WO2012093397A2 - Device and method for continuous energy expenditure measurement - Google Patents

Device and method for continuous energy expenditure measurement Download PDF

Info

Publication number
WO2012093397A2
WO2012093397A2 PCT/IL2012/000011 IL2012000011W WO2012093397A2 WO 2012093397 A2 WO2012093397 A2 WO 2012093397A2 IL 2012000011 W IL2012000011 W IL 2012000011W WO 2012093397 A2 WO2012093397 A2 WO 2012093397A2
Authority
WO
WIPO (PCT)
Prior art keywords
std
nom
nol
now
nos
Prior art date
Application number
PCT/IL2012/000011
Other languages
French (fr)
Other versions
WO2012093397A3 (en
Inventor
Miki Raviv
Nir Dotan
Original Assignee
Miki Raviv
Nir Dotan
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Miki Raviv, Nir Dotan filed Critical Miki Raviv
Publication of WO2012093397A2 publication Critical patent/WO2012093397A2/en
Publication of WO2012093397A3 publication Critical patent/WO2012093397A3/en

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/48—Other medical applications
    • A61B5/4866—Evaluating metabolism
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
    • A61B5/026—Measuring blood flow
    • A61B5/0295—Measuring blood flow using plethysmography, i.e. measuring the variations in the volume of a body part as modified by the circulation of blood therethrough, e.g. impedance plethysmography
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
    • A61B5/1118—Determining activity level
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
    • A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
    • A61B5/6802—Sensor mounted on worn items
    • A61B5/681—Wristwatch-type devices
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B2562/00—Details of sensors; Constructional details of sensor housings or probes; Accessories for sensors
    • A61B2562/02—Details of sensors specially adapted for in-vivo measurements
    • A61B2562/0219—Inertial sensors, e.g. accelerometers, gyroscopes, tilt switches

Definitions

  • continuous analysis of photoplethysmograph signal patterns measured at the digital artery area in the hand finger, or at the earlobe, together with signal data from a three-dimensional accelerometer, attached to the hand finger, can be used to identify wide range of activity patterns, including activity patterns where the hands are inactive.
  • a processor adapted to apply at least one set of rules to at least one of; i. the at least one signal;
  • the at least one device includes at least one photoplethysmographic (PPG) sensor adapted to provide PPG data of the subject over the period of time.
  • PPG photoplethysmographic
  • the system further includes a database of predefined metabolic equivalent of a task (MET)s.
  • MET metabolic equivalent of a task
  • the at least one of direct and indirect data is generated from a plurality of mammalian subjects.
  • the method further includes receiving both photoplethysmographic (PPG) and tri-axial acceleration data of the subject.
  • PPG photoplethysmographic
  • the at least one bodily tissue includes at least one of finger tissue and ear tissue.
  • the method further includes detecting at least one signal from at least one bodily tissue at a second location of the mammalian subject over a period of time.
  • the EEE is based on the MET value relevant to at least one activity type and an RMR of the subject. Moreover, according to an embodiment of the present invention, the method further includes identifying a type of movement of legs of the subject.
  • the method further includes detecting an increase in at least one of an amplitude and a frequency of a PPG signal thereby detecting the movement of the legs.
  • At least one device including at least one sensor adapted to detect at least one signal from at least one bodily tissue at a first location of the mammalian subject over a period of time;
  • a computer program product for energy expenditure determination of a mammalian subject including a computer-readable medium having program instructions embodied therein, which instructions, when read by a computer, cause the computer to;
  • the present invention provides systems and methods for energy expenditure determination of a mammalian subject, the system including at least one device comprising at least one sensor adapted to detect at least one signal from at least one bodily tissue at a first location and/or a second location of the mammalian subject over a period of time; and a processor adapted to apply at least one set of rules to at least one of the at least one signal and to a reduced data set associated with the at least one signal, thereby being configured to provide predictive values of energy expenditure of the mammalian subject over the period of time.
  • Fig. IB a simplified schematic illustration showing a communication device in the system for continuous energy expenditure measurement, in accordance with an embodiment of the present invention
  • FIG. 1C a simplified schematic illustration showing a system for continuous energy expenditure measurement comprising a combination of a ring device and an earring sensor, in accordance with an embodiment of the present invention
  • Fig. 2A is a simplified schematic illustration of a ring sensor on a finger, in accordance with an embodiment of the present invention
  • Fig. 3 is a schematic diagram of a method for development of an "if-then” rules set and its application in real-time for estimation of energy expenditure (EEE) calculation, in accordance with an embodiment of the present invention
  • Fig. 4A shows a scatter chart of a distribution of metabolic equivalent of task (MET), assigned by the method for each of the 15 seconds time segments. Time segments are clustered according to the actual physical activity (PA) types performed at this time segment S- Sitting, STA- Standing, W- Walking, R- Running , M- mixed activity . The box represents 50% of the results, and the line inside the box is the median, filled squares represent outliers, in accordance with an embodiment of the present invention;
  • PA physical activity
  • Fig. 5 is a graph of accumulated EEE as measured during the subjects activity by a tri-axial accelerometer (EEE PA cumulative) and a cumulative calculated EEE, based on heart pulse (EEE POLAR cumulative), in accordance with an embodiment of the present invention
  • Fig. 8 shows a scatter plot and regression line for prediction of pulse- based on an inter quartiles range, in accordance with an embodiment of the present invention
  • System 100 comprises a device, such as ring device 102 (or ring device 220, Fig. 2). Additionally or alternatively, system 100 comprises an ear sensor 240 (Fig. 2B). Ring device 102 comprises a PPG sensor 115.
  • the device and/or sensor comprise some or all of the following elements:- a wearable device 102, such as a ring, worn on a finger, or earring on an ear which both emits at least one signal to a bodily part and receives at least one signal from said bodily part; at least one diode 106, a switch 108, a power source, such as battery 110, a wireless transceiver 112, configured to receive signals from said battery and from a clock 116, as well as from the power source.
  • the transceiver is also adapted to transmit and/or receive data from a microprocessor 120.
  • the ring/earring sensor comprises an electrical power source 110, a three- dimensional accelerometer 118, at least one light source 101 and at least one light detection system 103 for measurement of a photoplethysmographic (PPG) signal.
  • PPG photoplethysmographic
  • Fig. IB a simplified schematic illustration shows a communication device 130 in system 100 for continuous energy expenditure measurement, in accordance with an embodiment of the present invention.
  • a human arm 210 typically comprises a wrist 202, a thumb 212 and four fingers 204, 206, 208 and 210.
  • the ring sensor 220 is subject to movement along three orthogonal axes: an X axis, 222, a Y axis 224 and a Z axis 226.
  • the earlobe PPG sensor is an earring worn on the earlobe containing: an electrical power source, light source and light detection system for measurement of photoplethysmographic signal, electronic circuits, microprocessor, transmitter and receiver. Both the earring PPG sensor and the ring acceleration sensor, are time synchronized between them. And the data generated by PPG sensor is transmitted to the acceleration sensor. And then the full activity profile can be identified by comparison to known activity patterns.
  • Another aspect of this invention is a similar device (not shown) and method that does not contain a Photoplethysmographic sensor, and is attached to the collar of a dog or a cat and is used for the evaluation of the energy expenditure of the animal.
  • Activity Profile Classification Algorithm is a similar device (not shown) and method that does not contain a Photoplethysmographic sensor, and is attached to the collar of a dog or a cat and is used for the evaluation of the energy expenditure of the animal.
  • the over all estimated energy expenditure during the day/week/month is calculated as a sum of the energy expenditure during all the time segments.
  • the output data acquired by the 4 sensors (acceleration in ⁇ , ⁇ , ⁇ axis and PPG sensor) during a certain time segment is analyzed at the microprocessor using a software that performs the classifying algorithm, the algorithm is based on calculation of parameters that describe the data recorded by each of the 4 sensors during a certain time segment.
  • the calculated parameters can be for example : average; standard deviation, coefficient of variance, median, inter-quartile range, integral over the time, minimum value , maximum value, number of times that the signal is crossing the median during a specific time segment.
  • -A set of predetermined if-and-only-if rule employed on the calculated parameters enable to classify the activity profile of the subject (e.g.
  • the method has a number of advantages compared with existing energy expenditure estimation methods:
  • the device is cheap and can be miniature in size
  • PA type is S is 0.255 (223 out of 873 cases)
  • PA type is not S
  • PA type is W
  • PA type is not M is 0.829 (724 out of 873 cases)
  • PA type is M is 0.171 (149 out of 873 cases)
  • P Partial agreement, identification of actual PA type L or S or STD as L, or S, or STD. Identification of actual PA type W or R or M as W, or R, or M.
  • Miss classification - the PA was identified as significantly different then actual e.g. L, S, STD, instead of W, R, M.
  • Example 2 Estimation of energy expenditure in free-living human using tri axial accelerometer worn as a ring finger.
  • the tri axial acceleration data was transmitted to a laptop computer that was carried by the subject.
  • MatLab software for each 15 seconds segment, the variables described in table 2 were calculated and the PA type was identified based on the rules described in example 1. if walking or running activities were identified the walking or running speed was calculated based on equation 3 and 4 accordingly described in example 1.
  • the RMR of the subject was calculated according to equation 1 and found to be 0.574 kcal/15 seconds
  • Figure 4 describes the distribution of METs assigned by the method for the actual different physical activity (PA) types. There is a clear association between the actual PA types and assigned METs values.
  • the EEE for each time segment was calculated by multiplication of the RMR with the MET value of PA that was identified by the classification rules and summed up.
  • Figure 5 describes the accumulated EEE as measured during the subject's activity by tri-axial accelerometer and the EEE based on heart pulse, demonstrating the similarity between the methods.
  • Example 3 Detection of legs movement intensity during cycling via ring finger PPG sensor.
  • Session A- The first session included a gradual increase of the intensity (RPM) over time in order to calibrate the system.
  • RPM intensity
  • Figure 6 describes RPM and subjects pulse increase over time.
  • Correlation coefficients ( R - Pearson) between the RPM, Pulse, Pulse + RPM and calculated parameters are described in tables 6 below: Table 6.
  • FIG. 4B shows a scatter chart of a distribution of metabolic equivalent of task (MET), assigned by the method for each of the 15 seconds time segments as in Fig. 4A, but excluding the time segments that were unidentified, in accordance with an embodiment of the present invention.
  • Fig. 5 there is shown a graph of accumulated EEE as measured during the subjects activity by a tri-axial accelerometer (EEE PA cumulative) and a cumulative calculated EEE, based on heart pulse (EEE POLAR cumulative), in accordance with an embodiment of the present invention.
  • EEE PA cumulative tri-axial accelerometer
  • EEE POLAR cumulative a cumulative calculated EEE, based on heart pulse
  • Fig. 6 shows a graph of a crank number of rotations per minute (RPM) corresponding increase in a subject's pulse over time, Pulse + RPM, as well as an inter-quartile range of a PPG signal over time, in accordance with an embodiment of the present invention.
  • RPM crank number of rotations per minute
  • Fig. 7 shows a scatter plot and regression line for prediction of RPM based on an inter quartiles range, in accordance with an embodiment of the present invention.
  • Fig. 9A is a graph of a crank number of rotations per minute (RPM) corresponding increase in a subject's pulse over time, Pulse + RPM, as well as an inter-quartile range of a PPG signal over time, in accordance with an embodiment of the present invention
  • Fig. 9B shows a scatter plot for prediction of inter-quartile range based on an inter quartiles range crank number of rotations per minute (RPM), in accordance with an embodiment of the present invention.
  • Fig. 10 shows a scatter plot of a correlation between inter-quartile range IQR and pulse, in accordance with an embodiment of the present invention.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Engineering & Computer Science (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Veterinary Medicine (AREA)
  • Physics & Mathematics (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Physiology (AREA)
  • Obesity (AREA)
  • Hematology (AREA)
  • Cardiology (AREA)
  • Dentistry (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

The present invention provides systems and methods for energy expenditure determination of a mammalian subject, the system including at least one device comprising at least one sensor adapted to detect at least one signal from at least one bodily tissue at a first location, and optionally at a second location, of the mammahan subject over a period of time; and a processor adapted to apply at least one set of rules to at least one of the at least one signal and to a reduced data set associated with the at least one signal, thereby being configured to provide predictive values of energy expenditure of the mammalian subject over the period of time.

Description

DEVICE AND METHOD FOR CONTINUOUS ENERGY EXPENDITURE
MEASUREMENT
FIELD OF THE INVENTION
The present invention relates generally to continuous monitoring of metabolic activity, and more specifically to methods and apparatus for continuous monitoring of energy expenditure.
BACKGROUND OF THE INVENTION
Metabolic syndrome is a risk factor for cardiovascular disease that is receiving much clinical attention [Grundy et al., 2004]. The dominant underlying developmental features of metabolic syndrome are abdominal obesity and insulin resistance. More than 50 million Americans are estimated to have metabolic syndrome. First-line treatment involves healthy lifestyle changes. The most important action is to balance dietary intake and calories burned. However, it can be difficult for people to monitor their level of physical activity over prolonged time periods, such as days, weeks or months.
The association between physical activity and positive health benefits has been well established [Blair et al, 1984 and Blair et al., 1995]. This has led to the Centers for Disease Control and Prevention and the American College of Sports Medicine recommendation that every US adult should perform 30 min of moderate-intensity physical activity on most, preferably all, days of the week [Pate et. al, 1995]. Although the benefits of regular, moderate-intensity physical activity have been shown, quantifying physical activity has proven to be a difficult task.
To date, there are few, if no, user-friendly, reliable and accurate ways to routinely assess metabolic physical activity and energy expenditure outside the laboratory in a free-living environment. This has significant ramifications for the subjects' weight management success. From the behavior-change literature [Dilley (1998); Klem (2000); and Schnool et al., (2001), it is well recognized that regular and accurate self-monitoring in the free-living environment can provide important feedback which increases self-awareness - the prerequisite for healthy decisionmaking and long-term lifestyle change.
True Energy Expenditure (TEE) of an individual during a certain period is very difficult to measure, and nearly all techniques make use of approximations of one kind or another. The gold standard today for TEE measurements is the use of metabolic carts, which indirectly analyze calorimetry. These metabolic carts measure the oxygen and carbon dioxide that a person inhales and exhales and from this, indirectly compute the calories burned during the period of measurement. Devices of this category differ from one another by 5-10% and differ even on repeated measurements of the same activity by around 5-10%. Most metabolic carts are rather large and bulky and are not suited for monitoring outside the laboratory setting.
On the other side of the spectrum, are devices such as pedometers, accelerometers, heart rate monitors, and arm bend multi-sensors device. These devices, when used for measuring energy expenditure, have one main limitation: none of them can be worn and function continuously for long time periods of up to months, without the need to attach and detach the sensor to the body via a chase strap (heart rate monitors), an arm band, or to a piece of clothing, such as to trousers or to a belt.
Some patent publications in the field include US5,964,701, US6,402,690, US6,699,199; US7,334,472 and US20070073178 Al. As can be seen from the prior art, there remains a need for a practical activity-monitoring platform based on wearable sensor, which is adapted to be worn for long time periods, without interfering with the daily activities, yet accurately measuring the energy expenditure of a subject. This kind of device would be very useful in order to balance dietary intake with calorific output.
SUMMARY OF THE INVENTION
It is an object of some aspects of the present invention to provide improved devices and methods for continuous monitoring of energy expenditure in a mammal.
In some embodiments of the present invention, improved methods and apparatus are provided for continuous monitoring of energy expenditure of humans.
The present invention is based on the discovery that the movements patterns of a human subject hand, as measured by a 3 dimensional accelerometer attached to its finger in a ring configuration and/or a bracelet configuration a wrist, during a specific time segment, can used to identify subject's activity profile (activity type and intensity).
The present invention further discloses that monitoring changes in photoplethysmograph signal patterns, measured proximal to a digital artery of a finger and/or in an earlobe, accurately reflects movement of other body parts (e.g. legs), and associated energy expenditure, even when the hand and finger are static.
Furthermore, the present invention discloses that energy expenditure of common human activities can be predicted from photoplethysmograph signal patterns, measured proximal to a digital artery of a finger and/or in an earlobe.
Thus, continuous analysis of photoplethysmograph signal patterns measured at the digital artery area in the hand finger, or at the earlobe, together with signal data from a three-dimensional accelerometer, attached to the hand finger, can be used to identify wide range of activity patterns, including activity patterns where the hands are inactive.
By using known general parameters regarding energy expenditure during a specific activity profile, together with individual parameters (such as age, weight, height or gender) that are used to estimate the resting metabolic rate of a subject, it is possible to calculate the estimated energy expenditure of a subject during the specific time segment. The over all estimated energy expenditure during the day is calculated as a sum of the energy expenditure during all the time segments.
There is thus provides according to an embodiment of the present invention, a system for energy expenditure determination of a mammalian subject, the system including; a. at least one device including at least one sensor adapted to detect at least one signal from at least one bodily tissue at a first location of the mammalian subject over a period of time; and
b. a processor adapted to apply at least one set of rules to at least one of; i. the at least one signal; and
ii. a reduced data set associated with the at least one signal;
c. thereby being configured to provide predictive values of energy expenditure of the mammalian subject over the period of time.
Additionally, according to an embodiment of the present invention, the at least one device is a ring and the at least one bodily tissue is within a finger.
Furthermore, according to an embodiment of the present invention, the at least one device is an earring device and the at least one bodily tissue is within an ear.
Moreover, according to an embodiment of the present invention, the at least one device includes at least one photoplethysmographic (PPG) sensor adapted to provide PPG data of the subject over the period of time.
Further, according to an embodiment of the present invention, the at least one device includes at least one 3D accelerometer adapted to provide tri-axial acceleration data of the subject over the period of time.
Yet further, according to an embodiment of the present invention, the system further includes a remote communication device including a transceiver for bidirectional communication with the at least one device.
Additionally, according to an embodiment of the present invention, the at least one device includes two devices.
Further, according to an embodiment of the present invention, the two devices include a ring device and an ear device.
Yet further, according to an embodiment of the present invention, the two devices each include a transceiver for communication with the remote communication device.
Additionally, according to an embodiment of the present invention, the remote communication device is a cellular phone.
Moreover, according to an embodiment of the present invention, the remote communication device is a computer.
According to another embodiment of the present invention, the computer is a portable computer.
According to a further embodiment of the present invention, the system further includes a database of predefined metabolic equivalent of a task (MET)s.
Additionally, according to an embodiment of the present invention, the system is further adapted to develop the at least one set of rules from at least one of direct and indirect data from at least one of a PPG sensor and a 3D accelerometer.
Furthermore, according to an embodiment of the present invention, the system further includes data mining software for manipulating the at least one of direct and indirect data.
Moreover, according to an embodiment of the present invention, the at least one of direct and indirect data is generated from a plurality of mammalian subjects.
Notably, according to an embodiment of the present invention, the first location is substantially static relative to a major limb of the mammalian subject.
There is thus provided according to an additional embodiment of the present invention, a method for energy expenditure determination of a mammalian subject, the method including;
a) detecting at least one signal from at least one bodily tissue at a first location of the mammalian subject over a period of time; and b) applying at least one set of rules to at least one of;
i. the at least one signal; and
ii. a reduced data set associated with the at least one signal;
thereby providing predictive values of energy expenditure of the mammalian subject over the period of time.
Moreover, according to an embodiment of the present invention, n the at least one bodily tissue is within a finger.
Additionally, according to an embodiment of the present invention, the at least one bodily tissue is within an ear.
Furthermore, according to an embodiment of the present invention, the detecting step includes receiving photoplethysmographic (PPG) data of the subject over the period of time.
Moreover, according to an embodiment of the present invention, the detecting step includes receiving tri-axial acceleration data of the subject over the period of time.
Further, according to an embodiment of the present invention, the method further includes transmitting data from at least one device to a remote communication device.
Yet further, according to an embodiment of the present invention, the method further includes receiving both photoplethysmographic (PPG) and tri-axial acceleration data of the subject.
Moreover, according to an embodiment of the present invention, the at least one bodily tissue includes at least one of finger tissue and ear tissue.
Additionally, according to an embodiment of the present invention, the energy expenditure is an estimation of energy expenditure (EEE).
Furthermore, according to an embodiment of the present invention, the method further includes applying data from a database of predefined metabolic equivalent of a task (MET)s.
Additionally, according to an embodiment of the present invention, the method further includes generating the at least one set of rules from at least one of direct and indirect data from at least one of a PPG sensor and a 3D accelerometer.
Moreover, according to an embodiment of the present invention, the method further includes applying data mining software to manipulate the at least one of direct and indirect data.
Additionally, according to an embodiment of the present invention, the at least one of direct and indirect data is generated from a plurality of mammalian subjects.
Moreover, according to an embodiment of the present invention, the first location is substantially static relative to a major limb of the mammalian subject.
Furthermore, according to an embodiment of the present invention, the method further includes detecting at least one signal from at least one bodily tissue at a second location of the mammalian subject over a period of time.
Additionally, according to an embodiment of the present invention, the EEE is based on the MET value relevant to at least one activity type and an RMR of the subject. Moreover, according to an embodiment of the present invention, the method further includes identifying a type of movement of legs of the subject.
Furthermore, according to an embodiment of the present invention, the method further includes detecting an increase in at least one of an amplitude and a frequency of a PPG signal thereby detecting the movement of the legs.
There is thus provided according to an additional embodiment of the present invention, a system for identification of an activity type in a mammalian subject, the system including;
a) at least one device including at least one sensor adapted to detect at least one signal from at least one bodily tissue at a first location of the mammalian subject over a period of time; and
b) a processor adapted to apply at least one set of rules to at least one of;
i. the at least one signal; and
ii. a reduced data set associated with the at least one signal;
thereby being configured to identify different activity types responsive to at least one output of the processor.
There is thus provided according to an additional embodiment of the present invention, a computer program product for energy expenditure determination of a mammalian subject, the product including a computer-readable medium having program instructions embodied therein, which instructions, when read by a computer, cause the computer to;
a) detect at least one signal from at least one bodily tissue at a first location of the mammalian subject over a period of time; and b) apply at least one set of rules to at least one of;
i. the at least one signal; and
ii. a reduced data set associated with the at least one signal;
thereby providing predictive values of energy expenditure of the mammalian subject over the period of time. Furthermore, according to an embodiment of the present invention, the product further uses the at least one set of rules to classify an activity profile of the subject over a plurality of time segments in the period of time.
Additionally, according to an embodiment of the present invention, the product further combines data from said activity profile with a relevant MET value associated with the activity profile and further calculates a resting metabolic rate (RMR) of the subject, thereby estimating the energy expenditure of the subject over the period of time.
Furthermore, according to an embodiment of the present invention, the product applies an equation:
RMR (kcal/time) x MET x time = EEE (kcal)
to calculate the EEE.
The present invention provides systems and methods for energy expenditure determination of a mammalian subject, the system including at least one device comprising at least one sensor adapted to detect at least one signal from at least one bodily tissue at a first location and/or a second location of the mammalian subject over a period of time; and a processor adapted to apply at least one set of rules to at least one of the at least one signal and to a reduced data set associated with the at least one signal, thereby being configured to provide predictive values of energy expenditure of the mammalian subject over the period of time.
The present invention will be more fully understood from the following detailed description of the preferred embodiments thereof, taken together with the drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention will now be described in connection with certain preferred embodiments with reference to the following illustrative figures so that it may be more fully understood.
With specific reference now to the figures in detail, it is stressed that the particulars shown are by way of example and for purposes of illustrative discussion of the preferred embodiments of the present invention only and are presented in the cause of providing what is beheved to be the most useful and readily understood description of the principles and conceptual aspects of the invention. In this regard, no attempt is made to show structural details of the invention in more detail than is necessary for a fundamental understanding of the invention, the description taken with the drawings making apparent to those skilled in the art how the several forms of the invention may be embodied in practice.
In the drawings:
Fig. 1A a simplified schematic illustration showing a system for continuous energy expenditure measurement comprising a ring device, in accordance with an embodiment of the present invention;
Fig. IB a simplified schematic illustration showing a communication device in the system for continuous energy expenditure measurement, in accordance with an embodiment of the present invention;
Fig. 1C a simplified schematic illustration showing a system for continuous energy expenditure measurement comprising a combination of a ring device and an earring sensor, in accordance with an embodiment of the present invention;
Fig. 2A is a simplified schematic illustration of a ring sensor on a finger, in accordance with an embodiment of the present invention;
Fig. 2B is a simplified schematic illustration of an earlobe sensor on an earlobe, in accordance with an embodiment of the present invention;
Fig. 3 is a schematic diagram of a method for development of an "if-then" rules set and its application in real-time for estimation of energy expenditure (EEE) calculation, in accordance with an embodiment of the present invention; Fig. 4A shows a scatter chart of a distribution of metabolic equivalent of task (MET), assigned by the method for each of the 15 seconds time segments. Time segments are clustered according to the actual physical activity (PA) types performed at this time segment S- Sitting, STA- Standing, W- Walking, R- Running , M- mixed activity . The box represents 50% of the results, and the line inside the box is the median, filled squares represent outliers, in accordance with an embodiment of the present invention;
Fig. 4B shows a scatter chart of a distribution of metabolic equivalent of task (MET), assigned by the method for each of the 15 seconds time segments as in Fig. 4A, but excluding the time segments that were unidentified, in accordance with an embodiment of the present invention;
Fig. 5 is a graph of accumulated EEE as measured during the subjects activity by a tri-axial accelerometer (EEE PA cumulative) and a cumulative calculated EEE, based on heart pulse (EEE POLAR cumulative), in accordance with an embodiment of the present invention;
Fig. 6 is a graph of a crank number of rotations per minute (RPM) corresponding increase in a subject's pulse over time, Pulse + RPM, as well as an inter-quartile range of a PPG signal over time, in accordance with an embodiment of the present invention;
Fig. 7 shows a scatter plot and regression line for prediction of RPM based on an inter quartiles range, in accordance with an embodiment of the present invention;
Fig. 8 shows a scatter plot and regression line for prediction of pulse- based on an inter quartiles range, in accordance with an embodiment of the present invention;
Fig. 9A is a graph of a crank number of rotations per minute (RPM) corresponding increase in a subject's pulse over time, Pulse + RPM, as well as an inter-quartile range of a PPG signal over time, in accordance with an embodiment of the present invention;
Fig. 9B shows a scatter plot for prediction of inter-quartile range based on an inter quartiles range crank number of rotations per minute (RPM), in accordance with an embodiment of the present invention; and Fig. 10 shows a scatter plot of a correlation between IQR and pulse, in accordance with an embodiment of the present invention.
In all the figures similar reference numerals identify similar parts.
DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS
In the detailed description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be understood by those skilled in the art that these are specific embodiments and that the present invention may be practiced also in different ways that embody the characterizing features of the invention as described and claimed herein.
Accelerometers are used to characterize the intensity and duration of physical activity. An accelerometer is a device that measures proper acceleration, the acceleration experienced relative to freefall. Single- and multi-axis models are available to detect magnitude and direction of the acceleration as a vector quantity, and can be used to sense orientation, vibration and shock.
In the past 15 years, micromachined portable accelerometers have been used in the research field to characterize the intensity and duration of physical activity (PA), and their output has been used for the estimation of energy expenditure (EEE) [Chen et al., 1997, Donahoo et al., 2004, Heil, 2006; and Rothney et al, 2007].
There is some discussion with relation to the optimal placement of an accelerometer sensor for EEE on a mammalian body. Accelerometer devices are typically worn on the hip with the aim of capturing displacement of the subject's center of mass, which is generally associated with moderate- to high-intensity activities, and accounts for the largest EE differences from a baseline.
However wearing the sensor on the hip, or using an armband to attach the sensor to the arm, is not practical for continues measurement of physical activity (PA) over long periods of time. It cannot be worn during sleep hours, during self caring activities, such as during a shower or bath. Furthermore, the need to attach and detach the device to/from the body every few hours reduces user adherence.
The present invention provides systems and methods for continuous monitoring of energy expenditure, using data measured in human subjects.
Reference is now made to Fig. 1A, which is a simplified schematic illustration showing a system 100 for continuous energy expenditure measurement comprising a ring device 102, in accordance with an embodiment of the present invention.
System 100 is constructed and configured to provide at least one of continuous energy expenditure measurement and data predicting continuous energy expenditure.
System 100 comprises a device, such as ring device 102 (or ring device 220, Fig. 2). Additionally or alternatively, system 100 comprises an ear sensor 240 (Fig. 2B). Ring device 102 comprises a PPG sensor 115. The device and/or sensor comprise some or all of the following elements:- a wearable device 102, such as a ring, worn on a finger, or earring on an ear which both emits at least one signal to a bodily part and receives at least one signal from said bodily part; at least one diode 106, a switch 108, a power source, such as battery 110, a wireless transceiver 112, configured to receive signals from said battery and from a clock 116, as well as from the power source. The transceiver is also adapted to transmit and/or receive data from a microprocessor 120.
The method of the invention comprises placing of a ring on a finger 206 or an earring sensor 240 on an earlobe of a subject respectively.
The ring/earring sensor comprises an electrical power source 110, a three- dimensional accelerometer 118, at least one light source 101 and at least one light detection system 103 for measurement of a photoplethysmographic (PPG) signal.
One or more filters 122 may be used to filter the signals received by the at least one light detection system 103. The sensor may comprise further electronic circuitry 107, as may be required (not shown). The transceiver may, according to some embodiments, be replaced by one or more separate transmitters and receivers (not shown). The transceiver may further receive and/or send signals via an antenna 114 to one or more remote locations.
Turning now to Fig. IB, a simplified schematic illustration shows a communication device 130 in system 100 for continuous energy expenditure measurement, in accordance with an embodiment of the present invention.
Communication device, such as a computer or cellular telephone, typically comprises an antenna 138 for receiving signals from the device 102 and/or sensor 240, a wireless transceiver 136, adapted to transmit signals from the antenna and to a microprocessor 134 and vice versa, and a display 132 in communication with the microprocessor 134, as is known in the art.
Reference is now made to Fig. 1C, which a simplified schematic illustration showing a system 140 for continuous energy expenditure measurement comprising a combination of a ring motion sensor 150 and an earring device 100 (similar or identical to ring device 100 of Fig. 1A, in accordance with an embodiment of the present invention. Both the ring motion sensor 150 and earring device 100 can communicate with each other. Additionally or alternatively ring motion sensor 150 and earring device 100 are constructed and configured to communicate with a host unit 170.
A ring motion sensor 150 typically comprises some or all of: an electrical power source, such as a battery 154, a tliree-dimensional accelerometer 156, at least one light source (not shown) and at least one light detection system (not shown) for measurement of a photoplethysmographic (PPG) signal. Ring motion sensor 150 further comprises a transceiver 162 may, according to some embodiments, be replaced by one or more separate transmitters and receivers (not shown). The transceiver may further receive and/or send signals via an antenna 152 to one or more remote locations or devices. Ring motion sensor 150 further comprises a microprocessor 158 and a clock 160. The functions of each of the elements herein are well understood to persons skilled in the art.
Ring motion sensor 150 comprises a three-dimensional accelerometer 156 may be replaced by a wrist sensor 1 9 (not shown) worn on a wrist 202 (Fig. 2) and comprises a three-dimensional accelerometer 156.
Host unit 170 is constructed and configured to communicate with at least one of the ring motion sensor and the earring device. The host unit typically comprises some or all of: an electrical power source, such as a battery 178. Host unit 170 further comprises a transceiver 184 may, according to some embodiments, be replaced by one or more separate transmitters and receivers (not shown). The transceiver may further receive and/or send signals via an antenna 172 to one or more remote locations or devices (100, 140). Host unit 170 further comprises a microprocessor 182 and a clock 174. Moreover, most host units are equipped with additional connectivity/storage devices, such as, but not limited to, a USB interface 176. The functions of each of the elements herein are well understood to persons skilled in the art.
Reference is now made to Fig. 2A, which is a simplified schematic illustration of a ring sensor 220 on a finger 206, in accordance with an embodiment of the present invention. According to some alternative embodiments the ring sensor, may be worn on another appropriate bodily part, such as, but not limited to, a wrist (202 not shown) or a thumb 212.
A human arm 210, typically comprises a wrist 202, a thumb 212 and four fingers 204, 206, 208 and 210. The ring sensor 220 is subject to movement along three orthogonal axes: an X axis, 222, a Y axis 224 and a Z axis 226.
During the activity of the subject throughout the day, the acceleration data generated by movement of the subject's hand, and photoplethysmographic signal data are measured continuously. The acceleration data together with the photoplethysmographic signal data are analyzed by the microprocessor in order to classify the subjects activity profile (e.g. lying , sitting, standing, walking, running, swimming, cycling) and its intensity, including activity patterns where the hands are not active. The classification is performed by comparing the pattern of the signals to previously measured signal pattern of activities that were recorded before. If the comparison yields identification of similar patterns, the activity type may be recognized if the correlation coefficient associated therewith is considered to be in a significant range. The pattern comparison is performed by calculation a set of descriptor parameters to each time segment, (e.g. average, standard deviation, median, inter quartile range (the interquartile range (IQR) is the distance between the 75th percentile and the 25th percentile) and by applying a set of classification "if then" rules, that were developed using data mining techniques, on the calculated descriptors the activity pattern and intensity can be identified. This is shown in Fig. 3 hereinbelow.
This classification data is used to calculate an estimation of the energy expenditure of the subject during a specific time segment.
Quantitative information obtained from the subject can be saved or/and transmitted to the outer recording device using wireless transmission technologies.
Reference is now made to Fig. 2B, which is a simplified schematic illustration of an earlobe sensor 240 on an earlobe 246, in accordance with an embodiment of the present invention. The earring sensor typically comprises an emission side 250 comprising at least one light source 248, such as an LED, as well as a detecting side 242 comprising at least one photodiode 244. Further details of the earring sensor are shown in Fig. 1C.
Hand acceleration is measured by an accelerometer in a ring or bracelet worn by the user, the ring/bracelet containing: an electrical power source, 3 dimensional accelerometer, electronic circuits, microprocessor, transmitter and receiver. Both the earring PPG sensor and the ring acceleration sensor, are synchronized between them. The data generated by PPG sensor is transmitted to the acceleration sensor.
Additionally or alternatively, the photoplethysmographic signal can be measured by an earring worn on the earlobe containing: an electrical power source, light source and light detection system for measurement of photoplethysmographic signal, electronic circuits, microprocessor as described hereinabove.
Turning to Fig. 3, there is seen a schematic diagram of a method 300 for development of an "if-then" rules set and its application in real-time for estimation 301 of energy expenditure (EEE) calculation, in accordance with an embodiment of the present invention.
Method for real-time for estimation 301 of energy expenditure (EEE) calculation 301 comprises a number of steps:
a) A 3D-data obtaining step 302, including obtaining data in three dimensions from 3D-accelerometer 156 (Fig. 1C);
b) A PPG data obtaining step 304 from PPG sensor 100.
Thereafter, a data reduction step 306 is performed, including application of various algorithms to the data to remove noise and outliers (see description hereinbelow) and/or other filters known in the art, thereby forming a reduced data set.
In a determining an activity and/intensity step 308, predetermined rules are applied to the reduced data set from step 306. The reduced data set is compared with historical profiles stored in the microprocessor. If a statistically reliable match is made between the reduced data step and at least one historical profile, then an activity profile and/or intensity may be defined in this step. For example, a "sitting profile" may be determined (see Fig. 4A- item "S").
Thereafter, in an EEE calculation step 310, the EEE of that person may be calculated.
For example, patent attorney SEL, has a mixed profile for hours 6 am- 9am, a sitting profile from 9 am until 11:30 am, a mixed profile from 11 :30am until 12 pm, a sitting profile from 12 pm until 18:00 pm, a walking profile from 18:00 pm until 18:30 pm, a mixed profile from 18:30 pm until 21:45 pm, a sitting profile from 21:45 pm until 22:45 pm and a lying profile from 22:45 pm until 6 am.
The patent attorney's fixed characteristics are: female, body mass index=21.8, age=52, weight 63 kg, height 1.70 m.
These fixed characteristics, coupled with the identified activity profiles from step 308, summed over the period of, for example, 24 hours, enable the determination of the subject's EEE in step 310.
For example, on days where the patent attorney's calorific intake is 2200 Kcal/day, but her resting metabolic rate is 1700 Kcal/day, she must add activities in order not to gain weight.
Central requirement of a sensor for EEE is the importance of long-term wearability and reliable sensor attachment to a bodily part. Since continuous monitoring is required, the device must be non-invasive and worn at all times. A ring worn on the finger configuration for the sensor is a natural choice. Because of the low weight and small size rings, which are generally worn without removal, more often than watches.
However, locating an accelerometer away from the body mass center, at the tip of the arm poses two major challenges in the estimation EE of the subject based on the accelerometer measurements:
a) the tip of the arm can move irrelatively to the movement of the body mass center hence not allowing reliable EEE; and
b) the legs can be very active contributing to the energy expenditure while the tip of the arm does not move a lot (such as during cycling). This invention provides a method for analysis of the data from a ring and/or braceletsensor that overcomes the above challenges and thus enable continuous measurement of EEE based on movement patterns of the tip of the arm as measured by a three-axis accelerometer worn as ring on the finger or bracelet on a wrist. This invention based on the discovery that it is possible to identify movement pattern of the hand that are unique to routine activity profiles of human in free-living environment (e.g. lying , sitting, standing, walking, running, swimming etc.). These patterns can be used to classify and identify the subject's activity type and intensity.
In other embodiments of this invention, the acceleration data is measured using a ring worn by the user on the hand finger and photoplethysmographic (PPG) data is measured using a PPG sensor embedded in an earring that measures PPG signal in the earlobe tissue. The acceleration data and PPG data are then analyzed together and enable classification of activity type and its intensity.
Photoplethysmographic ring sensors
Photoplethysmographic (PPG) finger-ring sensors are described in U.S. Pat. No. 5,964,701, issued Oct. 12, 1999, and in U.S. Pat. No 6,699,199 issued March 2, 2004. Photoplethysmographic finger-ring sensors may be employed, for example, for monitoring such physiological parameters as blood flow, blood constituent concentration, and pulse rate. A simple finger ring sensor is described with reference to FIG. 1C. As shown in FIG. 1C, one or more photo diodes and one light-emitting diodes (LEDs) are imbedded in a ring. The LEDs may emit light in the visible or infrared, and may be particularly chosen to emit light at one or more specified wavelengths. The light passes from the LED to the photo diodes through the finger tissue. Changes in the output signal of the sensor are related to changes in the blood volume in the blood vessel (mainly the digital artery) in the finger tissue.
The measured blood volume fluctuations are the sum of to pressure changes caused by the heart contractions, and pressure waves caused due to movement of the body parts. The pulse of an individual at rest may be detected as a periodic change in the sensor output. However, the main limitation of this technology is related to the fact that movement of the hand and the finger cause fluctuations in the blood volume of the tissue that in turn are causing fluctuations in the signal (noise). The noise is additive to the signal fluctuation due to pressure changes caused by the hart contraction. Thus, many inventions are related to methods for filtering movement artifacts in order to be able measure pulse, or blood gas constituents using such sensors.
The present invention provides a system and method which enable detection of leg movement, without moving the hand, which cause fluctuations in the signal measured in the finger ring sensor. This is due to volume changes in the finger tissue caused by from blood pressure wave originating from the movement of the legs. This fluctuation can be used as an indicator to the fact that although the hands of the subject are not moving, his legs are moving. This data can be used for identifying activity profile in which the hands are not moving but the legs are, e.g. cycling.
Measured blood volume fluctuations are the sum of pressure changes caused by heart contractions, and pressure waves caused due to movement of the other body parts. Both heart rates, and movement of body organs are increased during physical activity of an individual. Moreover, there are activity profiles where body organs movement is slow, not causing significant pressure waves, but still, heart rate will increase due to the energy consumed, causing the over all signal fluctuation of the PPG signal to increase.
U.S. Pat. No 6,699,199, issued on March 2, 2004, teaches the use of PPG signal measured from a specific organ as an indicator of movement.
In the present invention, it has been discovered that changes in photoplethysmograph signal patterns, measured in a region proximal to a digital artery in a finger reflects movement of other body parts (e.g. legs) even when the hands are static.
Thus, continuous analysis of the photoplethysmograph signal patterns measured at the digital artery area in the hand finger, together with readout of three- dimensional accelerometer attached to the hand finger, can be analyzed and used to identify wide range of activity patterns, including activity patterns where the hands are not active.
Method 300 for development of an "if-then" rules set is now described and comprises a number of steps: a) A 3D-data obtaining step 318, including obtaining data in three dimensions from 3D-accelerometer 156 (Fig. 1C);
b) A PPG data obtaining step 316 from PPG sensor 100.
Thereafter, a data reduction step 320 is performed, including application of various algorithms to the data to remove noise and outliers (see description hereinbelow) and/or other filters known in the art, thereby forming a reduced data set.
In a data mining step 322, data mining techniques, known in the art are used for development of a set of rules for classification of rules to define "profiles" for various activities, such as running, walking, lying, sitting etc., as well as various activities intensities. These rules are determined, inter alia, from actual activity profiles measured in groups of volunteers in an activity profile determining step 324.
The output of step 322, is a set of "if and only if rules 328, which are stored in one or more microprocessors 120, 158, 182 in system 140 (Fig. 1C). In a determining an activity and/intensity step 308, predetermined rules are applied to the reduced data set from step 306. The reduced data set is compared with historical profiles stored in the microprocessor. If a statistically reliable match is made between the reduced data step and at least one historical profile, then an activity profile and/or intensity may be defined in this step. For example, a "sitting profile" may be determined (see Fig. 4A- item "S").
Rules 328 are fed into step 308 of the method for real-time for estimation 301 of energy expenditure (EEE).
Additionally or alternatively, data from a predefined database of METs 326 are fed into step 310 of the method for real-time for estimation 301 of energy expenditure (EEE).
The metabolic equivalent of task (MET), or simply metabolic equivalent, is a physiological measure expressing the energy cost of physical activities[l] and is defined as the ratio of metabolic rate (and therefore the rate of energy consumption) during a specific physical activity to a reference metabolic rate, set by convention to 3.5 ml 02 kg-l-min-l or equivalently 1 kcal-kg-l- h-1 or 4.184 kJ-kg-l- h-1.
Fig. 4A shows a scatter chart of a distribution of metabolic equivalent of task (MET), assigned by the method for each of the 15 seconds time segments. Time segments are clustered according to the actual physical activity (PA) types performed at this time segment S- Sitting, STA- Standing, W- Walking, R- Running , M- mixed activity . The box represents 50% of the results, and the line inside the box is the median, filled squares represent outliers in accordance with an embodiment of the present invention.
The present invention provides methods which show that changes in photoplethysmograph signal patterns, measured in the earlobe can accurately reflect movement of other body parts (e.g. legs) even when the hand and finger are static. Therefore, continuous analysis of the photoplethysmograph signal patterns measured at the earlobe, together with readout of a three-dimensional accelerometer attached to a finger, can be analyzed and used to identify wide range of activity patterns, including activity patterns where the hands are not active.
The earlobe PPG sensor is an earring worn on the earlobe containing: an electrical power source, light source and light detection system for measurement of photoplethysmographic signal, electronic circuits, microprocessor, transmitter and receiver. Both the earring PPG sensor and the ring acceleration sensor, are time synchronized between them. And the data generated by PPG sensor is transmitted to the acceleration sensor. And then the full activity profile can be identified by comparison to known activity patterns.
Devices that contain PPG sensor and an acceleration sensor were the output of the sensors is analyzed tighter have been described: US2007299330A and US66991 9 describe a PPG sensor and acceleration sensor on a ring shape monitor, US2008177162 describes a headset in a tongue shape that is inserted into the ear and contain PPG sensor and acceleration sensor. In all the above described devices, the PPG sensor and the acceleration sensor are attached to the same member and are mounted onto the same peace of material, since the acceleration data is used to filter out PPG signal motion artifacts due to movement of the device. However No one have described a device with an acceleration sensor and PPG sensor that there data is analyzed together, and they are not mounted the same body member or on the same peace of material. Identification of activity type and intensity
US6571200 to by Mault J.R. describes an apparatus and a method for monitoring caloric expenditure resulting from body activity. The method is based on two steps, a learning mode and operation mode. During the learning mode the specific subject perform different activities in different rates while the subject's caloric expenditure is measured by an external gas analyzer device, and the activity rate is measured by accelerometer and by heart rate monitor. Specific caloric expenditure coefficient for each activity rate is then calculated. In the operational mode the calorie expenditure is measured based on input from body activity sensor and the hart rate sensor. However, Mault J.R. does not teach how the activity type (e.g. running, walking , swimming , cycling ect.) can be identified during the operational mode.
In the present invention, methods and systems for identification of the subjects activity type and its intensity are provided. This data is used to calculate the caloric expenditure of the subject based on known coefficients.
Calorie Estimated Energy Expenditure Based on RMR and METS
A unit of metabolic equivalent of a task, (MET), is defined as the ratio of a person's working metabolic rate relative to the resting metabolic rate (RMR) during performing a certain task or a certain activity profile. An estimation of person's calorie consumption can be easily calculated using this METS values. The persons RMR (kcal/day) can be estimated based on his age (years) , height (cm) and mass (kg) as given by following equation [MD Mifflin, ST St Jeor, LA Hill, BJ Scott, SA Daugherty and YO Koh American Journal of Clinical Nutrition, A new predictive equation for resting energy expenditure in healthy individuals Vol 51, 241-247]:
Equation 1
Figure imgf000023_0001
where s is +5 for males and -161 for female METS values correlate with oxygen requirements. Starting with 1, which is the least amount of activity (such as resting), the values increase with the amount of activity. For example, running at 9.7 km/h has a METS value of 10. Standard tables exist that provide METS values for a wide range of exercises and activities [Ainsworth, B.E., Haskell, W.L., et al., Compendium of Physical Activities: An update of activity codes and MET intensities. Med Sci Sports Exerc, 2000; 32 (Suppl):S498-S516, (2000)]. Hence, based on the user activity profile, suitable MET value is matched, based on its estimated or measured RMR, an estimation of his energy expenditure of a subject during the specific time segment can be calculated using the following equation:
Equation 2
RMR (kcal/time) x MET x time = EEE (kcal)
The over all estimated energy expenditure during the day/week/month is calculated as a sum of the energy expenditure during all the time segments.
A novel device and method for continues measurement of estimated energy expenditure (EEE) of human in free-leaving environment. The method is based on using a ring shape sensing device that is worn on the hand fingers of the subject as a ring. The device contains: 3 axis (Χ,Υ,Ζ) accelerometer; a Photoplethysmographic sensor (PPG); an electric power source like a battery; electronic circuits; embedded microprocessor used to run an algorithm for data analysis and determination of activity scenario, its intensity, and calculation of EEE; a time counting unit, an electronic memory for data storage; a switch for turning the electric power of the devise, and a radio transmitter and receiver . The sensing device is water sealed. See Fig. 1A. The data be analyzed data may be transmitted to a display unit (e.g computer, cellular telephone) See Fig IB.
The device of this invention can be also in the shape on an earring, which is attached to the earlobe, The device contains: a three axis (Χ,Υ,Ζ) accelerometer; a Photoplethysmographic sensor (PPG) where the light source is on one side of the earlobe, the light goes through the earlobe to the light sensor that is on the other side of the earlobe ; an electric power source like a battery; electronic circuits; embedded microprocessor used to run an algorithm for data analysis and determination of activity scenario, its intensity, and calculation of EEE; a time counting unit, an electronic memory for data storage; a switch for turning the electric power of the devise, and a radio transmitter and receiver. The sensing device is water sealed.
The device of this invention can be also made with sensors that are attached to two different body members: a PPG sensor attached to the earlobe of the subject as an earring, a ring accelerometer worn on the hand finger, and a host units for data analysis and display, see figure 1C for detailed description. Continues analysis of the photoplethysmograph signal patterns measured at the earlobe, together with readout of 3 dimensional accelerometer attached to the hand finger is analyzed and used to identify wide range of activity patterns, including activity patterns where the hands are not active.
The earlobe PPG sensor is an earring worn on the earlobe containing: an electrical power source, light source and light detection system for measurement of photoplethysmographic signal, electronic circuits, microprocessor, transmitter and receiver, a switch for turning the electric power of the devise. The sensing device is water sealed. Both the earring PPG sensor and the ring acceleration sensor, are time synchronized between them using at least one clock (See Fig 1C). The data generated by PPG sensor is transmitted to the acceleration sensor or to the host. The sensors data is analyzed by the microprocessor and activity profile is identified using a classification algorithm.
Three-axis accelerometer that can be used for this invention is a miniature electronic component that is sensitive to acceleration fluctuations in the three axes of movement - x, y, z. Values of the acceleration in each axis are sent on-demand through digital communication to the host controller - typically a microprocessor. The accelerometer is mounted on the device in a way the one of its axis is parallel to the axis of the finger, see Fig. 2.
Another aspect of this invention is a similar device (not shown) and method that does not contain a Photoplethysmographic sensor, and is attached to the collar of a dog or a cat and is used for the evaluation of the energy expenditure of the animal. Activity Profile Classification Algorithm.
Development of Activity Profile Classification Rules Set (Figure 3 right section)
The development of rules is done according to the following process:
i) Recoding of the data from the ring and or earring sensors ( Χ,Υ,Ζ, and PPG) worn by multiple individuals while performing a set of activities (e.g., sleeping, sitting, driving, standing, walking at different speeds, running in different speeds, swimming, cycling, self caring, house maintenances, dancing and the like) at various intensities. In another embodiment of this invention, the EE of the subject is measured using gas analysis-based methods, known in the art (see references) in order to quantify the intensity of the activity during specific time segment.
ii) Calculation of the description parameters (average; standard deviation, coefficient of variance, median, inter-quartile range, integral over the time, minimum value, maximum value, number of times that the signal is crossing the median during a specific time segment) for data recorded during each time segment (e,g each 15 seconds, or 60 seconds), and building a database including the activity type, intensity/speed, and the calculated parameters, for each time segment for each individual.
iii) Appling data mining techniques and software for identifying "if and only if rules for prediction of activity types and intensity based on the calculated parameters of a certain time segment. Such data mining can be done with a commercial soft ware such as WizWhy™ ( Wizsoft inc. Tel Aviv, Israel) for identification of "if and only if prediction rules. Another data mining techmque that can be used for classification is neuronal network analysis, decisions Logistic regression can be used to build a regression model for prediction of continues variable like rurining or walking speed.
iv) Writing a computer program which uses the set of rules to classify the activity profile and intensity of each time segment. Once the activity profile is known, the relevant MET value of this activity profile and the calculate RMR of the subject are used for an estimation of the subject's energy expenditure during the specific time segment using the following equation:
RMR (kcal/time) x MET x time = EEE (kcal)
The over all estimated energy expenditure during the day/week/month is calculated as a sum of the energy expenditure during all the time segments.
EEE Calculation (Figure 3 left section)
The output data acquired by the 4 sensors (acceleration in Χ,Υ,Ζ axis and PPG sensor) during a certain time segment (e.g 10 seconds, or 15 second, or 30 second, or 60 seconds) is analyzed at the microprocessor using a software that performs the classifying algorithm, the algorithm is based on calculation of parameters that describe the data recorded by each of the 4 sensors during a certain time segment. The calculated parameters can be for example : average; standard deviation, coefficient of variance, median, inter-quartile range, integral over the time, minimum value , maximum value, number of times that the signal is crossing the median during a specific time segment. -A set of predetermined if-and-only-if rule employed on the calculated parameters enable to classify the activity profile of the subject (e.g. sleeping, laying, sitting, standing, walking running, cycling , swimming ect.) and to determine its intensity. The energy expenditure of the subject during the specific time segment is the estimated based on the user activity profile, the MET value of the activity profile, its estimated or measured RMR. An estimation of the subject's energy expenditure during the specific time segment can be calculated using the following equation:
RMR (kcal/time) x MET x time = EEE (kcal)
The overall estimated energy expenditure during the day/week/month is calculated as a sum of the energy expenditure during all the time segments.
The EEE value of the certain time segment is then recoded in the memory device. The data is then transmitted to receiver units that can display the information to the subject upon its will.
Novel Features and Advantages: The method has a number of advantages compared with existing energy expenditure estimation methods:
a) the device is cheap and can be miniature in size;
b) the device used in this method can be worn for very long time periods ( months) without interfering the user daily life and there for can be used by people in a free-living environment.
c) It can be used for EEE even when the hands are not moving.
EXAMPLES
Example 1: Development of Physical Activity (PA) type Classification Algorithm.
Acceleration data of 4 individuals perfuming set of activities was recorded using a Wireless Sensing Triple Axis Reference design (ZSTAR), a demo for wireless demonstration of the 3-axis accelerometer MMA7260QT Sensor made by FreeScale Inc. The sensor was mounted on a ring finger worn by the subjects with the X axis parallel to the finger axis. Acceleration data was sampled at 30 Hz and transmitted to a personal computer were it was recoded as G values for x, y and z axis. Each individual have performed the activities described in Table 1.
Table 1
Figure imgf000028_0001
Activity description Activity type Activity
Duration
(minutes):
Standing-making coffee Stand 5
Standing -filling papers Stand 5
Mopping the floor, moderate Mix stand/walk 3
Mopping the floor vigorous Mix stand/walk 2
Walk Up stairs Walk 2
Walk Down stairs Walk 2
Sitting - Driving Sit 10
Walking in the street Walk 5
Walking on a treadmill at 2km/hour Walk 3
Walking on a treadmill at 4km/hour Walk 2
Walking on a treadmill at 6 km/hour Walk 3
Running on a treadmill at 6 km/hour Run 2
Running on a treadmill at 8 km/hour Run 2
Running on a treadmill at 10 km/hour Run 2
Basket ball play, moderate intensity Mix walk/run 3
Basket ball play, high intensity Mix walk/run 3
The data was divided to 873 15 seconds time segments, by using MatLab (MathWorks, 3 Apple Hill Drive Natick, MA 01760-2098 Natick, USA) software the following variables were calculated for each of the 15 seconds time segments (Table 2):
Table 2
Variable Axis (x,y,z) Variable code X
Median V6
Integral X V7
Maximum Peak X V8
Inter-quartile range X V9
Skew X V10
Kurtosis X VI 1
Standard deviation X V12
Coefficient of variance (CV) in segment X V13
Summation of signal power above 0.7Hz X V14
Summation of signal power below 0.7Hz X V15
Number of crossings the median in a segment X V16
CV of the Number of crossings the median of a X
segment in segments current segment , the one
before it and the one after it (n, n-1 and n+1) V17
Median y V18
Integral y V19
Maximum Peak y V20
Inter-quartile range y V21
Skew y V22
Kurtosis y V23
Standard deviation y V24
Coefficient of variance (CV) in segment y V25
Summation of signal power above 0.7Hz y V26
Summation of signal power below 0.7Hz y V27
Number of crossings the median in a segment y V28
CV of the Number of crossings the median of a y
segment in segments current segment , the one
before it and the one after it (n, n-1 and n+1) V29
Median z V30
Integral z V31
Maximum Peak z V32 Inter-quartile range z V33
Skew z V34
Kurtosis z V35
Standard deviation z V36
Coefficient of variance (CV) in segment z V37
Summation of signal power above 0.7Hz z V38
Summation of signal power below 0.7Hz z V39
Number of crossings the median in a segment z V40
CV of the Number of crossings the median of a z
segment in segments current segment, the one
before it and the one after it (n, n-1 and n+1) V41
A data base including the physical activity (PA) types: lying (L), sitting (S), standing (STD), walking (W), running (R), mixed activity ( ), walking /running speed (km/hour), and all variables (V6-V41) for the 873 time segments was compiled. WizWhy data mining software (V4.06, WizSoft Inc.) was used to identify all the significant "if-and-only-if rules for identification of PA types based on the variables.
Rules for identification of lying (L)
The following conditions explain when PA type is not L
1) V33 is 0.06 up to 3.32 (average = 0.56 )
2) V21 is 0.01 up to 0.07 (average = 0.03 )
and V33 is 0.00 up to 0.06 (average = 0.02 )
and V39 is 360.84 up to 435.13 (average = 396.85 )
3) VI 8 is -1.55 up to -0.67 (average = -0.92 )
4) VI 1 is 3.46 up to 14.63 (average = 6.17 )
and V38 is 4.80 up to 106.49 (average = 59.76 )
and V40 is 25.00 up to 105.00 (average = 73.30 )
5) V14 is 111.18 up to 397.99 (average = 194.49 )
and V23 is 2.45 up to 193.77 (average = 9.77 )
and V36 is 0.05 up to 0.17 (average = 0.13 ) ) VI 1 is 3.46 up to 14.65 (average = 6.93 )
and V15 is 193.54 up to 614.20 (average = 354.97 )
and V17 is 0.06 up to 0.32 (average = 0.18 )
) V22 is -0.61 up to -0.24 (average = -0.38 )
and V40 is 106.00 up to 165.00 (average = 125.62 )
) V35 is 1.04 up to 2.73 (average = 2.30 )
When at least one of the conditions holds, the probability that
PA type is not L
is 0.989 (843 out of 852 cases)
When all the conditions do not hold, the probability that
PA type is L
is 0.952 (20 out of 21 cases)
The total number of cases explained by the set of conditions: 863 The total number of cases in the data: 873
Success rate: 0.989 (863 / 873)
The primary probability that:
PA type is not L is 0.967 (844 out of 873 cases)
PA type is L is 0.033 (29 out of 873 cases)
Improvement Factor: 2.900 (0.033 / 0.011)
Rules for identification of sitting (S)
If-and-only-if Rule 1 (out of 2)
The following conditions explain when
PA type is Sitting
1) V7 is 107.72 up to 458.79 (average = 274.75 )
and V8 is 0.49 up to 3.50 (average = 1.51 )
and V37 is -204.60 up to 0.38 (average = -2.04 )
2) V15 is 71.69 up to 229.71 (average = 161.80 )
and V17 is 0.17 up to 1.03 (average = 0.44 )
and V32 is -0.86 up to -0.06 (average = -0.53 )
3) V6 is 0.24 up to 0.97 (average = 0.63 )
and V12 is 0.01 up to 0.20 (average = 0.08 )
and V28 is 113.00 up to 213.00 (average = 145.47 )
4) V24 is 0.19 up to 0.47 (average = 0.28 )
and V31 is -400.02 up to -272.93 (average = -328.70 )
5) V10 is -6.84 up to -0.75 (average = -2.23 )
and VI 5 is 257.56 up to 474.63 (average = 370.79 ) and V37 is -5.00 up to 0.26 (average = -0.79 )
6) V6 is 0.24 up to 0.97 (average = 0.61 )
and V10 is -0.64 up to 7.66 (average = 0.44 ) and V26 is 2.69 up to 126.71 (average = 59.34 )
7) V6 is 0.27 up to 1.00 (average = 0.66 )
and VI 6 is 87.00 up to 110.00 (average = 97.06 ) and V37 is -204.60 up to 0.03 (average = -5.09 )
When at least one of the conditions holds, the probability that
PA type is S
is 0.672 (221 out of 329 cases)
When all the conditions do not hold, the probability that PA type is not S
is 0.996 (542 out of 544 cases)
The total number of cases explained by the set of conditions:
The total number of cases in the data: 873
Success rate: 0.874 (763 / 873)
The primary probability that:
PA type is S is 0.255 (223 out of 873 cases)
PA type is not S is 0.745 (650 out of 873 cases)
Improvement Factor: 2.027 (0.255 / 0.126)
If-and-only-if ule 2 (out of 2)
The following conditions explain when
PA type is not S
V31 is -263.71 up to 625.18 (average = 143.10 )
and V32 is -0.02 up to 5.49 (average = 1.84 )
VI 1 is 1.93 up to 3.41 (average = 2.63 )
and V29 is 0.12 up to 1.42 (average = 0.42 )
and V35 is 3.48 up to 4.77 (average = 3.97 )
V8 is 0.09 up to 0.47 (average = 0.29 )
and VI 7 is 0.00 up to 0.11 (average = 0.05 )
VI 1 is 1.09 up to 3.45 (average = 2.61 )
and V21 is 0.19 up to 2.77 (average = 0.72 )
and V31 is -261.97 up to 625.18 (average = 184.43 )
VI 1 is 11.51 up to 164.96 (average = 29.80 )
and V22 is -7.89 up to -1.24 (average = -2.82 )
and V32 is -0.02 up to 5.49 (average = 1.77 )
VI 5 is 25.47 up to 222.38 (average = 112.95 )
and V29 is 0.12 up to 0.30 (average = 0.17 ) and V41 is 0.02 up to 0.11 (average = 0.06 )
7) V8 is 3.53 up to 3.81 (average = 3.66 )
8) V16 is 69.00 up to 86.00 (average = 79.57 )
and V39 is 490.14 up to 911.40 (average = 611.29 )
9) VI 1 is 1.99 up to 3.45 (average = 2.73 )
and V20 is -0.95 up to -0.34 (average = -0.56 )
and V34 is 0.12 up to 0.82 (average = 0.32 )
10) Vl l is 11.18 up to 164.96 (average = 20.77 )
and VI 6 is 35.00 up to 66.00 (average = 56.48 ) and V23 is 4.10 up to 76.41 (average = 11.88 )
11) VI 6 is 9.00 up to 66.00 (average = 48.34 )
and V23 is 1.09 up to 3.15 (average = 2.49 )
and V31 is -261.53 up to 551.25 (average = 169.85 )
12) V20 is -0.84 up to -0.35 (average = -0.55 )
and V35 is 2.04 up to 3.43 (average = 2.80 )
and V40 is 60.00 up to 117.00 (average = 83.50 )
13) V21 is 0.20 up to 0.92 (average = 0.46 )
and V22 is 0.56 up to 2.26 (average = 1.13 )
and V23 is 3.97 up to 25.94 (average = 8.49 )
14) V20 is -0.93 up to -0.34 (average = -0.56 )
and V40 is 60.00 up to 117.00 (average = 84.66 )
When at least one of the conditions holds, the probability that
PA type is not S
is 0.880 (644 out of 732 cases)
When all the conditions do not hold, the probability that
PA type is S
is 0.957 (135 out of 141 cases)
The total number of cases explained by the set of conditions:
779
The total number of cases in the data: 873 Success rate: 0.892 (779 / 873)
The primary probability that:
PA type is not S is 0.745 (650 out of 873 cases)
PA type is S is 0.255 (223 out of 873 cases)
Improvement Factor: 2.372 (0.255 / 0.108)
Rules for identification of Standing (STD)
The following conditions explain when
PA type is not STD
V28 is 48.00 up to 213.00 (average = 86.53 ) and V40 is 39.00 up to 216.00 (average = 84.07 )
V38 is 2.43 up to 162.92 (average = 91.86 ) and V39 is 210.48 up to 463.21 (average = 323.42 )
V14 is 639.36 up to 1,940.37 (average = 1,035.87 )
V29 is 0.00 up to 0.10 (average = 0.06 ) and V35 is 4.96 up to 92.58 (average = 12.20 )
V19 is -675.57 up to -388.96 (average = -454.29 )
VI 1 is 1.42 up to 9.44 (average = 3.70 ) and V22 is -1.01 up to 0.09 (average = -0.39 ) and V26 is 70.02 up to 149.97 (average = 112.97 )
V10 is 0.13 up to 1.04 (average = 0.52 ) and V17 is 0.00 up to 0.11 (average = 0.04 )
V16 is 71.00 up to 208.00 (average = 115.38 ) and V32 is -0.14 up to 0.36 (average = 0.13 ) and V37 is -1.50 up to -0.31 (average = -0.51 )
V6 is 0.24 up to 0.85 (average = 0.56 ) and V27 is 14.68 up to 243.50 (average = 149.38 ) and V33 is 0.00 up to 0.21 (average = 0.07 )
V7 is -65.84 up to 131.07 (average = 48.98 ) and V21 is 1.06 up to 2.38 (average = 1.47 )
V8 is 4.46 up to 5.30 (average = 5.11 ) and V34 is 0.04 up to 3.02 (average = 0.60 )
V10 is -5.22 up to -0.53 (average = -1.39 ) and V31 is -431.78 up to -273.85 (average = -334.13 )
V10 is 0.15 up to 0.98 (average = 0.60 ) and V39 is 18.36 up to 135.78 (average = 100.96 )
V6 is 0.31 up to 0.87 (average = 0.65 ) and V7 is 140.84 up to 406.99 (average = 271.36 ) and V12 is 0.62 up to 0.98 (average = 0.75 )
15) V7 is -181.34 up to 135.65 (average = -19.85 )
and V10 is -0.50 up to 0.13 (average = -0.19 )
and V40 is 48.00 up to 177.00 (average = 81.02 )
16) V7 is 163.67 up to 410.05 (average = 304.29 )
and V9 is 0.17 up to 0.20 (average = 0.19 )
and V40 is 45.00 up to 142.00 (average = 91.71 )
17) V8 is 1.28 up to 4.38 (average = 3.00 )
and V16 is 72.00 up to 160.00 (average = 101.43 ) and V17 is 0.12 up to 0.15 (average = 0.14 )
18) V10 is -1.91 up to -0.52 (average = -0.94 )
and VI 8 is -0.87 up to -0.23 (average = -0.68 )
and V26 is 430.43 up to 1,995.54 (average = 1,174.00 )
When at least one of the conditions holds, the probability that
PA type is not STD
is 0.9 1 (713 out of 766 cases)
When all the conditions do not hold, the probability that
PA type is STD
is 0.963 (103 out of 107 cases)
The total number of cases explained by the set of conditions:
816
The total number of cases in the data: 873
Success rate: 0.935 (816 / 873)
The primary probability that:
PA type is not STD is 0.821 (717 out of 873 cases) PA type is STD is 0.179 (156 out of 873 cases)
Improvement Factor: 2.737 (0.179 / 0.065) Rules for identification of walking fW)
If-and-only-if Rule 1 (out of 2)
The following conditions explain when
PA type is W
V14 is 67.75 up to 608.17 (average = 255.95 )
and V19 is -675.57 up to -379.43 (average = -450.71 ) and V21 is 0.09 up to 0.80 (average = 0.35 )
V6 is -0.44 up to -0.11 (average = -0.21 )
and V9 is 0.09 up to 0.64 (average = 0.34 )
and V10 is 0.14 up to 2.61 (average = 0.71 )
V6 is -0.42 up to -0.11 (average = -0.22 )
and V10 is -0.50 up to 0.12 (average = -0.14 )
and V27 is 409.45 up to 747.41 (average = 473.24 )
V8 is 0.88 up to 3.65 (average = 1.52 )
and VI 5 is 16.10 up to 96.46 (average = 69.76 )
and V21 is 0.12 up to 0.75 (average = 0.41 )
When at least one of the conditions holds, the probability that
PA type is W
is 0.991 (213 out of 215 cases)
When all the conditions do not hold, the probability that
PA type is not W
is 0.991 (652 out of 658 cases)
The total number of cases explained by the set of conditions:
The total number of cases in the data: 873
Success rate: 0.991 (865 / 873)
The primary probability that:
PA type is W is 0.251 (219 out of 873 cases) PA type is not W is 0.749 (654 out of 873 cases)
Improvement Factor: 27.375 (0.251 / 0.009)
If-and-only-if Rule 2 (out of 2)
The following conditions explain when
PA type is not W
V6 is -0.09 up to 1.12 (average = 0.58 )
and VI 5 is 99.01 up to 667.67 (average = 377.67 )
and V20 is -0.39 up to 5.54 (average = 1.20 )
V21 is 0.01 up to 0.08 (average = 0.03 )
V30 is -1.02 up to 0.20 (average = -0.45 )
V27 is 14.68 up to 307.63 (average = 187.34 )
V10 is -6.84 up to -0.52 (average = -1.37 )
and V18 is -0.92 up to 0.48 (average = -0.37 )
When at least one of the conditions holds, the probability that
PA type is not W
is 0.978 (654 out of 669 cases)
When all the conditions do not hold, the probability that
PA type is W
is 1.000 (204 out of 204 cases)
The total number of cases explained by the set of conditions:
The total number of cases in the data: 873
Success rate: 0.983 (858 / 873)
The primary probability that:
PA type is not W is 0.749 (654 out of 873 cases)
PA type is W is 0.251 (219 out of 873 cases) Improvement Factor: 14.600 (0.251 / 0.017)
Rules for identification of Running (R)
If-and-only-if Rule 1 (out of 2)
The following conditions explain when
PA type is R
1) V6 is 0.44 up to 1.12 (average = 0.76 )
and V9 is 1.00 up to 3.36 (average = 2.10 )
and V16 is 69.00 up to 106.00 (average = 81.68 )
When at least one of the conditions holds, the probability that
PA type is R
is 0.951 (97 out of 102 cases)
When all the conditions do not hold, the probability that
PA type is not R
is 1.000 (771 out of 771 cases)
The total number of cases explained by the set of conditions:
The total number of cases in the data: 873
Success rate: 0.994 (868 / 873)
The primary probability that:
PA type is R is 0.111 (97 out of 873 cases)
PA type is not R is 0.889 (776 out of 873 cases)
Improvement Factor: 19.400 (0.111 / 0.006)
If-and-only-if Rule 2 (out of 2)
The following conditions explain when PA type is not R
V9 is 0.01 up to 0.98 (average = 0.35 )
VI 7 is 0.04 up to 1.19 (average = 0.26 )
and V35 is 2.74 up to 190.58 (average = 9.33 )
and V41 is 0.03 up to 1.12 (average = 0.28 )
V6 is -0.59 up to 0.42 (average = 0.03 )
V16 is 4.00 up to 68.00 (average = 51.72 )
VI 1 is 2.87 up to 164.96 (average = 7.63 )
and VI 7 is 0.04 up to 0.82 (average = 0.20 )
and V27 is 407.57 up to 1,118.20 (average = 545.27 )
When at least one of the conditions holds, the probability that
PA type is not R
is 1.000 (776 out of 776 cases)
When all the conditions do not hold, the probability that
PA type is R
is 1.000 (97 out of 97 cases)
The total number of cases explained by the set of conditions:
The total number of cases in the data: 873
Success rate: 1.000 (873 / 873)
The primary probability that:
PA type is not R is 0.889 (776 out of 873 cases)
PA type is R is 0.111 (97 out of 873 cases)
Improvement Factor: l.#IO (0.111 / 0.000) Rules for identification of mixed activity CM)
If-and-only-ifRule 1 (out of 2)
The following conditions explain when
PA type is M
1) VI 1 is 3.46 up to 33.54 (average = 8.23 )
and V20 is 0.55 up to 5.54 (average = 2.31 )
and V31 is -48.68 up to 582.23 (average = 222.33 )
2) V21 is 0.29 up to 0.91 (average = 0.54 )
and V27 is 31.04 up to 160.20 (average = 110.76 ) and V28 is 88.00 up to 163.00 (average = 112.77 )
3) VI 6 is 41.00 up to 76.00 (average = 57.07 )
and V20 is 0.55 up to 4.34 (average = 1.81 )
and V28 is 89.00 up to 142.00 (average = 107.19 )
When at least one of the conditions holds, the probability that
PA type is M
is 0.679 (144 out of 212 cases)
When all the conditions do not hold, the probability that
PA type is not M
is 0.992 (656 out of 661 cases)
The total number of cases explained by the set of conditions:
800
The total number of cases in the data: 873
Success rate: 0.916 (800 / 873)
The primary probability that:
PA type is M is 0.171 (149 out of 873 cases)
PA type is not M is 0.829 (724 out of 873 cases)
Improvement Factor: 2.041 (0.171 / 0.084) If-and-only-if Rule 2 (out of 2)
The following conditions explain when
PA type is not M
V20 is -0.95 up to 0.54 (average = -0.18 )
VI 1 is 1.09 up to 3.45 (average = 2.68 )
and V13 is -22.43 up to 5.11 (average = 0.48 ) and V19 is -464.11 up to -15.64 (average = -275.25 )
V39 is 219.80 up to 462.69 (average = 336.60 ) and V40 is 8.00 up to 65.00 (average = 43.81 )
V14 is 2.55 up to 198.84 (average = 104.90 )
V16 is 79.00 up to 208.00 (average = 123.57 ) and V32 is -1.00 up to 0.67 (average = -0.09 )
V10 is 0.42 up to 3.98 (average = 1.10 )
and V16 is 29.00 up to 76.00 (average = 61.43 ) and V19 is -458.73 up to 0.81 (average = -285.79 )
V27 is 171.60 up to 602.37 (average = 357.64 ) and V28 is 7.00 up to 53.00 (average = 37.38 )
V6 is -0.03 up to 0.61 (average = 0.45 )
and V22 is -1.80 up to -0.24 (average = -0.83 ) and V31 is -417.07 up to -64.22 (average = -288.29 )
V15 is 25.47 up to 135.62 (average = 84.76 ) and V17 is 0.06 up to 0.34 (average = 0.15 ) and V39 is 218.74 up to 463.21 (average = 316.59 )
V6 is -0.44 up to -0.08 (average = -0.20 ) and V29 is 0.00 up to 0.05 (average = 0.03 )
V6 is 0.61 up to 0.79 (average = 0.71 )
and V25 is -1.19 up to 135.56 (average = 4.79 ) and V29 is 0.42 up to 1.00 (average = 0.58 )
VI 6 is 45.00 up to 76.00 (average = 64.89 ) and V17 is 0.04 up to 0.06 (average = 0.05 )
V18 is -1.22 up to -0.92 (average = -1.01 ) and V27 is 417.52 up to 596.59 (average = 492.71 )
When at least one of the conditions holds, the probability that
PA type is not M
is 0.989 (724 out of 732 cases)
When all the conditions do not hold, the probability that
PA type is M
is 1.000 ( 141 out of 141 cases)
The total number of cases explained by the set of conditions:
865
The total number of cases in the data: 873
Success rate: 0.991 (865 / 873)
The primary probability that:
PA type is not M is 0.829 (724 out of 873 cases)
PA type is M is 0.171 (149 out of 873 cases)
Improvement Factor: 18.625 (0.171 / 0.009)
Evolution of the algorithm performance
Table 3 provides result of raw prediction raw - agreement with actual Physical Activity (PA) type performed. L= lying. S=Sitting, STD= Standing, W=Walking in different speeds 2-6 km/hour, R= Running in different speeds 4-10 km/hour. M= mixed activity runnmg/walking (e.g. basket ball playing).
Y= Yes, the PA was identified, full agreement.
P= Partial agreement, identification of L or S or STD for actual PA type L, or S, or STD. Identification of W or R or M for actual PA type W, or R, or M.
Unknown (U)- no PA type was identified.
Misclassification (M) - the PA was identified as significantly different from actual e.g. L, S, STD, instead of W, R, M. Table 3
Figure imgf000046_0001
Ag ree
Ru
n PA me
No. type L Prediction S Prediction STD Prediction W Prediction R Prediction M Prediction nt
46 M NoL NoS No STD NoW NoR M Y
47 M NoL NoS No STD NoW NoR M Y
48 M NoL NoS No STD NoW NoR M Y
49 M NoL NoS No STD NoW No R M Y
50 M NoL NoS No STD NoW NoR M Y
51 M NoL NoS No STD No W NoR M Y
52 M NoL NoS No STD NoW No R M Y
53 M No L S No STD NoW NoR NoM M
54 M NoL NoS No STD NoW NoR NoM U
55 M NoL NoS No STD NoW NoR M Y
56 M NoL NoS No STD No W NoR M Y
57 M NoL NoS No STD NoW NoR M Y
58 M NoL NoS No STD NoW NoR M Y
59 M NoL NoS No STD NoW No R M Y
60 M NoL NoS No STD NoW NoR M Y
61 M NoL NoS No STD NoW NoR M Y
62 M No L NoS No STD NoW NoR M Y
63 M NoL NoS No STD W NoR M Y
64 M NoL NoS No STD NoW NoR M Y
65 M NoL NoS No STD NoW NoR M Y
66 M NoL NoS No STD NoW NoR M Y
67 M L NoS No STD W No R M Y
68 M NoL NoS No STD NoW NoR M Y
69 M NoL NoS No STD NoW NoR M Y
70 M NoL NoS No STD NoW No R M Y
71 M NoL NoS No STD NoW NoR M Y
72 M NoL NoS No STD NoW NoR M Y
73 M No L No S No STD No W No R M Y
74 M NoL NoS No STD NoW NoR M Y
75 M NoL NoS No STD NoW NoR M Y
76 M NoL NoS No STD NoW No R Y
77 M L NoS No STD NoW NoR M P
78 M NoL NoS No STD NoW NoR M Y
79 M NoL No S No STD No W No R NoM u
80 M NoL NoS No STD NoW NoR M Y
81 M NoL NoS No STD NoW NoR M Y
82 M NoL No S No STD NoW NoR M Y
83 R NoL NoS No STD NoW R NoM Y
84 R NoL NoS No STD NoW R NoM Y
85 R No L NoS No STD No W R NoM Y
86 R NoL NoS No STD NoW R NoM Y
87 R NoL NoS No STD NoW R NoM Y
88 R NoL NoS No STD No W R NoM Y
89 R NoL NoS No STD NoW R NoM Y
90 R NoL NoS No STD NoW R M Y
91 R NoL No S No STD NoW R NoM Y
92 R NoL NoS No STD NoW R NoM Y Ag ree
Ru
n PA me
No. type L Prediction S Prediction STD Prediction W Prediction R Prediction M Prediction nt
93 No L NoS No STD NoW R NoM Y
94 R NoL NoS No STD No W R NoM Y
95 R NoL NoS No STD NoW R NoM Y
96 R No L NoS No STD NoW R NoM Y
97 R NoL No S No STD NoW R NoM Y
98 R NoL NoS No STD NoW R NoM Y
99 R NoL NoS No STD No W R NoM Y
100 R NoL No S No STD NoW R NoM Y
101 R NoL NoS No STD NoW R NoM Y
102 R No L NoS No STD NoW R NoM Y
103 R No L No S No STD NoW R NoM Y
104 R No L NoS No STD NoW R NoM Y
105 R NoL NoS No STD NoW R NoM Y
106 R NoL NoS No STD NoW R NoM Y
107 R NoL NoS No STD NoW R NoM Y
108 R NoL NoS No STD NoW R NoM Y
109 R NoL NoS No STD NoW R NoM Y
110 R NoL NoS No STD NoW R NoM Y
111 R NoL No No STD NoW R NoM Y
112 R NoL NoS No STD NoW R NoM Y
113 R NoL NoS No STD NoW R NoM Y
114 R NoL NoS No STD No W R NoM Y
115 R NoL NoS No STD NoW R NoM Y
116 R NoL NoS No STD NoW R NoM Y
117 R No L No S No STD NoW R NoM Y
118 R NoL NoS No STD NoW R NoM Y
119 R NoL NoS No STD NoW R NoM Y
120 R No L No S No STD NoW R NoM Y
121 R No L NoS No STD NoW R NoM Y
122 R NoL NoS No STD NoW R No M Y
123 R NoL No S No STD No R NoM Y
124 R NoL NoS No STD NoW R NoM Y
125 R NoL NoS No STD NoW R NoM Y
126 R No L NoS No STD NoW R No M Y
127 R No L NoS No STD NoW R NoM Y
128 R NoL NoS No STD NoW R NoM Y
129 R No L NoS No STD NoW R NoM Y
130 R NoL NoS No STD NoW R NoM Y
131 S No L S No STD NoW NoR NoM Y
132 S NoL s No STD NoW No No M Y
133 s NoL s No STD NoW NoR NoM Y
134 s L s No STD No W NoR NoM Y
135 s NoL s No STD No W NoR No M Y
136 s No L s No STD NoW NoR NoM Y
138 s No L s No STD NoW NoR NoM Y
139 s L s No STD NoW NoR NoM Y
140 s NoL s No STD NoW NoR NoM Y Ag ree
Ru
n PA me
No. type L Prediction S Prediction STD Prediction W Prediction R Prediction M Prediction nt
142 S NoL S No STD NoW NoR NoM Y
143 s NoL S No STD NoW NoR No M Y
144 s No L s No STD NoW NoR NoM Y
145 s NoL s No STD NoW NoR NoM Y
146 s No L s No STD NoW NoR NoM Y
147 s No L s No STD NoW NoR NoM Y
148 s NoL s No STD NoW NoR NoM Y
149 s No L s STD NoW NoR NoM Y
150 s L s No STD NoW NoR NoM Y
151 s No L s No STD NoW NoR NoM Y
152 s NoL s No STD NoW NoR NoM Y
153 s NoL s No STD NoW NoR No Y
154 s NoL s No STD NoW NoR NoM Y
155 s NoL s No STD NoW NoR NoM Y
156 s NoL s No STD NoW NoR NoM Y
157 s L s No STD NoW NoR NoM Y
158 s NoL s No STD NoW NoR NoM Y
159 s L s No STD NoW NoR NoM Y
160 s NoL s No STD NoW NoR NoM Y
161 s NoL s No STD NoW NoR NoM Y
165 s NoL s No STD NoW NoR NoM Y
166 s L s STD NoW NoR NoM Y
167 s NoL s No STD NoW NoR NoM Y
168 s L s No STD NoW NoR NoM Y
169 s NoL s No STD NoW No R NoM Y
171 s NoL s No STD NoW NoR NoM Y
172 s NoL s No STD NoW NoR NoM Y
173 s NoL s No STD NoW NoR No M Y
174 s L s No STD NoW NoR NoM Y
175 s NoL s No STD No W NoR NoM Y
176 s NoL s No STD NoW NoR NoM Y
178 s NoL s No STD NoW NoR NoM Y
181 s NoL s No STD NoW NoR NoM Y
182 s No L s No STD NoW NoR NoM Y
183 s NoL s No STD NoW NoR NoM Y
184 s NoL s No STD NoW NoR NoM Y
186 s NoL s No STD No W NoR NoM Y
187 s NoL s No STD NoW NoR NoM Y
188 s NoL s No STD NoW NoR NoM Y
189 s NoL s No STD No W NoR NoM Y
190 s NoL s No STD NoW NoR NoM Y
191 s NoL s No STD No W NoR NoM Y
192 s NoL s No STD No W NoR NoM Y
193 s L s No STD NoW NoR NoM Y
194 s NoL s No STD NoW NoR NoM Y
195 s NoL s No STD No W NoR NoM Y
196 s NoL s No STD NoW NoR NoM Y Ag ree
Ru
n PA me No. type L Prediction S Prediction STD Prediction W Prediction R Prediction M Prediction nt
197 S L S No STD NoW NoR NoM Y
198 S NoL S No STD No W No R NoM Y
199 s NoL s No STD NoW NoR NoM Y
200 s NoL s No STD No W NoR NoM Y
201 s NoL s No STD No W NoR NoM Y
202 s NoL s No STD NoW NoR NoM Y
205 s NoL s STD No W No R NoM Y
207 s NoL s No STD NoW NoR NoM Y
208 s L s No STD NoW NoR NoM Y
209 s No L s No STD NoW NoR NoM Y
212 s NoL s No STD NoW NoR NoM Y
213 s NoL s No STD NoW No R NoM Y
214 s No L s No STD NoW NoR NoM Y
215 s NoL s No STD NoW NoR NoM Y
216 s L s No STD NoW NoR NoM Y
217 s NoL s No STD NoW NoR No M Y
218 s L s No STD NoW NoR NoM Y
219 s L s No STD NoW NoR NoM Y
220 s L s No STD NoW NoR NoM Y
221 s NoL s No STD NoW No NoM Y
222 s L s No STD NoW NoR NoM Y
225 s NoL s No STD No W NoR NoM Y
226 s No L s No STD W NoR NoM Y
227 s NoL s No STD NoW NoR NoM Y
229 s NoL s No STD No W NoR NoM Y
230 s NoL s No STD NoW NoR NoM Y
231 s NoL s No STD NoW NoR NoM Y
232 s No L s No STD NoW NoR NoM Y
233 s NoL s No STD NoW NoR NoM Y
234 s NoL s No STD NoW NoR NoM Y
236 s No L s No STD No W No No M Y
237 s NoL s No STD NoW NoR NoM Y
238 s L s No STD NoW NoR NoM Y
239 s NoL s No STD NoW NoR NoM Y
240 s NoL s No STD NoW NoR NoM Y
242 s NoL s STD NoW NoR NoM Y
243 s No L s No STD NoW No R NoM Y
177 s NoL NoS No STD NoW NoR NoM u
185 s NoL NoS No STD NoW NoR NoM u
210 s NoL No S No STD NoW NoR NoM u
241 s NoL NoS No STD NoW NoR NoM u
137 s NoL NoS STD NoW NoR NoM P
141 s NoL No S STD NoW NoR NoM P
162 s NoL NoS STD NoW NoR NoM P
163 s L NoS No STD NoW NoR NoM P
164 s NoL No S STD NoW No NoM P
170 s NoL NoS STD NoW NoR NoM P Ag ree
Ru
n PA me No. type L Prediction S Prediction STD Prediction W Prediction R Prediction M Prediction nt
179 s NoL S STD No W NoR NoM P
180 s NoL S STD NoW NoR NoM P
203 s NoL S STD No W NoR NoM P
204 s NoL NoS STD No W NoR NoM P
206 s NoL NoS STD No W NoR NoM P
211 s NoL NoS STD NoW NoR NoM P
223 s NoL NoS STD No W NoR NoM P
224 s L NoS No STD NoW NoR NoM P
228 s No L NoS STD NoW NoR NoM P
235 s NoL NoS STD NoW NoR NoM P
244 STD NoL NoS STD NoW NoR NoM Y
246 STD NoL S STD No W NoR NoM Y
247 STD NoL NoS STD NoW NoR NoM Y
250 STD NoL NoS STD NoW NoR NoM Y
252 STD NoL NoS STD W NoR NoM Y
253 STD NoL No S STD NoW No R NoM Y
254 STD NoL NoS STD NoW NoR NoM Y
255 STD NoL NoS STD NoW NoR NoM Y
256 STD NoL S STD NoW NoR NoM Y
258 STD NoL NoS STD NoW NoR NoM Y
259 STD NoL NoS STD NoW NoR NoM Y
262 STD NoL NoS STD NoW No R NoM Y
263 STD NoL NoS STD NoW NoR NoM Y
264 STD NoL NoS STD NoW NoR NoM Y
265 STD NoL NoS STD No W NoR No M Y
266 STD No L NoS STD NoW NoR NoM Y
267 STD NoL NoS STD NoW NoR NoM Y
268 STD NoL NoS STD NoW NoR NoM Y
270 STD NoL NoS STD NoW NoR NoM Y
271 STD NoL NoS STD NoW NoR NoM Y
272 STD NoL S STD No W NoR NoM Y
274 STD NoL NoS STD NoW NoR NoM Y
275 STD No L NoS STD NoW NoR NoM Y
276 STD NoL NoS STD No W No R NoM Y
277 STD NoL NoS STD NoW NoR NoM Y
279 STD NoL S STD NoW NoR NoM Y
283 STD NoL No S STD NoW NoR NoM Y
284 STD NoL NoS STD NoW NoR NoM Y
285 STD NoL NoS STD NoW NoR NoM Y
288 STD NoL No S STD No No R NoM Y
291 STD NoL NoS STD NoW NoR NoM Y
292 STD NoL NoS STD NoW NoR NoM Y
293 STD No L S STD NoW NoR NoM Y
298 STD NoL S STD NoW No R No M Y
300 STD NoL S STD NoW NoR NoM Y
305 STD NoL NoS STD NoW NoR NoM Y
312 STD NoL NoS STD NoW NoR NoM Y Ag ree
Ru
n PA me
No. type L Prediction S Prediction STD Prediction W Prediction R Prediction M Prediction nt
313 STD No L S STD NoW NoR NoM Y
314 STD NoL NoS STD No W No R NoM Y
315 STD NoL NoS STD NoW No R NoM Y
316 STD No L S STD No W NoR NoM Y
249 STD No L NoS No STD NoW No R NoM u
251 STD NoL NoS No STD No W NoR NoM u
269 STD No L NoS No STD NoW NoR NoM u
278 STD NoL NoS No STD NoW No R NoM u
281 STD No L NoS No STD NoW NoR NoM u
287 STD NoL No S No STD NoW NoR NoM u
289 STD NoL NoS No STD NoW NoR NoM u
290 STD No L NoS No STD NoW NoR NoM u
295 STD No L NoS No STD NoW NoR NoM u
299 STD NoL NoS No STD NoW NoR NoM u
302 STD NoL NoS No STD NoW NoR NoM u
304 STD No L NoS No STD NoW No NoM u
306 STD NoL NoS No STD NoW NoR NoM u
308 STD NoL NoS No STD NoW NoR NoM u
309 STD NoL NoS No STD NoW No R NoM u
311 STD NoL NoS No STD NoW NoR NoM u
257 STD NoL S No STD NoW NoR NoM P
260 STD NoL S No STD NoW No R No M P
261 STD NoL S No STD NoW NoR NoM P
273 STD NoL S No STD NoW NoR NoM P
280 STD NoL S No STD NoW NoR NoM P
282 STD NoL S No STD NoW NoR NoM P
294 STD NoL S No STD NoW NoR NoM P
296 STD L S No STD No W NoR NoM P
301 STD No L s No STD NoW NoR NoM P
303 STD L s No STD NoW NoR NoM P
310 STD NoL s No STD NoW NoR NoM P
245 STD NoL NoS No STD NoW NoR M M
248 STD L NoS No STD W NoR NoM M
286 STD NoL NoS No STD W No R NoM M
297 STD NoL NoS STD NoW NoR M M
307 STD NoL NoS STD NoW NoR M M
317 No L NoS No STD W NoR NoM Y
318 W NoL NoS No STD W NoR NoM Y
319 W NoL NoS No STD W NoR NoM Y
320 W NoL No S No STD W NoR NoM Y
321 W NoL NoS No STD W NoR NoM Y
322 W NoL NoS No STD W NoR NoM Y
323 W NoL S No STD W No R No M Y
324 w NoL NoS No STD w NoR NoM Y
325 w NoL NoS No STD w NoR NoM Y
326 w NoL No S No STD w No R NoM Y
327 w NoL No S No STD w NoR NoM Y Ag ree
Ru
n PA me
No. type L Prediction S Prediction STD Prediction W Prediction R Prediction M Prediction nt
328 w NoL NoS No STD W NoR No M Y
329 w NoL NoS No STD W No R NoM Y
330 w No L NoS No STD w NoR NoM Y
331 w NoL NoS No STD w NoR NoM Y
332 w NoL NoS No STD w NoR NoM Y
333 w NoL NoS No STD w NoR NoM Y
334 w NoL NoS No STD w NoR NoM Y
335 w No L No S No STD w NoR NoM Y
336 w NoL NoS No STD w NoR NoM Y
337 w NoL No S No STD w NoR NoM Y
338 w NoL NoS No STD w NoR NoM Y
339 w NoL NoS No STD w NoR NoM Y
340 w NoL No S No STD w NoR NoM Y
342 w NoL NoS No STD w NoR NoM Y
343 w NoL NoS No STD w NoR NoM Y
344 w NoL NoS No STD w NoR NoM Y
345 w No L NoS No STD w NoR NoM Y
346 w NoL NoS No STD w No R NoM Y
347 w No L No S No STD w NoR NoM Y
348 w NoL NoS No STD w NoR NoM Y
349 w NoL NoS No STD w NoR NoM Y
350 w No L No S No STD w No R NoM Y
351 w NoL NoS No STD w NoR NoM Y
352 w NoL NoS No STD w NoR NoM Y
353 w NoL NoS No STD w No R NoM Y
354 w NoL NoS No STD w NoR NoM Y
355 w NoL NoS No STD w NoR NoM Y
356 No L NoS No STD w NoR NoM Y
357 w No L NoS No STD w NoR NoM Y
358 w NoL NoS No STD w NoR NoM Y
359 w NoL No S No STD w NoR NoM Y
360 w NoL NoS No STD w NoR NoM Y
361 w NoL NoS No STD w NoR NoM Y
362 w NoL No S No STD w NoR NoM Y
363 w NoL NoS No STD w NoR NoM Y
364 w NoL NoS No STD w NoR NoM Y
365 w No L NoS No STD w NoR No M Y
366 w NoL NoS No STD w NoR No M Y
367 w NoL NoS No STD w NoR NoM Y
368 NoL No S No STD w NoR NoM Y
369 w NoL NoS No STD w NoR M Y
370 w NoL NoS No STD w NoR NoM Y
371 w No L No S No STD w No R NoM Y
372 w NoL NoS No STD w NoR NoM Y
373 w NoL NoS No STD w NoR NoM Y
374 w NoL No S No STD w NoR NoM Y
376 w NoL NoS No STD w NoR NoM Y Ag ree
Ru
n PA me
No. type L Prediction S Prediction STD Prediction W Prediction R Prediction M Prediction nt
377 W NoL NoS No STD W NoR NoM Y
378 W NoL NoS No STD W No R NoM Y
379 w NoL NoS No STD w NoR NoM Y
380 w NoL NoS No STD w NoR NoM Y
381 w NoL NoS No STD w NoR NoM Y
382 w NoL NoS No STD w NoR NoM Y
383 w No L NoS No STD w No R NoM Y
384 w NoL NoS No STD w NoR NoM Y
385 w NoL NoS No STD w NoR NoM Y
386 w NoL No S No STD w NoR NoM Y
387 w NoL NoS No STD w NoR NoM Y
388 w No L NoS No STD w NoR NoM Y
391 w NoL NoS No STD w No R No M Y
392 w NoL NoS No STD w NoR NoM Y
393 w NoL NoS No STD w NoR NoM Y
394 w NoL No S No STD w NoR NoM Y
395 w NoL NoS No STD w NoR NoM Y
397 w NoL NoS No STD w NoR NoM Y
398 No L No S No STD w NoR NoM Y
399 w NoL No S No STD w NoR No M Y
400 w NoL NoS No STD w NoR NoM Y
401 w NoL NoS No STD w NoR NoM Y
402 w NoL NoS No STD w NoR NoM Y
403 w NoL NoS No STD w NoR NoM Y
404 w NoL No S No STD w NoR NoM Y
405 w NoL NoS No STD w NoR NoM Y
406 w NoL NoS No STD w NoR NoM Y
407 w NoL No S No STD w No R NoM Y
408 w NoL NoS No STD w NoR NoM Y
409 w NoL NoS No STD w NoR NoM Y
410 w NoL No S No STD w No NoM Y
411 w NoL NoS No STD w NoR NoM Y
412 w NoL NoS No STD w NoR NoM Y
413 w No L No S No STD w NoR NoM Y
414 w NoL NoS No STD w NoR NoM Y
415 w NoL NoS No STD w NoR NoM Y
416 w No L No S No STD w No R NoM Y
417 w NoL NoS No STD w NoR NoM Y
418 w NoL NoS No STD w NoR NoM Y
419 w NoL No S No STD No NoM Y
420 w NoL NoS No STD w NoR NoM Y
421 w NoL NoS No STD w NoR NoM Y
422 w No L No S No STD w No R NoM Y
423 w NoL NoS No STD w NoR NoM Y
424 w NoL NoS No STD w NoR NoM Y
425 w NoL No S No STD w No R NoM Y
426 w NoL NoS No STD w NoR No M Y Ag ree
Ru
n PA me
No. type L Prediction S Prediction STD Prediction W Prediction R Prediction M Prediction nt
427 W NoL NoS NoSTD W NoR No M Y
429 W NoL NoS No STD W NoR NoM Y
430 w NoL NoS NoSTD w NoR NoM Y
431 w NoL No S NoSTD w NoR NoM Y
432 w NoL NoS NoSTD w NoR NoM Y
433 w NoL NoS NoSTD w No R NoM Y
434 w NoL NoS NoSTD w NoR NoM Y
435 w NoL NoS NoSTD w NoR NoM Y
436 w NoL NoS NoSTD w No R NoM Y
437 w NoL NoS NoSTD w No M Y
341 w NoL NoS NoSTD NoW NoR NoM u
375 w NoL NoS STD NoW NoR NoM u
390 w NoL No S NoSTD NoW NoR No M u
428 w No L S NoSTD W NoR NoM P
389 w NoL S NoSTD NoW NoR NoM M
396 w NoL No S STD NoW No R No M M
Table 4. Algorithm performance summary table
Figure imgf000055_0001
_L= lying. S=Sitting, STD= Standing, W=Walking in different speeds 2-6 km/hour, R= Running in different speeds 4-10 km/hour. M= mixed activity running/walking (e.g. basket ball playing) Full agreement =The PA was identified correctly.
P= Partial agreement, identification of actual PA type L or S or STD as L, or S, or STD. Identification of actual PA type W or R or M as W, or R, or M.
Unknown- no PA type was identified
Miss classification - the PA was identified as significantly different then actual e.g. L, S, STD, instead of W, R, M.
Development of Estimation of walking and running speed
Logistics regression analysis was performed to identify the variable that associate with the walking or running speed and to build a model enabling to predict walking or running speed based on the variables.
Based on logistic regression analysis the equation for estimation of walking speed is :
Equation 3
Walking speed (km/hour) =2.09+14.66* V6 +11.39*V21-0.0403 *V7- 0.00708*V31-3.55*V33
Based on logistic regression analysis the equation for estimation of ruiming speed is:Equation 4
Running speed (km/hour)
4.21+1.17*V12+4.62*V18+2.49*V24+0.003253*V38
I. Example 2: Estimation of energy expenditure in free-living human using tri axial accelerometer worn as a ring finger.
Methods: A male subject aged 45 years, height 172 , weight 80 kg , wearing a tri-axial accelerometer mount on a ring as described in example 1, and a calorimeter based on pulse measurement by a chest strap detector ( CS-200 ,POLAR ) performed a set various activities, in an out side environment: standing, sitting walking in different speeds, ixinning in different speeds , as well as mixed run/walk activities . The tri axial acceleration data was transmitted to a laptop computer that was carried by the subject. Using MatLab software, for each 15 seconds segment, the variables described in table 2 were calculated and the PA type was identified based on the rules described in example 1. if walking or running activities were identified the walking or running speed was calculated based on equation 3 and 4 accordingly described in example 1.
The RMR of the subject was calculated according to equation 1 and found to be 0.574 kcal/15 seconds
The METs value used for estimated energy expenditure calculation of the different activity type are as described in the table 5:
Table 5- Met Values table
Figure imgf000057_0001
Un identified PA 1.2
Figure 4 describes the distribution of METs assigned by the method for the actual different physical activity (PA) types. There is a clear association between the actual PA types and assigned METs values.
The EEE for each time segment was calculated by multiplication of the RMR with the MET value of PA that was identified by the classification rules and summed up.
Figure 5 describes the accumulated EEE as measured during the subject's activity by tri-axial accelerometer and the EEE based on heart pulse, demonstrating the similarity between the methods.
II. Example 3: Detection of legs movement intensity during cycling via ring finger PPG sensor.
Two sessions were preformed in order to demonstrate how the PPG signal measured in the hand finger (by a ring sensor) is correlates to the legs movement while the hand were not moving. A human subject was riding on a stationary bicycle with fixed gear, the rounds per minutes (RPM) of the bicycle crank, and the pulse (measured by chest strap sensor, POPLAR) were recorded every 15 seconds during the session. PPG sensor was attached to the right hand ring finger and the sensor readout data was sampled at 200Hz and recorded on personal computer. For each 15 seconds time segment the parameters described in table 4 were calculated for the recorded PPG data.
Session A- The first session included a gradual increase of the intensity (RPM) over time in order to calibrate the system. Figure 6 describes RPM and subjects pulse increase over time. Correlation coefficients ( R - Pearson) between the RPM, Pulse, Pulse + RPM and calculated parameters are described in tables 6 below: Table 6. Session A, correlation between RPM, Pulse and PPG signal parameters.
Figure imgf000059_0001
High positive correlation (R=0.82, p<0.001) was fund between parameters related to the distribution range of the PPG signal, Inter-quartile range (IQR) and standard deviation (SD), and the bicycle crank RPM. IQR and SD correlates to the pulse as well since the pulse also increased gradually as RPM increases. Linear regression analysis was used to build an equation for estimation of RPM and pulse based on IQR, see Figs.7 and 8.
Session B- The second session include a various changes in the RPM both increase and decrease of RPM. Figure 8 describes crank RPM, subjects pulse changes over time as well as the IQR for the relevant time segments. It is clear that the IQR increases when RPM increases and vice versa. Correlation coefficients ( R - Pearson) between the RPM, Pulse, Pulse + RPM and calculated parameters described in tables 7 below: Table 7. Session B, correlation between RPM, Pulse and PPG signal parameters.
High positive correlation (R=0.80, pO.001) was fund between IQR and the bicycle crank RPM (Fig. 9 and table 7). However, in this session there was no correlation to the pulse (Fig. 10 and table 7) since the pulse increase only as a delayed reaction to the increase in the intensity of the legs movement.
Conclusions: parameters related to the distribution range of the PPG signal measured in hand finger, especially IQR, have high positive correlation to the intensity of the leg movement. Increase in the signal IQR indicate increase in the intensity of the leg movement and vice versa. Therefore PPG signal can be used to identify activity profiles where that the data provided by 3 axis accelerometer worn as a ring finger does not sufficient.
Reference is also made now to Fig. 4B, which shows a scatter chart of a distribution of metabolic equivalent of task (MET), assigned by the method for each of the 15 seconds time segments as in Fig. 4A, but excluding the time segments that were unidentified, in accordance with an embodiment of the present invention. In Fig. 5, there is shown a graph of accumulated EEE as measured during the subjects activity by a tri-axial accelerometer (EEE PA cumulative) and a cumulative calculated EEE, based on heart pulse (EEE POLAR cumulative), in accordance with an embodiment of the present invention.
Additionally, Fig. 6 shows a graph of a crank number of rotations per minute (RPM) corresponding increase in a subject's pulse over time, Pulse + RPM, as well as an inter-quartile range of a PPG signal over time, in accordance with an embodiment of the present invention.
Fig. 7 shows a scatter plot and regression line for prediction of RPM based on an inter quartiles range, in accordance with an embodiment of the present invention.
Fig. 8 shows a scatter plot and regression line for prediction of pulse- based on an inter quartiles range, in accordance with an embodiment of the present invention.
Fig. 9A is a graph of a crank number of rotations per minute (RPM) corresponding increase in a subject's pulse over time, Pulse + RPM, as well as an inter-quartile range of a PPG signal over time, in accordance with an embodiment of the present invention;
Fig. 9B shows a scatter plot for prediction of inter-quartile range based on an inter quartiles range crank number of rotations per minute (RPM), in accordance with an embodiment of the present invention.
Fig. 10 shows a scatter plot of a correlation between inter-quartile range IQR and pulse, in accordance with an embodiment of the present invention.
REFERENCES
Blair SN, Goodyear NN, Gibbons LW, and Cooper KH. Physical fitness and incidence of hypertension in healthy normotensive men and women. JAMA 252: 487- 490, 1984.
Blair SN, Kohl HW, Barlow CE, Paffenbarger RS Jr, Gibbons LW, and Macera CA. Changes in physical fitness and all-cause mortality. JAMA 273: 1093- 1098, 1995.
Chen KY, Sun M. Improving energy expenditure estimation by using a triaxial accelerometer. J Appl Physiol 83: 2112-2122, 1997.
Dilley JW; (1998, Aug): Self-reflection as a tool for behavior change. Focus, 13 (9): 5-6.
Donahoo WT, Levine JA, Melanson EL. Variability in energy expenditure and its components. Curr Opin Clin Nutr Metab Care 7: 599-605, 2004.
Grundy, S.M., Brewer, H.B., et al., Definition of metabolic syndrome: report of the National, Heart, Lung, and Blood Institute/American Heart Association conference on scientific issues related to definition. Circulation. 2004; 109, pp.433- 438. (2004).
Klem ML; (2000, Nov): Successful losers. The habits of individuals who have maintained long-term Weight loss. Minnesota Medicine, 83 (11): 43-45.
Pate RR, Pratt M, Blair SN, Haskell WL, Macera CA, Bouchard C, Buchner D, Ettinger W, Heath GW, King AC, Kriska AM, Leon AS, Marcus BH, Morris J, Paffenbarger RS Jr, Patrick K, Pollock ML, Rippe JM, Sallis JF, and Wilmore JH. Physical activity and public health A recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. JAMA 273: 402-407, 1995.
Schnool R; Zimmerman BJ; (2001, Sept): Self-regulation training enhances dietary self-efficacy and Dietary fiber consumption. Journal of the American Dietetic Association, 101 (9): 1006-1011. The references cited herein teach many principles that are applicable to the present invention. Therefore the full contents of these publications are incorporated by reference herein where appropriate for teachings of additional or alternative details, features and/or technical background.
It is to be understood that the invention is not limited in its application to the details set forth in the description contained herein or illustrated in the drawings. The invention is capable of other embodiments and of being practiced and carried out in various ways. Those skilled in the art will readily appreciate that various modifications and changes can be applied to the embodiments of the invention as hereinbefore described without departing from its scope, defined in and by the appended claims.

Claims

1) A system for energy expenditure determination of a mammalian subject, the system comprising;
a. at least one device comprising at least one sensor adapted to detect at least one signal from at least one bodily tissue at a first location of the mammalian subject over a period of time; and b. a processor adapted to apply at least one set of rules to at least one of:
i. said at least one signal; and
ii. a reduced data set associated with said at least one signal; thereby being configured to sum data of a plurality of identified activity profiles of said subject over a plurality of time segments in said period of time to provide predictive values of energy expenditure of a whole body of said mammalian subject over said period of time.
2) A system according to claim 1, wherein said at least one device is a ring device and said at least one bodily tissue is within a finger.
3) A system according to claim 1, wherein said at least one device is an earring and said at least one bodily tissue is within an ear.
4) A system according to claim 1, wherein said at least one device comprises at least one photoplethysmographic (PPG) sensor adapted to provide PPG data of said subject over said period of time.
5) A system according to claim 1, wherein said at least one device comprises at least one 3D accelerometer adapted to provide tri-axial acceleration data of said subject over said period of time.
6) A system according to claim 4 or 5, further comprising a remote communication device comprises a transceiver for bidirectional communication with said at least one device.
7) A system according to claim 6, wherein said at least one device comprises two devices.
8) A system according to claim 7, wherein said two devices comprise a ring device and an ear device.
9) A system according to claim 7, wherein said two devices each comprise a transceiver for communication with said remote communication device.
10) A system according to claim 9, wherein said remote communication device is a cellular phone.
11) A system according to claim 9, wherein said remote communication device is a computer.
12) A system according to claim 11, wherein said computer is a portable computer.
13) A system according to claim 1, further comprising a database of data of predefined metabolic equivalent of a task (MET)s.
14) A system according to claim 13, wherein said system is further adapted to develop said at least one set of rules from at least one of direct and indirect data from at least one of a PPG sensor and a 3D accelerometer.
15) A system according to claim 14, wherein said system further comprises data mining software for manipulating said at least one of direct and indirect data.
16) A system according to claim 15, wherein said at least one of direct and indirect data is generated from a plurality of mammalian subjects.
17) A system according to claim 1, wherein said first location is substantially static relative to a major limb of said mammalian subject.
18) A method for energy expenditure determination of a mammalian subject, the method comprising;
a. detecting at least one signal from at least one bodily tissue at a first location of the mammalian subject over a period of time; and b. applying at least one set of rules to at least one of: i. said at least one signal; and
ii. a reduced data set associated with said at least one signal; thereby providing predictive values of energy expenditure of said mammalian subject over said period of time.
19) A method according to claim 18, wherein said at least one bodily tissue is within a finger.
20) A method according to claim 18, wherein said at least one bodily tissue is within an ear.
21) A method according to claim 18, wherein said detecting step comprises receiving photoplethysmographic (PPG) data of said subject over said period of time.
22) A method according to claim 1, wherein said detecting step comprises receiving tri-axial acceleration data of said subject over said period of time.
23) A method according to claim 21 or 22, further comprising transmitting data from at least one device to a remote communication device.
24) A method according to claim 23, further comprising receiving both photoplethysmographic (PPG) and tri-axial acceleration data of said subject.
25) A method according to claim 24, wherein said at least one bodily tissue comprises at least one of finger tissue and ear tissue.
26) A method according to claim 18, wherein said energy expenditure is an estimation of energy expenditure (EEE).
27) A method according to claim 26, further comprising applying data from a database of predefined metabolic equivalent of a task (MET)s.
28) A method according to claim 18, further comprising generating said at least one set of rules from at least one of direct and indirect data from at least one of a PPG sensor and a 3D accelerometer.
29) A method according to claim 28, further comprising applying data mining software to manipulate said at least one of direct and indirect data.
30) A method according to claim 29, wherein said at least one of direct and indirect data is generated from a plurality of mammalian subjects.
31) A method according to claim 18, wherein said first location is substantially static relative to a major limb of said mammalian subject.
32) A method according to claim 27, wherein said EEE is based on said MET value relevant to at least one activity type and an RMR of said subject.
33) A method according to claim 18, further comprising identifying an activity type of said subject.
34) A method according to claim 33, further comprising detecting an increase in at least one of an amplitude and a frequency of a PPG signal thereby defining said activity type of said subject.
35) A method according to claim 18, further comprising detecting at least one signal from at least one bodily tissue at a second location of the mammalian subject over said period of time.
36) A system for identification of an activity type in a mammalian subject, the system comprising;
a. at least one device comprising at least one sensor adapted to detect at least one signal from at least one bodily tissue at a first location of the mammalian subject over a period of time; and b. a processor adapted to apply at least one set of rules to at least one of:
i. said at least one signal; and
ii. a reduced data set associated with said at least one signal; thereby being configured to identify different activity types responsive to at least one output of said processor.
37) A computer program product for energy expenditure determination of a mammalian subject, the product comprising a computer-readable medium having program instructions embodied therein, which instructions, when read by a computer, cause the computer to:
a) detect at least one signal from at least one bodily tissue at a first location of the mammalian subject over a period of time; and b) apply at least one set of rules to at least one of:
i. said at least one signal; and
ii. a reduced data set associated with said at least one signal; thereby providing predictive values of energy expenditure of said mammalian subject over said period of time.
38) A computer program product according to claim 37, wherein said product further uses the at least one set of rules to classify an activity profile of said subject over a plurality of time segments in said period of time.
39) A computer program product according to claim 38, wherein said product further combines data from said activity profile with a relevant MET value associated with said activity profile and further calculates a resting metabolic rate (RMR) of the subject thereby estimating said energy expenditure of said subject over said period of time.
40) A computer program product according to claim 39, wherein said product applies an equation:
RMR (kcal/time) x MET x time = EEE (kcal)
to calculate said EEE.
41) A method according to claim 18, further comprising identifying a type of movement of legs of said subject.
PCT/IL2012/000011 2011-01-09 2012-01-09 Device and method for continuous energy expenditure measurement WO2012093397A2 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US201161431018P 2011-01-09 2011-01-09
US61/431,018 2011-01-09

Publications (2)

Publication Number Publication Date
WO2012093397A2 true WO2012093397A2 (en) 2012-07-12
WO2012093397A3 WO2012093397A3 (en) 2015-06-18

Family

ID=46457776

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/IL2012/000011 WO2012093397A2 (en) 2011-01-09 2012-01-09 Device and method for continuous energy expenditure measurement

Country Status (1)

Country Link
WO (1) WO2012093397A2 (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016041073A1 (en) * 2014-09-17 2016-03-24 2352409 Ontario Inc. Device and method for monitoring fat balance
WO2016061668A1 (en) * 2014-10-23 2016-04-28 2352409 Ontario Inc. Device and method for identifying subject's activity profile
US9616291B2 (en) 2014-11-19 2017-04-11 Suunto Oy Wearable sports monitoring equipment with context determination capabilities and relating method
EP3057672A4 (en) * 2013-10-14 2017-10-25 NIKE Innovate C.V. Fitness training system for merging energy expenditure calculations from multiple devices

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0934021A2 (en) * 1996-10-24 1999-08-11 Massachusetts Institute Of Technology Patient monitoring finger ring sensor
US8452388B2 (en) * 2009-02-27 2013-05-28 Kimberly-Clark Worldwide, Inc. Apparatus and method for assessing vascular health
US8956293B2 (en) * 2009-05-20 2015-02-17 Sotera Wireless, Inc. Graphical ‘mapping system’ for continuously monitoring a patient's vital signs, motion, and location

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3057672A4 (en) * 2013-10-14 2017-10-25 NIKE Innovate C.V. Fitness training system for merging energy expenditure calculations from multiple devices
US10136840B2 (en) 2013-10-14 2018-11-27 Nike, Inc. Fitness training system for merging energy expenditure calculations from multiple devices
US11045114B2 (en) 2013-10-14 2021-06-29 Nike, Inc. Fitness training system for merging energy expenditure calculations from multiple devices
US11564597B2 (en) 2013-10-14 2023-01-31 Nike, Inc. Fitness training system for merging energy expenditure calculations from multiple devices
WO2016041073A1 (en) * 2014-09-17 2016-03-24 2352409 Ontario Inc. Device and method for monitoring fat balance
WO2016061668A1 (en) * 2014-10-23 2016-04-28 2352409 Ontario Inc. Device and method for identifying subject's activity profile
US9616291B2 (en) 2014-11-19 2017-04-11 Suunto Oy Wearable sports monitoring equipment with context determination capabilities and relating method

Also Published As

Publication number Publication date
WO2012093397A3 (en) 2015-06-18

Similar Documents

Publication Publication Date Title
US10327674B2 (en) Biometric monitoring device with immersion sensor and swim stroke detection and related methods
US10178973B2 (en) Wearable heart rate monitor
JP6464113B2 (en) Energy consumption
WO2016061668A1 (en) Device and method for identifying subject&#39;s activity profile
CN112546602B (en) Activity monitoring device for evaluating exercise intensity
US20190269353A1 (en) Gps accuracy refinement using external sensors
US9568492B2 (en) Fitness monitoring device with altimeter and gesture recognition
JP6502361B2 (en) System and method for estimating cardiovascular fitness of a person
US20170188864A1 (en) Spherical handheld biosensor array device for physiological status monitoring
JP2019000653A (en) Calculating pace and energy expenditure from athletic movement attributes
CN111629661A (en) Non-invasive hydration detection system and apparatus
WO2012093397A2 (en) Device and method for continuous energy expenditure measurement
De Pessemier et al. Health monitoring and user profiling for sports activities: evaluating heart rate measurements and activity recognition
JP2024018876A (en) Information processing system, server, information processing method, program, and learning model
Patancheru A Wearable Prototype of Reflective Sensor for Non Invasive Measurement of Heart Rate

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 12732445

Country of ref document: EP

Kind code of ref document: A2

NENP Non-entry into the national phase in:

Ref country code: DE

122 Ep: pct app. not ent. europ. phase

Ref document number: 12732445

Country of ref document: EP

Kind code of ref document: A2