Multi-modal measurement device and system
Through the joint analysis and machine learning model of EG, PG and IG sensors, the problems of complex operation of existing equipment and insufficient signal quality of wearable devices are solved, and efficient and simple body health assessment is achieved.
Patent Information
- Application Number
- CN202380083315.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-12-01
- Filing Date
- 2023-12-01
- Publication Date
- 2025-07-04
AI Technical Summary
Existing complex measurement equipment is complex in health assessments, has high resource requirements and is not suitable for long-term home monitoring, and wearable devices are difficult to provide high-quality biosignal measurements.
The surface potential difference sensor (EG), surface vibration sensor (PG) and impedance sensor (IG) were used to jointly analyze the EG signal, PG signal and IG signal to derive high-quality biological signal data and evaluate it in combination with machine learning models.
It achieves a high-quality body part health assessment while simplifying operations, suitable for both home and clinical settings, improving measurement accuracy and reliability.
Smart Images
Figure CN120265202A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of monitoring the activity of a body part of a subject. Background Art
[0002] Monitoring the activity of a body part is essential for assessing the health or well-being of a tested subject, such as a human, and for determining possible medical procedures to be administered to them. Such procedures can be beneficial and are prescribed to address or evaluate specific issues during a health / well-being assessment or after a medical procedure for evaluating the outcome of the procedure or the recovery process therefrom.
[0003] Sophisticated measurement equipment is used to achieve high-precision measurements and generally results in a higher confidence in the measurement endpoints. Such sophisticated measurement equipment requires a high operational workload, a professional / trained operator, and a large amount of resources and space to operate. Additionally, such sophisticated measurement equipment is not practical for long-term monitoring of patients, for example, in a home environment.
[0004] On the other hand, there are wearable devices, such as smartwatches, which measure certain bio-signal activities with greater convenience while compromising the quality of the measured signals and the modalities being measured.
[0005] Therefore, there is a need in the art for a measurement device that facilitates convenient operation while providing high-quality measurement and assessment of the health and / or condition of a body part of a monitored subject. Summary of the Invention
[0006] According to one aspect of the present disclosure, there is provided a device for measuring and / or monitoring bio-signals of a body part of a subject, the device including a surface potential difference sensor, electrogram EG, configured to measure the electrical activity of the body part and provide at least one EG signal indicative of the electrical activity of the body part, a surface vibration sensor, phonogram PG, configured to measure the acoustic activity of the body part and provide a PG signal indicative of the acoustic activity of the body part, and an impedance sensor, impedancegram IG, configured to measure the electrical impedance of the body part and provide an IG signal indicative of the electrical impedance of the body part.
[0007] Advantageously, a device including an EG sensor, a PG sensor, and an IG sensor enables the combined analysis of the EG signal, the PG signal, and the IG signal to derive at least one data based on all of the EG signal, the PG signal, and the IG signal.
[0008] As used herein, the term "subject" refers to a body to which measurements are applied, such as a human or an animal, such as a mammal, a reptile, or a bird. The terms "human", "subject", "user" (to whom the device is applied, such as a patient or a subject), "mammal", "body", "animal", "reptile", and "bird" are used interchangeably.
[0009] As used herein, the phrase "acoustic activity of a body part" refers to mechanical activity, vibrational activity, gas flow activity, fluid flow activity, etc., or any combination thereof.
[0010] Advantageously, a device including an EG sensor, a PG sensor, and an IG sensor enables the combined analysis of EG signals, PG signals, and IG signals to derive at least one predefined data based on all of the EG signals, PG signals, and IG signals.
[0011] As used herein, the term "datum" refers to information that is understandable by a human and / or a machine, which directly or indirectly indicates the existing condition or state or predicted condition of the monitored subject, its body part and / or its multiple body parts, and optionally the functional interaction between multiple body parts.
[0012] As used herein, the terms "datum", "data", "parameter", "feature", and "condition" respectively refer to the terms "predefined data", "predefined data", "predefined parameter", "predefined feature", and "predefined condition", and are interchangeable with the terms "predefined data", "predefined parameter", "predefined feature", and "predefined condition", such that the type of "datum", "data", "parameter", "feature", or "condition" that is derived or predicted is known or is part of a possible result within a known range. According to some aspects, the predefined nature of the "datum", "data", "parameter", "feature", or "condition" refers to an abnormality detected therein, in which case the abnormality attribute itself is predefined.
[0013] As used herein, the terms "combined" or "combining" with respect to the analysis of EG signals, PG signals, and IG signals mean performing the analysis such that the analysis result is based on all of the EG signals, PG signals, and IG signals, and excluding one of the EG signals, PG signals, and IG signals does not result in the same analysis result. For example, the combined analysis of EG signals, PG signals, and IG signals has a different result from the analysis of any one or two of the EG signals, PG signals, and IG signals in combination. The difference in the results of the combined analysis of all EG signals, PG signals, and IG signals can be a difference in confidence values, such as statistical significance, noise level, etc., and / or can be a difference in the actual values of the analysis results.
[0014] According to some embodiments, joint analysis can be achieved by inputting all of the EG signal, PG signal, and IG signal into a computational model for generating a result based on all of the EG signal, PG signal, and IG signal. Alternatively, according to some embodiments, one or more of the EG signal, PG signal, and IG signal are input into a first computational model, and the result thereof is inserted into another computational model that receives at least some of the other signals as input for deriving data based on all of the EG signal, PG signal, and IG signal. According to some embodiments, the other computational model receives the output from an earlier computational model as input.
[0015] According to some embodiments, the joint analysis is based on additional inputs. According to some embodiments, the additional inputs include information related to the subject, such as age, gender, height, weight, clinical status, clinical history, medications, comorbidities, genetic information, or other demographic information. According to some embodiments, the additional inputs include signals obtained from additional sensors other than the IG sensor, EG sensor, or PG sensor. According to some embodiments, the additional inputs are provided from an external device or sensor, and / or are provided by a user of the device, such as via an input device facilitated by the device.
[0016] As used herein, a computational model refers to a program that is executed / executable on a computer, configured to receive certain inputs, apply certain calculations or algorithms, and generate at least one output. The computational model can be a machine learning algorithm, a filtering operation, a derivation model, a convolutional model, etc.
[0017] According to some aspects, examples of data derived from the joint analysis of all of the EG signal, PG signal, and IG signal include data related to cardiac phase annotations, such as isovolumetric contraction, ventricular contraction, ventricular ejection, isovolumetric relaxation, ventricular filling, diastole, atrial contraction, etc. Advantageously, such data can be used to analyze cardiac function. According to some embodiments, cardiac function can be used to determine cardiac-related clinical conditions or pathologies, such as hypertrophy, dilation, reduced valve function, stenosis, sclerosis, insufficiency, etc.
[0018] According to some embodiments, the data can be used to estimate the heart rate, for example, in the range of 40 to 300 heartbeats per minute, and can be used for cardiac rhythm analysis, detection of arrhythmias, stress detection, and identification of cardiac abnormalities. According to some embodiments, the rhythm analysis indicates a rhythm that can be classified as normal, atrial fibrillation, supraventricular tachycardia, bradycardia, heart block, ventricular fibrillation, etc., which indicates the function of the valves and identifies cardiac abnormalities.
[0019] According to some embodiments, the data can be used for artificial valve function analysis, for example, by providing an indication of one or more of valve timing, closure, leaflet stiffness, thrombosis, etc., which may be advantageous for estimating valve function for early intervention and treatment.
[0020] According to some embodiments, the data can be used for stroke volume estimation, for example, in the range of 50 to 100 ml, which indicates the function of the valve and is used to identify cardiac abnormalities.
[0021] According to some aspects, examples of data derived from the joint analysis of all signals among the EG signal, PG signal, and IG signal include data related to respiratory phase annotations, such as inspiration, end-inspiration, expiration, end-expiration, etc. Advantageously, such data can be used as a basis for respiratory function analysis, a basis for compensating for respiratory modulation in other biological signals, and a basis for analyzing respiratory modulation in other biological signals. According to some embodiments, the derived data of the device is modulated based on respiratory activity.
[0022] According to some embodiments, the data can be used for respiratory rate estimation, for example, a respiratory rate in the range of 4 to 40 breaths per minute, which can be used for rhythm analysis of the lungs, identifying respiratory abnormalities, etc.
[0023] As used herein, the term "sensor" refers to a module configured to sense or measure a specified property, particularly surface potential difference, surface vibration, and impedance, where the sensing activity is performed directly by the sensor, such as when the sensor includes an interface with the user's body, or indirectly, such as when the sensor is configured to be connected to an interface medium (such as an electrode) placed on the user's body.
[0024] According to some embodiments, the interface medium includes an optical sensor, such as a light electrode, configured to measure changes in electromagnetic waves emitted, reflected, and / or absorbed at certain wavelengths.
[0025] According to some embodiments, one or more of the sensors in the device include an interface medium configured to be placed on the user's body, such as on a certain part of the user's skin.
[0026] According to some embodiments, one or more of the sensors include a connector configured to enable electrical, capacitive coupling, magnetic, and / or mechanical connection with the interface medium. Advantageously, such a configuration facilitates the use of replaceable interface media, such as replaceable electrodes for improving hygiene and / or measurement quality.
[0027] As used herein, the term "EG sensor" refers to a sensor configured to measure the potential difference between two or more electrodes placed on a user's body. The EG sensor may include sensors that function similarly to an electromyogram (EMG) sensor or an electrocardiogram (EEG) sensor.
[0028] According to some aspects, the EG sensor includes or is configured to be connected to 2 electrodes, 3 electrodes, 4 electrodes, up to 10 electrodes or more. According to some embodiments, the EG sensor includes or is configured to be connected to an electrode array, which can advantageously facilitate vector analysis and / or surface potential mapping.
[0029] According to some aspects, the IG sensor includes or is configured to be connected to 2 electrodes, 3 electrodes, 4 electrodes, up to 10 electrodes or more. According to some embodiments, the IG sensor includes or is configured to be connected to an electrode array, which can advantageously facilitate vector analysis and / or surface potential mapping.
[0030] According to some embodiments, the IG sensor and the EG sensor are connected to a shared electrode. According to some embodiments, the shared EG / IG electrode is used to acquire EG and IG signals. According to some embodiments, the signals obtained from the shared EG / IG electrode are frequency multiplexed, i.e., the shared electrode acquires an EG signal at a specified EG frequency / frequency range and an IG signal at a specified IG frequency / frequency range. According to some embodiments, the specified EG frequency or frequency range has a lower frequency than the specified IG frequency.
[0031] According to some embodiments, the specified EG frequency or frequency range is from 0 Hz to 1 kHz, while the specified IG frequency or frequency range is higher than 1 kHz, higher than 10 kHz, higher than 20 kHz or higher than 50 kHz, thus facilitating <aq< frequency multiplexing.
[0032] According to some embodiments, at least temporarily, there is an overlap between the specified EG frequency range and the specified IG frequency range, and a time-division multiplexing scheme is used to acquire both the EG signal and the IG signal using the shared electrode, thus requiring sequential signal acquisition. For example, the specified EG frequency range is from 0 Hz to 1 kHz, and the specified IG frequency range is from 0 kHz to 100 kHz or 200 kHz. According to some embodiments, time-domain multiplexing is advantageous for facilitating bioimpedance spectroscopy (BIS).
[0033] According to some embodiments, all electrodes connected to or integrated with the IG sensor and the EG sensor are shared EG / IG electrodes. According to some embodiments, at least some of the electrodes connected to or integrated with the IG sensor and the EG sensor are shared EG / IG electrodes, while one or more of the electrodes connected to or integrated with the IG sensor or the EG sensor are dedicated electrodes.
[0034] As used herein, the term "IG sensor" refers to a sensor configured to measure the conductivity property between two or more electrodes placed on a user's body, such as by providing a reference signal at one or more predetermined frequencies and for evaluating the impedance between two or more electrodes. According to some aspects, the electrodes for obtaining the EG signal or the IG signal are made of one or more of metal, carbon, cloth, foam, tag, and / or wet / gel electrodes. According to some embodiments, the type of electrode for obtaining the EG signal is determined based on the measured condition. For example, a cloth electrode can be used for long-term wear, allowing normal skin breathing and stretching with movement, and a foam electrode can be suitable for short-term use and is beneficial for providing high adhesion.
[0035] According to some embodiments, one or more electrodes in the EG or IG sensor can be used as a reference electrode.
[0036] As used herein, the term "PG sensor" refers to a sensor configured to measure vibrations / stethoscopies, such as acoustic and / or mechanical vibrations caused or induced by the activity of one or more body parts of a user or by other factors, other factors such as body fluids (including blood, urine, gastric acid, saliva, etc.), non-body fluids (including swallowed water, etc.), non-body solids (including swallowed food, etc.), body gases (including digestive by-products, etc.), non-body gases (including air in the lungs, etc.), artificial components (such as implanted components, including prosthetic heart valves, joints, catheters, etc.).
[0037] According to some embodiments, the PG sensor includes or is connected to an interface configured to obtain and optionally amplify mechanical / acoustic vibrations / stethoscopies (such as a membrane) and convert the obtained vibrations / stethoscopies into an electrical PG signal. According to some embodiments, the PG sensor is configured to convert the obtained vibrations / stethoscopies into an electrical PG signal using an accelerometer, a gyroscope, a magnetometer, a piezoresistive sensor, a capacitor sensor using a pressure capacitor (e.g., a parallel plate capacitor configured to convert sound pressure into displacement and serve as the moving plate of the capacitor), a piezoelectric sensor, a solid-state acoustic detector, a microelectromechanical system (MEMS), etc.
[0038] According to some embodiments, the body part is an implanted artificial part, and the PG sensor is configured to measure an acoustic signal indicative of the state or condition of the body part. According to some embodiments, the body part is an implanted artificial part, and the EG sensor and / or the IG sensor are configured to measure an EG signal and / or an IG signal indicative of the activity or condition of the implanted artificial part.
[0039] The position of the device, in particular the position of the electrode / interface, is selected based on the body part being measured. According to some embodiments, the electrode / interface is configured to obtain a specified measurement non-invasively, i.e., without penetrating the user's dermal tissue. According to some embodiments, the electrode / interface is minimally invasive to the user's dermal tissue. For example, the electrode / interface may include microneedles or a microneedle grid.
[0040] According to some aspects, the device is placed on the anterior chest wall at the left side of the sternum at the 2nd, 3rd, and 4th ICSs, or at the midclavicular line on the left at the 4th ICS, or at the right side of the sternum at the 2nd ICS. Such placement is useful for facilitating cardiac analysis.
[0041] According to some aspects, the device is placed on the anterior chest wall or the posterior chest wall on either side of the sternum, possibly between the clavicular rib and the last rib. Such placement is useful for facilitating pulmonary analysis.
[0042] According to some aspects, the device is placed near the joint under examination such that there is a proper mechanical coupling mainly with the PG sensor and for measuring impedance for fluid and tissue analysis. In such use, the strip is useful for holding the device in the desired position.
[0043] According to some aspects, the device is placed near the right lumbar region or the left lumbar region, the umbilical region, the right iliac region or the left iliac region, or the lower abdominal region to facilitate digestive analysis.
[0044] To hold the device in the desired position, self-adhesive electrodes can be advantageously used. In addition to or instead of the adhesion of the electrodes, bands, tethers, suction cups (with or without active vacuum) can be used to position the device in place, it can be held in position by hand, or it can be fabric-integrated. The device can include adhesion members around the electrodes and / or in the space between the electrodes, or the device can be covered by a plaster or other adhesive such that the device is at least partially or completely covered by the plaster, by applying a separate glue between the device and the skin, by using an additional double-sided adhesive as an interface, by placing an interface member on the skin that stays on the skin while the device can be attached to and removed from the interface, etc.
[0045] As used herein, the term "indication" refers to the quality of information derivable from a signal affected by the activity or state of a body part being monitored. In particular, a signal can indicate the activity of multiple body parts because the monitored / target body part resides near other body parts and may be affected by their activity. Additionally, even when the activity of the measured body part is not directly affected by the activity of other body parts, the activity of other body parts may affect the measured signal. For example, the electrical activity associated with the movement of skeletal muscle typically affects the measurement of the potential difference measured by an EG electrode configured to measure the electrical activity of the heart.
[0046] According to some embodiments, the device is a handheld device configured to be placed briefly at a specific location for rapid assessment or prediction of a condition.
[0047] According to some embodiments, the device includes a socket for mechanically connecting to an external electronic device such as a mobile device.
[0048] According to some embodiments, the device is wearable on a user's body. According to some embodiments, the device includes or is configured to be attached to an adhesive medium for attaching the device to the user's body. According to some embodiments, attachment of the device to the user's body is achieved through the adhesive properties of one or more interfaces / electrodes connected to or integrated with the device.
[0049] According to some embodiments, an attachment mechanism is used to attach / wear the device on the user's body, such as a strip that holds the device in a desired location / position by applying a mechanical force in the direction of the user's body.
[0050] According to some embodiments, the device is integrated into or connectable to a fabric wearable by the user.
[0051] According to some embodiments, the body part is a biological body part. According to some aspects, the body part includes one or more of the heart, lungs, arteries, veins, and / or valves.
[0052] As used herein, a body part refers to tissues in a user's body, combinations of tissues, implanted tissues, implanted devices, implanted artificial organs, or combinations thereof. Generally, a body part affects certain physiological functions such as cardiac function, respiratory function, neurological function, digestive function, vascular function, structural function, etc.
[0053] According to some aspects, the device further includes a communication unit configured to obtain EG signals from an EG sensor, PG signals from a PG sensor, and IG signals from an IG sensor, and transmit the signals or processed information derived from one or more of the signals to an external communication unit.
[0054] According to some aspects, one or more of the signals sent to the external communication unit are processed at the device before being sent. The processing of the signals can include one or more of analog-to-digital conversion, digital-to-analog conversion, noise reduction, high-frequency filtering, low-frequency filtering, band filtering, modulation, adjusting the amplitude / gain of the signal, encryption, compression, dimensionality reduction, adaptive filtering, machine learning model inference, etc.
[0055] According to some embodiments, the device includes a signal processing unit implemented in hardware, software, or a combination of hardware and software, configured to process one or more of the EG signal, PG signal, or IG signal.
[0056] According to some aspects, the device includes a memory configured to store the acquired EG signal, PG signal, and IG signal. According to some aspects, the memory is a non-transitory memory. According to some embodiments, the memory is configured to store the signals at least until they communicate with an external device wirelessly or via a wired connection. According to some embodiments, the memory is configured to store one or more signals in an unprocessed state (such as storing the original samples of the signal) or a processed state (such as filtered, modulated signal, machine learning model inference, etc.).
[0057] According to some aspects, the memory is configured to store the output of the joint analysis or the results of intermediate steps thereof. According to some embodiments, the memory is a digital memory configured to store digital signals. According to some embodiments, the memory includes analog memory elements configured to store analog signals.
[0058] As used herein, unless otherwise stated or implied, the term signal refers to a digital signal of an analog signal or a digital signal of an analog signal in the time domain, frequency domain, or other domain (such as the z-domain, s-domain, etc.), whether in its original format or a transformed signal. According to some embodiments, the signal is transformed from the time domain to the frequency domain by, for example, performing a Fourier transform on the signal using a fast Fourier transform function. According to some embodiments, the signal undergoes a transformation using one or more of the following transformation models / techniques: Fourier, cosine, wavelet, Hilbert, Laplace, Walsh-Hadamard, z, cepstrum, and / or Haar.
[0059] According to some embodiments, the device is configured to perform a joint analysis. For example, the device can be an independent device configured to perform the analysis independently of an external computing device.
[0060] According to some embodiments, the device transmits the acquired EG signal, PG signal, and IG signal, or a preprocessed version thereof, to an external computing device, such as a server, a mobile device, a computer, a smartwatch, another wearable electronic device, a base station, a relay device, etc., where the external computing device is configured to obtain the EG signal, PG signal, and IG signal and perform a joint analysis, or relay the signal and / or the processed information related to or based on the signal to an additional external computing device for further processing, analysis, or display. In such a configuration, the device essentially performs the function of signal acquisition, while the external device performs all or most of the joint analysis. According to some embodiments, the preprocessed version of the signal or signals includes intermediate parameters extracted / derived from one or more signals.
[0061] According to some embodiments, preprocessing the signal includes applying normalization to the signal or the derivative of the signal, e.g., general normalization such that the signal is within a certain range of values, or normalization for preparing the signal as an input to be fed into a machine learning model.
[0062] According to some embodiments, normalization is performed within a certain segment or stage of the signal. According to some embodiments, the normalization of one signal is performed independently of other signals. According to some embodiments, the normalization of one signal is performed based on the value of one or more other signals or in relation to the value of one or more other signals. According to some embodiments, normalization is applied to features extracted from one or more signals.
[0063] According to some embodiments, preprocessing the signal includes applying time-domain adjustment, e.g., by applying dynamic time warping, stretching the signal in time, compressing the signal in time, and / or shifting the signal in time.
[0064] According to some embodiments, the relay device is a proprietary relay device configured to facilitate secure communication with the device, e.g., by using a specific communication protocol, encrypted communication, etc.
[0065] According to some embodiments, the device is configured to perform preliminary processing / analysis of the acquired signals and transmit the intermediate analysis results to an external device for joint analysis.
[0066] According to some embodiments, the device can operate in a mobile setting, such as in an ambulance, on a ship, on an airplane, etc., or in a remote setting, such as in a user's home.
[0067] According to some embodiments, the device can be operated in a clinical environment, such as for preoperative, perioperative, postoperative, or ICU settings, in a mobile setting, such as by a general practitioner, in a care setting, such as in an assisted living facility or an elderly facility, in an emergency setting, such as in a public place, in a sports environment, such as for sports medicine, in a support and first response environment, such as operated by firefighters, military, police, etc., in an affiliated doctor's office, such as in school nursing, disaster relief, rural health, developing countries, correctional facilities, etc.
[0068] As used herein, the term "OP" refers to a "surgery" or another therapeutic procedure administered to a user, such as a drug treatment, an invasive surgery, or a non-invasive intervention, such as an intervention using ultrasound, magnetic fields, and / or electromagnetic waves or another medical intervention.
[0069] In mobile and remote settings, it is advantageous that an important part of the joint analysis is performed by an external device, such as a server or computer at a caregiver's clinic or at a hospital.
[0070] According to some aspects, the device and the external device are connected via a wire. According to some aspects, the device and the external device are wirelessly connected. According to some aspects, the device and the external device are connected via a short-range communication connection, such as Bluetooth, Zigbee, wireless LAN, etc. According to some aspects, the device and the external device are connected through a computer network. According to some aspects, the communication between the device and the external device is carried via a telecommunications infrastructure. According to some embodiments, at least some of the joint analysis is performed using a cloud computing system.
[0071] According to some embodiments, the device is configured to be used in a clinical environment, such as in a doctor's clinic or in a hospital. In such a case, the device can be connected to an external device either via a wire or wirelessly. According to some embodiments, the device is also configured to be powered by an external device, i.e., the device is configured to receive power for performing its operations from an external device, for example, through a wired connection with the external device.
[0072] According to some embodiments, the device is configured to be battery-powered. According to some embodiments, the battery is a rechargeable battery. According to some embodiments, the battery is a replaceable battery. According to some embodiments, the device is powered by wireless power transfer. According to some embodiments, the device is powered using an energy harvesting mechanism, such as a mechanical energy harvesting mechanism, an electromagnetic energy harvesting mechanism, a thermal energy harvesting mechanism, etc., where the harvested energy is obtained from the subject's body and / or the device's environment. According to some embodiments, the energy harvesting mechanism is based on radio waves, body movement, body temperature, ambient light, radio waves, body fluids, pressure changes, environmental radiation, fluid or gas flow, etc.
[0073] According to some embodiments, the external device includes one or more of a telephone, a computer, a tablet, a smart watch, a server, a cloud computing system, etc.
[0074] According to some aspects, the wearable device further includes a processor configured to obtain an EG signal from an EG sensor, a PG signal from a PG sensor, an IG signal from an IG sensor, and jointly analyze the EG signal, the PG signal, and the IG signal to allow evaluation of a predefined condition of a body part or a subject by deriving at least one data based on all of the EG signal, the PG signal, and the IG signal.
[0075] According to some aspects, deriving at least one data based on all of the EG signal, the PG signal, and the IG signal includes detecting a feature preferably including at least one of a distance between two predefined points of interest, a rate of occurrence of an event, or a value at a point of interest.
[0076] According to some aspects, deriving at least one data based on all of the EG signal, the PG signal, and the IG signal includes detecting a predefined feature preferably including at least one of a distance between two predefined points of interest, a rate of occurrence of an event, or a value at a point of interest.
[0077] As used herein, the "feature" of the term "predefined feature" refers to an attribute or combination of attributes derived individually from one or more signals by considering the signal(s) without considering other signals, or an attribute or combination of attributes derived from a combination of two or more signals.
[0078] Predefined features include, for example, attributes of a signal related to its intensity at certain points, positions of local or global maxima / minima, differences between values of two or more points, rate of change, statistical attributes, etc., or one or more associated signals derived therefrom, such as time derivatives of first-order, second-order, or higher-order signals. Such features can be obtained from the signal or the associated signals derived therefrom in the time domain or the frequency domain.
[0079] According to some embodiments, predefined features are derived using transforms or transform coefficients of one or more of the signals, statistical moments, extrema, percentiles, model fitting and coefficients of model fitting (such as polynomials, exponentials, geometries, distributions, series, rationals, splines, autoregressions, etc.), signal morphology (such as rise and / or fall times), area, segmentation between and within signals, area ratios between the rising and falling parts of a signal (e.g., threshold crossings, magnitudes of feature points, their ratios and norms, etc.).
[0080] According to some embodiments, analysis and models based on complexity / entropy features such as fractal dimension, approximate entropy, sample entropy, conditional entropy, fuzziness, etc. are used to derive predefined features.
[0081] According to some embodiments, analysis and models based on time series analysis are used to derive predefined features such as long-range correlation, correlation, and templates.
[0082] According to some embodiments, analysis and models based on one or more other features are used to derive predefined features.
[0083] According to some aspects, a convolutional model is used to derive certain features from a signal or from a combination of signals.
[0084] Joint analysis of EG signals, PG signals, and IG signals is performed and is configured to allow the assessment of a predefined condition of a body part or a subject. The phrase "allow the assessment" relates to the nature that the extracted / derived data / data is associated with, correlated with, and / or indicative of a predefined condition.
[0085] As used herein, the term "predefined condition" refers to the current / existing / principal state or condition of a body part such as the current state or condition of the heart, or refers to the expected / predicted state of a body part at a future point, after a certain time period, and / or after a certain scheduled procedure such as a diagnostic or therapeutic procedure.
[0086] According to some aspects, the predefined condition is the predicted response or state of a body part in the case of administering a therapeutic procedure to a user such as a drug treatment, an invasive operation, or a non-invasive intervention such as an intervention using ultrasound, magnetic fields, and / or electromagnetic waves. According to some embodiments, the term "therapeutic procedure" may refer to one or more of physical therapy, psychotherapy, or other forms of treatment.
[0087] According to some aspects, the existing or suspected predefined condition relates to the clinical / pathological condition of a user or a body part.
[0088] According to some aspects, the predefined condition may relate to both an existing condition and a predicted condition. For example, the predefined condition may relate to the detection of a pathological condition of a body part and its predicted progression. For example, the predefined condition may be related to the detection of a pathological condition of a body part and to the predicted response of the body part or the user to one or more interventions or to the avoidance / absence of one or more interventions.
[0089] According to some aspects, the predefined condition is the assessment and / or prediction of the general state or health of a user.
[0090] According to some embodiments, the predicted predefined conditions include an estimation of pre- or post-operative risks, including, for example, peri-operative and post-operative morbidity and mortality, shock, bleeding, wound infection, deep vein thrombosis, pulmonary embolism, pulmonary embolism, pulmonary complications, urinary retention, heart failure, myocardial infarction, etc. The conditions can be used to evaluate the patient's pre-operative state to determine whether the OP risk is acceptable, or if pre-operative rehabilitation is needed, what pre-, peri- and post-operative paths are suitable for the patient, etc. According to some embodiments, pre-operative use can involve the prediction of post-operative complications in the context of pre-operative risk assessment, for example, the probability of 30-day mortality or morbidity.
[0091] According to some embodiments, the pre-operative application of the device on a patient / subject involves the steps of: acquiring IG, PG, and EG sensor data / signals, extracting features therefrom, inputting the extracted features into a pre-learned or trained model, and receiving a risk score or probability of post-operative general adverse events, optionally, with a confidence interval, or a probability of a specific post-operative adverse outcome with a confidence interval.
[0092] According to some embodiments, the device is used to indicate heart conditions, such as stenosis, infarction, valvular insufficiency, etc., wherein the device is applied to a patient to acquire IG, PG, EG sensor data, extract features, input the extracted features into a pre-learned / trained model, and receive a yes / no answer or probability, preferably with a confidence interval, and / or a probability of a specific lesion present, preferably with a confidence interval, if the heart is healthy.
[0093] According to some embodiments, the device is used to indicate lung / respiratory conditions, such as COPD, acute infection, etc., wherein the device is applied to a patient to acquire IG, PG, and EG sensor data, extract features, input the features into a pre-learned / trained model, and receive a yes / no answer or probability with a confidence interval and / or a stage estimation and functional status of the lung.
[0094] According to some embodiments, the device is used to estimate the functional state / condition of an artificial heart valve, such as proper closure or leakage, by applying the device to a patient / user, acquiring IG, PG, and EG sensor data, extracting features, inputting the features into a pre-learned / trained model, and receiving an estimation of the heart valve state, including leaflet flexibility, leakage, closure characteristics, thrombosis, etc.
[0095] According to some embodiments, the device is used to indicate or predict the failure of an artificial heart valve, such as sclerosis of leaflets, thromboembolic events, etc., by repeatedly applying the device to a patient / user, acquiring IG, PG, and EG sensor data, extracting features, inputting the features into a pre-learned / trained model, and obtaining a wear estimate and using known early changes in the data to project the future wear, possible failure, or trajectory of functional impairment.
[0096] According to some embodiments, the device is used to estimate the post-intervention trajectory of a user from whom IG, PG, and EG sensor data are acquired, extract features, and input the features into a pre-learned model, possibly together with additional patient information such as age, gender, height, weight, medications, known comorbidities, etc., to estimate the possible outcome of the intervention before its application.
[0097] According to some embodiments, the device is used to generate personalized treatment, e.g., after a cardiac event, during COPD monitoring, etc., by applying the device one or periodically after an adverse event, after the diagnosis of a chronic disease, or after an intervention, to track disease or treatment success based on the data acquired by the system or to tailor the treatment specifically to the needs of the patient / user.
[0098] According to some embodiments, the processor is configured to detect at least one point of interest or segment in one signal based on another detected point of interest or segment in another signal. According to some embodiments, the processor is configured to detect a segment in one signal and use the detected segment to indicate an area of interest in another signal to detect a point of interest therein.
[0099] According to some embodiments, a first point of interest or first segment is detected in a first signal, based on which a second point of interest or second segment is detected in a second signal, and based on which a third point of interest or third segment is detected in a third signal.
[0100] According to some embodiments, the point of interest in the EG signal is the R peak. This point can be detected by identifying the point with the maximum absolute amplitude in the segment between prominent absolute signal derivative values.
[0101] According to some embodiments, the point of concern in the PG signal is the onset of S1. It can be detected by identifying the start of a segment with high signal energy in the frequency range of the S1 heart sound.
[0102] According to some embodiments, the segment of interest in the PG signal is the S3 sound segment, which is caused by a low-frequency transient vibration occurring in the early diastole at the end of the rapid diastolic filling phase of the right or left ventricle. According to some embodiments, the S3 segment is detected by identifying the start and offset of a segment with high signal energy in the frequency range associated with the S3 heart sound.
[0103] According to some embodiments, the interesting segment in the PG signal is the S4 sound segment, which, if present, is caused by atrial contraction. According to some embodiments, the S4 sound segment is detected by identifying the start and end of a segment having high signal energy in the frequency range associated with the S4 heart sound.
[0104] According to some embodiments, the interesting segment is the cardiac cycle in the EG signal. This segment can be detected by identifying the R peaks of two consecutive cardiac cycles.
[0105] According to some embodiments, the processor is configured to detect two or more points of interest in the same signal. For example, the processor is configured to detect two or more maxima and / or minima in the signal.
[0106] As used herein, unless otherwise specified, the term "point of interest" refers to a "predefined point of interest" such that it refers to a point of specific known clinical interest and / or a point of specific statistical or computational interest.
[0107] As used herein, unless otherwise specified, the terms "segment", "section", and "phase" are used interchangeably and refer to a specific time range, frequency range, or another one-dimensional or multi-dimensional section in a signal or its transform, and these terms can be interchanged with the terms "predefined segment", "predefined section", and "predefined phase", respectively.
[0108] The term "predefined" refers to a segment, section, phase, point, condition, and / or feature that has clinical and / or computational significance, or that has a known meaning (such as a medically defined point in a signal related to a known activity or property of a body part), or that has an unknown meaning (such as a feature extracted / derived internally by a computational model in the case of being presented or not presented to a user or operator, such as a feature extracted / derived by a convolutional layer in a machine learning model).
[0109] According to some aspects, a "predefined" attribute refers to a classification of one or more of a plurality of known categories. According to some aspects, the predefined attribute includes one or more predicted trajectories of one or more in a signal within a time after measurement, such as a projected corridor or a potential predicted value.
[0110] The term "predefined" also includes that the attributes of a segment, section, phase, point, condition, and / or feature are abnormal and do not conform to expected values or attributes. In this case, the abnormality itself is considered predefined.
[0111] According to some aspects, joint analysis includes performing time synchronization between different signals. According to some aspects, the time synchronization is sub-millisecond synchronization. According to some aspects, time synchronization of the signals is performed by timestamping each of the signals with a signal at a predefined frequency and aligning the timings of the different signals using an interpolation scheme. According to some embodiments, the predefined frequency signal has a frequency higher than 1 kHz. According to some embodiments, synchronization is performed by synchronizing sampling events obtained from two or more signals or between two or more signals.
[0112] According to some aspects, joint analysis includes performing point and / or segment detection based on one or more of an EG signal, a PG signal, or an IG signal to detect points of interest and / or segments, and analyzing other signals based on the detected points of interest or segments.
[0113] According to some aspects, performing segment detection includes segmenting a signal into predefined sections or predefined phases of predefined biological events, such as respiratory cycles, cardiac cycles, and / or circadian cycles, or another biological event that is voluntary or involuntary.
[0114] According to some aspects, analyzing other signals based on the segmentation includes identifying characteristic attributes in one or more of the other signals based on the detected points and / or segments.
[0115] As used herein, the term "biological event" refers to an activity in a user's body that is voluntarily or involuntarily affected by the body part being monitored or by other body parts.
[0116] In some aspects, the biological event is a repetitive cyclic / rhythmic event, such as a respiratory or cardiac cycle or a part thereof. The term cycle or rhythm refers to the repetitive or patterned nature of the event and / or its sequential attributes, e.g., the expected sequence of phases in a cardiac cycle, the expected sequence of a respiratory cycle, the expected sequence of a circadian cycle, etc., which typically form a pattern associated with the activity. Such repetitive events can be involuntary, such as a cardiac cycle or an involuntary respiratory cycle, or at least partially involuntary, such as a guided respiratory activity.
[0117] In some aspects, the biological event is an aperiodic event, such as skeletal muscle movement, coughing, swallowing, blinking, digestion-related activities, etc.
[0118] According to some embodiments, one model or multiple models are employed to detect points of interest or segments, such as a point detection model or a segmentation model. According to some embodiments, one or more of the models correspond to machine learning models.
[0119] According to some embodiments, a machine learning model is used to perform the derivation of at least one data based on all of the EG signal, PG signal, and IG signal in a joint analysis. According to some embodiments, the machine learning model is a supervised machine learning model trained based on a labeled input and output training dataset. According to some embodiments, the machine learning model is an unsupervised machine learning model trained based on unlabeled or raw data.
[0120] According to some embodiments, for example, based on signals and / or data obtained from a subject, a machine learning model is trained on a device.
[0121] According to some embodiments, in a joint analysis, at least one data is derived based on all of the EG signal, PG signal, and IG signal by inputting all of the EG signal, PG signal, and IG signal into a single model for data derivation.
[0122] According to some embodiments, in a joint analysis, at least one data is derived based on all of the EG signal, PG signal, and IG signal by inputting the EG signal, PG signal, and IG signal into multiple models for data derivation, wherein the output of at least one model is inserted as an input into another model, thereby forming a multi-stage derivation. According to some embodiments, in the multi-stage derivation, information (such as segments, features, and / or points) in one signal of one model is used to derive based on the derived information (such as segments, features, and / or points) in another signal of another model.
[0123] According to some embodiments, segments detected in one signal are used to detect points of interest in another signal. According to some embodiments, points of interest detected in one signal are used to detect segments in another signal. According to some embodiments, points of interest detected in one signal are used to detect another point of interest in another signal. According to some embodiments, segments detected in one signal are used to detect another segment in another signal.
[0124] According to some embodiments related to cardiac monitoring, an R is detected in the EG signal, based on which an S1 is detected in the PG signal, and based on which a B is detected in the IG signal. The parameter "R" in the EG signal refers to the R wave in the QRS complex and represents the electrical stimulation when it passes through the main part of the ventricular wall. The parameter "S1" in the PG signal refers to the first heart sound amplitude / energy, and the parameter "B" in the IG signal represents the opening of the aortic valve.
[0125] According to some embodiments, the points of interest in the EG signal include one or more of the Q point, R point, S point, J point, T point, and / or U point.
[0126] According to some embodiments, a segment or region of interest in an EG signal includes one or more of a PR segment, a P wave segment, a QRS segment, an RR segment, an ST segment, a TO segment, a TP segment, a QT segment, or any segment between two points of interest in the EG signal.
[0127] According to some embodiments, a point of interest in an IG signal includes one or more of a point A, a point B, a point C, a point X, a point Y, and / or a point O.
[0128] According to some embodiments, a segment or region of interest in an IG signal includes one or more segments of any segment defined by two points of interest in the IG signal.
[0129] According to some embodiments, the EG signal is used to detect an R peak or a sharp R peak point, and the R peak is used as a landmark. The S1 point in the PG can also be detected in a noisy environment. In an exemplary implementation, a method based on probability and distribution is used to detect the S1 point, which uses the R peak as a reference point in time. Additionally, the T wave extracted from the EG signal can also be used to define a time period or boundary, further enhancing the robustness of S1 detection by providing a rationality window. According to some embodiments, once the R and S1 are identified, the B point in the cardiac IG curve is identified based on the R and S1, for example, using a distribution method based on the timing between the R peak, the S1 peak, and the B point, advantageously enabling the detection of the B point even when the IG signal is weak or its signal-to-noise ratio is low.
[0130] According to some embodiments, the R peak point is detected using the EG signal, and then the B point and / or the C point are detected from the IG signal. Exemplarily, the PG signal is used to determine a region of interest, enhance the measurement accuracy of other signals, and / or determine the quality of the EG and / or IG signals.
[0131] According to some embodiments, in the joint analysis, points P, Q, R, S, and / or T are identified / detected from the EG signal, points S1 and / or S2 are detected from the PG signal, and points A, B, C, V, X, Y, and / or O are detected from the IG signal. According to some embodiments, the cardiac cycle is at least segmented into the following two phases: ventricular contraction (systolic phase) and ventricular relaxation (diastolic phase). According to some embodiments, the segment of isovolumic contraction (also referred to as pre-ejection period) is identified between the Q point of the EG signal and the B point of the IG signal. According to some embodiments, the segment starting from the B point of the IG signal and ending at the start of S2 that marks ventricular ejection (also referred to as left ventricular ejection time) in the PG signal is identified. According to some embodiments, the segment starting at the Q point of the EG signal and ending at the X point of the IG that depicts the total systolic time is identified. According to some embodiments, the ventricular isovolumic relaxation segment is identified, which starts at the start of S2 in the PG signal (aortic valve closure) and ends at the O point of the IG signal (mitral valve opening). According to some embodiments, the segment delimiting atrial contraction is identified between the P point of the EG signal (atrial depolarization) and the start of S2 in the PG signal.
[0132] According to some embodiments, points of interest (such as peaks, valleys, and zero-crossing points in the respiratory signal) are used to identify phases of inspiration, such as end-inspiration, expiration, and end-expiration segments.
[0133] In certain cases, especially in signals acquired in the presence of added noise, such as due to interference or activity attributed to the user or their surroundings, the detection of the start of S1 in the PG signal is challenging. According to some embodiments, in the EG signal, the R peak corresponds to the depolarization of the ventricle that causes a large ventricular muscle contraction, while the C point in the IG corresponds to the maximum ventricular contraction. Since the start of S1 corresponds to the closure of the mitral valve, it temporally resides between the R peak in the EG and the C point in the IG. Thus, detecting the R and C points is more robust than detecting S1 in a high-noise scenario due to the high-frequency content and large amplitude. Therefore, in some embodiments, when identifying R and C, the segment defined therebetween is specifically searched for the start of S1, which helps in its more robust detection.
[0134] According to some embodiments, since the A point of the IG signal is after the P point and before the start of S1, the local minimum in the IG signal in the segment from the P point of the EG signal to the start of S1 in the PG signal is used to identify the A point of the IG signal.
[0135] The X point in the IG signal and the start point S2 in the PG signal are adjacent in the time domain because the X point in the IG signal relates to the cessation of blood flow caused by the closure of the aortic valve, which also generates the S2 sound. However, both the X point in the IG signal and the start point S2 in the PG signal may not be easily detectable in a high-noise environment. Advantageously, according to some embodiments, the positions of the X point in the IG signal and the start point S2 in the PG signal are estimated by using the C-to-P segment and the probability position information within this segment. Additionally, according to some embodiments, the close relationship between the X point in the IG signal and the start point S2 in the PG signal is used to increase the robustness of the detection of the precise position by leveraging their redundant position information.
[0136] According to some embodiments, the Q point marking the start of ventricular depolarization is used to identify the segment of interest, while the X point marks the closure of the aortic valve, thereby defining the Q-X segment. According to some embodiments, the S1 sound / PG segment is identified based on the Q-X segment, particularly within the time frame defined by the Q-X segment.
[0137] According to some embodiments, the joint analysis of signals is utilized by combinatorially determining the heart rate derived from the EG signal and the heart rate derived from the IG signal to evaluate the higher reliability of heart rate measurement.
[0138] According to some embodiments, the heart rate derived from the EG signal, the heart rate of the IG signal, and the heart rate of the PG signal are jointly analyzed to estimate the heart rate with high reliability.
[0139] According to some embodiments, the time difference between consecutive EG R peaks is used to derive the beat-to-beat EG heart rate time series, and a second heart rate time series is calculated from consecutive IG C points. According to some embodiments, both heart rate time series are modulated by respiratory sinus arrhythmia and are thus jointly used for respiratory rate estimation.
[0140] According to some embodiments, the time difference between consecutive EG R peaks is used to derive the beat-to-beat EG heart rate time series, a second heart rate time series is calculated from consecutive IG C points, and a third heart rate time series is calculated from consecutive S1 points and consecutive S2 points, where all three heart rate time series are modulated by respiratory sinus arrhythmia and are thus jointly used for respiratory rate estimation.
[0141] According to some embodiments, the left ventricular ejection time (LVET) is derived from the IG signal and the PG signal between the B point and the start of S1. According to some embodiments, the pre-ejection period (PEP) is measured from the Q point of the EG signal to the B point of the IG signal. Thus, according to some embodiments, the total systolic time (TST) is calculated as the sum of LVET and PEP.
[0142] According to some embodiments, the device is configured for long-term monitoring of a user. For example, the device is configured to be worn or applied to the user for a period of more than 6 hours.
[0143] Advantageously, long-term monitoring can indicate events that occur rarely or can indicate trends in detected conditions. For example, long-term monitoring can be used after a surgical or therapeutic procedure to detect possible failures after the procedure and / or to detect the recovery process after the procedure. In other examples, long-term monitoring can be used for home monitoring to provide long-term information about the user's condition to a care provider, for example, for long-term follow-up.
[0144] According to some embodiments, the device is configured for medium-term monitoring of a user. For example, the device is configured to be worn or applied to the user for a period of 10 minutes to 6 hours. According to some embodiments, the device is used for short-term monitoring. For example, the device is configured to be worn or applied to the user for a period of 1 second to 10 minutes.
[0145] Advantageously, medium-term and short-term monitoring can be used to evaluate the user's condition before or after a therapeutic procedure, for ambulance settings, for general ward monitoring, etc.
[0146] According to some embodiments, the device is configured to be used periodically, for example, once an hour, once a day, once every predefined period, or once every predefined event, such as a follow-up visit to a clinic.
[0147] According to some embodiments, at least one data is derived using information obtained in one or more previous uses of the device. According to some embodiments, at least one data indicates a change in the condition of the user or body part by detecting a change in information or value between different uses.
[0148] According to some aspects, joint analysis includes deriving data from each of the signals in the signal, where the data derived from different signals includes redundancy therebetween. According to some aspects, the redundancy in the data derived from different signals includes data (such as values) for deriving the same parameter (such as heart rate) multiple times, where each time a different signal or signal combination is used to derive the data.
[0149] For example, data redundancy can be achieved by determining heart rate based on a PG signal, determining heart rate based on an IG signal, and determining heart rate based on an EG signal.
[0150] According to some aspects, the joint analysis further includes estimating the signal quality or confidence parameter / indicator of one of the EG signal, PG signal, or IG signal by comparing data derived from one selected from the EG signal, PG signal, or IG signal with data derived from one or more of the other signals as reference signals. According to some aspects, determining the reference timing data from the reference signal and estimating the signal quality or confidence parameter / indicator of the selected signal among the EG signal, PG signal, or IG signal is achieved by determining the timing data of the selected signal and comparing the timing data of the selected signal with the reference timing data from the reference signal.
[0151] According to some aspects, determining the reference timing data from the reference signal and estimating the signal quality or confidence parameter / indicator of one selected signal among the EG signal, PG signal, or IG signal is achieved by: estimating first data based on a physiological parameter from the reference signal, estimating second data based on the same physiological parameter from the selected signal, and determining a difference value between the data estimated from the selected signal and the data estimated from the reference signal.
[0152] As used herein, the term confidence parameter or confidence metric refers to an indicator related to the certainty of a model regarding a determined or predicted condition or data. In some embodiments, the confidence parameter corresponds to a confidence interval such that the confidence, in a statistical sense, relates to the probability or certainty that a detected or predicted value is correct or that a measured value corresponds to an actual value. For example, a confidence interval with a 95% confidence level indicates that the actual value will be within the upper and lower limit values specified by the confidence interval with 95% certainty. In some embodiments, the estimated signal quality or confidence parameter indicates the noise level or the distribution of possible values around or including a measured value, derived value, or predicted value.
[0153] According to some embodiments, the heart rate derived from the EG is used as a reference value for the heart rates derived from the IG and PG signals, such that if the IG or PG heart rate is different from the EG heart rate, a high difference indicates low signal quality in these signals and thus reduces the confidence in the derived features.
[0154] According to some embodiments, the correlation between signals is modeled, for example, by using a transfer function. According to some embodiments, the IG signal is estimated from the EG signal. For example, the estimation error based on the difference between the measured IG signal and the estimated IG signal indicates the signal quality.
[0155] According to some embodiments, the self-similarity of a periodic signal (e.g., a segment of the cardiac cycle within the signal) is used to estimate signal quality. Such that for example if a signal segment in one or more signals corresponding to a periodic segment is similar to a previous and / or subsequent periodic segment in the (one or more) signals in terms of the signal values of the features, this indicates high signal quality. On the other hand, if the signal segment is different from the previous and / or subsequent periodic signal segments, this indicates poor signal quality and thus a reduced confidence level.
[0156] According to some aspects, the joint analysis includes determining a first physiological parameter from one or more of an EG signal, a PG signal, or an IG signal, determining a second physiological parameter from one or more of the EG signal, the PG signal, or the IG signal, and deriving a third physiological parameter based on the first physiological parameter and the second physiological parameter. According to some aspects, the first parameter, the second parameter, and the third parameter are different from each other. According to some aspects, the first parameter, the second parameter, and the third parameter are the same parameter. According to some aspects, the first parameter and the second parameter are the same parameter, and the third parameter is a different parameter.
[0157] As used herein, the term "physiological parameter" is used interchangeably with the terms "body parameter" and "biological parameter", and refers to a parameter that indicates the physiological state or activity of a user or a body part of the user, such as blood pressure, body temperature, respiratory rate, heart rate, blood oxygen saturation, and various electrophysiological signals, which at least partially represent the operation of the subject's body, and these parameters indicate or can be used to evaluate or monitor the health state of the user.
[0158] According to some aspects, a feature derived from one signal in the signals is derived based on another feature derived from another signal.
[0159] According to some aspects, a feature corresponds to a value corresponding to a characteristic of the signal. According to some aspects, a feature is provided as an input to an input layer of a machine learning model (such as an input layer of a neural network).
[0160] According to some aspects, the joint analysis further includes providing the derived multiple features to a detection and / or prediction model, where the detection and / or prediction model preferably corresponds to a machine learning model, and generating a score or value indicating the probability of a predefined condition of a body part from the detection and / or prediction model.
[0161] According to some embodiments, one or more of the models include supervised machine learning models, such as random forest, XGBoost, logistic regression, artificial neural network, recurrent network, deep learning model, probability model, statistical model, Bayesian model, support vector machine, and / or kernel method.
[0162] According to some embodiments, one or more of the models are anomaly detection models, such as replicator neural networks, autoencoders, variational autoencoders, long short-term memory neural networks, hidden Markov models, and / or fuzzy logic models.
[0163] According to some embodiments, one or more of the models include clustering models, such as centroid models, distribution models, density models, subspace models, group models, and / or graph models.
[0164] According to some embodiments, one or more of the models include pattern recognition models, such as models using classification, clustering, ensemble, and / or regression methods.
[0165] According to some embodiments, an anonymized training set is used to train one or more of the models.
[0166] According to some embodiments, the system is configured to detect a specific heart pathology, such as aortic stenosis. According to some embodiments, during the training phase of one or more computational models applied to a device or system, the computational models are applied to a plurality of patients with known indications of aortic stenosis and patients without such indications, generating an annotated data set that includes at least EG, IG, and PG data and annotations of the presence or absence of aortic stenosis. The EG, IG, and PG data are then preprocessed and cleaned, which includes filtering artifacts from movement and removing crosstalk between different biosignals. According to some embodiments, using joint signal analysis, segment and point detection is performed to identify feature points and segments in the signals, such as QRS complexes, heart sounds, respiratory phases, and phases in the cardiac cycle. From the identified points and segments, one or more intra-signal and inter-signal features, such as amplitude, distance, transform, and coefficients, are extracted and then fed together with the annotations of the presence or absence of aortic stenosis into a supervised machine learning algorithm, thereby creating a model representing the relationship between the features from the EG, IG, and PG signals and the results. Optionally, according to some embodiments, additional patient information (such as one or more of age, gender, height, weight, comorbidities, medications, previous interventions, etc.) can be used as input. When the device / system is applied to a new user / patient, the EG, IG, and PG signals are measured / recorded, and the same signal processing and feature extraction methods as in the training phase are applied. When those features and optionally additional features are presented to the pre-learned / trained model, a result will be generated by inference, which is interpreted as the presence or absence of aortic stenosis within a certain confidence interval.
[0167] According to some embodiments, signals, features, and optionally patient information are combined with a set of 30-day postoperative outcomes (such as heart failure, stroke, surgical site infection, etc.) during a training phase. According to some embodiments, during an inference phase, the system is applied preoperatively to a patient to assess the risk of an intervention by providing a risk score or a specific probability of an adverse outcome.
[0168] According to some aspects, a system for measuring and / or monitoring biosignals of a body part of a subject is provided. The system includes a wearable device according to the above aspects, a system communication unit adapted to communicate with the communication unit of the wearable device to obtain an EG signal, a PG signal, and an IG signal therefrom, and a processor. The processor is configured to obtain the EG signal, the PG signal, and the IG signal from the system communication unit and jointly analyze the EG signal, the PG signal, and the IG signal to allow assessment of a predefined condition of the body part by deriving at least one data based on all of the EG signal, the PG signal, and the IG signal. Preferably, wherein deriving at least one data based on all of the EG signal, the PG signal, and the IG signal includes detecting a feature that preferably includes at least one of a distance between two predefined points of interest, a rate of occurrence of an event, or a value at a point of interest.
[0169] According to some aspects, the joint analysis performed by the processor in the system corresponds to the joint analysis described with respect to the aspects provided above.
[0170] According to some aspects, a system for measuring and / or monitoring biosignals of a body part of a subject is provided. The system includes a processor and a device. The device includes a surface potential difference sensor configured to measure the electrical activity of the body part and provide at least one EG signal indicative of the electrical activity of the body part, an electro-gram (EG), a surface vibration sensor configured to measure the acoustic activity of the body part and provide a PG signal indicative of the acoustic activity of the body part, a phono-gram (PG), and an impedance sensor configured to measure the impedance of the body part and provide an IG signal indicative of the impedance of the body part, an impedance-gram (IG), and a communication circuit configured to transmit signals from the EG sensor, the PG sensor, and the IG sensor to the processor. According to some aspects, the processor is configured to obtain the EG signal, the PG signal, and the IG signal from the communication circuit and jointly analyze the EG signal, the PG signal, and the IG signal to allow assessment of a predefined condition of the heart by deriving at least one data based on all of the EG signal, the PG signal, and the IG signal.
[0171] According to some aspects, the joint analysis performed by the processor in the system corresponds to the joint analysis described with respect to the aspects provided above.
[0172] According to some embodiments, the system includes additional sensors. According to some embodiments, the sensors are embedded in the device. According to some embodiments, the sensors are not embedded in the device.
[0173] According to some embodiments, the additional sensors include one or more of a body temperature sensor, an ambient temperature sensor, a pulse oximeter, a glucose sensor, a volatile compound sensor, a barometer, an accelerometer, a plethysmograph, and the like.
[0174] According to some aspects, the joint analysis further includes deriving data based on all of the EG signal, the PG signal, and the IG signal and on one or more of the signals obtained by the additional sensors.
[0175] According to some embodiments, the additional sensors include one or more inertial measurement unit (IMU) sensors used in the device, such as an accelerometer, a gyroscope, a magnetometer, etc., which advantageously facilitate the detection of movement and thus one or more of motion artifact processing and / or compensation, for example, providing additional input for respiratory signal estimation based on chest movement as a respiratory signal, for example, providing additional input for cardiac diagnosis by using ballistocardiogram or gyroscope, providing additional input for joint analysis, such as movement angle, travel distance, rotation angle, etc., providing an estimate of activities (such as steps, energy consumption, posture, etc.), facilitating cardiac and respiratory analysis in combination with activities, and providing posture estimation for posture-related signal analysis.
[0176] According to some embodiments, the additional sensors include one or more photoplethysmography (PPG) sensors, such as pulse wave, tissue oxygenation, and perfusion, which facilitate joint analysis by indicating the pulse arrival time that can be used to determine the distance between the R peak and the start of the PPG, and can be used to estimate cardiac timing parameters, vascular stiffness, and can provide insights into hemodynamic parameters, bleeding, etc. Additionally, the PPG sensor can be used to estimate local tissue oxygenation as a representative of oxygen delivery, and tissue perfusion can be used as an indicator of circulatory parameters.
[0177] According to some embodiments, the additional sensors include a temperature sensor, such as a body temperature sensor, whose signal indicates a systemic effect, such as inflammation or fever. Additionally, temperature measurement can be used for perioperative temperature management of the patient / user.
[0178] Certain embodiments of the present disclosure may include some, all, or none of the above advantages. One or more technical advantages can be readily understood by those skilled in the art from the accompanying drawings, description, and claims included herein. Additionally, although specific advantages have been listed above, various embodiments may include all, some, or none of the listed advantages.
[0179] In addition to the above-exemplified aspects and embodiments, other aspects and embodiments will become apparent by reference to the drawings and by study of the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS
[0180] Examples of illustrative embodiments are described below with reference to the drawings. In the drawings, the same structures, elements, or components that appear in more than one drawing are generally labeled with the same numerals in all the drawings in which they appear. Alternatively, elements or parts that appear in more than one figure may be labeled with different numerals in the different figures in which they appear. The dimensions of the components and features shown in the drawings are generally chosen for convenience and clarity of presentation and are not necessarily shown to scale. The drawings are listed below.
[0181] Figure 1 Devices according to some embodiments are schematically illustrated;
[0182] Figure 2 Devices communicating with external devices according to some embodiments are schematically illustrated;
[0183] Figure 3 Devices having an integrated processor according to some embodiments are schematically illustrated;
[0184] Figure 4 Systems including devices operable in a mobile setting according to some embodiments are schematically illustrated;
[0185] Figure 5 Patches having an interface according to some embodiments are schematically illustrated;
[0186] Figure 6a Wearable systems including a device and a strip according to some embodiments are schematically illustrated;
[0187] Figure 6b Wearable systems including a device attached to a shirt according to some embodiments are schematically illustrated;
[0188] Figure 7a Handheld devices according to some embodiments are schematically illustrated;
[0189] Figure 7b Systems including a device having a docking socket for a mobile device according to some embodiments are schematically illustrated;
[0190] Figure 8 Methods for measuring and / or monitoring biosignals of a body part of a subject according to some embodiments are schematically illustrated;
[0191] Figure 9a The staging architecture of a model for jointly analyzing signals according to some embodiments is schematically illustrated;
[0192] Figure 9b Schematically shows a parallel architecture of a model for jointly analyzing signals according to some embodiments; and
[0193] Figure 10 Schematically shows a graph including IG, PG, and EG signals according to some embodiments. Detailed Description
[0194] In the following description, various aspects of the present disclosure will be described. For purposes of explanation, specific configurations and details are set forth in order to provide a thorough understanding of the different aspects of the present disclosure. However, it will be apparent to those skilled in the art that the present disclosure may be practiced without the specific details presented herein. Additionally, well-known features may be omitted or simplified so as not to obscure the present disclosure.
[0195] Now refer to Figure 1 , which schematically shows a device 100 according to some embodiments. The device 100 includes an EG sensor 120, an IG sensor 140, and a PG sensor 160, and optionally, the device includes one or more additional sensors 180.
[0196] Each of the EG sensor 120, the IG sensor 140, the PG sensor 160, and optionally the additional sensor 180 is integrated with or connectable to an interface for obtaining signals from a subject's body. According to some embodiments, the EG sensor 120 and the IG sensor 140 are connected to a shared set of electrodes (not shown), while the PG sensor 160 is connected to an interface such as a membrane.
[0197] The EG sensor 120 is a surface potential difference sensor configured to measure the electrical activity of a body part and provide / generate at least one EG signal indicative of the electrical activity of the body part.
[0198] The IG sensor 140 is an impedance sensor configured to measure the impedance of a body part and provide an IG signal indicative of the impedance of the body part.
[0199] The PG sensor 160 is a surface vibration sensor configured to measure the acoustic activity of a body part and provide a PG signal indicative of the acoustic activity of the body part.
[0200] The device 100 is configured to obtain EG signals, IG signals, and PG signals for facilitating the joint analysis of the signals to allow the evaluation of the condition of a body part or a subject by deriving at least one data based on all of the EG signals, IG signals, and PG signals.
[0201] According to some embodiments, the additional sensor 180 is a predefined / preconfigured sensor, such as a temperature sensor, a blood oxygen measurement sensor, etc.
[0202] Reference is now made to Figure 2 , which schematically shows a device 200 communicating with an external device 270 according to some embodiments. Similar to Figure 1 the device 100, the device 200 includes an EG sensor 120, an IG sensor 140, and a PG sensor 160, and optionally, the device includes one or more additional sensors 180. Additionally, the device 200 includes a communication unit or module 210, configured to obtain EG signals, IG signals, and PG signals and transmit the signals to an external communication unit or module 272 of the external device 270.
[0203] The external device 270 includes an external communication unit or module, configured to receive EG signals, IG signals, and PG signals from the communication unit or module 210 of the device 200 and provide the signals to a processor 274. In this embodiment, the processor 274 is configured to perform at least a part of the joint analysis, however, such a configuration is only optional.
[0204] Optionally, the external device 270 communicates with or is adapted to communicate certain operation information, such as calibration information, operation commands, wake-up commands, sleep commands, etc., to the device 200 to control the operation of the device 200 and / or confirm the operation state, such as the connection establishment state.
[0205] Optionally, the device 200 includes an internal processor 220, configured to perform certain preprocessing procedures on one or more of the EG signals, IG signals, and PG signals.
[0206] According to some embodiments, the device 200 can be connected to an external additional sensor 190 either wired or wirelessly.
[0207] Reference is now made to Figure 3 , which schematically shows a device 300 having an integrated processor 320 according to some embodiments.
[0208] Similar to Figure 1 the device 100 and Figure 2 the device 200, the device 300 includes an EG sensor 120, an IG sensor 140, and a PG sensor 160, and optionally, the device includes one or more additional sensors 180. The integrated processor 320 is configured to obtain EG signals, IG signals, and PG signals from the sensors and perform a joint analysis of the sensors. In addition, the integrated processor 320 is configured to obtain one or more additional signals from one or more additional sensors 180 (if any).
[0209] The device 300 further includes a user interface 330 for obtaining input from a user and providing output to the user. The user interface includes one or more of a display, a touch display, a light indicator, a sound indicator, a microphone, a vibration mechanism, one or more buttons, and the like.
[0210] In some embodiments, the user interface (particularly the output mechanisms such as the display, the light indicator, the sound indicator, and / or the vibration mechanism) is configured to provide biofeedback to the user based on one or more of the EG signal, the IG signal, and the PG signal, or the result of their combined analysis or a partial analysis thereof.
[0211] Optionally, the device 300 further includes a communication unit or module 310 configured to communicatively connect to a communication module of other devices, for example, for receiving operation or configuration information, and / or for transmitting information related to one or more of the EG signal, the IG signal, and the PG signal, or the result of their combined analysis or a partial analysis thereof.
[0212] In the following embodiments, references to a device (such as device 201) may be interchangeable with references to device 200 as in Figure 2 and may also be interchangeable with references to device 300 corresponding to Figure 3 or device 100 corresponding to Figure 2
[0213] Now refer to Figure 4 , which schematically shows a system 400 including a device 200 operable in a mobile setting according to some embodiments. In a mobile setting, the device 200 and the external device 270 may move relative to each other, and they may move away from each other. As shown, the device 200 may be operated in a moving vehicle (such as an ambulance 420), while the external device is away from the device 200 and may be operated in a clinic 410. According to some embodiments, the device 200 communicates with the external device 270 directly or through a repeater in the ambulance 420. As Figure 4 shown, the device 200 communicates wirelessly with the external device 270 through a cellular network 430 or a dedicated wireless network.
[0214] Optionally, the device 200 or the repeater in the ambulance 420 communicates wirelessly directly with the external device 270 or the repeater in the clinic 410. Optionally, the server 440 communicates with the device 200 for receiving one or more signals or information related to one or more signals, and further optionally, the server 440 includes or is connected to a server processor 442 configured to perform a combined analysis or a partial analysis thereof.
[0215] According to some embodiments, the communication between the device 200 and the external device 270 includes one or more of a cellular network (directly or via a mobile device), via a wireless network, or via a wired network, via the Internet and / or via an intranet.
[0216] According to some embodiments, the device 200 is located in the user's home, while the external device 270 is located in an ambulance, a hospital, or a clinic with a caregiver.
[0217] According to some embodiments, the device 200 is located in an ambulance, while the external device 270 is located in a hospital. According to some embodiments, the device 200 is located in a doctor's office, while the external device 270 is located in a hospital or an ambulance with a caregiver.
[0218] Now refer to Figure 5 , which schematically shows a patch 500 with an interface according to some embodiments. According to some embodiments, the patch is integrated with a device (such as devices 100, 200, 300) or removably attached to the device, wherein the patch 500 includes an interface configured to engage with the subject's body and provide signals to the sensors. As Figure 5 shown, the first electrode 510 and the second electrode 520 are both electrodes configured to facilitate EG measurement and IG measurement. In other words, the first electrode 510 and the second electrode 520 are shared IG / EG electrodes.
[0219] The patch 500 further includes a surface vibration interface configured to facilitate PG measurement, such as an acoustic membrane 530.
[0220] Optionally, the patch further includes an additional electrode 540, which is a shared IG / EG electrode, a dedicated IG electrode, or a dedicated EG electrode.
[0221] Optionally, the patch includes an interface for providing signals to additional sensors, such as a light electrode 550 or a thermistor 560.
[0222] The interface is connected or connectable to the corresponding sensor.
[0223] According to some embodiments, the surface of the device or patch attachable or applicable to a subject has a length of up to 12 cm and a width of up to 12 cm. According to some embodiments, the surface of the device or patch attachable or applicable to a subject has a length of 6 cm and a width of 6 cm. According to some embodiments, the patch has a depth / height of at most 1 cm, and preferably 0.6 cm.
[0224] According to some embodiments, one or more electrodes, particularly the interfaces for IG and / or EG sensors, have a diameter of 24 mm.
[0225] According to some embodiments, the interface of the PG sensor has a diameter of 30 mm.
[0226] According to some embodiments, the device is configured to be wearable by a subject. According to some embodiments, one or more of the following attachment methods are used to attach the device to the user / subject: a tether, such as a wired connection to the sensor, a suction cup with active vacuum or passive suction, the device being fully or partially covered with tape, a separate gluing member between the device and the skin, a double-sided adhesive, using an attachment on the skin that is configured to reversibly hold the device or patch in place, a tattoo / e-ink, etc.
[0227] Now refer to Figure 6a , which schematically shows a wearable system 600 including a device 201 and a strip 620 according to some embodiments. The strip 620 is configured to hold the device 201 in place and is pressed against the skin of the subject to facilitate the measurement of IG, PG, and EG signals.
[0228] According to some embodiments, the strip 620 is configured to be worn around the upper torso, limb, middle torso, lower torso, neck, hip, or head of the subject based on the target body part or physiological system to be monitored / measured.
[0229] According to some embodiments, the device 201 or patch is configured to be attached to the strip by the pressure exerted by the strip 620 and the friction between the strip 620 and the device 201. According to some embodiments, the dorsal side of the device 201 and / or the corresponding internal part of the strip 620 includes a high-friction material, such as rubber or silicone.
[0230] According to some embodiments, the device 201 or patch is attached to the strip 620 using adhesion. According to some embodiments, the device 201 or patch is attached to the strip 620 using a reversible attachment mechanism, such as a magnetic attachment, a socket and a mating part, a hook-and-loop fastening mechanism (such as Velcro®), etc.
[0231] Now refer to Figure 6b , which schematically shows a wearable system 601 including a device 201 attached to a textile article, particularly a shirt 640, according to some embodiments.
[0232] According to some embodiments, the device 201 or patch is attached to or integrated into a textile article (such as a shirt 640) using adhesion. According to some embodiments, the device 201 or patch is attached to a textile article, such as a shirt 640, using a reversible attachment mechanism, such as a magnetic attachment, a socket and a mating part, a hook-and-loop fastening mechanism (such as Velcro®), etc.
[0233] According to some embodiments, the textile article is a shirt, hat, scarf, pants, underwear, braid, gloves, socks, and / or vest.
[0234] Now refer to Figure 7a , which schematically shows a handheld device 700 according to some embodiments. The device 201 or patch is attached to or attachable to a handle 720, which facilitates positioning the device 201 to a specific area on a subject's body for performing measurements or monitoring. According to some embodiments, a user interface including an input device and / or an output device is integrated into the handle 720.
[0235] Now refer to Figure 7b , which schematically shows a system 701 including a device 201 having a docking socket 740 for a mobile device 760 according to some embodiments. The socket 740 is configured to hold the mobile device 760 (such as a smart phone), and the connector 750 is configured to facilitate communication between the device 201 and the mobile device 760. According to some embodiments, the functions of the mobile device 760 are similar to those of the external device 300 proposed above. According to some embodiments, the user interface of the mobile device 760 can be used to view information about the results of one or more signals or joint analysis. According to some embodiments, the user interface of the mobile device 760 is configured to facilitate control of the operation or functions of the device 201.
[0236] According to some embodiments, the mobile device 760 is configured to relay information between the device 201 and the external device 300 using a communication module in the mobile device 760.
[0237] Now refer to Figure 8 , which schematically shows an algorithm, processor configuration, or method 800 for measuring and / or monitoring bio-signals of a body part of a subject. Initially, optionally, demographic and / or medical information of the subject is acquired 802, the device / patch is applied at a target location 804 on the subject's body, and EG, IG, and PG signals are acquired 806 and preprocessed 808.
[0238] According to some embodiments, the preprocessing 808 of the signals involves applying one or more of filtering, noise reduction, transformation, modulation, smoothing, derivation, and / or normalization.
[0239] Then, the preprocessed signals along with the demographic and / or medical information are fed into a computational model 810 for joint analysis, and inferences of the results are performed 812 from the computational model.
[0240] According to some embodiments, the acquisition of the EG, IG, and PG signals includes obtaining a baseline measurement of the EG, IG, and PG signals and then obtaining a snapshot of the new EG, IG, and PG signals. According to some embodiments, the acquisition of the snapshots of the EG, IG, and PG signals is continuous, and method 800 operates continuously in a single forward feed mode. According to other embodiments, the results are fed back to an intermediate stage in the method to create a closed loop, for example, by feeding previously inferred information back into the model as its input, or by triggering the acquisition of new EG, IG, and PG signals based on the previously inferred results.
[0241] According to some embodiments, method 800 forms an "inference stage" in the model training method, which includes obtaining baseline data and time series snapshot data of the EG, IG, and PG signals, then, performing preprocessing of the EG, IG, and PG signals and using the information for physiological parameter estimation and feature extraction, then, generating a model by deriving a generalized representation of the relationship between the features and the results based on the previously annotated data, and finally, using the model in the inference stage based on method 800.
[0242] According to some embodiments, a model is used for joint analysis such that all EG, IG, and PG signals and additional information (such as demographic and / or medical information) are fed into the model for performing joint analysis.
[0243] According to other embodiments, several models are used in various architectures for performing joint analysis.
[0244] Now refer to Figure 9a , which schematically shows a staged architecture 900 of a model for joint analysis of signals according to some embodiments. Initially, a first signal (such as an EG signal) is fed into a first model 902, then, the result of the first model 902 is fed together with a second signal (such as an IG signal) into a second model 904, then, the result of the second model 904 and optionally the result of the first model 902 are fed together with a third signal (such as a PG signal) into a third model 906 for deriving data based on the joint analysis of the combination of all EG, IG, and PG signals and optionally demographic and / or medical information.
[0245] According to some embodiments, the results of the intermediate stage / model are also used as data and output to the user.
[0246] Now refer to Figure 9b, which schematically shows a parallel architecture 901 of a model for joint analysis of signals according to some embodiments. A first signal such as an EG signal is fed to a fourth model 912, a second signal such as an IG signal is fed to a fifth model 914, a third signal such as a PG signal is fed to a sixth model 916, and then the results of the fourth model 912, the fifth model 914, and the sixth model 916 are fed to a seventh model 918 for deriving data based on a joint analysis of the combination of all EG, IG, and PG signals and optionally demographic and / or medical information fed to one or more models.
[0247] According to some embodiments, one or more of the intermediate models, in particular models 902, 904, 906, 912, 914, 916, and 918 as described above, may include multiple stacked models.
[0248] According to some embodiments, one or more of the above-mentioned models, in particular models 902, 904, 906, 912, 914, 916, and 918, may receive more than one of the IG, PG, or EG signals.
[0249] According to some embodiments, joint analysis is performed such that the analysis of the third signal is performed based on the analysis result of the second signal, and the analysis of the second signal is performed based on the analysis result of the first signal. According to different embodiments, the first signal, the second signal, and the third signal mutually exclusively correspond to the IG, PG, and EG signals in different orders.
[0250] According to some embodiments, joint analysis is performed such that the analysis of the third signal is performed based on the results of the analysis of the second signal and the first signal or based on the joint analysis of the second signal and the first signal. According to different embodiments, the first signal, the second signal, and the third signal mutually exclusively correspond to the IG, PG, and EG signals in different orders.
[0251] Now refer to Figure 10 , which schematically shows a graph 1000 including IG, PG, and EG signals according to some embodiments. The EG signal is analyzed to extract points P, Q, R, S, and / or T, the PG signal is analyzed to extract S1 start, S2 start, S3 start (not shown), S4 start (not shown), and the IG signal is analyzed to extract points A, B, C, X, Y, and / or O.
[0252] According to some embodiments, the interesting segment in the PG signal is the S3 sound segment, which is caused by a low-frequency transient vibration occurring in the early diastole at the end of the rapid diastolic filling phase of the right or left ventricle. According to some embodiments, the S3 segment is detected by identifying the start and end of a segment having high signal energy in the frequency range associated with the S3 heart sound.
[0253] According to some embodiments, the interesting segment in the PG signal is the S4 sound segment, which, if present, is caused by atrial contraction. According to some embodiments, the S4 sound segment is detected by identifying the start and end of a segment having high signal energy in the frequency range associated with the S4 heart sound.
[0254] The P point corresponds to atrial depolarization, the A point corresponds to atrial contraction, the R point corresponds to ventricular depolarization, the start of S1 corresponds to the closure of the mitral valve, the B point corresponds to the opening of the aortic valve, the C point corresponds to maximum ventricular contraction or maximum systolic flow, the start of S2 and the X point correspond to the closure of the aortic valve, the Y point corresponds to the closure of the pulmonary valve, and the O point corresponds to the opening of the mitral valve.
[0255] According to some embodiments, the joint analysis includes determining a segment based on two detected points, where both points are extracted from the analysis of the same signal.
[0256] According to some embodiments, the segment includes one or more of an atrial contraction segment from the P point to the S point or the start of S1, an isovolumetric contraction segment from the R point to the B point, a ventricular ejection segment from the B point to the start of S2 or the X point, an isovolumetric relaxation segment from the start of S2 or the X point to the O point, a ventricular contraction from the S point or the start of S1 to the start of S2 or the X point and / or a ventricular relaxation segment from the start of S2 or the X point to the S point.
[0257] According to some embodiments, the joint analysis includes determining a segment based on two detected points, where one point is extracted from the analysis of one signal and the other point is extracted from another signal.
[0258] At least a portion of the analysis of the EG signal, PG signal, and / or IG signal is based on the analysis of two other signals, particularly on detected points or segments in the other signals.
[0259] Unless otherwise specifically stated, as will be apparent from the following discussion, it should be understood that throughout the specification discussion, the use of terms such as "processing", "computing", "operating", "determining", "estimating", etc. refers to the actions and / or processes of a computer or computing system or similar electronic computing device that manipulates and / or transforms data represented as physical quantities (such as electronic quantities) within the registers and / or memories of the computing system into other data similarly represented as physical quantities within the memories, registers, or other such information storage, transmission, or display devices of the computing system. Further, throughout the specification, the term "plurality" may be used to describe two or more components, devices, elements, parameters, etc.
[0260] Embodiments of the present disclosure may include apparatuses for performing the operations herein. The apparatus may be specially constructed for the desired purpose, or it may include a general-purpose computer selectively activated or reconfigured by a computer program stored in the computer. Such a computer program may be stored in a computer-readable storage medium, such as but not limited to any type of disk, including floppy disks, optical disks, CD-ROMs, magneto-optical disks, read-only memory (ROM), random access memory (RAM), electrically programmable read-only memory (EPROM), electrically erasable and programmable read-only memory (EEPROM), magnetic or optical cards, or any other type of medium suitable for storing electronic instructions and capable of being coupled to a computer system.
[0261] In addition to the foregoing explanations, the following aspects, including the enumerated aspects 1 to 15, are also relevant to the present disclosure and are part of the specification and shall not be confused with the appended claims.
[0262] 1. An apparatus (100, 200, 201, 300) for measuring and / or monitoring a bio-signal of a body part of a subject, the apparatus (100, 200, 201, 300) comprising:
[0263] - A surface potential difference sensor, especially an electrogram EG sensor (120), configured to measure the electrical activity of the body part and provide at least one EG signal indicative of the electrical activity of the body part,
[0264] - A surface vibration sensor, especially a phonogram PG sensor (160), configured to measure the acoustic activity of the body part and provide a PG signal indicative of the acoustic activity of the body part, and
[0265] - An impedance sensor, especially an impedance plethysmogram IG sensor (140), configured to measure the impedance of the body part and provide an IG signal indicative of the impedance of the body part,
[0266] Preferably, wherein the body part includes the heart and / or the lungs.
[0267] 2. The apparatus (100, 200, 201, 300) according to aspect 1, further comprising a communication unit (210), the communication unit (210) being configured to obtain the EG signal from the EG sensor (120), the PG signal from the PG sensor (160), and the IG signal from the IG sensor (140), and transmit the signals to an external communication unit (272).
[0268] 3. The apparatus (100, 200, 201, 300) according to aspect 1, further comprising:
[0269] A processor (220), configured to:
[0270] Obtain an EG signal from an EG sensor (120),
[0271] Obtain a PG signal from a PG sensor (160),
[0272] Obtain an IG signal from an IG sensor (140), and
[0273] Jointly analyze the EG signal, the PG signal, and the IG signal to allow the evaluation of a predefined condition of a body part or a subject by deriving at least one data based on all of the EG signal, the PG signal, and the IG signal,
[0274] Preferably, wherein deriving at least one data based on all of the EG signal, the PG signal, and the IG signal includes detecting features, and the features preferably include at least one of a distance between two predefined points of interest, an incidence rate of an event, or a value at a point of interest.
[0275] 4. The device (100, 200, 201, 300) according to aspect 3, wherein the joint analysis includes performing point and / or segment detection based on one or more of the EG signal, the PG signal, or the IG signal to detect points of interest and / or segments, and analyzing other signals based on the detected points of interest or segments. Optionally, wherein performing segment detection includes segmenting the signal into predefined segments or predefined phases of predefined biological events, such as respiratory cycles, cardiac cycles, and / or circadian cycles.
[0276] Preferably, wherein analyzing other signals based on the segmentation includes identifying characteristic attributes in one or more of the other signals based on the detected points and / or segments.
[0277] 5. The device (100, 200, 201, 300) according to aspect 3 or 4, wherein the joint analysis includes deriving data from each of the signals, and the data derived from different signals includes redundancy therebetween.
[0278] Optionally, wherein the joint analysis further includes estimating a signal quality or a confidence parameter of a signal selected from the EG signal, the PG signal, or the IG signal by comparing data derived from a signal selected from the EG signal, the PG signal, or the IG signal with data derived from one or more of the other signals as reference signals, preferably by:
[0279] Determining reference timing data from the reference signal, determining the timing data of the selected signal, and comparing the timing data of the selected signal with the reference timing data from the reference signal, or by
[0280] Estimate first data based on a physiological parameter from a reference signal, estimate second data based on the same physiological parameter from a selected signal, and determine a deviation value between the data estimated from the selected signal and the data estimated from the reference signal.
[0281] 6. The device (100, 200, 201, 300) according to any one of aspects 3 to 5, wherein the joint analysis includes:
[0282] Determine a first physiological parameter from one or more of an EG signal, a PG signal, or an IG signal,
[0283] Determine a second physiological parameter from another or others of an EG signal, a PG signal, or an IG signal, and
[0284] Derive a third physiological parameter based on the first physiological parameter and the second physiological parameter,
[0285] wherein the first parameter, the second parameter, and the third parameter are different from each other,
[0286] wherein the first parameter, the second parameter, and the third parameter are the same parameter, or
[0287] wherein the first parameter and the second parameter are the same parameter, and the third parameter is a different parameter.
[0288] 7. The device (100, 200, 201, 300) according to any one of aspects 3 to 6, wherein a feature derived from one of the signals is derived based on another feature derived from another signal.
[0289] 8. The device (100, 200, 201, 300) according to aspect 7, wherein the joint analysis further includes:
[0290] Provide the derived multiple features to a detection and / or prediction model, wherein the detection and / or prediction model preferably corresponds to a machine learning model, and
[0291] Generate a score or value indicating the probability of a predefined condition of a body part from the detection and / or prediction model.
[0292] 9. A system for measuring and / or monitoring a biological signal of a body part of a subject, the system comprising:
[0293] The wearable device (100, 200, 201, 300) according to aspect 2,
[0294] A system communication unit (272) adapted to communicate with a communication unit (210) of the wearable device (100, 200, 201, 300) to obtain an EG signal, a PG signal, and an IG signal from the communication unit (210), and
[0295] Processors (274, 442), configured to:
[0296] Obtain an EG signal, a PG signal, and an IG signal from a system communication unit (272), and
[0297] Jointly analyze the EG signal, the PG signal, and the IG signal to allow for the assessment of a predefined condition of a body part by deriving at least one data based on all of the EG signal, the PG signal, and the IG signal,
[0298] Preferably, wherein deriving at least one data based on all of the EG signal, the PG signal, and the IG signal includes detecting a feature, which preferably includes at least one of a distance between two predefined points of interest, an incidence rate of an event, or a value at a point of interest.
[0299] 10. The system according to aspect 9, wherein the joint analysis includes one or more of the following:
[0300] Performing point and / or segment detection based on one or more of the EG signal, the PG signal, or the IG signal to detect points of interest and / or segments, and analyzing other signals based on the detected points of interest or segments. Optionally, wherein performing segment detection includes segmenting the signal into predefined segments or predefined phases of a predefined biological event, such as a respiratory cycle, a cardiac cycle, and / or a circadian rhythm cycle. Preferably, wherein analyzing other signals based on the segmentation includes identifying characteristic attributes in one or more of the other signals based on the detected points and / or segments, and / or
[0301] Deriving data from each of the signals, wherein the data derived from different signals includes redundancy therebetween. Optionally, wherein the joint analysis further includes estimating a signal quality or a confidence parameter of the EG signal, the PG signal, or the IG signal by comparing data derived from one signal selected from the EG signal, the PG signal, or the IG signal with data derived from one or more of the other signals as reference signals. Preferably by:
[0302] Determining reference timing data from a reference signal, determining timing data of the selected signal, and comparing the timing data of the selected signal with the reference timing data from the reference signal, or by
[0303] Estimating a first data based on a physiological parameter from a reference signal, estimating a second data based on the same physiological parameter from the selected signal, and determining a deviation value between the data estimated from the selected signal and the data estimated from the reference signal.
[0304] 11. The system according to aspect 9 or 10, wherein the joint analysis includes:
[0305] Determine a first physiological parameter from one or more of an EG signal, a PG signal, or an IG signal,
[0306] Determine a second physiological parameter from one or more of the other of the EG signal, the PG signal, or the IG signal, and
[0307] Derive a third physiological parameter based on the first physiological parameter and the second physiological parameter,
[0308] wherein the first parameter, the second parameter, and the third parameter are different from each other,
[0309] wherein the first parameter, the second parameter, and the third parameter are the same parameter, or
[0310] wherein the first parameter and the second parameter are the same parameter, and the third parameter is a different parameter.
[0311] 12. A system for measuring and / or monitoring a biological signal of a body part of a subject, the system comprising:
[0312] A processor (220, 274, 442), and
[0313] A device (100, 200, 201, 300), comprising:
[0314] - A surface potential difference sensor (120), electrogram EG, configured to measure the electrical activity of the body part and provide at least one EG signal indicative of the electrical activity of the body part,
[0315] - A surface vibration sensor (160), phonogram PG, configured to measure the acoustic activity of the body part and provide a PG signal indicative of the acoustic activity of the body part, and
[0316] - An impedance sensor (140), impedancegram IG, configured to measure the impedance of the body part and provide an IG signal indicative of the impedance of the body part, and
[0317] A communication circuit (210, 272), the communication circuit being configured to transmit signals from the EG sensor, the PG sensor, and the IG sensor to the processor (220, 274, 442),
[0318] wherein the processor (220, 274, 442) is configured to:
[0319] Obtain the EG signal, the PG signal, and the IG signal from the communication circuit (210, 272), and
[0320] Jointly analyze the EG signal, the PG signal, and the IG signal to allow for the evaluation of a predefined condition of the heart by deriving at least one data based on all of the EG signal, the PG signal, and the IG signal,
[0321] Preferably, the body part includes the heart and / or the lungs.
[0322] 13. The system according to aspect 12, wherein the joint analysis includes one or more of the following:
[0323] Performing point and / or segment detection based on one or more of the EG signal, the PG signal, or the IG signal to detect points of interest and / or segments, and analyzing other signals based on the detected points of interest or segments. Optionally, wherein performing segment detection includes segmenting the signal into predefined segments or predefined phases of predefined biological events, such as respiratory cycles, cardiac cycles, and / or circadian rhythm cycles. Preferably, wherein analyzing other signals based on the segmentation includes identifying characteristic attributes in one or more of the other signals based on the detected points and / or segments, and / or
[0324] Deriving data from each of the signals, wherein the data derived from different signals includes redundancy therebetween.
[0325] Optionally, wherein the joint analysis further includes estimating the signal quality or confidence parameter of one of the EG signal, the PG signal, or the IG signal by comparing the data derived from one of the signals selected from the EG signal, the PG signal, or the IG signal with the data derived from one or more of the other signals as reference signals. Preferably by:
[0326] Determining reference timing data from the reference signal, determining the timing data of the selected signal, and comparing the timing data of the selected signal with the reference timing data from the reference signal, or by
[0327] Estimating first data according to the physiological parameter from the reference signal, estimating second data according to the same physiological parameter from the selected signal, and determining the deviation value between the data estimated from the selected signal and the data estimated from the reference signal.
[0328] 14. The system according to any one of aspects 11 to 13, wherein the joint analysis includes:
[0329] Determining a first physiological parameter from one or more of the EG signal, the PG signal, or the IG signal.
[0330] Determining a second physiological parameter from another or more of the EG signal, the PG signal, or the IG signal, and
[0331] Deriving a third physiological parameter based on the first physiological parameter and the second physiological parameter.
[0332] Wherein the first parameter, the second parameter, and the third parameter are different from each other.
[0333] wherein, the first parameter, the second parameter, and the third parameter are the same parameter, or
[0334] wherein, the first parameter and the second parameter are the same parameter, and the third parameter is a different parameter.
[0335] 15. The system according to any one of aspects 11 to 14, wherein a feature derived from one of the signals is derived based on another feature derived from another signal, and
[0336] Optionally, wherein the joint analysis further includes:
[0337] providing the derived multiple features to a detection and / or prediction model, wherein the detection and / or prediction model preferably corresponds to a machine learning model, and
[0338] generating, from the detection and / or prediction model, a score or value indicating the probability of a predefined condition of a body part.
[0339] The present disclosure may be described in the general context of computer-executable instructions, such as program modules, executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The present disclosure may also be practiced in a distributed computing environment where tasks are performed by remote processing devices linked through a communication network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
[0340] According to some embodiments, one or more computational models may be selected and / or adjusted based on a group or hierarchical classification of users, e.g., based on one or more of the following information: age, gender, height, weight, clinical status, clinical history, medications, comorbidities, genetic information, or other demographic information.
[0341] The terms used herein are for the purpose of describing particular embodiments only and are not limiting. As used herein, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that when used in this specification, the terms "comprises" or "comprising" specify the presence of the stated features, integers, steps, operations, elements, or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or groups thereof.
[0342] Although many exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain modifications, additions, and sub-combinations thereof. Accordingly, the following appended claims and the claims hereafter introduced are intended to be construed to include all such modifications, additions, and sub-combinations within their scope.
[0343] Reference Signs
[0344] 100, 200, 201, 300 Devices
[0345] 120 EG Sensor
[0346] 140 IG Sensor
[0347] 160 PG Sensor
[0348] 180 Additional Sensor
[0349] 190 External Additional Sensor
[0350] 210 Communication Module
[0351] 220 Internal Processor
[0352] 270 External Device
[0353] 272 External Communication Module
[0354] 274 Processor
[0355] 310 Communication Module
[0356] 320 Integrated Processor
[0357] 330 User Interface
[0358] 400 System
[0359] 410 Clinic
[0360] 420 Ambulance
[0361] 430 Cellular Network
[0362] 440 Server
[0363] 442 Server Processor
[0364] 500 Patch
[0365] 510 First Electrode
[0366] 520 Second Electrode
[0367] 530 Acoustic Membrane
[0368] 540 Additional Electrode
[0369] 550 Optical Electrode
[0370] 560 Thermistor
[0371] 600, 601 Wearable System
[0372] 620 Band
[0373] 640 Shirt
[0374] 700, 701 Systems
[0375] 720 Handle
[0376] 740 Socket
[0377] 750 Connector
[0378] 760 Mobile Device
[0379] 800 Method
[0380] 802, 804, 806, 808, 810, 812 Method Steps
[0381] 900 Staging Architecture
[0382] 901 Parallel Structure
[0383] 902 First Model
[0384] 904 Second Model
[0385] 906 Third Model
[0386] 912 Fourth Model
[0387] 914 Fifth Model
[0388] 916 Sixth Model
[0389] 918 Seventh Model
[0390] 1000 Curve Chart
Claims
1. A device (100, 200, 201, 300) for measuring and / or monitoring a bio-signal of a body part of a subject, the device (100, 200, 201, 300) comprising: - An electrograph (EG) sensor (120) configured to measure the electrical activity of the body part and provide at least one EG signal indicative of the electrical activity of the body part, - A phonograph (PG) sensor (160) configured to measure the acoustic activity of the body part and provide a PG signal indicative of the acoustic activity of the body part, - An impedance graph (IG) sensor (140) configured to measure the electrical impedance of the body part and provide an IG signal indicative of the electrical impedance of the body part, and - A processor (220) configured to: Obtain the EG signal from the EG sensor (120), Obtain the PG signal from the PG sensor (160), Obtain the IG signal from the IG sensor (140), and Jointly analyze the EG signal, the PG signal and the IG signal to allow evaluation of a predefined condition of the body part or the subject by deriving at least one data based on all of the EG signal, the PG signal and the IG signal, Wherein the processor is further configured to perform the joint analysis by performing segment detection based on one or more of the EG signal, the PG signal or the IG signal to detect an interest segment and analyzing other signals based on the detected interest segment, wherein performing segment detection includes segmenting the signal into predefined segments or predefined phases of predefined bio-events, such as respiratory cycles, cardiac cycles and / or circadian rhythm cycles, and wherein analyzing the other signals based on the segmentation includes identifying characteristic attributes in one or more of the other signals based on the detected segments, And / or the processor is further configured to perform the joint analysis using a machine learning model by providing the EG signal, the PG signal and the IG signal as inputs to the machine learning model.
2. The device (100, 200, 201, 300) according to claim 1, wherein, The joint analysis includes deriving data from each of the signals, wherein the data derived from different signals includes redundancy therebetween, Optionally, wherein the joint analysis further includes estimating a signal quality or a confidence parameter of a signal selected from the EG signal, the PG signal or the IG signal by comparing data derived from a signal selected from the EG signal, the PG signal or the IG signal with data derived from one or more of the other signals as a reference signal, preferably by: Determining reference timing data from the reference signal, determining the timing data of the selected signal, and comparing the timing data of the selected signal with the reference timing data from the reference signal, or by Estimating a first data according to a physiological parameter from the reference signal, estimating a second data according to the same physiological parameter from the selected signal, and determining a deviation value between the data estimated from the selected signal and the data estimated from the reference signal.
3. The device (100, 200, 201, 300) according to any one of claims 1 or 2, wherein, The joint analysis includes: determining a first physiological parameter from one or more of the EG signal, the PG signal, or the IG signal, determining a second physiological parameter from one or more of the other of the EG signal, the PG signal, or the IG signal, and deriving a third physiological parameter based on the first physiological parameter and the second physiological parameter, wherein the first parameter, the second parameter, and the third parameter are different from each other, wherein the first parameter, the second parameter, and the third parameter are the same parameter, or wherein the first parameter and the second parameter are the same parameter, and the third parameter is a different parameter.
4. The device (100, 200, 201, 300) according to any one of claims 1 to 3, wherein, A feature derived from one of the signals is derived based on another feature derived from another signal.
5. The device (100, 200, 201, 300) according to claim 4, wherein, The joint analysis further includes: providing the derived plurality of features to a detection and / or prediction model, wherein the detection and / or prediction model preferably corresponds to a machine learning model, and generating from the detection and / or prediction model a score or value indicative of the probability of the predefined condition of the body part.
6. A system for measuring and / or monitoring biological signals of a body part of a subject, the system comprising: a wearable device (100, 200, 201, 300), comprising: - an electrograph EG sensor (120) configured to measure the electrical activity of the body part and provide at least one EG signal indicative of the electrical activity of the body part, - a phonogram PG sensor (160) configured to measure the acoustic activity of the body part and provide a PG signal indicative of the acoustic activity of the body part, and - an impedance graph IG sensor (140) configured to measure the impedance of the body part and provide an IG signal indicative of the impedance of the body part, a system communication unit (272) adapted to communicate with a communication unit (210) of the wearable device (100, 200, 201, 300) to obtain the EG signal, the PG signal, and the IG signal therefrom, and a processor (274, 442) configured to: obtain the EG signal, the PG signal, and the IG signal from the system communication unit (272), and jointly analyze the EG signal, the PG signal, and the IG signal to allow evaluation of a predefined condition of the body part by deriving at least one data based on all of the EG signal, the PG signal, and the IG signal, wherein the processor is further configured to perform joint analysis by performing segment detection based on one or more of the EG signal, the PG signal, or the IG signal to detect an interesting segment and analyzing other signals based on the detected interesting segment, wherein performing segment detection includes segmenting the signal into predefined segments or predefined phases of predefined biological events, such as respiratory cycles, cardiac cycles, and / or circadian rhythm cycles, and wherein analyzing the other signals based on the segmentation includes identifying characteristic attributes in one or more of the other signals based on the detected segments and / or the processor is further configured to perform the joint analysis using the machine learning model by providing the EG signal, the PG signal, and the IG signal as inputs to the machine learning model.
7. The system according to claim 6, wherein The joint analysis further includes: deriving data from each of the signals, wherein the data derived from different signals includes redundancy therebetween, optionally, wherein the joint analysis further includes estimating a signal quality or a confidence parameter of one of the EG signal, the PG signal, or the IG signal by comparing data derived from one of the EG signal, the PG signal, or the IG signal selected with data derived from one or more of the other signals as a reference signal, preferably by: determining reference timing data from the reference signal, determining timing data of the selected signal, and comparing the timing data of the selected signal with the reference timing data from the reference signal, or by estimating first data based on a physiological parameter from the reference signal, estimating second data based on the same physiological parameter from the selected signal, and determining a deviation value between the data estimated from the selected signal and the data estimated from the reference signal.
8. The system according to claim 6 or 7, wherein The joint analysis includes: determining a first physiological parameter from one or more of the EG signal, the PG signal, or the IG signal, determining a second physiological parameter from one or more of the other of the EG signal, the PG signal, or the IG signal, and deriving a third physiological parameter based on the first physiological parameter and the second physiological parameter, wherein the first parameter, the second parameter, and the third parameter are different from each other, wherein the first parameter, the second parameter, and the third parameter are the same parameter, or wherein the first parameter and the second parameter are the same parameter, and the third parameter is a different parameter.
9. A system for measuring and / or monitoring a biosignal of a body part of a subject, the system comprising: a processor (220, 274, 442), and a device (100, 200, 201, 300), comprising: - a surface potential difference sensor (120), an electrogram EG, configured to measure an electrical activity of the body part and provide at least one EG signal indicative of the electrical activity of the body part, - a surface vibration sensor (160), a phonogram PG, configured to measure an acoustic activity of the body part and provide a PG signal indicative of the acoustic activity of the body part, and - an impedance sensor (140), an impedancegram IG, configured to measure an impedance of the body part and provide an IG signal indicative of the impedance of the body part, and a communication circuit (210, 272), configured to transmit the signals from the EG sensor, the PG sensor, and the IG sensor to the processor (220, 274, 442), wherein the processor (220, 274, 442) is configured to: Obtain the EG signal, the PG signal, and the IG signal from the communication circuit (210, 272), and jointly analyze the EG signal, the PG signal, and the IG signal to allow for the assessment of a predefined condition of the heart by deriving at least one data based on all of the EG signal, the PG signal, and the IG signal, wherein the joint analysis includes performing segment detection based on one or more of the EG signal, the PG signal, or the IG signal to detect segments of interest and analyzing other signals based on the detected segments of interest. Optionally, wherein performing segment detection includes segmenting the signal into predefined segments or predefined phases of predefined biological events, such as respiratory cycles, cardiac cycles, and / or circadian rhythm cycles. Preferably, wherein analyzing the other signals based on the segmentation includes identifying characteristic attributes in one or more of the other signals based on the detected segments, preferably, wherein the body part includes the heart and / or the lungs.
10. The system according to claim 9, wherein, The joint analysis includes: Deriving data from each of the signals, wherein the data derived from different signals includes redundancy therebetween, optionally, wherein the joint analysis further includes estimating the signal quality or confidence parameter of a signal selected from the EG signal, the PG signal, or the IG signal by comparing the data derived from a signal selected from the EG signal, the PG signal, or the IG signal with the data derived from one or more of the other signals as a reference signal. Preferably by: Determining reference timing data from the reference signal, determining the timing data of the selected signal, and comparing the timing data of the selected signal with the reference timing data from the reference signal, or by Estimating a first data based on a physiological parameter from the reference signal, estimating a second data based on the same physiological parameter from the selected signal, and determining a deviation value between the data estimated from the selected signal and the data estimated from the reference signal.
11. The system according to any one of claims 9 or 10, wherein, The joint analysis includes: Determining a first physiological parameter from one or more of the EG signal, the PG signal, or the IG signal, Determining a second physiological parameter from another or more of the EG signal, the PG signal, or the IG signal, and Deriving a third physiological parameter based on the first physiological parameter and the second physiological parameter, wherein the first parameter, the second parameter, and the third parameter are different from each other, wherein the first parameter, the second parameter, and the third parameter are the same parameter, or wherein the first parameter and the second parameter are the same parameter, and the third parameter is a different parameter.
12. The system according to any one of claims 9 to 11, wherein, A feature derived from one of the signals is derived based on another feature derived from another signal, and optionally, wherein the joint analysis further includes: Providing the derived multiple features to a detection and / or prediction model, wherein the detection and / or prediction model preferably corresponds to a machine learning model, and Generating a score or value indicating the probability of the predefined condition of the body part from the detection and / or prediction model.