer, and the semiconducting polymers from Cambridge Display Technologies Ltd. (CDT) were subsequently blade-coated to form the emissive layer of the OLEDs. After all the layers were blade-coated, the samples were transferred to a thermal evaporator in a glovebox to evaporate calcium (99.5%, Strem Chemicals) and aluminum (99.999%, ACI Alloys) to finish the OLED stack. The shape of the top electrode determined the shape of the OLED emission. Each pixel was then encapsulated by face sealing using a UV curable epoxy (Delo Katiobond LP612) and plastic film (PQA1) on top. The details of the OLED fabrication process is described in details by Khan et al. (Khan, Y. et al. A flexible organic reflectance oximeter array. Proc. Natl. Acad. Sci. 115, E11015-E11024 (2018).).
The OPDs were printed on top of planarized PEN substrates (TeiJin PQA1) using blade coating techniques. The substrate was placed in a vacuum with 40 μL of heptadecaflouropolymer for 20 min to render substrate hydrophobic. A stainless steel stencil with cutouts of the desired PEDOT:PSS area was placed on top of the substrate, then treated for 1.2 min of oxygen plasma in a Diener Nano plasma system. Next, PEDOT:PSS was blade-coated, and the substrate was then annealed for 10 min at 120° C. 1:2 CDT Donor:PC70BM in 95:5 3,3′,5,5′-Tetramethylbenzidine:Benzyl Benzoate (TMB:BB) was blade-coated, and the substrate was then annealed for 1.5 h at 120° C. Finally, aluminum cathode at a base pressure of 5.10−6 Torr at a rate of 3-5 As−1 was evaporated to finish the OPD stack. The details of the OPD fabrication process is described in details by Khan et al. (Khan, Y. et al. A flexible organic reflectance oximeter array. Proc. Natl. Acad. Sci. 115, E11015-E11024 (2018).).
A base PEN substrate was used to assemble the sensor. Inkjet-printed silver (ANP DGP 40LT-15C) traces were used to connect the OLEDs and the OPDs keeping an emitter-detector spacing of 2, 4, or 6 mm. Then fabricated red and NIR OLEDs and OPD pixels were connected to the traces using silver paste. The silver traces were connected to a flat flexible cable (FFC) connector for interfacing the sensor to the control electronics. To calibrate the PPG sensors, a calibration platform composed of a silicon photodiode (Hamamatsu 52387 series) and a red LED (5 mm clear red LED, 660 nm peak emission) were used. A two-step calibration was used to account for the batch-to-batch device variability of the OLEDs and the OPDs. In the first step, the OLEDs were calibrated using the silicon photodiode by operating the OLEDs at 10 mA cm−2 and recording the photodiode current. Each OLED was then calibrated to the maximum current measured in a batch of devices. In the second step, the OPDs were calibrated by recording the OPD current while running the solid-state red LED at 0.7 mA drive current. Similar to the OLEDs, each OPD was then calibrated to the maximum OPD current measured in a batch of devices. The obtained values were then used together with the measured PPG signal to calculate the calibrated signal magnitude.
A Texas Instruments analog front end (AFE4490) was used to sequentially drive the OLEDs and read out the OPD signal. The two channels were selected using analog switches (Analog Devices ADG1608). The AFE was controlled with an Arduino Due microcontroller. The OLEDs were driven at 10 mA cm−2 with a 9V battery in push-pull mode. A two-stage OPD gain circuitry was used to amplify the photocurrent. 100 kΩ feedback resistor was used in the first stage, and unity gain was used in the second stage. Finally, the data was collected using a Universal Serial Bus (USB) interface and was processed using a custom in-house software.
PPG data was collected from the two channels using the control electronics. Then, Template matching and inverse-variance weighting algorithms were used to obtain weighted PPG signals from multi-channel PPG. The weight assignment for the TM algorithm was calculated by Eq. 3, and the weight assignment for the IVW algorithm was calculated by Eq. 4. Then, the heart rate was calculated by timing the distances in between peaks or troughs of the weighted PPG signal.
U.S. Patent Application Publication No. 2017/0156651 A1, which is incorporated herein by reference, discloses various aspects of PPG and oximetry measurements, including reflectance-based measurements, as well as useful PPG device materials. All references, including publications, patent applications, and patents, cited herein are hereby incorporated by reference to the same extent as if each reference were individually and specifically indicated to be incorporated by reference and were set forth in its entirety herein.
The use of the terms “a” and “an” and “the” and “at least one” and similar referents in the context of describing the invention (especially in the context of the following claims) are to be construed to cover both the singular and the plural, unless otherwise indicated herein or clearly contradicted by context. The use of the term “at least one” followed by a list of one or more items (for example, “at least one of A and B”) is to be construed to mean one item selected from the listed items (A or B) or any combination of two or more of the listed items (A and B), unless otherwise indicated herein or clearly contradicted by context. The terms “comprising,” “having,” “including,” and “containing” are to be construed as open-ended terms (i.e., meaning “including, but not limited to,”) unless otherwise noted. Recitation of ranges of values herein are merely intended to serve as a shorthand method of referring individually to each separate value falling within the range, unless otherwise indicated herein, and each separate value is incorporated into the specification as if it were individually recited herein. All methods described herein can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The use of any and all examples, or exemplary language (e.g., “such as”) provided herein, is intended merely to better illuminate the invention and does not pose a limitation on the scope of the invention unless otherwise claimed. No language in the specification should be construed as indicating any non-claimed element as essential to the practice of the invention.
Various embodiments are described herein. Variations of those embodiments may become apparent to those of ordinary skill in the art upon reading the foregoing description. The inventors expect skilled artisans to employ such variations as appropriate, and the inventors intend for the embodiments to be practiced otherwise than as specifically described herein. Accordingly, this specification includes all modifications and equivalents of the subject matter recited in the claims appended hereto as permitted by applicable law. Moreover, any combination of the above-described elements in all possible variations thereof is encompassed by the disclosure unless otherwise indicated herein or otherwise clearly contradicted by context.
This patent application is a continuation of PCT Application No. PCT/US2020/053478 by Yasser T. Khan et al., entitled “ORGANIC MULTI-CHANNEL OPTOELECTRONIC SENSORS FOR WEARABLE HEALTH MONITORING,” filed Sep. 30, 2020, which claims priority to U.S. Provisional Patent Application No. 62/908,219 by Yasser T. Khan, entitled “ORGANIC MULTI-CHANNEL OPTOELECTRONIC SENSORS FOR WEARABLE HEALTH MONITORING,” filed Sep. 30, 2019, each of which is incorporated herein by reference in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| 62908219 | Sep 2019 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | PCT/US2020/053478 | Sep 2020 | US |
| Child | 17691437 | US |
