Entropy-based heart sound tracking

Abstract

Systems and methods for identifying and tracking heart sound components are disclosed. An exemplary medical device system includes a data receiver circuit to receive heart sound information; and heart sound recognition circuitry to generate representative heart sound segments within a cardiac cycle, such as an ensemble average of a plurality of heart sound segments taken from a plurality of cardiac cycles. The heart sound recognition circuit can divide representative heart sound fragments into a plurality of heart sound data windows; respectively calculating a spectral entropy value for each of the plurality of heart sound data windows; and using the calculated spectral entropy value to determine one or more heart sound components including the S2 component. A physiological event detector may detect a cardiac event using the identified one or more heart sound components.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit of U.S. Provisional Application No. 63 / 432,649, filed on December 14, 2022, which is incorporated herein by reference in its entirety. Technical Field

[0003] The present invention relates generally to medical systems and, more particularly, to systems, devices, and methods for identifying and tracking heart sounds from a subject. Background Art

[0004] Heart sounds are generally associated with the mechanical vibrations of the heart and the flow of blood through it. Heart sounds recur with each cardiac cycle and are separated and categorized based on the activity associated with the vibrations. Historically, heart sounds were assessed manually, using only the audible portion of the vibrations. Devices are now capable of assessing the full spectrum of heart vibrations, including both audible and subaudible components, so the term "sound" in this disclosure refers to the full spectrum of vibrations. Typically, heart sounds sensed from a subject may include multiple components within the cardiac cycle, including the first heart sound (S1), second heart sound (S2), third heart sound (S3), or fourth heart sound (S4). S1 is associated with the vibrations produced by the heart during the tension of the mitral valve. S2 is produced by the closure of the aortic and pulmonary valves and marks the beginning of diastole. S3 is produced by the vibrations in early diastole, corresponding to passive ventricular filling during diastole when blood rushes into the ventricles. S4 is produced by the vibrations in late diastole, corresponding to active ventricular filling when the atria contract and push blood into the ventricles. In healthy subjects, S3 is usually weak and S4 is rarely audible. However, a pathological S3 or S4 may be higher-pitched and louder.

[0005] Heart sounds have been used to assess the systolic and diastolic function of the heart. Systole is the contraction of the heart, or a period of contraction, which causes blood to be forced out of the heart (such as the ventricles) and into the aorta and pulmonary artery. Diastole is the relaxation of the heart, or a period of relaxation, during which blood flows back into the heart (such as the ventricles). Patients with heart disease may have deteriorated systolic or diastolic function. For example, congestive heart failure (CHF) occurs when the heart is unable to supply enough blood to maintain a healthy physiological state.

[0006] Implantable medical devices (IMDs) have been used to monitor patients with heart disease, such as to detect cardiac events that lead to worsening heart failure (WHF). IMDs can sense physiological signals from patients and deliver electrical stimulation therapy to improve cardiac function in CHF patients. Frequent patient monitoring via IMDs can help identify patients at increased risk of developing future heart failure events, ensure timely treatment, reduce heart failure hospitalizations, improve patient prognosis, and reduce healthcare costs. Summary of the Invention

[0007] Ambulatory medical devices (AMDs), such as implantable medical devices (IMDs), subcutaneous medical devices, wearable medical devices or other external medical devices, can be used to monitor heart disease patients. AMDs can sense the electrical activity or mechanical activity of the heart via sensing electrodes and / or physiological sensors, and detect cardiac events such as arrhythmias or WHF. IMDs can include a pulse generator that can generate electrical stimulation therapy and deliver it to the heart or other excitable tissues (e.g., neural targets) to restore or improve cardiac function in CHF patients or to correct arrhythmias. For example, detection of arrhythmias can trigger cardiac pacing (pacing) or electric shocks, or detection of WHF events can trigger electrical stimulation therapy, such as resynchronization therapy (CRT) to correct cardiac asynchrony in heart failure patients.

[0008] AMD can use the heart sounds detected from the patient to detect cardiac events. For example, S1 and / or S2 can be used to detect arrhythmias, such as supraventricular tachycardia or ventricular tachycardia. Pulmonary fluid accumulation in CHF patients may lead to increased ventricular filling pressure and diastolic dysfunction, thereby triggering a pathologically louder S3. The powerful atrial contraction carried out to overcome the abnormally stiff ventricles in CHF patients can produce a strong S4. Therefore, monitoring S3 or S4 can help determine whether a patient has diastolic dysfunction, detect WHF events, or assess the risk of a patient developing future WHF.

[0009] Mobile heart sound detection involves placing a heart sound sensor at an extracutaneous, subcutaneous, submuscular, intramuscular, or substernal location at or near the heart. A heart sound sensor, such as an accelerometer, can be included within an IMD for implantation, or associated with an implantable lead for epicardial or endocardial placement. The S1 and S2 heart sounds typically have frequencies in the range of approximately 10-250 Hz. With wide interpersonal variability and frequency shifts secondary to technical equipment, S2 typically has a higher frequency than S1. For example, most of the S1 power typically falls within approximately 10-50 Hz, and most of the S2 power typically falls within approximately 20-70 Hz. The early diastolic sound, S3, produced by the rapid filling of the dilated ventricle, and the late diastolic sound, S4, produced by the contraction of the left atrium against the non-compliant left ventricle, when present, typically have lower intensity and lower frequency.

[0010] Heart sound components (e.g., S1, S2, S3, or S4) are conventionally detected using a time-domain, amplitude-based method that includes detecting the maximum signal amplitude or its variability within a heart sound detection window, such as the peak heart sound signal power of the heart sound signal or the root mean square (RMS) value of the heart sound signal. The timing of the maximum amplitude is then identified as the timing position of the heart sound component. However, there may often be multiple peaks within the heart sound signal of a single heart sound window, and the main peak (the peak with the largest amplitude) may not always represent the actual heart sound component. Therefore, the time-domain, amplitude-based method may sometimes mistakenly identify a pseudo-peak as the target heart sound component. On the other hand, the time-domain, amplitude-based method may be sensitive to electromagnetic or physiological noise or interference. At a low signal-to-noise ratio (SNR), the timing information of heart sound components such as S1 or S2 may not be accurately determined. This may cause further errors in the detection of other heart sound components (e.g., S3 or S4) or heart sound-based cardiac time intervals, which are measured with reference to the timing of the S1 or S2 components, such as the pre-ejection period (PEP, the interval between the onset of the QRS and S1), the systolic timing interval (STI, the interval between the onset of the QRS and S2), the left ventricular ejection time (LVET, the interval between S1 and S2), or the diastolic timing interval (DTI, the interval between S2 and the onset of the QRS of the next cardiac cycle). For cardiac event detection that depends on the timing of heart sounds and / or cardiac time intervals (e.g., WHF events), detection performance may be reduced. For at least these reasons, the inventors have recognized that, among other things, time-domain, amplitude-based methods may not be ideal identification means in at least some circumstances (such as at low SNRs), and that there remains an unmet need for more robust heart sound component identification and more reliable cardiac event detection using the identified heart sound components.

[0011] The present invention discusses systems, devices, and methods for determining and tracking cardiac sound components based on the spectral entropy of cardiac sound signals. An exemplary medical device system includes: a data receiver circuit for receiving cardiac sound information; and a cardiac sound recognition circuit for generating a representative cardiac sound segment within a cardiac cycle, such as an ensemble average of multiple cardiac sound segments taken from multiple cardiac cycles. The cardiac sound recognition circuit can divide the representative cardiac sound segment into multiple cardiac sound data windows; calculate a spectral entropy value for each of the multiple cardiac sound data windows; and use the calculated spectral entropy values ​​to determine one or more cardiac sound components including the S2 component. The medical device system may include a physiological event detector for detecting cardiac events using the determined one or more cardiac sound components.

[0012] Example 1 is a medical device system comprising: a data receiver circuit configured to receive cardiac sound information; and a cardiac sound recognition circuit configured to: generate a representative cardiac sound segment within a cardiac cycle using at least a portion of the received cardiac sound information; divide the representative cardiac sound segment into a plurality of cardiac sound data windows; calculate a spectral entropy value for each of the plurality of cardiac sound data windows; and determine one or more cardiac sound components using the calculated spectral entropy values, the one or more cardiac sound components including an S2 component.

[0013] In Example 2, the subject matter described in Example 1 optionally includes a heart sound recognition circuit, which can be configured to: divide a representative heart sound segment into multiple heart sound data windows using a sliding time window; generate a spectral entropy time series by concatenating the calculated spectral entropy values ​​according to the corresponding time series of the multiple heart sound data windows; and use the spectral entropy time series to determine one or more heart sound components, the one or more heart sound components including the S2 component.

[0014] In Example 3, the subject matter according to any one or more of Examples 1 to 2 optionally includes a portion of the received heart sound information, which may include multiple heart sound segments respectively taken from multiple cardiac cycles, wherein the heart sound identification circuit is configured to: time-align the multiple heart sound segments relative to corresponding reference points; and use an ensemble average of the multiple time-aligned heart sound segments to generate a representative heart sound segment.

[0015] In Example 4, the subject matter of Example 3 optionally includes a data receiver circuit that can be configured to receive cardiac electrical signals sensed simultaneously with cardiac sound information over multiple cardiac cycles, wherein the corresponding reference points include ventricular activation in multiple cardiac cycles of the simultaneously sensed cardiac electrical signals.

[0016] In Example 5, the subject matter according to any one or more of Examples 1 to 4 optionally includes a data receiver circuit that can be configured to receive information related to the instantaneous heart rate sensed simultaneously with the heart sound information over multiple cardiac cycles, wherein the portion of the received heart sound information over the multiple cardiac cycles used to generate the representative heart sound segment corresponds to substantially the same instantaneous heart rate or is within a predetermined heart rate range.

[0017] In Example 6, the subject matter of Example 2 may optionally include the heart sound identification circuit being configured to determine the S2 component based on a local minimum of a portion of the spectral entropy time series within the S2 detection window.

[0018] In Example 7, the subject matter of Example 6 optionally includes a data receiver circuit that can be configured to receive (i) a cardiac electrical signal and (ii) information about an instantaneous heart rate, both of which are sensed simultaneously with the heart sound information over multiple cardiac cycles, wherein the heart sound identification circuit is configured to use ventricular activation on the cardiac electrical signal and the instantaneous heart rate to determine the S2 detection window.

[0019] In Example 8, the subject matter according to any one or more of Examples 1 to 7 optionally includes a heart sound recognition circuit, which can be configured to: generate one or more heart rate or rhythm-dependent heart sound spectral entropy templates and store them in a memory, each heart rate or rhythm-dependent heart sound spectral entropy template including a spectral entropy time series of heart sounds at a corresponding heart rate or rhythm; and further use the one or more heart rate or rhythm-dependent heart sound spectral entropy templates to determine one or more heart sound components.

[0020] In Example 9, the subject matter described in Example 8 optionally includes a data receiver circuit that can be configured to receive information related to the instantaneous heart rate or heart rhythm sensed simultaneously with the heart sound information, wherein the heart sound identification circuit is configured to: select one of one or more stored heart rate or heart rhythm-dependent heart sound spectral entropy templates, which has a corresponding heart rate or heart rhythm that matches the instantaneous heart rate or heart rhythm; and use the selected stored heart rate or heart rhythm-dependent heart sound spectral entropy template to determine one or more heart sound components.

[0021] In Example 10, the subject matter described in Example 9 optionally includes a heart sound recognition circuit, which can be configured to: modify the spectral entropy time series of a representative heart sound segment using a selected stored heart rate or rhythm-dependent heart sound spectral entropy template; and determine one or more heart sound components using the modified spectral entropy time series of the representative heart sound segment.

[0022] In Example 11, the subject matter according to Example 10 optionally includes that, in order to modify the spectral entropy time series of the representative heart sound segment, the heart sound recognition circuit is configured to calculate: (i) the average of the spectral entropy time series of the representative heart sound segment and (ii) the spectral entropy time series in a selected stored heart rate or rhythm dependent heart sound spectral entropy template.

[0023] In Example 12, the subject matter according to any one or more of Examples 10 to 11 optionally includes a heart sound recognition circuit that can be configured to update a selected stored heart rate or rhythm dependent heart sound spectral entropy template in a memory using a modified spectral entropy time series of a representative heart sound segment.

[0024] In Example 13, the subject matter according to any one or more of Examples 1 to 12 optionally includes a heart sound recognition circuit that can be further configured to detect an S3 component or an S4 component from a representative heart sound segment based at least in part on the determined timing information of the S2 component.

[0025] In Example 14, the subject matter according to Example 13 optionally includes a heart sound recognition circuit, which can be further configured to: determine a confidence level of the determined S2 component; and detect the S3 component or the S4 component when the determined confidence level exceeds a threshold.

[0026] In Example 15, the subject matter according to any one or more of Examples 1 to 14 may optionally include a physiological event detector configured to detect a cardiac event using the determined one or more heart sound components.

[0027] Example 16 is a method for identifying cardiac sound components, comprising: receiving cardiac sound (HS) information; generating a representative cardiac sound segment within a cardiac cycle using at least a portion of the received cardiac sound information; dividing the representative cardiac sound segment into a plurality of cardiac sound data windows; calculating a spectral entropy value for each of the plurality of cardiac sound data windows; and using the calculated spectral entropy values ​​to determine one or more cardiac sound components, the one or more cardiac sound components including an S2 component.

[0028] In Example 17, the subject matter of Example 16 optionally includes receiving information related to the instantaneous heart rate sensed simultaneously with the heart sound information over multiple cardiac cycles, wherein the portion of the received heart sound information over the multiple cardiac cycles used to generate the representative heart sound segment corresponds to substantially the same instantaneous heart rate or is within a predetermined heart rate range.

[0029] In Example 18, the subject matter described in Example 17 optionally includes determining one or more heart sound components, which may include: generating a spectral entropy time series by concatenating the calculated spectral entropy values ​​according to the corresponding time series of multiple heart sound data windows; and determining one or more heart sound components based on a local minimum value of a portion of the spectral entropy time series within the S2 detection window, the one or more heart sound components including the S2 component.

[0030] In Example 19, the subject matter according to any one or more of Examples 17 to 18 optionally includes: generating one or more heart rate or rhythm-dependent heart sound spectral entropy templates and storing them in a memory, each heart rate or rhythm-dependent heart sound spectral entropy template including a spectral entropy time series of heart sounds at a specific heart rate or rhythm; and further using the one or more heart rate or rhythm-dependent heart sound spectral entropy templates to determine one or more heart sound components.

[0031] In Example 20, the subject matter described in Example 19 optionally includes: receiving information related to the instantaneous heart rate or rhythm sensed simultaneously with the heart sound information over multiple cardiac cycles; selecting one of one or more stored heart rate or heart rhythm-dependent heart sound spectral entropy templates having a corresponding heart rate or heart rhythm that matches the instantaneous heart rate or heart rhythm; and using the selected stored heart rate or heart rhythm-dependent heart sound spectral entropy template to determine one or more heart sound components.

[0032] In Example 21, the subject matter according to any one or more of Examples 19 to 20 optionally includes: generating a spectral entropy time series by concatenating the calculated spectral entropy values ​​according to the corresponding time series of multiple heart sound data windows; modifying the spectral entropy time series of a representative heart sound segment using a selected stored heart rate or rhythm-dependent heart sound spectral entropy template; determining one or more heart sound components using the modified spectral entropy time series of the representative heart sound segment; and updating the selected stored heart rate or rhythm-dependent heart sound spectral entropy template in the memory using the modified spectral entropy time series of the representative heart sound segment.

[0033] In Example 22, the subject matter according to any one or more of Examples 16 to 21 may optionally include detecting a cardiac event using the determined one or more heart sound components.

[0034] Heart sound recognition and tracking based on spectral entropy as described in the present disclosure can improve the functionality of mobile medical devices or medical diagnostic systems or devices that use heart sounds to detect physiological events (such as arrhythmia episodes or WHF events) or to evaluate cardiac function or diagnose cardiac conditions. Compared with conventional time-domain, amplitude-based heart sound recognition, the spectral entropy-based method quantifies the complexity (or non-uniformity) of the power spectrum amplitude of the heart sound signal or a portion thereof (such as within the cardiac cycle). High spectral entropy indicates that the signal energy is more evenly distributed across a wide frequency range in the frequency domain, while low spectral entropy indicates that the uniformity of the signal energy distribution is low. A broadband signal (such as a white noise signal) is characterized in that the signal energy is almost evenly distributed over a wide frequency range and therefore has a large spectral entropy. In contrast, most of the signal energy of a narrowband signal is concentrated in a narrow frequency band and is distributed less evenly and therefore has a smaller spectral entropy. Certain heart sound components (such as S1 and S2) are narrowband signals. For example, the S1 power typically falls between about 10-50 Hz, and the S2 power typically falls between about 20-70 Hz. Therefore, S1 and S2 typically have lower spectral entropy values. Heart sound detection based on spectral entropy as described herein combines the advantages of frequency domain spectrum analysis and Shannon entropy, both of which can be more robust to random noise and interference from various sources than time domain, amplitude-based heart sound detection methods. In particular, the spectral entropy-based method is more immune to occasional time domain high amplitude noise, which is more likely to cause false detection when using time domain, amplitude-based heart sound detection methods. As the detection accuracy of heart sound components and, in particular, the timing information of such heart sound components is improved, more reliable and efficient detection of cardiac events (such as WHF or arrhythmias) and improved patient prognosis can be achieved.

[0035] This disclosure provides part of the teachings of this application and is not intended to be exclusive or exhaustive of the subject matter. Further details about this subject matter are found in the detailed description and the appended claims. Other aspects of the present disclosure will be apparent to those skilled in the art upon reading and understanding the following detailed description and consulting the accompanying drawings, each of which should not be construed as limiting. The scope of the present disclosure is defined by the appended claims and their legal equivalents. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Various embodiments are shown by way of example in the illustrations of the accompanying drawings. Such embodiments are exemplary and are not intended to be exclusive or exhaustive of the present subject matter.

[0037] Figure 1 An example of a patient management system and portions of an environment in which the system may operate are generally shown.

[0038] Figure 2An example of a heart sound-based physiological event detection system configured to identify heart sound components and detect physiological events using the identified heart sound components is generally shown.

[0039] Figure 3A-Figure 3B An example of a heart sound recognition circuit that identifies heart sound components from a spectral entropy time series is shown.

[0040] Figure 4 An example of generating a spectral entropy time series from a representative heart sound segment and detecting heart sound components therefrom is generally shown.

[0041] Figure 5 An overlay of multiple spectral entropy time series is shown, each spectral entropy time series being generated from a corresponding representative heart sound segment.

[0042] Figures 6A-6D Examples of heart sound tracking based on spectral entropy at different signal-to-noise ratios (SNRs) of heart sound signals are shown.

[0043] Figure 7 An example of detecting worsening heart failure (WHF) based at least in part on spectral entropy-based identification and tracking of the S3 heart sound is generally shown.

[0044] Figure 8 An example of a method for identifying and tracking heart sound components from a heart sound signal and detecting physiological events using the identified heart sound components is generally shown.

[0045] Figure 9 A block diagram generally illustrates an example machine upon which any one or more of the techniques (eg, methodologies) discussed herein may be performed. DETAILED DESCRIPTION

[0046] Disclosed herein are systems, devices, and methods for tracking cardiac sound components (such as S1, S2, S3, or S4 components) based on the spectral entropy of cardiac sound signals. An exemplary medical device system includes: a data receiver circuit for receiving cardiac sound information, and a cardiac sound recognition circuit for generating a representative cardiac sound segment within a cardiac cycle, such as an ensemble average of multiple cardiac sound segments taken from multiple cardiac cycles. The cardiac sound recognition circuit can divide the representative cardiac sound segment into multiple cardiac sound data windows; calculate a spectral entropy value for each of the multiple cardiac sound data windows; and use the calculated spectral entropy values ​​to determine one or more cardiac sound components including the S2 component. A physiological event detector can use the identified one or more cardiac sound components to detect cardiac events.

[0047] Figure 1An example patient management system 100 and portions of an environment in which the patient management system 100 can operate are generally shown. The patient management system 100 can perform a range of activities, including remote patient monitoring and diagnosis of disease conditions. Such activities can be performed proximately to the patient 101, such as in the patient's home or office; through a central server, such as in a hospital, clinic, or doctor's office; or through a remote workstation, such as a secure wireless mobile computing device.

[0048] The patient management system 100 may include one or more mobile medical devices, an external system 105, and a communication link 111 that provides for communication between the one or more mobile medical devices and the external system 105. The one or more mobile medical devices may include an implantable medical device (IMD) 102, a wearable medical device (WMD) 103, or one or more other implantable, leadless, subcutaneous, external, wearable, or mobile medical devices configured to monitor, sense, or detect information from the patient 101, determine physiological information about the patient, or provide one or more therapies to treat various conditions of the patient, such as one or more cardiac or non-cardiac conditions (e.g., dehydration, sleep-disordered breathing, etc.).

[0049] In one example, IMD 102 may include one or more conventional cardiac rhythm management devices implanted in the patient's chest, having a lead system including one or more transvenous, subcutaneous, or non-invasive leads or catheters to position one or more electrodes or other sensors (e.g., heart sound sensors) in, on, or around the heart of patient 101 or at one or more other locations in the patient's chest, abdomen, or neck. In another example, IMD 102 may include, for example, a monitor implanted subcutaneously in the chest of patient 101, including a housing containing circuitry and, in some examples, one or more sensors such as a temperature sensor.

[0050] IMD 102 may include evaluation circuitry configured to detect or determine specific physiological information of patient 101, or configured to determine one or more conditions or provide information or alerts to a user, such as patient 101 (e.g., a patient), a clinician, or one or more other caregivers or processes, such as described herein. In an example, IMD 102 may be an implantable cardiac monitor (ICM) configured to collect cardiac information (optionally along with other physiological information) from a patient. IMD 102 may alternatively or additionally be configured as a therapeutic device configured to treat one or more medical conditions of patient 101. Therapy may be delivered to patient 101 via a lead system and associated electrodes or using one or more other delivery mechanisms. Therapy may include delivering one or more medications to patient 101, such as using IMD 102 or one or more other ambulatory medical devices. In some examples, therapy may include cardiac resynchronization therapy for correcting dyssynchrony and improving cardiac function in a patient with heart failure. In other examples, IMD 102 may include a drug delivery system, such as a drug infusion pump, for delivering medication to a patient for managing arrhythmias or complications resulting from arrhythmias, hypertension, hypotension, or one or more other physiological conditions. In other examples, IMD 102 may include one or more electrodes configured to stimulate the patient's nervous system or provide stimulation to muscles in the patient's airway, etc.

[0051] The WMD 103 may include one or more wearable or external medical sensors or devices (e.g., an automatic external defibrillator (AED), a Holter monitor, a patch-based device, a smart watch, a smart accessory, a wrist-worn or finger-worn medical device such as a finger-based photoplethysmography sensor, etc.).

[0052] In an example, the IMD 102 or WMD 103 may include or be coupled to an implantable or wearable sensor for sensing heart sound signals, and include a heart sound recognition circuit for identifying one or more heart sound components (such as S1, S2, S3, or S4) based on a spectral entropy time series derived from the sensed heart sound signals. Also included in the IMD 102 or WMD 103 is a heart sound-based event detector circuit that can detect physiological events (e.g., an arrhythmia episode, or a heart failure exacerbation (WHF) event) based at least on heart sound metrics of the detected one or more heart sound components. Examples of such heart sound metrics may include the amplitude or timing of the heart sound components relative to a fiducial point within the cardiac cycle. In some examples, at least a portion of the heart sound recognition circuit and / or the heart sound-based event detector circuit can be implemented in and by the external system 105.

[0053] External system 105 may include a dedicated hardware / software system, such as a programmer, a remote server-based patient management system, or alternatively, a system primarily defined by software running on a standard personal computer. External system 105 may manage patient 101 through IMD 102 or one or more other mobile medical devices connected to external system 105 via communication link 111. In other examples, IMD 102 may be connected to WMD 103 via communication link 111, or WMD 103 may be connected to external system 105 via communication link 111. For example, this may include programming IMD 102 to perform one or more of acquiring physiological data, performing at least one self-diagnostic test (such as a self-diagnostic test for device operating status), analyzing physiological data, or optionally delivering or adjusting therapy for patient 101. Additionally, external system 105 may send information to or receive information from IMD 102 or WMD 103 via communication link 111. Examples of information may include real-time or stored physiological data from patient 101, diagnostic data (such as detection of patient hydration status, hospitalization status, response to therapy delivered to patient 101), or device operating status of IMD 102 or WMD 103 (e.g., battery status, lead impedance, etc.). Communication link 111 may be an inductive telemetry link, a capacitive telemetry link, or a radio frequency (RF) telemetry link, or wireless telemetry based on, for example, "strong" Bluetooth or IEEE 802.11 Wireless Fidelity "Wi-Fi" interface standards. Other configurations and combinations of patient data source interfaces are also possible.

[0054] The external system 105 may include an external device 106 located near one or more mobile medical devices and a remote device 108 located relatively remote from the one or more mobile medical devices and communicating with the external device 106 via a communication network 107. Examples of external device 106 may include a medical device programmer. Among other possible functions, the remote device 108 may be configured to evaluate the collected patient or patient information and provide alert notifications. In an example, the remote device 108 may include a centralized server that serves as a central hub for storing and analyzing the collected data. The server may be configured as a single, multiple, or distributed computing and processing system. The remote device 108 may receive data from multiple patients. This data may be collected by one or more mobile medical devices in addition to other data acquisition sensors or devices associated with the patient 101. The server may include a memory device for storing the data in a patient database. The server may include alert analyzer circuitry for evaluating the collected data to determine whether specific alert conditions are met. Meeting the alert conditions may trigger the generation of an alert notification, such as one or more human-perceivable user interfaces. In some examples, the alert condition may alternatively or additionally be evaluated by one or more mobile medical devices, such as implantable medical devices. By way of example, the alert notification may include a web page update, a phone call or pager, an email, SMS, a text or "instant" message, and a message to the patient and a direct notification to emergency services and clinicians simultaneously. Other alert notifications are also possible. The server may include an alert prioritizer circuit configured to prioritize the alert notifications. For example, alerts for a detected medical event may be prioritized using a similarity metric between physiological data associated with the detected medical event and physiological data associated with historical alerts.

[0055] The remote device 108 may additionally include one or more locally configured clients or remote clients securely connected to the server via the communication network 107. Examples of clients may include personal desktop computers, laptop computers, mobile devices, or other computing devices. System users (such as clinicians or other qualified medical professionals) may use the client to securely access stored patient data compiled in a database in the server and select and prioritize patients and alerts for healthcare provision. In addition to generating alert notifications, the remote device 108 (including the server and interconnected clients) may also execute a follow-up plan by sending a follow-up request to one or more mobile medical devices, or by sending a message or other communication as a compliance notification to the patient 101 (e.g., the patient), the clinician, or an authorized third party.

[0056] The communication network 107 can provide wired or wireless interconnectivity. In an example, the communication network 107 can be based on the Transmission Control Protocol / Internet Protocol (TCP / IP) network communication specification, although other types or combinations of networking implementations are also possible. Similarly, other network topologies and arrangements are also possible.

[0057] One or more of the external device 106 or the remote device 108 can output the detected physiological events to a system user, such as a patient or clinician, or to a process, such as an instance of a computer program executable in a microprocessor. In an example, the process can include automatically generating a recommendation for anti-arrhythmic therapy or a recommendation for further diagnostic testing or treatment. In an example, the external device 106 or the remote device 108 can include a corresponding display unit for displaying physiological or functional signals, or for issuing an alert, alarm, emergency call, or other form of warning that a cardiac arrhythmia has been detected. In some examples, the external system 105 can include an external data processor configured to analyze physiological or functional signals received by one or more mobile medical devices and confirm or deny the detection of an arrhythmia. Computationally intensive algorithms (such as machine learning algorithms) can be implemented in the external data processor to retrospectively process data to detect arrhythmias.

[0058] Parts of one or more mobile medical devices or external systems 105 can be implemented using hardware, software, firmware or a combination thereof. Parts of one or more mobile medical devices or external systems 105 can be implemented using a dedicated circuit that can be constructed or configured to perform one or more functions, or can be implemented using a general-purpose circuit that can be programmed or otherwise configured to perform one or more functions. Such general-purpose circuits can include a microprocessor or a portion thereof, a microcontroller or a portion thereof, or a programmable logic circuit, a memory circuit, a network interface, and various components for interconnecting these components. For example, among other aspects, a "comparator" can include an electronic circuit comparator that can be constructed to perform a specific function of comparing two signals, or a comparator can be implemented as a part of a general-purpose circuit that can be driven by a code that instructs a part of a general-purpose circuit to perform a comparison between two signals. A "sensor" can include an electronic circuit that is configured to receive information and provide an electronic output representing such received information.

[0059] The therapeutic device 110 can be configured to send information to or receive information from one or more mobile medical devices or an external system 105 using a communication link 111. In an example, one or more mobile medical devices, external device 106, or remote device 108 can be configured to control one or more parameters of the therapeutic device 110. The external system 105 can allow programming of one or more mobile medical devices and can receive information about one or more signals acquired by the one or more mobile medical devices, such as information that can be received via the communication link 111. The external system 105 can include a local external implantable medical device programmer. The external system 105 can include a remote patient management system that can monitor the patient's status or adjust one or more treatments, such as from a remote location.

[0060] Figure 2 Generally, an example of a heart sound-based physiological event detection system 200 is shown that can identify and track heart sound components and use the identified heart sound components to detect physiological events, such as an arrhythmia episode or a heart failure exacerbation (WHF) event. The physiological event detection system 200 can include one or more of a data receiver circuit 210, a control circuit 220, a user interface 230, and a therapy circuit 240. At least a portion of the physiological event detection system 200 can be implemented in an IMD 102, a WMD 103, or an external system 105, such as one or more of an external device 106 or a remote device 108.

[0061] The data receiver circuit 210 can receive physiological information from the patient. In an example, the data receiver circuit 210 can include a sense amplifier circuit configured to sense physiological signals from the patient via a physiological sensor (such as an implantable, wearable, or otherwise mobile sensor or electrode associated with the patient). The sensor can be incorporated into a mobile device (such as the IMD 102 or WMD 103) or otherwise associated therewith. In some examples, the physiological signals sensed from the patient can be stored in a storage device, such as an electronic medical record (EMR) system. The data receiver circuit 210 can receive the physiological signals from the storage device, such as in response to a user command or a trigger event. By way of example and not limitation, Figure 2As shown, the data receiver circuit 210 may include a heart sound sensing circuit 212, a heart electrical signal sensing circuit 214, and a heart rate circuit 216. The heart sound sensing circuit 212 may receive heart sound information, such as a heart sound signal sensed from a patient. In an example, the heart sound sensing circuit 212 may be coupled to a heart sound sensor to sense body motion / vibration signals indicative of heart vibrations, which are associated with or indicative of heart sounds. The heart sound sensor may take the form of an accelerometer, an acoustic sensor, a microphone, a piezoelectric-based sensor, or other vibration or acoustic sensor. The accelerometer may be a single-axis, dual-axis, or three-axis accelerometer. Examples of accelerometers may include a flexible piezoelectric crystal (e.g., quartz) accelerometer or a capacitive accelerometer manufactured using microelectromechanical systems (MEMS) technology. The heart sound sensor may be included in the IMD 102 or WMD 103, or disposed on a lead, such as part of a lead system associated with the IMD 102 or WMD 103. In an example, an accelerometer (or other sensor type) can sense an epicardial or endocardial acceleration (EA) signal from a portion of the heart, such as on the endocardial or epicardial surface of one of the left ventricle, right ventricle, left atrium, or right atrium. The EA signal can include components corresponding to various heart sound components, such as one or more of S1, S2, S3, or S4.

[0062] The cardiac electrical signal sensing circuit 214 can sense cardiac electrical signals. Examples of cardiac electrical signals can include a surface electrocardiogram (ECG) sensed from electrodes placed on the body surface, a subcutaneous ECG sensed from electrodes placed subcutaneously, and an intracardiac electrogram (EGM) sensed from one or more implantable electrodes. The sensing electrodes can be included in the IMD 102 or WMD 103, or communicatively coupled thereto. The heart rate circuit 216 can detect the patient's heart rate, such as from a signal sensed by the cardiac electrical signal sensing circuit 214. Alternatively, the heart rate (or pulse rate) can be detected from a cardiac mechanical signal. As will be discussed further below, one or more of the cardiac electrical signal or heart rate information can be used by the control circuit to construct a representative heart sound segment from which a spectral entropy time series can be derived.

[0063] In some examples, the data receiver circuit 210 can receive other physiological or functional signals, including, for example, body activity signals, posture signals, chest or cardiac impedance signals, arterial pressure signals, pulmonary artery pressure signals, left atrial pressure signals, RV pressure signals, LV coronary pressure signals, coronary blood temperature signals, blood oxygen saturation signals, heart sound signals, physiological responses to activity, apnea-hypopnea index, one or more respiratory signals (such as respiratory rate signals or tidal volume signals), brain natriuretic peptide (BNP), blood tests, sodium and potassium levels, blood glucose levels, and other biomarkers and biochemical markers.

[0064] The control circuit 220 can identify and track heart sound components and detect physiological events based on at least one of the heart sound components. The control circuit 220 can be implemented as part of a microprocessor circuit, which can be a dedicated processor such as a digital signal processor, an application-specific integrated circuit (ASIC), a microprocessor, or other type of processor for processing information including physical activity information. Alternatively, the microprocessor circuit can be a general-purpose processor that can receive and execute an instruction set for performing the functions, methods, or techniques described herein.

[0065] Control circuitry 220 may include a circuit set that includes one or more other circuits or subcircuits, such as heart sound recognition circuitry 222 and physiological event detector 224. These circuits may perform the functions, methods, or techniques described herein, either individually or in combination. In an example, the hardware of the circuit set may be immutably designed to perform a specific operation (e.g., hardwired). In an example, the hardware of the circuit set may include variably connected physical components (e.g., execution units, transistors, simple circuits, etc.) that include a computer-readable medium that is physically modified (e.g., magnetically, electrically, by removable placement of constant mass particles, etc.) to encode instructions for a specific operation. When the physical components are connected, the underlying electrical properties of the hardware components are changed, for example, from an insulator to a conductor or vice versa. The instructions enable embedded hardware (e.g., an execution unit or a loading mechanism) to create a member of the circuit set in hardware via the variably connected components to perform a portion of a specific operation when in operation. Thus, when the device is operating, the computer-readable medium is communicatively coupled to the other components of the circuit set members. In an example, any one of the physical components may be used in more than one member of more than one circuit set. For example, in operation, an execution unit may be used in a first circuit of a first circuit set at one point in time and reused at a different time by a second circuit in the first circuit set or by a third circuit in the second circuit set.

[0066] The heart sound identification circuit 222 can pre-process the heart sound signals received from the heart sound sensing circuit 212. In an example, the heart sound identification circuit 222 can include a filter or filter bank to remove or attenuate one or more of low-frequency signal baseline wander, high-frequency noise, or other unwanted frequency content. In an example, the heart sound segments can be bandpass filtered to a frequency range of approximately 5-90 Hz or approximately 9-90 Hz. In an example, the filter can include a second-order or higher-order differentiator configured to calculate a second-order or higher-order differential of the heart sound signal.

[0067] Heart sound identification circuit 222 can identify heart sound components from the preprocessed heart sound signal, optionally also using the heart electrical signal and heart rate information provided by heart electrical signal sensing circuit 214 and heart rate circuit 216, respectively. Heart sound components can be identified using a spectral entropy time series of representative segments of the heart sound signal. Figure 3A-Figure 3B An embodiment of the heart sound identification circuit 222 is shown by way of example and not limitation, such as Figure 3A and Figure 3B The heart sound recognition circuits 322A and 322B are shown respectively. Figure 3A As shown, the heart sound identification circuit 322A includes a heart sound ensemble circuit 323, a spectral entropy circuit 326, and a heart sound component detector 328. The heart sound ensemble circuit 323 can generate a heart sound ensemble that includes multiple segments of the received heart sound signal, such as heart sound segments within a corresponding multiple cardiac cycles. In an example, the heart sound ensemble circuit 323 can use a cardiac electrical signal (received from the cardiac electrical signal sensing circuit 214) that is sensed simultaneously with the heart sound signal. The heart sound ensemble circuit 323 can detect a cardiac cycle as a time interval between two consecutive cardiac activations (e.g., R waves or QRS complexes) on an ECG or intracardiac EGM signal. The heart sound ensemble circuit 323 can align the multiple heart sound segments relative to corresponding reference points for multiple cardiac cycles. The reference points can include, for example, ventricular activations (e.g., R waves) in multiple cardiac cycles of the ECG or EGM signal.

[0068] In some examples, the heart sound ensemble circuit 323 can filter the received heart sound signals, such as heart sound segments in multiple cardiac cycles, and select a subset of heart sound segments that meet specified criteria to form a heart sound ensemble. The screening criteria can include one or more of the patient's activity level, heart rate, respiratory rate, or time of day, among other conditions. In an example, the heart sound ensemble circuit 323 can select a subset of heart sound segments whose corresponding heart rates or cycle lengths fall within a specified range, such as substantially the same cycle length within a margin of + / - 100 milliseconds, or substantially the same heart rate within a margin of + / - 5 beats per minute. In another example, the heart sound ensemble circuit 323 can select a subset of heart sound segments whose corresponding physical activity levels fall within a specified range, such as substantially the same physical activity levels within a specified margin, or select a subset of heart sound segments whose corresponding respiratory rates fall within a specified range, such as substantially the same respiratory rates within a specified margin. In yet another example, the heart sound signals can be sensed during a specified time period of a day. The heart sound ensemble circuit 323 may select a subset of heart sound segments that are sensed during substantially the same time period of a day within a specified margin.

[0069] Heart sound ensemble circuitry 323 may align the selected subsets of heart sound signal segments to form a heart sound ensemble. The alignment may be relative to an R-wave on an ECG signal sensed simultaneously with the heart sound within the corresponding cardiac cycle. Heart sound ensemble circuitry 323 may use the aligned subsets of the selected heart sound signal segments to generate a representative heart sound segment within the cardiac cycle. In an example, the representative heart sound segment may be an ensemble average of the aligned subsets of the selected heart sound signal segments.

[0070] The spectral entropy circuit 326 can use representative heart sound segments to generate a spectral entropy time series. Spectral entropy quantifies the complexity (or non-uniformity) of the power spectrum amplitude of a heart sound segment in the cardiac cycle. A high spectral entropy indicates that the signal energy is more evenly distributed across a wide frequency range in the frequency domain, while a low spectral entropy indicates that the uniformity of the signal energy distribution is low. The signal energy of a broadband signal is almost evenly distributed over a wide frequency range and therefore has a large spectral entropy. In contrast, most of the signal energy of a narrowband signal is concentrated in a narrow frequency band and therefore has a smaller spectral entropy. In order to determine the spectral entropy time series, the spectral entropy circuit 326 can apply a sliding time window to a representative (e.g., ensemble-averaged) heart sound segment and generate multiple heart sound data windows; calculate a corresponding spectral entropy value for each of the multiple heart sound data windows; and concatenate the calculated spectral entropy values ​​at the corresponding time sequence of the heart sound data windows to form a spectral entropy time series.

[0071] The heart sound component detector 328 can identify one or more heart sound components from the spectral entropy time series. Some heart sound components (such as S1 and S2) are narrowband signals. For example, the S1 power falls roughly within approximately 10-50 Hz, and the S2 power falls roughly within approximately 20-70 Hz. Such heart sound components generally have corresponding spectral entropy values ​​that are lower than other parts of the heart sound signal. The heart sound component detector 328 can identify S1 based on a local minimum of a portion of the spectral entropy time series within the S1 detection window. Similarly, the heart sound component detector 328 can identify S2 based on a local minimum of a portion of the spectral entropy time series within the S2 detection window. The S1 detection window and the S2 detection window can have user-programmable positions (e.g., the start or end of the detection window) and window lengths. In an example, the S1 detection window can start 50 milliseconds (msec) after the detected R wave on the ECG signal and have a duration of 300 msec. The S2 detection window can start at a specified offset after the detected R wave or S1 heart sound. In some examples, the position and length of the S1 detection window and the S2 detection window can be determined using the spectral entropy time series of the heart sound signals collected from the patient population. In some examples, the heart sound component detector 328 can compare the spectral entropy time series with a predetermined threshold and detect local minima from the portion of the spectral entropy time series that is below the predetermined threshold (which may correspond to S1 and / or S2). S1 and S2 can be differentially identified based on the corresponding timing of the detected local minima. Figure 4 and Figure 5 Examples of generating spectral entropy time series and using them to detect heart sound components are discussed.

[0072] The heart sound component detector 328 can detect other heart sound components, such as S3 or S4 components, from the spectral entropy time series using the methods described above for detecting S1 and / or S2 components. Alternatively or additionally, the heart sound component detector 328 can detect S3 or S4 components based on the detected S1 and / or S2 components. For example, the heart sound component detector 328 can detect S3 within an S3 detection window using the spectral entropy time series or the heart sound energy signal. The S3 detection window can be defined based on the R wave timing or the timing of the detected S2. In an example, the S3 detection window can begin at the S2 timing or at a specified offset after the detected S2 (e.g., approximately 50-125 msec after the S2 timing). The S3 detection window can have a duration of approximately 125 msec. In some examples, the offset or S3 window duration can be a function of a physiological variable, such as heart rate. For example, the offset can be inversely proportional to the heart rate, such that the S3 detection window can begin at a smaller offset after S2 at higher heart rates. In another example, the heart sound component detector 328 can detect S4 within an S4 detection window using a spectral entropy time series or a heart sound energy signal. Because S4 typically occurs temporally after S3 and before S1 of the next cardiac cycle, the S4 detection window can be defined as starting at S3 or at a specified offset after S3 and ending at the R wave or S1 heart sound of the next cardiac cycle.

[0073] In some examples, the heart sound component detector 328 can determine a confidence level for the identification of a heart sound component, such as a confidence level for the S1 timing or S2 timing (relative to a reference point, such as an R wave in the same cardiac cycle). In an example, the confidence level can be determined based on the spectral entropy amplitude in the S1 detection window or the S2 detection window relative to a spectral entropy floor (such as a spectral entropy value for a portion of the heart sound signal outside the S1 detection window and the S2 detection window). If the spectral entropy amplitude in the S1 detection window or the S2 detection window is below the spectral entropy floor by more than a predetermined margin, a high confidence level for the identification of S1 or S2 can be determined. Conversely, if the spectral entropy amplitude in the S1 detection window or the S2 detection window is within a predetermined margin of the spectral entropy floor, a low confidence level for the identification of S1 or S2 can be determined. In some cases, abnormal conduction or arrhythmias (e.g., bigeminy and tripeminy) can cause high variability in heart sounds even at substantially the same heart rate due to the delayed hemodynamic effects of the previous beat. Thus, when such an arrhythmia or abnormal conduction is detected, a lower confidence in the timing of S1 or S2 can be determined. In some examples, the timing of S1 or S2 can be tracked over time, and a consistent pattern of spectral entropy troughs over time can indicate a high confidence in S1 or S2, while a less consistent pattern may suggest a low confidence. The heart sound component detector 328 can adjust the detection of other heart sound components (e.g., S3 or S4) based on the confidence in the detection of S1 or S2. For example, the heart sound component detector 328 can detect S3 or S4 in the same cardiac cycle (as the detected S2) only when the confidence in the timing of S2 exceeds a threshold.

[0074] Figure 3B A heart sound recognition circuit 322B is shown, which also includes a heart sound ensemble circuit 323, a spectral entropy circuit 326, and a heart sound component detector 328. The heart sound recognition circuit 322B may additionally include a template matching circuit 325, which is communicatively coupled to a storage device (e.g., a memory) that stores a database 330 of one or more heart rate or rhythm dependent heart sound spectral entropy templates. In an example, the heart rate dependent heart sound spectral entropy templates may be generated using heart sound data from a patient population (such as a plurality of patients with similar medical conditions or demographic characteristics). Each of the templates may include one or more of the following: a spectral entropy time series SE T, spectral entropy values ​​of heart sound components (such as S1 and S2), or the timing of S1 and S2 relative to a reference point (such as an R wave), all of which can be determined at a specific heart rate or heart rate range. The heart rate-dependent heart sound spectral entropy template can be constructed as a lookup table or an association map, in which a correspondence between the heart rate (or heart rate range) and the corresponding template information (such as the spectral entropy time series, the S1 and S2 timing, etc.) is established. The template matching circuit 325 can receive information about the instantaneous heart rate sensed simultaneously with the heart sound information over multiple cardiac cycles from the heart rate circuit 216, and query the database 330 to identify a heart rate matching template having a corresponding heart rate or heart rate range that matches the instantaneous heart rate sensed simultaneously from the stored heart rate-dependent HS spectral entropy templates. The heart rate matching template may include one or more of the following: a spectral entropy time series SE for heart rate matching T (HR), heart rate-matched spectral entropy values ​​of heart sound components (such as S1 and S2), or heart rate-matched S1 and S2 timings.

[0075] The heart sound component detector 328 can use the heart rate matching template to detect the heart sound component. In an example, the heart sound component detector 328 can detect the heart sound component using only the heart rate matching template. For example, instead of using the representative heart sound segment (generated by the heart sound ensemble circuit 323) to identify S1 or S2, the heart sound component detector 328 can directly identify S1 or S2 from the heart rate matching template (or retrieve information about it). If the heart rate matching template stores the S1 or S2 timing of the heart rate match, the timing can be retrieved and used as an approximation of the S1 or S2 timing. This approximation using previously stored S1 and S2 timing information can reduce computing time and improve the efficiency of S1 or S2 tracking. In some examples, the heart sound component detector 328 can additionally detect other heart sound components such as S3 from the representative heart sound segment based on at least the S1 or S2 timing derived or retrieved from the heart rate matching template.

[0076] In another example, the spectral entropy circuit 326 may apply a sliding time window to the representative heart sound segment to generate a plurality of heart sound data windows, and generate a spectral entropy time series SE, as described above with reference to FIG. Figure 3A The spectral entropy circuit 326 can then generate the heart rate matched spectral entropy time series SE stored in the heart rate matched spectral entropy template. T (HR) to modify the spectral entropy time series SE. Examples of the modified SE, SE′ may include (i) the spectral entropy time series SE of the representative heart sound segment and (ii) the heart rate-matched spectral entropy time series SE stored in the heart rate-matched spectral entropy template T (HR) The heart sound component detector 328 can then use the same Figure 3A The described technique detects heart sound components from the modified spectral entropy time series SE'.

[0077] The template updating circuit 340 can use the modified spectral entropy time series SE' to update the heart rate matching spectral entropy template in the database 330. For example, the heart rate matching spectral entropy time series SE T (HR), which is part of the heart rate matched spectral entropy template stored in the database 330, may be replaced by the modified spectral entropy time series SE'. The updated spectral entropy template may be stored in the database 330 for future use.

[0078] In some examples, the database 330 may store one or more rhythm-dependent heart sound spectral entropy templates. Unlike the heart rate-dependent heart sound spectral entropy templates generated using heart sound data at a specific heart rate or heart rate range (e.g., from a patient group with similar medical conditions or demographic characteristics), the heart rhythm-dependent heart sound spectral entropy templates may be generated using heart sound data from a patient group when the patient experiences a specific type of heart rhythm (such as atrial fibrillation, atrial flutter, ventricular premature beats, bigeminy, tripeminy, ventricular arrhythmia, etc.). The heart rhythm-dependent heart sound spectral entropy templates may be constructed as a lookup table or an association map. Figure 3B Similar to the described heart rate-based template matching, in an example, the template matching circuit 325 can receive information about the heart rhythm during the period of collecting heart sound data from the data receiver circuit 210, or determine the heart rhythm using a heart signal sensed simultaneously with the heart sound information received from the heart electrical signal sensing circuit 214. The template matching circuit 325 can query the database 330 to identify a heart rhythm matching template having a corresponding heart rhythm that matches the heart rhythm type of the simultaneously sensed heart signal from the stored heart rhythm-dependent HS spectral entropy templates. The heart sound component detector 328 can then use the heart rhythm matching template to detect the heart sound component in a manner similar to that described above with respect to using the heart rate matching template.

[0079] Return Reference Figure 2, the physiological event detector 224 can use one or more of the detected heart sound components to detect a physiological event. In some examples, the physiological event detector 224 can use the detected heart sound components to generate a heart sound metric, and use the heart sound metric, optionally together with other physiological information obtained from the patient, to detect a physiological event. Examples of heart sound metrics can include the intensity of the heart sound component (e.g., amplitude or signal energy under the curve) or one or more heart sound-based cardiac timing intervals, such as the pre-ejection period (PEP) (such as measured between the onset of the QRS and the S1 heart sound), the systolic timing interval (STI) (such as measured between the onset of the QRS complex and the S2 heart sound on the ECG), the left ventricular ejection time (LVET) (such as measured as the interval between the S1 heart sound and the S2 heart sound), or the diastolic timing interval (DTI) (such as between the S2 heart sound and the onset of the subsequent QRS complex on the ECG). These heart sound-based cardiac timing intervals can be related to the cardiac contractility or diastolic function of the heart. Heart sound metrics may also include a PEP / LVET ratio, an STI / DTI ratio, an STI / cycle length (CL) ratio, or a DTI / CL ratio or other composite metrics.

[0080] In an example, the physiological event detector 224 can track heart sound components over time and generate heart sound trends, such as S1 amplitude trends, S2 amplitude trends, S1 timing trends, S2 timing trends, or trends of heart sound-based cardiac timing intervals (e.g., PEP trends, STI trends, LVET trends, DEI trends, etc.). Figures 6A-6D An example of tracking heart sound components over time is further described. Based on the heart sound trends, the physiological event detector 224 can generate a cardiac function indicator that indicates myocardial contractility, cardiac synchrony, and cardiac hemodynamics. In an example, the physiological event detector 224 can use one or more of the measures of S1 intensity, S2 intensity, or STI to detect arrhythmia episodes, or distinguish different arrhythmias (e.g., atrial tachycardia arrhythmias, supraventricular tachycardia arrhythmias, or ventricular tachycardia arrhythmias). For example, a decrease in S1 intensity can indicate decreased cardiac contractility, and a decrease in S2 intensity can indicate decreased cardiac output, both of which can be used to detect arrhythmias and deterioration of cardiac hemodynamics during arrhythmias. In another example, a heart sound metric (such as S3 intensity) can be used to detect WHF. An increase in S3 intensity indicates decreased ventricular compliance and deteriorated diastolic function, thereby indicating the occurrence of WHF. Additionally or alternatively, a decrease in S1 strength or a decrease in STI may indicate poor cardiac contractility or reduced electromechanical coupling, which is indicative of the occurrence of WHF. Figure 7An example of detecting WHF based on heart sound components identified and tracked using a spectral entropy based approach is further described.The physiological event detector 224 may additionally or alternatively detect respiratory, renal, neurological, etc. medical conditions based on heart sound metrics.

[0081] The user interface 230 may include an input unit and an output unit. In an example, at least a portion of the user interface 230 may be implemented in the external system 105. The input unit may receive user input for programming the data receiver circuit 210 and the control circuit 220, such as for sensing heart sound signals, calculating spectral entropy, detecting heart sound components, generating heart sound metrics, or detecting parameters for physiological events. The input unit may include a keyboard, on-screen keyboard, mouse, trackball, touchpad, touch screen, or other pointing or navigation device. The output unit may include a display for displaying, among other things, sensed heart sound signals, representative heart sound segments, spectral entropy time series, heart sound metrics, information about detected physiological events, and any intermediate measurements or calculations. The output unit may also present a therapeutic titration protocol and recommended therapy to the user, such as via a display unit, including changes in parameters in the therapy provided by the implanted device, prescriptions for obtaining the implanted device, initiation or changes in medication therapy, or other treatment options for the patient. The output unit may include a printer for printing a hard copy of the information that may be displayed on the display unit. The signals and information can be presented in a table, chart, diagram, or any other type of textual, tabular, or graphical presentation format. The presentation of output information can include audio or other media formats. In an example, the output unit can generate an alert, alarm, emergency call, or other form of warning to signal the system user about the detected medical event.

[0082] The therapy circuit 240 can be configured to deliver therapy to the patient, such as in response to a detected physiological event. The therapy can be preventive or therapeutic in nature, such as for modifying, restoring, or improving a patient's neurological, cardiac, or respiratory function. Examples of therapy can include electrical stimulation therapy delivered to the heart, neural tissue, or other target tissue, cardioversion therapy, defibrillation therapy, or drug therapy (including delivering a drug to the patient). In some examples, the therapy circuit 240 can modify an existing therapy, such as adjusting stimulation parameters or drug dosage.

[0083] Figure 4 An example of generating a spectral entropy time series from a representative (such as an ensemble average) heart sound segment 410 and detecting heart sound components therefrom is generally shown. The ensemble average heart sound segment 410 can be divided into a plurality of heart sound data windows 420A, 420B, 420C, ..., 420N, such as by using a sliding time window that moves from the beginning to the end of the ensemble average heart sound segment 410. Although Figure 4The heart sound data windows shown do not overlap with each other, but in some examples, at least some of the heart sound data windows may overlap with each other. The size (length) of the sliding window and the amount of overlap can each have a user-programmable value. The spectral entropy can be calculated for each heart sound data window using the following equation:

[0084] S m (w k )=|X m (w k )| 2 (1)

[0085]

[0086] Where equation (1) uses the heart sound data window w k The Fourier transform of X m (w k ) to calculate the heart sound data window w k The power spectrum S m (w k ); Equation (2) estimates the power spectrum S m (w k ) probability distribution; and equation (3) calculates the spectral entropy SE(m), where m represents the window index and L represents the total number of HS data windows.

[0087] Spectral entropy values ​​may be calculated for each of the heart sound data windows 420A, 420B, 420C, ..., 420N. The resulting spectral entropy values ​​430A, 430B, 430C, ..., 430N may be concatenated relative to the corresponding timing of the data windows 420A, 420B, 420C, ..., 420N (such as the center of the corresponding heart sound window shown herein) to form a spectral entropy time series 430. By way of example and not limitation, Figure 4 Only one spectral entropy value is shown for each heart sound data window. In some examples, a sequence of multiple spectral entropy values ​​can be calculated for each heart sound data window. In an example, the spectral entropy sequence can have the same temporal resolution as the original heart sound signal, such that when spectral entropy values ​​430A, 430B, 430C, ..., 430N are concatenated, the resulting spectral entropy time series 430 can have the same number of data points as the representative heart sound segment 410.

[0088] Figure 5Generally, an example of an overlay of a plurality of spectral entropy time series 520 is shown, each spectral entropy time series being generated from a corresponding representative (e.g., ensemble-averaged) heart sound segment, one of which is shown as a representative heart sound segment 510. The representative heart sound segments all correspond to substantially the same heart rate (within a margin of, for example, + / - 5 bpm) or within the same heart rate range. The spectral entropy time series can be generated using the techniques described above with reference to Figure 4 The method described is used to generate Figure 5 In the example shown, the heart sound signal is bandpass filtered to approximately 30-100 Hz, and the spectral entropy is calculated for a window of heart sound data created using a sliding window 20 samples long (at the original heart sound sampling rate), with an overlap of 16 samples. Other window lengths or overlaps may also be used. The spectral entropy time series 520 (each spectral entropy time series has a ratio of Figure 4 The spectral entropy time series 430 (shown as a finer data resolution) can be temporally aligned with the representative heart sound segment 510 relative to the corresponding timing of the heart sound data window. An S1 detection window 532 is then applied to the spectral entropy time series 520, and a local minimum 542 within the S1 detection window 532 is detected and identified as S1. Similarly, an S2 detection window 534 is applied to the spectral entropy time series 520, and a local minimum 544 within the S2 detection window 534 is detected and identified as S2. Since the multiple spectral entropy time series 520 are temporally aligned with the representative heart sound segment 510, an S1 amplitude 552 and an S2 amplitude 554 can each be determined from the representative heart sound segment 510 based on the locations of the local minimum 542 and the local minimum 544, respectively.

[0089] Figures 6A-6D Heart sound tracking based on spectral entropy at different signal-to-noise ratios (SNRs) of heart sound signals is shown. Figure 6A A phonocardiogram 610 at high SNR is shown, and Figure 6B A phonocardiogram 620 at low SNR is shown. The phonocardiogram represents a plurality of heart sound segments stacked together (along the y-axis, indexed by segment number). The plurality of heart sound segments are extracted from a plurality of cardiac cycles with varying heart rates and aligned relative to a reference point (the origin of the x-axis), such as the R wave of the corresponding cardiac cycle. Each heart sound segment (on the x-axis) has a color-coded or gray-coded signal amplitude to show how the amplitude changes over time. Figure 6B As shown, at low SNR, some heart sound components (such as S2) become blurred, which indicates a low and irregular S2 amplitude. Detecting S2 in such a low SNR environment can be challenging when using a time-domain, amplitude-based heart sound recognition method. Figure 6C shows a spectral entropy map 630 generated from the phonocardiogram 610, and Figure 6DA spectral entropy map 640 generated from the phonocardiogram 620 is shown. The spectral entropy map represents a plurality of spectral entropy time series, each spectral entropy time series calculated from a corresponding heart sound segment within the cardiac cycle, which are stacked along the y-axis and indexed by the heart sound segment number. Each spectral entropy time series (on the x-axis, time) has a color-coded or gray-coded spectral entropy value to show the change in spectral entropy over time. Heart sound components (particularly S1 and S2) can be detected from each spectral entropy time series based on, for example, local minima in the corresponding S1 detection window and S2 detection window, as described above with reference to FIG. Figure 2 as well as Figure 3A-Figure 3B To track the heart sound components, the heart sound components detected from multiple stacked spectral entropy time series can be connected to show the heart sound component trend, such as Figure 6C S2 Trend 632 or Figure 6D S2 trend in 642. Figure 6C and Figure 6D As shown, the entropy-based method described in this invention document can accurately identify heart sound components (such as S2) and track their changes over time under variable heart rates, even in the presence of low SNR in the original heart sound signal.

[0090] Figure 7An example of detecting worsening heart failure (WHF) based at least in part on spectral entropy-based identification and tracking of S3 heart sounds is generally shown. From each cardiac cycle, S3 can be detected based on the timing of S2 detected within the cardiac cycle using an entropy-based method as described above. In an example, S3 can be detected within an S3 detection window that starts at the S2 timing, or at a specified offset after the detected S2 (e.g., approximately 50-125msec after the S2 timing) and lasts for a duration of approximately 125msec. An S3 metric (such as the root mean square (RMS) of S3) can be calculated and tracked over time, as represented by a daily S3 trend 710. A WHF detection index can be calculated using the S3 RMS value (optionally together with other sensor information) by a weighted combination unit or other type of sensor fusion engine. In this example, daily S3 is tracked for approximately 90 days until the WHF detection index value reaches and exceeds a predetermined WHF detection threshold, at which time (time zero 720) a WHF event is considered detected and the user can be alerted. In an example, the WHF detection index can include the ratio of the intensity (e.g., amplitude or RMS value) of the S3 to S1 components. U.S. patent application Ser. No. 15 / 473,783 to Pramodsingh et al., entitled “SYSTEMS AND METHODS FOR DETECTING WORSENING HEART FAILURE,” describes techniques for detecting worsening cardiac events (such as WHF events) using information including heart sounds from multiple sensors, the disclosure of which is incorporated herein by reference in its entirety.

[0091] Figure 8 Generally, an example of a method 800 for identifying and tracking heart sound components from heart sound signals and using the identified heart sound components to detect physiological events is shown. The method 800 can be implemented and executed in a mobile medical device (such as an implantable or wearable medical device) or in a remote patient management system. In an example, the method 800 can be implemented and executed by an IMD 102 or WMD 103, an external system 105, or a heart sound-based physiological event detection system 200.

[0092] Method 800 begins at 810 by receiving physiological information of a subject, including heart sounds across multiple cardiac cycles. The heart sounds can be detected using a sensor associated with or included in a mobile device or a wearable device. In some examples, the heart sounds can be analyzed using an endocardial acceleration signal sensed from inside the heart. Other physiological information can also be received, including cardiac electrical signals (such as an electrocardiogram (ECG) or an intracardiac electrogram (EGM)), heart rate, signals indicating mechanical cardiac activity (including, for example, thoracic or cardiac impedance signals, arterial pressure signals, pulmonary artery pressure signals, left atrial pressure signals, RV pressure signals, LV pressure signals, heart sounds or endocardial acceleration signals, physiological responses to activity, an apnea-hypopnea index, one or more respiratory signals (such as a respiratory rate signal or a tidal volume signal), etc.

[0093] At 820, a representative heart sound segment can be generated using at least a portion of the received heart sound. An example of a representative heart sound segment is an ensemble average of multiple heart sound segments taken from corresponding cardiac cycles and time-aligned relative to corresponding reference points (such as R waves in multiple cardiac cycles of an ECG signal). In some examples, a subset of heart sound segments that meet specified conditions are selected and used to establish the representative heart sound segment. The selection criteria may include one or more of conditions such as patient activity level, heart rate, respiratory rate, or time of day. For example, a subset of heart sound segments corresponding to substantially the same cycle length (e.g., within a margin of + / - 100 milliseconds) or substantially the same heart rate (e.g., within a margin of + / - 5 beats per minute) can be selected and used to establish the representative heart sound segment. In another example, a subset of heart sound segments sensed during substantially the same time period of a day (e.g., within a specific margin) can be selected and used to establish the representative heart sound segment.

[0094] At 830, the representative heart sound segment may be divided into a plurality of heart sound data windows using a sliding time window. The heart sound data windows may not overlap with each other. Alternatively, at least some of the heart sound data windows may overlap with each other. The size (length) of the sliding window and the amount of overlap may each have a user-programmable value. Spectral entropy may be calculated for each heart sound data window using equations (1) to (3), such as those described above with reference to Figure 4 At 840 , the spectral entropy values ​​calculated for the heart sound data windows, respectively, may be concatenated relative to the corresponding time sequence of the data windows to form a spectral entropy time series.

[0095] At 850, one or more heart sound components including S1 or S2 components may be identified from the spectral entropy time series, such as using the method described above with reference to Figure 3A-Figure 3BThe described heart sound component detector 328. In an example, the S1 component can be identified as a local minimum of a portion of the spectral entropy time series within the S1 detection window. Similarly, the S2 component can be identified as a local minimum of a portion of the spectral entropy time series within the S2 detection window. The S1 and S2 time windows can have user-programmable positions and sizes (lengths). By way of example and not limitation, the S1 detection window can start 50 milliseconds (msec) after the R wave detected on the ECG signal and have a duration of 300 milliseconds. The S2 detection window can start at a specified offset after the detected R wave or S1 heart sound. Other heart sound components (such as S3) can be detected from the heart sound energy signal within a detection window (e.g., an S3 detection window) defined based on the timing of the S1 or S2 components. By way of example and not limitation, the S3 detection window can start at the S2 timing, or at a specified offset after the detected S2 (e.g., approximately 50-125 msec after the S2 timing) and last for a duration of approximately 125 msec.

[0096] In some examples, the identification of heart sound components (eg, S1 or S2) can be based on heart rate dependent heart sound spectral entropy templates previously generated and stored in a template database. Each template may include a spectral entropy time series SE T , spectral entropy values ​​of heart sound components (such as S1 and S2), or one or more of the S1 and S2 timings relative to a reference point (such as an R wave), all of which are determined at a specific heart rate or heart rate range. The instantaneous heart rate sensed simultaneously with the heart sound information over multiple cardiac cycles can then be used to query the template database to identify (or select) a heart rate matching template having a corresponding heart rate or heart rate range that matches the instantaneous heart rate. The selected heart rate matching template can then be used to detect one or more heart sound components. In an example, instead of the representative heart sound segments and spectral entropy time series generated at steps 820 and 840, respectively, only the heart rate matching template is used to generate various heart sound components. In another example, the spectral entropy time series generated from the representative HS segment at step 840 can be modified by the heart rate matching spectral entropy time series stored in the heart rate matching spectral entropy template. The modification can include, for example, taking the average of the spectral entropy time series of the representative heart sound segment and the heart rate matching spectral entropy time series stored in the heart rate matching spectral entropy template. Then, one or more heart sound components can be detected from the modified spectral entropy time series.In some examples, the heart rate-matched spectral entropy template in the database can be updated using the modified spectral entropy time series.

[0097] At 860, one or more heart sound components may be used to detect cardiac events, such as using physiological event detector 224. Physiological events may include indicators of myocardial contractility, cardiac synchrony, and cardiac hemodynamics, arrhythmia episodes, or WHF events. In some examples, heart sound metrics may be used with other sensor information to detect respiratory, renal, nervous system, and other medical conditions based on heart sound metrics generated from heart sound segments. In some examples, based on the detection of physiological events, recommendations may be generated and provided to the user. The recommendations may include one or more of further diagnostic tests to be performed, adjustments to one or more parameters used to detect physiological events, or treatments to be delivered, or adjustments to one or more treatment parameters. A system user may review and adjudicate the detected physiological events and reprogram one or more detection or therapy parameters, such as using user interface 230. In some examples, therapy may be delivered to the patient in response to the detected physiological events, such as via treatment circuit 240, such as Figure 2 Examples of therapy may include electrical stimulation therapy delivered to the heart, neural tissue, or other target tissue, cardioversion therapy, defibrillation therapy, or drug therapy (including drug delivery to tissue or organs). In some examples, an existing therapy or treatment plan may be modified to treat the detected arrhythmia, such as by modifying a patient follow-up schedule or adjusting stimulation parameters or drug dosage.

[0098] Figure 9 A block diagram generally illustrates an example machine 900 on which any one or more of the techniques (e.g., methods) discussed herein may be executed. Portions of this description may apply to the computing framework of various portions of the IMD 102, WMD 103, external system 105, or heart sound-based physiological event detection system 200.

[0099] In an alternative embodiment, the machine 900 can operate as a standalone device or can be connected (e.g., networked) to other machines. In a networked deployment, the machine 900 can operate as a server machine, a client machine, or both in a server-client network environment. In an example, the machine 900 can act as a peer machine in a peer-to-peer (P2P) (or other distributed) network environment. The machine 900 can be a personal computer (PC), a tablet PC, a set-top box (STB), a personal digital assistant (PDA), a mobile phone, a network appliance, a network router, a switch or a bridge, or any machine capable of executing instructions (sequentially or otherwise) specifying the actions to be taken by the machine. Furthermore, although only a single machine is shown, the term "machine" should also be construed to include any collection of machines that individually or jointly execute one (or more) sets of instructions to perform any one or more of the methodologies discussed herein, such as cloud computing, software as a service (SaaS), or other computer cluster configurations.

[0100] Examples as described herein may include, or may be operated by, logic or multiple components or mechanisms. A circuit set is a collection of circuits implemented as a tangible entity including hardware (e.g., simple circuits, gates, logic, etc.). Circuit set membership can flexibly change over time and with the variability of the underlying hardware. A circuit set includes members that, when operated, can perform a specified operation individually or in combination. In an example, the hardware of a circuit set can be immutably designed to perform a specific operation (e.g., hardwired). In an example, the hardware of a circuit set can include variably connected physical components (e.g., execution units, transistors, simple circuits, etc.) that include a computer-readable medium that has been physically modified (e.g., magnetically, electrically, by removable placement of constant mass particles, etc.) to encode instructions for a specific operation. When the physical components are connected, the underlying electrical properties of the hardware components are changed, for example, from an insulator to a conductor or vice versa. The instructions enable embedded hardware (e.g., execution units or loading mechanisms) to create members of the circuit set in hardware via variable connections that, when operated, perform parts of a specific operation. Thus, when the device is operating, the computer-readable medium element is communicatively coupled to the other components of the circuit set members. In an example, any one of the physical components can be used in more than one member of more than one circuit set. For example, during operation, an execution unit can be used in a first circuit in a first circuit set at one point in time and reused by a second circuit in the first circuit set or by a third circuit in the second circuit set at a different time.

[0101] The machine (e.g., a computer system) 900 may include a hardware processor 902 (e.g., a central processing unit (CPU), a graphics processing unit (GPU), a hardware processor core, or any combination thereof), a main memory 904, and a static memory 906, some or all of which may communicate with each other via an interconnect (e.g., a bus) 908. The machine 900 may also include a display unit 910 (e.g., a raster display, a vector display, a holographic display, etc.), an alphanumeric input device 912 (e.g., a keyboard), and a user interface (UI) navigation device 914 (e.g., a mouse). In an example, the display unit 910, the input device 912, and the UI navigation device 914 may be a touch screen display. The machine 900 may additionally include a storage device (drive unit) 916, a signal generating device 918 (e.g., a speaker), a network interface device 920, and one or more sensors 921, such as a global positioning system (GPS) sensor, a compass, an accelerometer, or other sensors. The machine 900 may include an output controller 928, such as a serial (e.g., Universal Serial Bus (USB)), parallel, or other wired or wireless (e.g., infrared (IR), near field communication (NFC), etc.) connection to communicate with or control one or more peripheral devices (e.g., a printer, a card reader, etc.).

[0102] The storage device 916 may include a machine-readable medium 922 on which is stored one or more sets of data structures or instructions 924 (e.g., software) that embody or are utilized by any one or more of the techniques or functionality described herein. The instructions 924 may also reside, completely or at least partially, within the main memory 904, static storage 906, or hardware processor 902 during execution of the instructions by the machine 900. In an example, one or any combination of the hardware processor 902, the main memory 904, the static storage 906, or the storage device 916 may constitute a machine-readable medium.

[0103] Although the machine-readable medium 922 is illustrated as a single medium, the term “machine-readable medium” may include a single medium or multiple media (e.g., a centralized or distributed database, and / or associated caches and servers) configured to store one or more instructions 924.

[0104] The term "machine-readable medium" may include any medium that is capable of storing, encoding, or carrying instructions for execution by the machine 900 and causing the machine 900 to perform any one or more of the techniques in this disclosure, or capable of storing, encoding, or carrying data structures used by or associated with such instructions. Non-limiting examples of machine-readable media may include solid-state memory, as well as optical and magnetic media. In an example, a mass-type machine-readable medium includes a machine-readable medium having a plurality of particles that have a constant (e.g., stationary) mass. Therefore, a mass-type machine-readable medium is not a transient propagating signal. Specific examples of mass-type machine-readable media may include: non-volatile memory, such as semiconductor memory devices (e.g., electrically programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM)) and flash memory devices; magnetic disks, such as internal hard disks and removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks.

[0105] The instructions 924 may also be transmitted or received using a transmission medium over a communication network 926 via a network interface device 920 utilizing any of a variety of transmission protocols, such as Frame Relay, Internet Protocol (IP), Transmission Control Protocol (TCP), User Datagram Protocol (UDP), Hypertext Transfer Protocol (HTTP), etc. Example communication networks may include a local area network (LAN), a wide area network (WAN), a packet data network (such as the Internet), a mobile telephone network (such as a cellular network), a plain old telephone (POTS) network, and a wireless data network (such as a wireless network). The Institute of Electrical and Electronics Engineers (IEEE) 802.11 series of standards, known as ), IEEE 802.16 family of standards), IEEE 802.15.4 family of standards, peer-to-peer (P2P) networks, and the like. In an example, the network interface device 920 may include one or more physical jacks (e.g., Ethernet jacks, coaxial jacks, or telephone jacks) or one or more antennas for connecting to the communications network 926. In an example, the network interface device 920 may include multiple antennas for wireless communication using at least one of single-input multiple-output (SIMO), multiple-input multiple-output (MIMO), or multiple-input single-output (MISO) technology. The term "transmission medium" shall be deemed to include any intangible medium capable of storing, encoding, or carrying instructions for execution by the machine 900, and including digital or analog communication signals or other intangible media to facilitate communication of such software.

[0106] Various embodiments are shown in the above drawings. One or more features from one or more of these embodiments may be combined to form other embodiments.

[0107] The method examples described herein may be at least partially machine or computer implemented. Some examples may include a computer-readable medium or machine-readable medium encoded with instructions that are operable to configure an electronic device or system to perform the methods described in the examples above. Implementations of such methods may include code, such as microcode, assembly language code, high-level language code, or the like. Such code may include computer-readable instructions for performing various methods. The code may constitute part of a computer program product. In addition, the code may be tangibly stored on one or more volatile or non-volatile computer-readable media during execution or at other times.

[0108] The foregoing detailed description is intended to be illustrative rather than limiting. The scope of the disclosure should, therefore, be determined with reference to the appended claims, along with the full scope of equivalents to which such claims are legally entitled.

Claims

1. A medical device system comprising: a data receiver circuit configured to receive heart sound information; and A heart sound recognition circuit, wherein the heart sound recognition circuit is configured to: generating a representative heart sound segment within a cardiac cycle using at least a portion of the received heart sound information; dividing the representative heart sound segment into a plurality of heart sound data windows; calculating a spectral entropy value for each of the plurality of heart sound data windows; as well as One or more heart sound components including the S2 component are determined using the calculated spectral entropy value.

2. The medical device system according to claim 1, wherein: The heart sound recognition circuit is configured to: Dividing the representative heart sound segment into the plurality of heart sound data windows using a sliding time window; generating a spectral entropy time series by concatenating the calculated spectral entropy values ​​according to corresponding time sequences of the plurality of heart sound data windows; and The one or more heart sound components including the S2 component are determined using the spectral entropy time series.

3. The medical device system according to any one of claims 1 to 2, wherein: The portion of the received heart sound information includes a plurality of heart sound segments respectively taken from a plurality of cardiac cycles, Wherein, the heart sound recognition circuit is configured as follows: Time-aligning the plurality of heart sound segments relative to corresponding reference points; and The representative heart sound segment is generated using an ensemble average of a plurality of time-aligned heart sound segments.

4. The medical device system according to claim 3, wherein: The data receiver circuit is configured to receive cardiac electrical signals sensed simultaneously with the heart sound information over a plurality of cardiac cycles, The corresponding reference points include: ventricular activations in multiple cardiac cycles of simultaneously sensed cardiac electrical signals.

5. The medical device system according to any one of claims 1 to 4, wherein: The data receiver circuit is configured to receive information about the instantaneous heart rate sensed simultaneously with the heart sound information over a plurality of cardiac cycles, The portion of the received heart sound information over a plurality of cardiac cycles used to generate the representative heart sound segment corresponds to substantially the same instantaneous heart rate or is within a predetermined heart rate range.

6. The medical device system according to claim 2, wherein: The heart sound recognition circuit is configured to determine the S2 component based on a local minimum of a portion of the spectral entropy time series within an S2 detection window.

7. The medical device system according to claim 6, wherein: The data receiver circuit is configured to receive (i) cardiac electrical signals and (ii) information regarding instantaneous heart rate, both of which are sensed simultaneously with the heart sound information over a plurality of cardiac cycles, The heart sound recognition circuit is configured to determine the S2 detection window using ventricular activation on the cardiac electrical signal and the instantaneous heart rate.

8. The medical device system according to any one of claims 1 to 7, wherein: The heart sound recognition circuit is configured to: generating one or more heart rate or rhythm dependent heart sound spectral entropy templates and storing them in a memory, each heart rate or rhythm dependent heart sound spectral entropy template comprising a spectral entropy time series of heart sounds at a corresponding heart rate or rhythm; as well as The one or more heart sound components are further determined using the one or more heart rate or heart rhythm dependent heart sound spectral entropy templates.

9. The medical device system according to claim 8, wherein: The data receiver circuit is configured to receive information about an instantaneous heart rate or rhythm sensed simultaneously with the heart sound information, Wherein, the heart sound recognition circuit is configured as follows: selecting one of the one or more stored heart rate or heart rhythm dependent heart sound spectral entropy templates having a corresponding heart rate or heart rhythm that matches the instantaneous heart rate or heart rhythm; and The one or more heart sound components are determined using a selected stored heart rate or heart rhythm dependent heart sound spectral entropy template.

10. The medical device system according to claim 9, wherein: The heart sound recognition circuit is configured to: modifying the spectral entropy time series of the representative heart sound segment using the selected stored heart rate or rhythm dependent heart sound spectral entropy template; and The one or more heart sound components are determined using the modified spectral entropy time series of the representative heart sound segment.

11. The medical device system according to claim 10, wherein: In order to modify the spectral entropy time series of the representative heart sound segment, the heart sound recognition circuit is configured to calculate: the average of: (i) the spectral entropy time series of the representative heart sound segment and (ii) the spectral entropy time series in the selected stored heart rate or rhythm dependent heart sound spectral entropy template.

12. The medical device system according to any one of claims 10 to 11, wherein: The heart sound recognition circuit is configured to update the selected stored heart rate or rhythm dependent heart sound spectral entropy template in the memory using the modified spectral entropy time series of the representative heart sound segment.

13. The medical device system according to any one of claims 1 to 12, wherein: The heart sound identification circuit is further configured to detect an S3 component or an S4 component from the representative heart sound segment based at least in part on the determined timing information of the S2 component.

14. The medical device system according to claim 13, wherein: The heart sound recognition circuit is further configured to: determining a confidence level for the determined S2 component; and When the determined confidence exceeds a threshold, the S3 component or the S4 component is detected.

15. The medical device system according to any one of claims 1 to 14, further comprising: A physiological event detector is configured to detect a cardiac event using the determined one or more heart sound components.

Images