System and method for obtaining health data using a neural network

• US 10,888,280 B2
• Filed: 02/17/2018
• Issued: 01/12/2021
• Est. Priority Date: 09/24/2016
• Status: Active Grant

• Current Assignee: Trilinear Bioventures LLC
• Original Assignee: Sanmina-SCI Corporation
• Inventors: Newberry, Robert Steven
• Primary Examiner(s): Koharski, Christopher
• Assistant Examiner(s): Moss, James
• Application Number:
  US15/898,580
• Publication Number:
  US 20180214088A1
• Time in Patent Office: 1,060 Days
• Field of Search: None
• CPC Class Codes: A61B 2560/0223 of calibration, e.g. protocols for calibrating sensors
• CPC Class Codes: A61B 5/0002 Remote monitoring of patients using telemetry, e.g. transmission of vital signals via a communication network A61B5/07 takes precedence; transmission systems for measured values G08C; transmission H04B; transmission of digital information H04L; wireless communication networks H04W
• CPC Class Codes: A61B 5/0022 Monitoring a patient using a global network, e.g. telephone networks, internet
• CPC Class Codes: A61B 5/0077 Devices for viewing the surface of the body, e.g. camera, magnifying lens
• CPC Class Codes: A61B 5/024 Detecting, measuring or recording pulse rate or heart rate A61B5/0205, A61B5/021 take precedence
• CPC Class Codes: A61B 5/02416 using photoplethysmograph signals, e.g. generated by infrared radiation A61B5/14552 takes precedence
• CPC Class Codes: A61B 5/0816 Measuring devices for examining respiratory frequency measuring frequency of electric signals G01R23/00
• CPC Class Codes: A61B 5/14532 for measuring glucose, e.g. by tissue impedance measurement
• CPC Class Codes: A61B 5/1455 using optical sensors, e.g. spectral photometrical oximeters
• CPC Class Codes: A61B 5/14551 for measuring blood gases
• CPC Class Codes: A61B 5/6817 Ear canal
• CPC Class Codes: A61B 5/6826 Finger
• CPC Class Codes: A61B 5/7225 Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation A61B5/30, A61B5/316 take precedence; transformers for specific applications H01F38/00; amplifiers H03F
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: A61B 5/7275 Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
• CPC Class Codes: A61B 5/743 Displaying an image simultaneously with additional graphical information, e.g. symbols, charts, function plots
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/30 for calculating health indices; for individual health risk assessment
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders
• CPC Class Codes: Y02A 90/10 Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Abstract

A photoplethysmography (PPG) circuit or non-contact camera obtains PPG signals at a plurality of wavelengths. A signal processing module obtains at least a first spectral response around a first wavelength and a second spectral response around a second wavelength. The signal processing device generates PPG input data using the PPG signals, wherein the PPG input data includes one or more parameters obtained from each of the first spectral response and the second spectral response. A neural network processing device generates an input vector including the PPG input data and determines an output vector including health data, wherein the health data includes for example, an oxygen saturation level, a glucose level or a blood alcohol level.

Claims

1. A device, comprising:

an optical circuit including;

a plurality of light emitting diodes configured to emit light at least at a first wavelength in a range of 370 nm to 410 nm and at a second wavelength equal to or greater than 660 nm;

at least one photodetector configured to detect photoplethysmography (PPG) signals in response to pulsating blood flow, wherein the PPG signals include a first spectral response obtained from light reflected at the first wavelength from skin tissue of a patient and a second spectral response obtained from light reflected at the second wavelength from the skin tissue of the patient;

a signal processing circuit configured to generate PPG input data using the first spectral response at the first wavelength in a range of 370 nm to 410 nm and the second spectral response at the second wavelength equal to or greater than 660 nm; and

a neural network processing device implementing a machine learning algorithm, wherein one or more parameters of the machine learning algorithm are determined using a training set, wherein the training set includes training PPG input data obtained from a healthy population and corresponding known glucose levels from the healthy population, wherein the training PPG input data includes spectral responses at the first wavelength and at the second wavelength from the healthy population and wherein the neural network processing device is configured to;

determine a glucose level in blood flow of the patient from the PPG input data including the first spectral response at the first wavelength and the second spectral response at the second wavelength.

2. The device of claim 1, wherein the second wavelength is in one of:

a visible range or an infrared (IR) range.

3. The device of claim 1, wherein the PPG input data includes:

a value L<sub>λ

1</sub> generated using the first spectral response, wherein the value L<sub>λ

1</sub> isolates the first spectral response due to pulsating blood flow; and

a value L<sub>λ

2</sub> generated using the second spectral response, wherein the value L<sub>λ

2</sub> isolates the second spectral response due to pulsating blood flow.

4. The device of claim 3, wherein the PPG input data further includes:

a value R<sub>λ

1,λ

2</sub> obtained from a ratio of the value L<sub>λ

1</sub> and the value L<sub>λ

2</sub>.

5. The device of claim 1, wherein the PPG input data includes:

a first AC component signal I_AC generated using the first spectral response;

or

a second AC component signal I_AC generated using the second spectral response.

6. The device of claim 1, wherein the PPG input data includes:

a plurality of systolic points and diastolic points generated using the first spectral response; and

a plurality of systolic points and diastolic points generated using the second spectral response.

7. The device of claim 1, wherein the PPG signals are generated by at least one of:

a PPG circuit or a non-contact camera.

8. The device of claim 1, wherein the neural network processing device is further configured to:

periodically receive an updated learning vector, wherein the updated learning vector is generated from an updated training set, wherein the updated training set includes updated PPG input data obtained using updated PPG signals at the first wavelength and at the second wavelength and corresponding known glucose concentration levels; and

reconfigure one or more parameters of the machine learning algorithm for determining the glucose level in blood flow of the patient using the updated learning vector.

9. The device of claim 1, wherein the device is implemented within a user device or is configured to communicate with a user device.

10. A device, comprising:

a signal processing circuit configured to;

receive photoplethysmography (PPG) signals, wherein the PPG signals include a first spectral response obtained from light at a first wavelength reflected from skin tissue of a patient and a second spectral response obtained from light reflected at a second wavelength reflected from skin tissue of the patient, wherein the first wavelength is between 370 nm and 410 nm and wherein the second wavelength is equal to or greater than 660 nm;

generate PPG input data using the first spectral response obtained from light reflected at the first wavelength and the second spectral response obtained from light reflected at the second wavelength; and

a neural network processing device configured to;

pre-configure one or more parameters using a learning vector generated from a training set, wherein the training set includes additional PPG input data obtained from a healthy population and corresponding known nitric oxide (NO) levels, wherein the additional PPG input data includes additional spectral responses at the first wavelength and at the second wavelength; and

determine an NO level in blood flow from the first spectral response obtained from light reflected at the first wavelength and the second spectral response obtained from light reflected at the second wavelength.

11. The device of claim 10, wherein the second wavelength includes a wavelength in the IR range.

12. The device of claim 10, wherein the first wavelength is 390 nm or 395 nm.

13. The device of claim 10, wherein the signal processing module is configured to generate the PPG input data using the first spectral response and the second spectral response by generating characteristic features related to the shape of the PPG waveform from each of the first spectral response and the second spectral response.

14. The device of claim 13, wherein the signal processing module is configured to generate PPG input data using the first spectral response and the second spectral response by generating from each of the first spectral response and the second spectral response;

one or more of;

a pulse shape, average distance between pulses, variance, instant energy information, or energy variance.

15. The device of claim 10, wherein the neural network processing device is further configured to correlate one or more parameters obtained from at least one of the first spectral response and the second spectral response to an oxygen saturation level in blood flow.

16. The device of claim 15, wherein the neural network processing device is further configured to correlate the one or more parameters obtained from the first spectral response and the second spectral response to the oxygen saturation level in blood flow.

17. The device of claim 10, wherein the neural network processing device is further configured to correlate one or more parameters obtained from at least one of the first spectral response and the second spectral response to a glucose level in blood flow.

18. A device, comprising:

a signal processing circuit including a plurality of light emitting diodes and at least one photodetector configured to;

obtain photoplethysmography (PPG) signals at a first wavelength from skin tissue of a patient and at a second wavelength from skin tissue of the patient, wherein the first wavelength is in a range of 370 nm-410 nm and wherein the second wavelength is in a visible or an infrared (IR) range;

generate PPG input data using the PPG signals obtained from the patient at the first wavelength in the range of 370 nm-410 nm and at the second wavelength in the visible or the IR range; and

a neural network processing device configured to;

obtain one or more parameters generated from a training set, wherein the training set includes PPG signals obtained from a healthy population at the first wavelength in a range of 370 nm to 410 nm and at the second wavelength in the visible or the IR range and corresponding known glucose levels obtained from the healthy population;

receive the PPG input data; and

determine a glucose level in blood flow of the patient from the PPG signals obtained from the patient at the first wavelength in the range of 370 nm to 410 nm and at the second wavelength in the visible or the IR range.

19. The device of claim 18, wherein the second wavelength includes one of:

660 nm, 880 nm, 940 nm, or 1050 nm.

20. The device of claim 18, wherein the neural network processing device is further configured to determine an NO level in the blood flow from the PPG input data.

21. The device of claim 18, wherein the signal processing module is configured to generate the PPG input data from the PPG signals by generating one or more of:

a pulse shape, average distance between pulses, variance, instant energy information, or energy variance.

22. The device of claim 18, wherein the signal processing module is configured to generate the PPG input data from the PPG signals by generating characteristic features related to a shape of waveforms of the PPG signals.

23. The device of claim 18, wherein the neural network processing device is further configured to:

periodically receive an updated learning vector, wherein the updated learning vector is generated from an updated training set, wherein the updated training set includes additional PPG input data obtained using additional PPG signals at the first wavelength in the range of 370 nm-410 nm and at the second wavelength in the visible or IR range and known glucose concentration levels; and

reconfigure one or more parameters for generating the glucose level in blood flow of the patient from the PPG input data.

24. The device of claim 18, wherein the device is implemented within a user device or is configured to communicate with a user device.