TISSUE HYDRATION MONITOR
1. A tissue hydration monitor, comprising:
- a housing configured to be disposed against a user'"'"'s skin;
a sensor module disposed within the housing, the sensor module comprising;
a plurality of LEDs configured to emit light toward the user'"'"'s skin at a plurality of different wavelengths; and
a detector configured to detect light from each of the plurality of LEDs transmitted and reflected through the user'"'"'s skin over a period of time and generate signals corresponding to an intensity of detected light at each of the different wavelengths;
a processor/controller module configured for receiving signals from the sensor module, executing an algorithm for generating a baseline hydration level based on the received signals, calculating a relative hydration level at time points within the period of time, and generating an output indicative of relative hydration at the time points; and
a power supply configured to provide power to the sensor module and the processor/controller module.
A tissue hydration monitor and method includes a sensor module having a plurality of LEDs positioned to emit a plurality of different wavelengths of light toward the user'"'"'s skin and a detector that detects light transmitted and reflected through the user'"'"'s skin to generate signals corresponding to an intensity of detected light at each of the different wavelengths. A processor/controller module generates a baseline hydration level based on the received signals, calculates a relative hydration level, and generates an output indicative of relative hydration personalized to the user. The housing is secured against the user'"'"'s skin by an adhesive patch or a strap.
- 1. A tissue hydration monitor, comprising:
a housing configured to be disposed against a user'"'"'s skin; a sensor module disposed within the housing, the sensor module comprising; a plurality of LEDs configured to emit light toward the user'"'"'s skin at a plurality of different wavelengths; and a detector configured to detect light from each of the plurality of LEDs transmitted and reflected through the user'"'"'s skin over a period of time and generate signals corresponding to an intensity of detected light at each of the different wavelengths; a processor/controller module configured for receiving signals from the sensor module, executing an algorithm for generating a baseline hydration level based on the received signals, calculating a relative hydration level at time points within the period of time, and generating an output indicative of relative hydration at the time points; and a power supply configured to provide power to the sensor module and the processor/controller module.
- View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14)
- 15. A method for monitoring hydration in a subject, the method comprising:
placing a sensor device against the skin of the subject, the sensor device comprising a plurality of LEDs configured to emit light toward the user'"'"'s skin at a plurality of different wavelengths; detecting light from each of the plurality of LEDs transmitted and reflected through the user'"'"'s skin over a period of time and generating intensity signals corresponding to light at each of the different wavelengths; generating a baseline hydration level from the intensity signals; calculating a relative hydration level relative to the baseline hydration level at a plurality of time points within the period of time; and generating an output indicative of relative hydration at the time points.
- View Dependent Claims (16, 17, 18, 19, 20, 21)
This application claims the benefit of the priority of U.S. Provisional Application No. 62/663,926, filed Apr. 27, 2018, which is incorporated herein by reference in its entirety.
The present invention relates to a tissue hydration monitor and more particularly to a tissue hydration monitoring system and method for personalized hydration monitoring.
Dehydration is a condition in which water in a living body decreases below the individual'"'"'s normal functioning level. Dehydration can occur when an individual exerts him or herself for extended periods of time with inadequate water intake to offset losses of internal water from respiration, perspiration, and waste removal, or the temperature rises to a point where an individual cannot produce enough sweat to maintain their normal body temperature. Persons that regularly exert themselves in low humidity and/or high temperature conditions and/or for extended periods of time are prone to experience dehydration or dehydration symptoms. Elderly persons and children are also especially prone to experience dehydration or dehydration symptoms.
In less severe cases of dehydration, an individual'"'"'s ability to perform tasks may begin to deteriorate. For example, in the case of endurance or professional athletes, an individual that becomes dehydrated by loss of as little as 2% body weight may begin to experience a decline in performance. Low levels of hydration lead to low blood volume, compromising circulation, nutrient exchange, hormone balance, and waste removal. When dehydrated, sodium levels in the blood decrease, resulting in hyponatremia, the first signs of which can include fatigue, headache, weakness and nausea. Additional manifestations include cramping, disorientation and confusion, swelling of extremities, and, in extreme cases, swelling of the brain. Cramping can be common in athletes and is a good key indicator that the body has depleted its electrolytes. Losses in excess of 5% of body weight can decrease the capacity of an individual to perform a task by as much as 30%.
There is currently no wearable sensor on the market that provides accurate and personalized hydration requirements. It is important for athletes to have a self-calibrated measurement of hydration needs because each athlete will have particular needs based on their diet, exercise and genetics. In order to maintain peak performance, it would be desirable to monitor an individual'"'"'s hydration level regularly, or even continuously, allowing the detection of fluctuations at early stages before performance levels are impacted, and especially before the person approaches a critical dehydration condition. The present invention is directed to a method and device that allows one to easily monitor and provide notification of hydration levels to permit early correction before dehydration occurs.
In an exemplary embodiment, a wearable optical device uses light emitting diodes that emit light and one or more photo diodes that detect light from the skin. In a preferred embodiment, the device will use at least 4 optical wavelengths, e.g., 740 nm, 850 nm, 940 nm, 1450 nm, and a single detector. Other wavelengths may be used, for example, at variations of +/−10 nm based on LED availability and technical adjustments. Ideally, the LEDs and photo diode will be positioned at least 1 cm away from each other with a light blocking partition in between to stop any light from traveling directly from an LED to detector without penetrating the skin. The depth the light will travel in tissue depends on the wavelength and the source-detector separation and tissue constituents, up to several centimeters deep.
The wearable device includes a wireless communication module, for example, BLUETOOTH® or similar wireless communication circuitry and corresponding software, to communicate with a smart phone, tablet, or computer in which an application (“app”) has been stored for entering settings and for receiving and storing data.
In one aspect of the invention, the system controller collects information from the user over a pre-determined time period to establish a personalized baseline from which subsequent performance is measured. Thresholds for generating alarms or other indicators of deviation may be pre-set by the system controller, or the user may tighten or expand the threshold range using the associated app. In one embodiment, the system controller, or the associated app on the user'"'"'s phone, tablet or computer, may execute a learning algorithm or other appropriate algorithm to continuously or periodically update the user'"'"'s personalized baseline. For example, as the user'"'"'s conditioning improves, or as the intensity of workouts increases, he or she may experience changes in hydration responses which should be taken into account when determining acceptable ranges.
The device may be fitted with a strap to allow the device to be worn on a wrist, upper or lower arm, ankle, calf, or other location on a limb. The strap should be sufficiently pliable and elastic to firmly press the sensor area against the skin to prevent outside light from entering the sensor, and to ensure that the light-blocking partition is effective in preventing light from the LEDs from being picked up directly by the detector without passing through the skin. In a preferred embodiment, the sensor is temporarily affixed to the skin using a non-irritating, skin-friendly pressure-sensitive adhesive (“PSA”). Such adhesives are commercially available from a number of sources for securing skin-bonded devices for periods of time of a few hours up to a week.
The present invention relies upon on near-infrared spectroscopy of tissues. This method is not novel and has been extensively researched. There have been previous patents filed using the near-infrared spectroscopy method to measure tissue hydration. However, the prior art either uses different wavelengths, different algorithms, or both, and does not provide a personalized calibration system for determining an optimal hydration range.
In one aspect of the invention, a tissue hydration monitor includes a housing configured to be disposed against a user'"'"'s skin; a sensor module disposed within the housing, the sensor module including: a plurality of LEDs configured to emit light toward the user'"'"'s skin at a plurality of different wavelengths; and a detector configured to detect light from each of the plurality of LEDs transmitted and reflected through the user'"'"'s skin over a period of time and generate signals corresponding to an intensity of detected light at each of the different wavelengths; a processor/controller module configured for receiving signals from the sensor module, executing an algorithm for generating a baseline hydration level based on the received signals, calculating a relative hydration level at time points within the period of time, and generating an output indicative of relative hydration at the time points; and a power supply configured to provide power to the sensor module and the processor/controller module. The period of time may include periods of user activity ranging from resting to exercising, wherein the baseline hydration range is generated at time points during resting and the relative hydration level is calculated at multiple time points during exercising. The plurality of different wavelengths are wavelengths absorbed by oxygenated hemoglobin, deoxygenated hemoglobin, lipid, and water and are preferably 740 nm, 850 nm, 940 nm and 1450 nm. An indicator responsive to the processor/controller module is provided for generating an alert indicative of a predetermined deviation from the baseline hydration level. The predetermined deviation is preferably a change of less than or equal to +2%. The indicator may be an LED disposed to emit light from a surface of the housing. A light shield is preferably disposed around the detector to prevent stray light from impinging upon the detector.
In some embodiments, an adhesive material is applied to the housing to removably seal the sensor to the user'"'"'s skin. The adhesive material may be in the form of a patch dimensioned to seal the perimeter of the housing against the user'"'"'s skin to produce a substantially light-tight seal. The patch may be formed from at least a breathable material and a water resistant material. In other embodiments, a strap may be attached to the housing for retaining the housing against the user'"'"'s skin.
The tissue hydration monitor may further include a communication module in communication with the processor/controller module for transmitting data to a remote mobile device for displaying the output. The mobile device may have an application installed therein with instructions for further processing of the output.
In another aspect of the invention, a method for monitoring hydration in a subject includes placing a sensor device against the skin of the subject, the sensor device comprising a plurality of LEDs configured to emit light toward the user'"'"'s skin at a plurality of different wavelengths; detecting light from each of the plurality of LEDs transmitted and reflected through the user'"'"'s skin over a period of time and generating intensity signals corresponding to light at each of the different wavelengths; generating a baseline hydration level from the intensity signals; calculating a relative hydration level relative to the baseline hydration level at a plurality of time points within the period of time; and generating an output indicative of relative hydration at the time points. The plurality of different wavelengths are wavelengths absorbed by oxygenated hemoglobin, deoxygenated hemoglobin, lipid, and water and may be 740 nm, 850 nm, 940 nm and 1450 nm. The method may include transmitting data to a remote mobile device for displaying the output and generating an alert indicative of a predetermined deviation from the baseline hydration level, where the predetermined deviation is a change of less than or equal to +2%. The period of time includes periods of subject activity ranging from resting to exercising, wherein the baseline hydration range is generated at time points during resting and the relative hydration level is calculated at multiple time points during exercising.
A Fourier transform of the 850 nm signal will provide the heart rate of the user. At 940 nm (and again at about 1200 nm and 1420 nm), lipid is the dominant absorptive chromophore, so the signal will provide a measure of lipid content of the tissue. Starting at around 1000 nm, with significant peaks at around 1480 nm and 1950 nm, water becomes the dominant absorber, providing a measure of the water content of the tissue. In the exemplary embodiment, light at 1450 nm (NIR) was selected, primarily for economic reasons, for detecting water absorption Using light with each of the four discrete wavelength ranges together will provide an indication of the scattering and melanin contribution.
Over the time and range scales of interest, the intensity of light received can be approximated by a linear combination of the absorption coefficient μa and the scattering coefficient μs.
where A1 and A2 are unknown constants that depend on the sensor geometry, electronics, and other factors that will not change appreciably over the time-frames of interest. The absorption coefficient μa has a complicated dependency on wavelength and tissue composition. Over the wavelengths of interest, the scattering coefficient can be approximated to be a linear function of wavelength.
Eq. 3 provides the total absorption coefficient for a generic tissue:
μa=BSμa,oxy+B(1−S)μa,deoxy+Wμa,waterFμa,fat+Mμa,melanosome+2.3Cbiliεbili+2.3βC εβC (3)
where S is the HGb oxygen saturation of mixed arterio-venous vasculature, B is the average blood volume fraction (fv,blood), W is the water content (fv,water), Bili is the bilirubin concentration (C(M)), βC is the β-carotene concentration (C(M)), F is the fat content (fv,fat), and M is the melanosome volume fraction (fv,melanosome), or alternatively, the concentration of melanin monomers (C(M)).
Each term can potentially depend on time and on the wavelength of light. Over the time-frame of interest, the last four terms will not change appreciably over time but on optical wavelength, so they can be lumped into a parameter D. In addition, we are primarily interested in estimating relative changes to hydration (W) from a baseline (starting) value, so W(ti) is separated into W1+ΔW(t). Writing that out with the explicit dependencies for time wavelength λ and time t,
μa(λ, t)=BS(t)μa,oxy(λ)+B(1−S(t))μa,deoxy(λ)+(W1+ΔW(t))μa,water(λ)+D(λ) (4)
The inventive device will measure the intensity of light at four wavelengths, denoted with λk and thousands of points in time, denoted with ti.
I(λk, ti)=A1(BS(ti)μa,oxy(λk)+B(1−S(ti))μa,deoxy(λ4)+(W1+ΔW(ti))μa,water(λk)+D(λk))+(s1+s2λk) (5)
S, the level of HGb oxygen saturation and will vary with each arterial pulse over a time frame of a second or less. Relative hydration ΔW will change over the time-frame of minutes or hours, and our goal is to track these changes. A variety of signal processing techniques may be applied to estimate ΔW(ti) in the presence of the other unknowns A1, A2, S(ti), W1, D, s1 and s2. One example would be one or more least-squares fits (regression) using all of the I(λk, ti) data recorded by the device over the period of minutes or hours. Temporal band-pass filtering of I(λk, ti) may be employed, if appropriate, to separate spectral components of I.
Relative hydration values may be used to generate the user'"'"'s personal hydration index (“PHI”). In an exemplary implementation, for the initial set of measurements (i.e., the “training data”), the first 3 min of a training session is used to establish an average intensity value for the sensors'"'"' four LEDs. Data is then collected for the duration of the training session. At the end of the training session, the average intensity value for the entire session is taken as the baseline for the next session. Notifications indicating “out of optimal range” can be initially pre-set at a threshold of 2% from baseline based on general agreement within the scientific literature that this deviation corresponds to onset of cognitive defects. In a preferred embodiment, notifications would be generated before a change of +/−2% occurs to allow sufficient time for the user to rehydrate before an out of range condition is reached. During a given training or activity session, i.e., a monitored time period, the inventive device will periodically measure and calculate relative hydration to determine the user'"'"'s relative hydration level at a number of time points, where the monitored period of time may include periods of activity ranging from resting to exercising. The type of exercise with vary with the individual, and may range from walking to intense athletic activity. At any given time point during the monitored time period, the inventive device may generate a notification if the measured relative hydration approaches or exceeds the predetermined threshold to prompt the user to re-hydrate. Notifications may occur at multiple time points during exercising, particularly in extended periods of intense activity.
Ideally, notifications will be generated prior to the point at which the user'"'"'s performance could be impacted by hydration status. In testing of the prototype, a sensitivity of 0.1% was achieved, which is 10× more sensitive than thirst (1-2%). With more data and more advanced analytics, the range may be personalized to allow the user to set his or her own ideal threshold to allow proactive hydration to maintain the user'"'"'s optimal hydration level throughout an activity. In other embodiments, signal processing techniques such as machine learning and data mining may be employed to derive a relationship between I(λk, ti) and ΔW (ti) to generate the user'"'"'s PHI and appropriate notification thresholds.
Referring first to the hydration monitoring system 400 diagram in
The device incorporates both novel hardware and software. As shown in
The battery 418 and BLUETOOTH® 424 will allow the device to operate and communicate wirelessly with the mobile application 432. The mini-USB port 416 will allow charging of the battery as well as wired data transfer. The voltage regulators, microcontroller, analog front-end will power and control the sensor. The battery life must be sufficient to drive the LEDs, data collection, indicator lights, and Bluetooth data transmission for at least 24 hrs.
In the exemplary implementation, light sources 409-412 are LEDs selected to emit light at each of 740 nm, 850 nm, 940 nm, and 1450 nm, for detecting absorption by oxygenated hemoglobin (HbO2), deoxygenated hemoglobin (Hb), lipids, and water, respectively. The specific combination of wavelengths selected are unique in that they collectively provide a comprehensive personalization of the measurements including fat content, skin color, tissue oxygenation, heart rate, and water content. It should be noted that the detection targets exhibit absorption peaks at other wavelengths, and that selection of appropriate combination of LEDS and wavelengths will be within the level of skill in the art. In general, any combination of wavelengths in the range of 600-2000 nm that is sufficient to provide blood flow, oxygen saturation, lipid and water content may be used. In one embodiment, the LED wavelengths can be changed to target the spectra of other or additional chromophores of interest. Different algorithms to normalize and process the data can also be applied to achieve similar/improved results. Various classification algorithms may be used to achieve similar/improved results as well.
The housing 404 will preferably be formed of an opaque (non-light-transmissive) plastic or polymer material, e.g., black or dark colored, to minimize stray light being reflected or transmitted. The housing may be formed of two halves, closely fitted together and sealed to minimize moisture intrusion into the interior of the device. In some embodiments, the housing 404 may be a rigid material, a rigid material with an elastomeric coating (for comfort and/or increased safety), or an elastomeric material with sufficient rigidity to provide protection for the electronic components enclosed therein. It may be desirable to fill the interior of the housing with a potting material or other protectant to ensure that the electronic components are well protected against moisture and impact. Each of LEDs 409-412, indicator light 420, and detector 415 will preferably be located behind transparent windows that are sealed to the interior of the housing to create a waterproof seal. The windows may be partially recessed within the surface, i.e., not flush with the contact surface 405, to serve as a light shield to minimize lateral light leakage into the detector.
The embodiments of
The data processing algorithms described above are used to generate the personal hydration index (PHI), which takes into account light scattering, melanin, lipid content, and blood flow as well as water content. Another parameter that may be measured and incorporated into the algorithms is heart rate. The algorithm may use the heart rate frequency of the water content to measure the arterial water content. When water is consumed, it is first absorbed into the blood stream from the stomach before being distributed to skeletal muscle. The lipid content can be used to normalize the PHI to athletes having a different Body Mass Index (BMI). Testing may be required to determine a maximum BMI at which the device may not be effective in providing an accurate measurement due to a thicker lipid layer. The algorithm to use the heart rate to isolate and ensure the water measurement is coming from arterial blood is novel. The algorithm to normalize the data using individualized scattering coefficients as well as lipid content is novel. The algorithm to combine all the measurements into a personalized hydration index is novel.
For use, the user will preferably place the device on an area of skin that directly overlies a major artery and which has a sufficiently flat and smooth area, e.g., free of folds, bends, wrinkles, or protruding scars, to allow the sensor to lay flat on the skin. Possible locations include the brachial artery on the inside of the bicep, the posterior tibial artery behind the calf, and femoral artery on the inside of the thigh. As previously described with reference to
Motion artifacts can be a major issue for light detection on the skin. Approaches to minimize these motion artifacts include ensuring the adhesive is applied to keep the sensor on the skin without separation. Use of a water absorbing material on the adhesive patch will help absorb sweat accumulated during exercise. For embodiments employing an adhesive patch, the patch should be formed of a breathable material to allow airflow and minimize moisture accumulation between the sensor and the skin. LED pulse frequencies may be set up as semi-random to avoid repeated frequencies that may contribute to motion artifacts, such as the heart rate. By using a moving average of the data points, it is possible to smooth out the data and exclude large artifacts due to exercise movement.
The calibration period for establishing the personal baseline should preferably be at least 12 hrs, during which the high and low limits of the PHI are calculated. The calibration procedure should preferably be performed on a day of relatively low movement by the user to allow for more accurate determination of the optimal hydration range. The LEDs can also be turned on and off at different frequencies to avoid heart rate synchronizing with any movement artifacts (heartbeat).
Another approach to establishing the baseline PHI can include employing a hybrid procedure: a long collection time (12 hrs), or using a pre-determined “standard PHI”, then continuously updating the baseline with each successive workout to gradually personalize it for the user. Using the PHI, relative changes in hydration during periods of exercise can be determined.
In some embodiments, for example, where the hydration monitors are used on team athletes, the data can be collected for monitoring by a coach or trainer, with identifying information intact, to allow the coach or trainer to identify individual players who may need to be called in from the activity to rehydrate. In other embodiments, the data can be stripped of identifying information and stored in a cloud for data mining purposes. Each user may be labeled using one or more tags that classify the user'"'"'s athletic grouping. For example, labels may include, but not be limited to, exercise of choice, resting heart rate, height, weight, gender, ethnicity, exercise duration, exercise intensity, athletic classification, e.g., amateur or professional, or other characteristics that can be used to distinguish among factors that may impact hydration and/or sensitivity to fluctuations in hydration.
Using these labels, various data mining algorithms may be used to classify the athlete as over-hydrated, dehydrated, or optimally hydrated based on their PHI. Possible data mining algorithms include learning machines (neural networks, support vector machines, Bayesian networks, genetic algorithms, etc.), statistics, clustering, regression, etc. The computational processing may either be done locally on the mobile application, or on a cloud computing system, in which measurements may be compared against populations of persons falling within the same or similar classifications based on data accumulated from other users and stored in a central database. Data analytics may include feature selection and machine learning algorithms to identify attributes that are most determinative of hydration. Once the classification(s) is/are made, one or more indicator LEDs may be activated to display, for example, red for measurements out of optimal range, green for inside optimal range, and yellow or orange for insufficient data/error or for nearing out of optimal range. The user can access more detailed information through the mobile application.
The more data that is collected, the more accurate the PHI will be, and classification algorithms will become more robust to indicate dehydration, over hydration, or optimal hydration. Measurements can be improved by applying more than one sensor on the body. By measuring in different areas of the body, it is possible to generate an overall total body average of hydration, rather than a point measurement. This technique may be useful in discarding outlier data due to artifacts at one location.
The use of more than one sensor will necessitate more than one adhesive patch (or strap) and an modified data processing algorithm to combine the measurements and tease out the artifacts. A combined total body personal hydration index (PHI) can be calculated from the multiple measurement sites.
In some situations, the user may not want to make continuous, real-time measurements during activity. In these situations, an adhesive or strap need not be applied. Instead, the user can simply place the sensor on dry skin for measurement periods, such as before and after a workout, to provide two measurement points. Assuming the calibration of the optimal hydration range has already been achieved, this simple measurement with provide data indicating how much water loss occurred during a given exercise session. A similar approach may be used if the user wishes to monitor his or her hydration during time outs or other interruptions in activity. To provide a practical illustration, it is not uncommon for athletes to suffer severe muscle cramping during football games or tennis matches. A quick measurement during a break in play could allow the athlete to quickly determine whether they are approaching dehydration and take preventive measures, before actual cramping occurs.
The software includes a self-calibration algorithm, using the high and low hydration range in at least one calibration day (24 hrs) before the device is ready to accurately determine a personalized optimal range, taking into account, tissue scattering, lipid content as well as water content. Once the optimal range is determined, the device will be able to determine percentage PHI changes, indicating over-hydration, dehydration, or optimal hydration during exercise. The more measurements the user logs, the more accurate the optimal range becomes and therefore the more accurate PHI predictions.
The user may decide to use only this or her own measurements, in which the processing can be processed in the mobile application, or he or she may choose to access data from other users on the cloud using cloud computing techniques as are known in the art. Ideally, once enough data is loaded onto the cloud, the athlete will be able to compare and contrast his/her athletic level and hydration needs against other athletes of similar categories.
Novel aspects of the inventive device and method include the specific preprocessing of the data to feed into classification algorithms as well as the option to either run the classification locally on the mobile application or using cloud computing.
Although the disclosure has been shown and described with respect to one or more implementations, equivalent alterations and modifications will occur to others skilled in the art based upon a reading and understanding of this specification and the annexed drawings. The disclosure includes all such modifications and alterations and is limited only by the scope of the following claims. In particular regard to the various functions performed by the above described components (e.g., elements, resources, etc.), the terms used to describe such components are intended to correspond, unless otherwise indicated, to any component which performs the specified function of the described component (e.g., that is functionally equivalent), even though not structurally equivalent to the disclosed structure which performs the function in the herein illustrated exemplary implementations of the disclosure.