WO2025210419A1

WO2025210419A1 - Identifying a particular health event using a single lead electrical signal sensed by a medical device

Info

Publication number: WO2025210419A1
Application number: PCT/IB2025/052399
Authority: WO
Inventors: Neha Jain; Neethu Lekshmi VASUDEVAN JALAJA; Sean R. LANDMAN; Maudeline R. DEUS
Original assignee: Medtronic Inc
Current assignee: Medtronic Inc
Priority date: 2024-04-05
Filing date: 2025-03-05
Publication date: 2025-10-09
Anticipated expiration: 2026-10-05

Abstract

An example system includes an implantable medical device configured to sense an electrocardiogram (ECG) signal of a patient via a single lead; and processing circuitry configured to: separate the ECG signal into a plurality of ECG segments corresponding to a plurality of heartbeat time periods; apply the plurality of ECG segments to a machine learning (ML) model; and determine whether one or more of the plurality of ECG segments indicate myocardial infarction, the model being trained on a plurality of training ECG segments corresponding to the plurality of heartbeat time periods.

Description

IDENTIFYING A PARTICULAR HEALTH EVENT USING A SINGLE LEAD ELECTRICAL SIGNAL SENSED BY A MEDICAL DEVICE

[0001] This application claims the benefit of U.S. Provisional Patent Application Serial No. 63/575,418, filed April 5, 2024, the entire content of which is incorporated herein by reference.

FIELD

[0002] The disclosure relates generally to medical device systems and, more particularly, medical device systems configured to identify a particular health event.

BACKGROUND

[0003] Medical devices may be used to monitor physiological signals of a patient. For example, some medical devices are configured to sense electrocardiogram (ECG) signals indicative of the electrical activity of the heart, or other electrogram (EGM) signals indicative of electrical activity of the patient, via electrodes. Some medical devices may be configured to deliver a therapy in conjunction with or separate from the monitoring of EGM signals.

[0004] Acute myocardial infarction (AMI) is very deadly, claiming millions of lives a year. Patients often ignore their symptoms or do not realize that they are having an AMI until it is too late. Some patients have silent AMI which increases their risk for deadly AMIs in the future.

SUMMARY

[0005] ECGs may also be used for left-ventricular (LV) systolic dysfunction diagnosis, e.g., for identification of simple abnormalities on an ECG, or classifying ejection fraction (EF) using a 12-lead ECG system. In some examples, this disclosure is directed to techniques for separating an ECG signal obtained from a single lead, such as from an implantable medical device (IMD), into a plurality of ECG segments and applying the separated ECG segments to a machine learning (ML) model to determine whether one or more of the plurality of ECG segments from the ECG signal indicate myocardial infarction, such as AMI. [0006] A single lead ECG signal may be continuously sensed and monitored by an IMD, e.g., autonomously on a periodic, triggered, or other basis. In this manner, the ECG signal may be used as a screening tool to identify early symptoms of myocardial infarction without a patient even going to hospital. Accordingly, medical intervention and/or treatment due to myocardial infarction may be applied sooner which may lower the risk of long-term complications of myocardial infarction and may reduce mortality and/or morbidity. In some examples, an alert may be triggered to a clinician to order a medical examination to confirm the myocardial infarction and then determine further therapeutic options.

[0007] Processing circuitry of an IMD and/or a system including the IMD may be configured to separate the ECG signal into a plurality of ECG segments corresponding to a plurality of heartbeat time periods and apply the plurality of ECG segments to a ML model to determine whether one or more of the plurality of ECG segments indicates myocardial infarction, or to transmit ECG signal and/or the plurality of ECG segments to other devices for analysis. For example, daily or other periodic ECG signal and/or ECG segment transmissions may allow the determination of whether one or more of the plurality of ECG segments indicates myocardial infarction in a cloud platform that may send an alert to a clinician computing device.

[0008] In some examples, separating an ECG signal into particular ECG segments, such as beginning at a particular R-wave of the ECG signal and ending a period of time after a following R-wave, the following R-wave being the immediate next R-wave after the particular R-wave, and applying these particular separated ECG segments to a ML model increases the robustness of the ML model in determining myocardial infarction based on ECG signal data, which may enable improved detection of myocardial infarction. This may help determine myocardial infarction or a classification of myocardial infarction with greater sensitivity and/or specificity than other techniques, or than would otherwise be possible with a single lead ECG.

[0009] In some examples in accordance with techniques of this disclosure, processing circuitry of a medical device system may apply an electrical signal, such as an ECG signal or an electroencephalogram (EEG) signal, obtained from a single lead, such as from an IMD, to a ML model to generate one or more predictions of whether the electrical signal indicates a particular health event and determining a particular morphology of the electrical signal is indicative of the particular health event based on application of an integrated gradient to the one or more predictions. In some examples, the processing circuitry may determine whether the particular health event occurred or predict the particular health event will occur based on the determined morphology of the electrical signal. In this manner, the electrical signal may be used as a screening tool to identify an occurrence or prediction of particular health event, such as AMI, without a patient even going to hospital. Accordingly, medical intervention and/or treatment due to a particular health event, such as AMI, may be applied sooner which may lower the risk of long-term complications of particular health event, such as AMI, and may reduce mortality and/or morbidity. In some examples, an alert may be triggered to a clinician to order a medical examination to confirm the particular health event and then determine further therapeutic options.

[0010] In some examples, processing circuitry of medical system may re-train the ML model based on the determined particular morphology of the electrical signal that is indicative of the particular health event. Processing circuitry determining a particular morphology of the electrical signal is indicative of the particular health event based on application of an integrated gradient to the one or more predictions may determine morphological features of the electrical signal that lead to a true prediction and/or morphological features of the electrical signal that lead to a false prediction. In some examples, processing circuitry may use the determined particular morphology of the electrical signal that is indicative of the particular health event to retrain, such as continuously or periodically, the ML model with features identified to mostly or only contribute to a true prediction. Retraining in this manner may improve the sensitivity and/or specificity of a ML model and may enable continuous and/or personalized retraining of the ML model, which may increase the sensitivity and/or specificity of a ML model to a particular positioning of an IMD in a patient or particularly anatomy/physiology of the patient.

[0011] In addition, a device described herein may sense an electrical signal, such as an ECG signal, of a patient continuously, hourly, and/or daily. Consequently, dynamic changes in whether an electrical signal or particular morphology of the electrical signal indicates a particular health event occurred, such as an ECG indicating AMI occurred, or predicts a particular health event, such as AMI, will occur within a short period of time, such as within 24 to 48 hours, may be tracked. Such tracking is not possible today, and will open up new possibilities for treatment recommendations.

[0012] In one example, this disclosure describes a system comprising an implantable medical device configured to sense an electrocardiogram (ECG) signal of a patient via a single lead; and processing circuitry configured to: separate the ECG signal into a plurality of ECG segments corresponding to a plurality of heartbeat time periods; apply the plurality of ECG segments to a machine learning (ML) model; and determine whether one or more of the plurality of ECG segments indicate myocardial infarction, the model being trained on a plurality of training ECG segments corresponding to the plurality of heartbeat time periods.

[0013] In another example, this disclosure describes a system comprising: an implantable medical device configured to sense an electrical signal of a patient via a single lead; and processing circuitry configured to: apply the electrical signal of the patient to a machine learning (ML) model to generate one or more predictions of whether the electrical signal indicates a particular health event; apply an integrated gradient to the one or more predictions; determine a particular morphology of the electrical signal that is indicative of the particular health event based on the application of the integrated gradient to the one or more predictions; and output an indication of the particular morphology of the electrical signal that is indicative of the particular health event.

[0014] In another example, this disclosure describes a method comprising: receiving, via an implantable medical device, an electrocardiogram (ECG) signal of a patient sensed via a single lead of the implantable medical device; separating the ECG signal into a plurality of ECG segments corresponding to a plurality of heartbeat time periods; applying the plurality of ECG segments to a machine learning (ML) model; and determining whether one or more of the plurality of ECG segments indicate myocardial infarction, the model being trained on a plurality of training ECG segments corresponding to the plurality of heartbeat time periods.

[0015] In another example, this disclosure describes a method comprising: receiving, via an implantable medical device, an electrical signal of a patient sensed via a single lead; applying the electrical signal of the patient to a machine learning (ML) model to generate one or more predictions of whether the electrical signal indicates a particular health event; applying an integrated gradient to the one or more predictions; determining a particular morphology of the electrical signal that is indicative of the particular health event based on the application of the integrated gradient to the one or more predictions; and outputting an indication of the particular morphology of the electrical signal that is indicative of the particular health event.

[0016] The summary is intended to provide an overview of the subject matter described in this disclosure. It is not intended to provide an exclusive or exhaustive explanation of the systems, device, and methods described in detail within the accompanying drawings and description below. Further details of one or more examples of this disclosure are set forth in the accompanying drawings and in the description below. Other features, objects, and advantages will be apparent from the description and drawings, and from the claims.

BRIEF DESCRIPTION OF THE DRAWINGS

[0017] FIG. 1 illustrates the environment of an example medical system in conjunction with a patient, in accordance with one or more techniques disclosed herein.

[0018] FIG. 2 is a functional block diagram illustrating an example configuration of the implantable medical device (IMD) of the medical system of FIG. 1, in accordance with one or more techniques disclosed herein.

[0019] FIG. 3 is a conceptual side-view diagram illustrating an example configuration of the IMD of FIGS. 1 and 2, in accordance with one or more techniques disclosed herein. [0020] FIG. 4 is a functional block diagram illustrating an example configuration of the external device of FIG. 1, in accordance with one or more techniques disclosed herein. [0021] FIG. 5 is a block diagram illustrating an example system that includes an access point, a network, external computing devices, such as a server, and one or more other computing devices, which may be coupled to the IMD and external device of FIGS. 1-4, in accordance with one or more techniques disclosed herein.

[0022] FIG. 6 illustrates an example of separating an ECG signal into a plurality of ECG segments, in accordance with one or more techniques disclosed herein.

[0023] FIG. 7 is a flow diagram illustrating an example technique for operating a system to determine whether one or more of a plurality of ECG segments of an ECG signal indicate a particular health event, in accordance with one or more techniques disclosed herein. [0024] FIG. 8 is a flow diagram illustrating an example technique for determining whether a particular morphology of an electrical signal is indicative of a particular health event, in accordance with one or more techniques disclosed herein.

[0025] FIGS. 9A-9D illustrate examples of determining whether particular morphologies of an ECG signal are indicative of a particular health event, in accordance with one or more techniques disclosed herein.

[0026] FIGS. 10 are conceptual diagrams illustrating example training processes for an artificial intelligence model, in accordance with examples of the current disclosure.

[0027] FIG. 11 is a conceptual diagram illustrating an example machine learning model configured to determine whether one or more of a plurality of ECG segments of an ECG signal indicate myocardial infarction or determine a particular morphology of an electrical signal is indicative of a particular health event.

[0028] Like reference characters denote like elements throughout the description and figures.

DETAILED DESCRIPTION

[0029] A variety of types of medical devices sense physiologic signals, such as cardiac EGM signals, EEG signals, or other EGM signals. In some examples, cardiac EGMs may also include electrocardiogram (ECGs or EKGs). Implantable medical devices (IMDs) may sense and monitor EGMs. The electrodes used by IMDs to sense EGMs are typically integrated with a housing of the IMD and/or coupled to the IMD via one or more elongated leads.

[0030] Example IMDs that monitor cardiac EGMs include pacemakers and implantable cardioverter-defibrillators, which may be coupled to intravascular or extravascular leads, as well as pacemakers with housings configured for implantation within the heart, which may be leadless. An example of pacemaker configured for intracardiac implantation is the Micra™ Transcatheter Pacing System, available from Medtronic, Inc. Some IMDs that do not provide therapy, e.g., implantable patient monitors, sense cardiac EGMs such as ECGs. Examples of such an IMD are the Reveal LINQ™ and LINQ II™ Insertable Cardiac Monitors (ICMs), available from Medtronic, Inc., which may be inserted subcutaneously. Such IMDs may facilitate relatively longer- term monitoring of patients during normal daily activities and may periodically transmit collected data to a network service, such as the Medtronic Carelink™ Network.

[0031] In some examples, the techniques herein include segmenting an ECG signal obtained from a single lead, such as from an IMD, into a plurality of ECG segments and applying the separated ECG segments to a ML model to determine whether one or more of the plurality of ECG segments from the ECG signal indicate myocardial infarction, such as AMI. The ECG signal may be used as a screening tool to identify myocardial infarction without a patient even going to hospital. In some examples, a single lead may be a single electrode combination or vector. In some examples, a single lead may include only electrodes on a housing of an ICM or other IMD.

[0032] In some examples, the techniques herein include applying an electrical signal (EGM signal), such as an ECG signal or an electroencephalogram (EEG) signal, obtained from a single lead, such as from an IMD, to a ML model to generate one or more predictions of whether the electrical signal indicates a particular health event and determining a particular morphology of the electrical signal is indicative of the particular health event based on application of an integrated gradient to the one or more predictions. In some examples, the techniques herein may include determining whether the particular health event occurred or predict the particular health event will occur based on the determined morphology of the electrical signal. In some examples, a single lead may be a single electrode combination or vector. In some examples, a single lead may include only electrodes on a housing of an ICM or other IMD.

[0033] FIG. 1 illustrates the environment of an example medical system 2 in conjunction with a patient 4, in accordance with one or more techniques of this disclosure. The example techniques may be used with an IMD 10, which may be in wireless communication with at least one of external device 12 and other devices not pictured in FIG. 1. In some examples, IMD 10 is implanted outside of a thoracic cavity of patient 4 (e.g., subcutaneously in the pectoral location illustrated in FIG. 1). IMD 10 may be positioned near the sternum near or just below the level of the heart of patient 4, e.g., at least partially within the cardiac silhouette. IMD 10 includes a plurality of electrodes (not shown in FIG. 1) and is configured to sense a cardiac EGM via the plurality of electrodes. In some examples, IMD 10 takes the form of the Reveal LINQ™ or LINQ II™ ICM, or another ICM similar to, e.g., a version or modification of, the Reveal LINQ™ or LINQ II™ ICMs.

[0034] External device 12 may be a computing device with a display viewable by the user and an interface for providing input to external device 12 (i.e., a user input mechanism). In some examples, external device 12 may be a notebook computer, tablet computer, workstation, one or more servers, cellular phone, personal digital assistant, or another computing device that may run an application that enables the computing device to interact with IMD 10. External device 12 is configured to communicate with IMD 10 and, optionally, another computing device (not illustrated in FIG. 1), via wireless communication. External device 12, for example, may communicate via near-field communication technologies (e.g., inductive coupling, NFC or other communication technologies operable at ranges less than 10-20 cm) and far-field communication technologies (e.g., RF telemetry according to the 802.11 or Bluetooth® specification sets, or other communication technologies operable at ranges greater than near-field communication technologies).

[0035] External device 12 may be used to configure operational parameters for IMD 10. External device 12 may be used to retrieve data from IMD 10. The retrieved data may include values of physiological parameters measured by IMD 10, indications of episodes of arrhythmia or other maladies detected by IMD 10, and physiological signals recorded by IMD 10. For example, external device 12 may retrieve cardiac ECG signal segments recorded by IMD 10. As discussed in greater detail below with respect to FIG. 5, one or more remote computing devices may interact with IMD 10 in a manner similar to external device 12, e.g., to program IMD 10 and/or retrieve data from IMD 10, via a network. [0036] Processing circuitry of medical system 2, e.g., of IMD 10, external device 12, and/or of one or more other computing devices, may be configured to perform the example techniques of this disclosure for determining whether one or more of the plurality of ECG segments from the ECG signal, indicate myocardial infarction, such as AMI. In some examples, the processing circuitry of medical system 2 may separate an ECG signal obtained from a single lead, such as from an IMD, into a plurality of ECG segments corresponding to a plurality of heartbeat time periods. In some examples, a respective heartbeat time period of the plurality of time periods begins at a particular R-wave of the ECG signal and ends a period of time after a following R-wave, the following R-wave being immediately after the particular R-wave. In some examples, the processing circuitry of medical system 2 may apply the plurality of ECG segments to a ML model to determine whether one or more of the plurality of ECG segments indicate myocardial infarction, the ML model being trained on a plurality of training ECG segments corresponding to the plurality of heartbeat time periods. In some examples, processing circuitry of medical system 2 may determine whether an amount of the plurality of ECG segments that indicate myocardial infarction satisfies a myocardial infarction threshold. In some examples, the myocardial infarction threshold may be at least 50% of the plurality of ECG segments indicating myocardial infarction. Although described in the context of examples in which IMD 10 that senses the ECG signal comprises an insertable cardiac monitor, example systems including one or more implantable or external devices of any type configured to sense an ECG signal may be configured to implement the techniques of this disclosure. [0037] Processing circuitry of medical system 2, e.g., of IMD 10, external device 12, and/or of one or more other computing devices, may be configured to perform the example techniques of this disclosure for applying an electrical signal, such as an ECG signal or an EEG signal, obtained from a single lead, such as from an IMD, to a ML model to generate one or more predictions of whether the electrical signal indicates a particular health event and determining a particular morphology of the electrical signal is indicative of the particular health event based on application of an integrated gradient to the one or more predictions. In some examples, processing circuitry of medical system 2 may output an indication of the particular morphology of the electrical signal that is indicative of the particular health event. In some examples, processing circuitry of medical system 2 may determine whether the particular health event occurred or predict the particular health event will occur based on the determined morphology of the electrical signal. In some examples, processing circuitry of medical system 2 may re-train the ML model based on the determined particular morphology of the electrical signal that is indicative of the particular health event.

[0038] FIG. 2 is a functional block diagram illustrating an example configuration of IMD 10 of FIG. 1 in accordance with one or more techniques described herein. In the illustrated example, IMD 10 includes electrodes 16A and 16B (collectively “electrodes 16”), antenna 26, processing circuitry 50, sensing circuitry 52, communication circuitry 54, storage device 56, switching circuitry 58, and sensors 62. Although the illustrated example includes two electrodes 16, IMDs including or coupled to more than two electrodes 16 may implement the techniques of this disclosure in some examples. [0039] Processing circuitry 50 may include fixed function circuitry and/or programmable processing circuitry. Processing circuitry 50 may include any one or more of a microprocessor, a controller, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or equivalent discrete or analog logic circuitry. In some examples, processing circuitry 50 may include multiple components, such as any combination of one or more microprocessors, one or more controllers, one or more DSPs, one or more ASICs, or one or more FPGAs, as well as other discrete or integrated logic circuitry. The functions attributed to processing circuitry 50 herein may be embodied as software, firmware, hardware or any combination thereof. [0040] Sensing circuitry 52 may be selectively coupled to electrodes 16 via switching circuitry 58, e.g., to select the electrodes 16 and polarity, referred to as the sensing vector, used to sense an electrical signal, such as an ECG signal or an EEG signal or other EGM signal, as controlled by processing circuitry 50. Sensing circuitry 52 may sense signals from electrodes 16, e.g., to produce an ECG signal, in order to facilitate monitoring the electrical activity of the heart or to produce an EEG signal, in order to facilitate monitoring the electrical activity of a brain. Sensing circuitry 52 also may monitor signals from sensors 62, which may include one or more accelerometers, pressure sensors, and/or optical sensors, as examples. In some examples, sensing circuitry 52 may include one or more filters and amplifiers for filtering and amplifying signals received from electrodes 16 and/or sensors 62.

[0041] Sensing circuitry 52 and/or processing circuitry 50 may be configured to detect cardiac depolarizations (e.g., P-waves of atrial depolarizations or R-waves of ventricular depolarizations) when the ECG signal amplitude crosses a sensing threshold. For cardiac depolarization detection, sensing circuitry 52 may include a rectifier, filter, amplifier, comparator, and/or analog-to-digital converter, in some examples. In some examples, sensing circuitry 52 may output an indication to processing circuitry 50 in response to sensing of a cardiac depolarization. In this manner, processing circuitry 50 may receive detected cardiac depolarization indicators corresponding to the occurrence of detected R-waves and P-waves in the respective chambers of heart. Processing circuitry 50 may use the indications of detected R-waves for separating the ECG signal into a plurality of ECG segments corresponding to a plurality of heartbeat time periods. For example, a respective heartbeat time period of the plurality of time periods may begin at a particular R-wave of the ECG signal and ends a period of time after a following R-wave, the following R-wave being immediately after the particular R-wave. In some examples, the plurality of time periods may begin at a particular R-peak of the particular R-wave of the ECG signal and ends a period of time after a following R-peak of the following R-wave, the following R-peak being immediately after the particular R-peak.

[0042] Sensing circuitry 52 may also provide one or more digitized electrical signals, such as ECG signals, to processing circuitry 50 for analysis, e.g., for use in cardiac rhythm discrimination, to separate an ECG signal into a plurality of ECG segments and apply the separated ECG segments to a ML model to determine whether one or more of the plurality of ECG segments from the ECG signal indicate myocardial infarction, and/or to apply an electrical signal to a ML model to generate one or more predictions of whether the electrical signal indicates a particular health event and determine a particular morphology of the electrical signal is indicative of the particular health event based on application of an integrated gradient to the one or more predictions. In some examples, processing circuitry 50 may store the digitized electrical signal in storage device 56. Processing circuitry 50 of IMD 10, and/or processing circuitry of another device that retrieves data from IMD 10, may analyze the electrical signal to determine whether one or more of the plurality of ECG segments from the ECG signal indicate myocardial infarction and/or generate one or more predictions of whether the electrical signal indicates a particular health event and determine a particular morphology of the electrical signal is indicative of the particular health event based on application of an integrated gradient to the one or more predictions according to the techniques of this disclosure.

[0043] Communication circuitry 54 may include any suitable hardware, firmware, software or any combination thereof for communicating with another device, such as external device 12, another networked computing device, or another IMD or sensor. Under the control of processing circuitry 50, communication circuitry 54 may receive downlink telemetry from, as well as send uplink telemetry to external device 12 or another device with the aid of an internal or external antenna, e.g., antenna 26. In addition, processing circuitry 50 may communicate with a networked computing device via an external device (e.g., external device 12) and a computer network, such as the Medtronic CareLink® Network. Antenna 26 and communication circuitry 54 may be configured to transmit and/or receive signals via inductive coupling, electromagnetic coupling, Near Field Communication (NFC), Radio Frequency (RF) communication, Bluetooth, WiFi, or other proprietary or non-proprietary wireless communication schemes.

[0044] In some examples, storage device 56 includes computer-readable instructions that, when executed by processing circuitry 50, cause IMD 10 and processing circuitry 50 to perform various functions attributed to IMD 10 and processing circuitry 50 herein. Storage device 56 may include any volatile, non-volatile, magnetic, optical, or electrical media, such as a random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), electrically-erasable programmable ROM (EEPROM), flash memory, or any other digital media. Storage device 56 may store, as examples, programmed values for one or more operational parameters of IMD 10 and/or data collected by IMD 10 for transmission to another device using communication circuitry 54. Data stored by storage device 56 and transmitted by communication circuitry 54 to one or more other devices may include digitized electrical signals, such as an ECG signal or an EEG signal, as examples.

[0045] FIG. 3 is a conceptual side-view diagram illustrating an example configuration of IMD 10 of FIGS. 1 and 2. In the example shown in FIG. 3, IMD 10 may include a leadless, subcutaneously-implantable monitoring device having a housing 15 and an insulative cover 76. Electrode 16A and electrode 16B may be formed or placed on an outer surface of cover 76. Circuitries 50-62, described above with respect to FIG. 2, may be formed or placed on an inner surface of cover 76, or within housing 15. In the illustrated example, antenna 26 is formed or placed on the inner surface of cover 76, but may be formed or placed on the outer surface in some examples. In some examples, one or more of sensors 62 may be formed or placed on the outer surface of cover 76. In some examples, insulative cover 76 may be positioned over an open housing 15 such that housing 15 and cover 76 enclose antenna 26 and circuitries 50-62, and protect the antenna and circuitries from fluids such as body fluids.

[0046] One or more of antenna 26 or circuitries 50-62 may be formed on the inner side of insulative cover 76, such as by using flip-chip technology. Insulative cover 76 may be flipped onto a housing 15. When flipped and placed onto housing 15, the components of IMD 10 formed on the inner side of insulative cover 76 may be positioned in a gap 78 defined by housing 15. Electrodes 16 may be electrically connected to switching circuitry 58 through one or more vias (not shown) formed through insulative cover 76. Insulative cover 76 may be formed of sapphire (i.e., corundum), glass, parylene, and/or any other suitable insulating material. Housing 15 may be formed from titanium or any other suitable material (e.g., a biocompatible material). Electrodes 16 may be formed from any of stainless steel, titanium, platinum, iridium, or alloys thereof. In addition, electrodes 16 may be coated with a material such as titanium nitride or fractal titanium nitride, although other suitable materials and coatings for such electrodes may be used.

[0047] FIG. 4 is a block diagram illustrating an example configuration of components of external device 12. In the example of FIG. 4, external device 12 includes processing circuitry 80, communication circuitry 82, storage device 84, and user interface 86.

[0048] Processing circuitry 80 may include one or more processors that are configured to implement functionality and/or process instructions for execution within external device 12. For example, processing circuitry 80 may be capable of processing instructions stored in storage device 84. Processing circuitry 80 may include, for example, microprocessors, DSPs, ASICs, FPGAs, or equivalent discrete or integrated logic circuitry, or a combination of any of the foregoing devices or circuitry. Accordingly, processing circuitry 80 may include any suitable structure, whether in hardware, software, firmware, or any combination thereof, to perform the functions ascribed herein to processing circuitry 80. [0049] Communication circuitry 82 may include any suitable hardware, firmware, software or any combination thereof for communicating with another device, such as IMD 10. Under the control of processing circuitry 80, communication circuitry 82 may receive downlink telemetry from, as well as send uplink telemetry to, IMD 10, or another device. Communication circuitry 82 may be configured to transmit or receive signals via inductive coupling, electromagnetic coupling, Near Field Communication (NFC), Radio Frequency (RF) communication, Bluetooth, WiFi, or other proprietary or non-proprietary wireless communication schemes. Communication circuitry 82 may also be configured to communicate with devices other than IMD 10 via any of a variety of forms of wired and/or wireless communication and/or network protocols.

[0050] Storage device 84 may be configured to store information within external device 12 during operation. Storage device 84 may include a computer-readable storage medium or computer-readable storage device. In some examples, storage device 84 includes one or more of a short-term memory or a long-term memory. Storage device 84 may include, for example, RAM, DRAM, SRAM, magnetic discs, optical discs, flash memories, or forms of EPROM or EEPROM. In some examples, storage device 84 is used to store data indicative of instructions for execution by processing circuitry 80. Storage device 84 may be used by software or applications running on external device 12 to temporarily store information during program execution.

[0051] Data exchanged between external device 12 and IMD 10 may include operational parameters. External device 12 may transmit data including computer readable instructions which, when implemented by IMD 10, may control IMD 10 to change one or more operational parameters and/or export collected data. For example, processing circuitry 80 may transmit an instruction to IMD 10 which requests IMD 10 to export collected data (e.g., digitized electrical signals, such as ECG signals or EEG signals) to external device 12. In turn, external device 12 may receive the collected data from IMD 10 and store the collected data in storage device 84. Processing circuitry 80 may implement any of the techniques described herein to analyze electrical signals, such as ECG signals or EEG signals, received from IMD 10, e.g., to determine whether one or more of a plurality of ECG segments from an ECG signal indicate myocardial infarction and/or apply the electrical signal to a ML model to generate one or more predictions of whether the electrical signal indicates a particular health event and determine a particular morphology of the electrical signal is indicative of the particular health event based on application of an integrated gradient to the one or more predictions.

[0052] A user, such as a clinician or patient 4, may interact with external device 12 through user interface 86. User interface 86 includes a display (not shown), such as a liquid crystal display (LCD) or a light emitting diode (LED) display or other type of screen, with which processing circuitry 80 may present information related to IMD 10, e.g., electrical signals, such as ECG signals or EEG signals. In addition, user interface 86 may include an input mechanism to receive input from the user. The input mechanisms may include, for example, any one or more of buttons, a keypad (e.g., an alphanumeric keypad), a peripheral pointing device, a touch screen, or another input mechanism that allows the user to navigate through user interfaces presented by processing circuitry 80 of external device 12 and provide input. In other examples, user interface 86 also includes audio circuitry for providing audible notifications, instructions or other sounds to the user, receiving voice commands from the user, or both.

[0053] FIG. 5 is a block diagram illustrating an example system that includes an access point 90, a network 92, external computing devices, such as a server 94, and one or more other computing devices 100A-100N (collectively, “computing devices 100”), which may be coupled to IMD 10 and external device 12 via network 92, in accordance with one or more techniques described herein. In this example, IMD 10 may use communication circuitry 54 to communicate with external device 12 via a first wireless connection, and to communicate with an access point 90 via a second wireless connection. In the example of FIG. 5, access point 90, external device 12, server 94, and computing devices 100 are interconnected and may communicate with each other through network 92. In some examples, one or more computing devices 100 may be cloud computing devices.

[0054] Access point 90 may include a device that connects to network 92 via any of a variety of connections, such as telephone dial-up, digital subscriber line (DSL), or cable modem connections. In other examples, access point 90 may be coupled to network 92 through different forms of connections, including wired or wireless connections. In some examples, access point 90 may be a user device, such as a tablet or smartphone, that may be co-located with the patient. IMD 10 may be configured to transmit data, such as electrical signals, such as ECG signals or EEG signals, to access point 90. Access point 90 may then communicate the retrieved data to server 94 via network 92.

[0055] In some cases, server 94 may be configured to provide a secure storage site for data that has been collected from IMD 10 and/or external device 12. In some cases, server 94 may assemble data in web pages or other documents for viewing by trained professionals, such as clinicians, via computing devices 100. One or more aspects of the illustrated system of FIG. 5 may be implemented with general network technology and functionality, which may be similar to that provided by the Medtronic CareLink® Network.

[0056] In some examples, one or more of computing devices 100 may be a tablet or other smart device located with a clinician, by which the clinician may program, receive alerts from, and/or interrogate IMD 10. For example, the clinician may access data collected by IMD 10 through a computing device 100, such as when patient 4 is in in between clinician visits, to check on a status of a medical condition. In some examples, the clinician may enter instructions for a medical intervention for patient 4 into an application executed by computing device 100, such as based on a status of a patient condition determined by IMD 10, external device 12, server 94, or any combination thereof, or based on other patient data known to the clinician. Device 100 then may transmit the instructions for medical intervention to another of computing devices 100 located with patient 4 or a caregiver of patient 4. For example, such instructions for medical intervention may include an instruction to change a drug dosage, timing, or selection, to schedule a visit with the clinician, or to seek medical attention. In further examples, a computing device 100 may generate an alert to patient 4 based on a status of a medical condition of patient 4, which may enable patient 4 proactively to seek medical attention prior to receiving instructions for a medical intervention. In this manner, patient 4 may be empowered to take action, as needed, to address his or her medical status, which may help improve clinical outcomes for patient 4.

[0057] In the example illustrated by FIG. 5, server 94 includes a storage device 96, e.g., to store data retrieved from IMD 10, and processing circuitry 98. Although not illustrated in FIG. 5 computing devices 100 may similarly include a storage device and processing circuitry. Processing circuitry 98 may include one or more processors that are configured to implement functionality and/or process instructions for execution within server 94. For example, processing circuitry 98 may be capable of processing instructions stored in memory 96. Processing circuitry 98 may include, for example, microprocessors, DSPs, ASICs, FPGAs, or equivalent discrete or integrated logic circuitry, or a combination of any of the foregoing devices or circuitry. Accordingly, processing circuitry 98 may include any suitable structure, whether in hardware, software, firmware, or any combination thereof, to perform the functions ascribed herein to processing circuitry 98. Processing circuitry 98 of server 94 and/or the processing circuity of computing devices 100 may implement any of the techniques described herein to analyze electrical signals, such as ECG signals or EEG signals, received from IMD 10.

[0058] Storage device 96 may include a computer-readable storage medium or computer-readable storage device. In some examples, memory 96 includes one or more of a short-term memory or a long-term memory. Storage device 96 may include, for example, RAM, DRAM, SRAM, magnetic discs, optical discs, flash memories, or forms of EPROM or EEPROM. In some examples, storage device 96 is used to store data indicative of instructions for execution by processing circuitry 98.

[0059] Although the techniques for determining whether one or more of the plurality of ECG segments from the ECG signal indicate myocardial infarction, such as AMI, and/or for applying an electrical signal to a ML model to generate one or more predictions of whether the electrical signal indicates a particular health event and determining a particular morphology of the electrical signal is indicative of the particular health event based on application of an integrated gradient to the one or more predictions are described herein primarily (e.g., with respect to FIGS. 6-9D) as being performed by processing circuitry 50 of IMD 10, such techniques may be performed, in whole or part, by processing circuitry of any one or more devices of system 2, such as processing circuitry 80 of external device 12, processing circuitry 98 of server 94, or processing circuitry of one or more computing devices 100.

[0060] In some examples, IMD 10 may be an insertable cardiac monitor (ICM) or other CIED (cardiac implantable electronic device) with ECG signal transmission capability. The ECG signal may be used as a screening tool to identify myocardial infarction without the patient 4 even going to hospital. In some examples, daily ECG signal transmission may allow the detection of myocardial infarction in a cloud platform that may send an alert to a clinician computing device.

[0061] In some examples, IMD 10 may obtain ECG signal data from a single lead configured to sense an ECG signal routinely, such as, but not limited or, hourly, once every 12 hours, daily, nightly, weekly, etc. In some examples, IMD 10 may determine whether a patient has myocardial infarction based on corresponding ECG data routinely, such as, but not limited or, hourly, daily, nightly, weekly, bi-weekly, etc. Other devices, such as external device 12, server 94, and computing devices 110, may similarly determine whether a patient has myocardial infarction based on transmissions of sensed ECG signals from IMD 10, e.g., a digitized segment of a number of minutes of ECG signals each day. In some examples, IMD 10 may transmit obtained ECG data, corresponding indication of whether one or more plurality of ECG segments of the ECG signal indicate myocardial infarction, and/or corresponding indication of myocardial infarction to external device 12. In some examples, IMD 10 may determine whether the ECG signal indicates myocardial infarction within real time. In some examples, IMD 10 may determine whether the ECG signal indicates myocardial infarction within a period of time of obtaining ECG signal data, such as, but not limited to, immediately, up to 1 hour, up to 6 hours, up to 12 hours, up to 1 day, up to 3 days, up to 1 week, up to 2 weeks, up to 1 month, etc. In some examples, ECG signal data may be obtained from an electronic health record (EHR) dataset.

[0062] FIG. 6 is a diagram illustrating an example technique for separating an ECG signal into a plurality of ECG segments. FIG. 7 is a flow diagram illustrating an example technique for medical system 2. In some examples, as shown in FIG. 6, processing circuitry 50 may separate an ECG signal 610 into a plurality of ECG segments corresponding to a plurality of heartbeat time periods (700). In some examples, a respective heartbeat time period 650 of the plurality of time periods may begin at a particular R-wave 620 of the ECG signal 610 and end at a particular period of time 640 after a following R-wave 630, the following R-wave 630 being immediately after the particular R-wave 620. In some examples, a period of time 645 after the following R-wave 630 may be within 20 sample points after the following R-wave 630. In some examples, if the ECG signal is 128 Hertz (Hz), there are 128 sample points per second in the ECG signal. In some examples, the period of time after the following R-wave 630 may be within 50 milliseconds after the following R-wave 630. In some examples, processing circuitry 50 may use an R-peaks of respective R-waves to separate the ECG signal 610 into a plurality of ECG segments. For example, a respective heartbeat time period 650 may begin at a particular R-peak of the particular R-wave 620 of the ECG signal 610 and end a period of time 640 after a following R-peak of the following R-wave 630, the following R-peak being immediately after the particular R-peak. In some examples, because a respective heartbeat time period 650 ends at a particular period of time 640 after a following R-wave, parts of a respective first heartbeat time period may at least partially overlap with parts of a subsequent second heartbeat time period in which the following R- wave 630 of the first heartbeat time period becomes the particular R-wave 620 of the subsequent second heartbeat time period.

[0063] In some examples, the processing circuitry 50 may apply the plurality of ECG segments to a ML model (710). In some examples, the plurality of ECG segments may be sequential. In some examples, the processing circuitry 50 may apply the plurality of ECG segments sequentially. In some examples, the processing circuitry 50 may determine an amount of the plurality of ECG segments that indicate myocardial infarction (720). to determine whether one or more of the plurality of ECG segments indicate myocardial infarction (720). In some examples, the ML model may be trained on a plurality of training ECG segments corresponding to the plurality of heartbeat time periods. In some examples, processing circuitry 50 may determine whether an amount of the plurality of ECG segments that indicate myocardial infarction satisfy a myocardial infarction threshold (730). In some examples, in response to the amount of the plurality of ECG segments that indicate myocardial infarction not satisfying the myocardial infarction threshold processing circuitry 50 may obtain additional ECG signal data (735), such as over a longer period of time or a period of time after initial ECG signal, and proceed back to (700) to separate the additional ECG signal data. In some examples, the myocardial infarction threshold may be at least 30% of the plurality of ECG segments indicating myocardial infarction. In some examples, the myocardial infarction threshold may be at least 50% of the plurality of ECG segments indicating myocardial infarction. In some examples, the myocardial infarction threshold may be at least 75% of the plurality of ECG segments indicating myocardial infarction. In some examples, processing circuitry 50 may be configured to determine the ECG signal indicates myocardial infarction in response to the amount of the plurality of ECG segments that indicate myocardial infarction satisfying the myocardial infarction threshold (740). In some examples, processing circuitry 50 may output an indication the ECG signal indicates myocardial infarction, such as AMI (750). For example, in response to determining the ECG signal indicates myocardial infarction, processing circuitry 50 may send an emergency alert notifying patient 4 to seek immediate medical attention, such as going to a hospital, and/or send an emergency alert to a clinician, first responder, or nearby bystanders that patient is suffering an AMI and needs medical attention.

[0064] In some examples, processing circuitry 50 separating the ECG signal 610 into the plurality of ECG segments, such as from a particular R-wave620 to a period of time 640 after a following R-wave 630, may enable a ML model to analyze an entire QRS complex as well as any ST elevations or depressions while also having the inputs to the ML model starting at a relatively same location (e.g., an R-wave such as R-wave 620). In some examples, separating an ECG signal into these particular ECG segments and applying these particular separated ECG segments to a ML model increases the robustness of the ML model in determining myocardial infarction based on ECG signal data, which may enable improved detection of myocardial infarction in real time.

[0065] In some examples, the ML model may be trained using a transfer learning approach. For example, the ML model may be trained on arrhythmia classification of a plurality of training ECG segments corresponding to the plurality of heartbeat time periods. In some examples, the ML model may be a convolutional neural network, such as a Residual Network.

[0066] In some examples, IMD 10 may obtain electrical signal data, such as an ECG signal or an EEG signal, from a single lead configured to sense an electrical signal routinely, such as, but not limited or, hourly, once every 12 hours, daily, nightly, weekly, etc. In some examples, IMD 10 may determine a particular morphology of the electrical signal is indicative of the particular health event, determine whether the particular health event occurred or predict the particular health event will occur based on the determined morphology of the electrical signal, and/or re-train a ML model based on the determined particular morphology routinely, such as, but not limited or, hourly, daily, nightly, weekly, bi-weekly, etc. Other devices, such as external device 12, server 94, and computing devices 110, may similarly determine a particular morphology of an electrical signal is indicative of the particular health event, determine whether the particular health event occurred or predict the particular health event will occur based on the determined morphology of the electrical signal, and/or re-train a ML model based on an electrical signal received from IMD 10, e.g., a digitized segment of a number of minutes of an electrical signal each day. In some examples, IMD 10 may transmit obtained electrical signal data, corresponding determined particular morphology of the electrical signal that is indicative of the particular health event, corresponding indication of the particular morphology of the electrical signal that is indicative of the particular health event, and/or determination the particular health event occurred or prediction the particular health event will occur based on the determined morphology of the electrical signal to external device 12. In some examples, IMD 10 may determine whether the electrical signal indicates a particular morphology of the electrical signal is indicative of the particular health event and/or an output indication the particular health event occurred or prediction the particular health event will occur based on the determined morphology of the electrical signal within real time. In some examples, IMD 10 may determine whether the electrical signal indicates a particular morphology of the electrical signal is indicative of the particular health event and/or an output indication the particular health event occurred or prediction the particular health event will occur based on the determined morphology of the electrical signal within a period of time of obtaining electrical signal data, such as, but not limited to, immediately, up to 1 hour, up to 6 hours, up to 12 hours, up to 1 day, up to 3 days, up to 1 week, up to 2 weeks, up to 1 month, etc. In some examples, electrical signal data may be obtained from an electronic health record (EHR) dataset.

[0067] FIG. 8 is a flow diagram illustrating an example technique of medical system 2. In some examples, as shown in FIG. 8, processing circuitry 50 may apply an electrical signal, such as an ECG signal or an EEG signal, obtained from a single lead, such as from an IMD, to a ML model to generate one or more predictions of whether the electrical signal indicates a particular health event (810). In some examples, processing circuitry 50 may separate the electrical signal into a plurality of segments and apply the plurality of segments to a ML model. For example, as shown in FIG. 6, processing circuitry 50 may separate an ECG signal into a plurality of ECG segments corresponding to a plurality of heartbeat time periods, a respective heartbeat time period of the plurality of time periods begins at a particular R-wave of the ECG signal and ends a period of time after a following R-wave, the following R-wave being immediately after the particular R-wave In some examples, the particular health event is a myocardial infarction or myocardial ischemia.

[0068] In some examples, processing circuitry 50 may determine a particular morphology of the electrical signal is indicative of the particular health event based on application of an integrated gradient to the one or more predictions (820). In some examples, integrated gradient may be an interpretability or explainability technique for deep neural networks to determine input feature importance to a ML model prediction. In some examples, processing circuitry 50 may apply an integrated gradient to compute the gradient of the ML model’s prediction output to the input features. For example, processing circuitry 50 may apply an integrated gradient to determine morphological features of the electrical signal that lead to a true prediction and/or morphological features of the electrical signal that lead to a false prediction and retrain the ML model with features identified to mostly or only contribute to a true prediction, which may help improve the ML model. In some examples, a particular morphology of the electrical signal that may be identified as indicative of the particular health event based on application of an integrated gradient to the one or more predictions may include at least one of ST elevation, ST depression, deep Q waves, or deep S waves. In some examples, particular morphology of the electrical signal that may be indicative of the particular health event based on application of an integrated gradient to the one or more predictions may additionally or alternatively include at least one of inverted T waves, short QRS, prolonged QRS, large P waves, or deviations between T to P region.

[0069] FIGS. 9A-9D show examples of processing circuitry 50 determining a particular morphology of an ECG signal 910 is indicative of the particular health event based on application of an integrated gradient to the one or more predictions. As shown in FIGS. 9A-9D, processing circuitry 50 may determine particular morphology features 920 of the ECG signal 910 are more indicative of a particular health event, such as myocardial infarction, while determining other morphology features 930 of the ECG signal 910 are less indicative of a particular health event, such as myocardial infarction. For example, as shown in FIG. 9B, processing circuitry 50 may determine particular morphology features 920 of the ECG signal 910 that are more indicative of a particular health event (e.g., AMI) include a deep Q wave and inverted T wave with little to no ST elevation or depression of the ECG signal 910 and may be indicative of AMI. For example, as shown in FIG. 9C, processing circuitry 50 may determine particular morphology features 920 of the ECG signal 910 that are more indicative of a particular health event (e.g., AMI) include an inverted P wave as well as an inverted T wave, with little to no ST elevation or depression of the ECG signal 910 and may be indicative of AMI occurring in within the next 24 to 48 hours. In addition, processing circuitry 50 may determine particular morphology features 920 of the ECG signal 910 that are more indicative of a particular health event (e.g., myocardial infarction or ischemia) include an ECG signal having P-wave shape differing from a “normal” P wave shape and was indicative of myocardial infarction or ischemia. For example, as shown in FIG. 9D, processing circuitry 50 may determine particular morphology features 920 of the ECG signal 910 that are more indicative of a particular health event (e.g., AMI) include an abnormal inverted T wave and deep Q wave of the ECG signal 910 and may be indicative of acute myocardial infarction occurring in within the next 24 to 48 hours. [0070] In some examples, processing circuitry 50 may output an indication of the particular morphology of the electrical signal that is indicative of the particular health event (830). In some examples, processing circuitry 50 may determine whether the particular health event occurred or predict the particular health event will occur based on the determined morphology of the electrical signal (835). For example, as shown in FIG. 9C, processing circuitry may determine an ECG signal 910 includes an inverted P wave as well as an inverted T wave, with little to no ST elevation or depression that re indicative of acute myocardial infarction occurring in within the next 24 to 48 hours and processing circuitry 50 may generate and output a prediction that the particular health event, such as myocardial infarction, will occur within a short period of time, such as within 24 or 48 hours. In some examples, processing circuitry 50 may also send emergency alerts based on the generated prediction. For example, processing circuitry 50 may send an emergency alert notifying patient to seek medical attention, such as going to a hospital, in response to generating a prediction an acute health event, such as AMI, will occur within the next 24 or 48 hours.

[0071] In some examples, processing circuitry 50 may re-train the ML model based on the determined particular morphology of the electrical signal that is indicative of the particular health event to continuously improve specificity and sensitivity of the ML model. In some examples, re-training the ML model using the determined particular morphology of the electrical signal that is indicative of the particular health event may further train the ML model on particular morphological features of the electrical signal to focus on and/or the particular morphological features of the electrical signal to avoid when determining whether the particular health event occurred or predict the particular health event will occur based on the determined morphology of the electrical signal. In some examples, re-training the ML model based on the data of the particular patient may help improve the accuracy of a ML model based the location and position of the IMD 10 in the body of the patient which may personalize the ML model to the particular IMD 10 and the particular patient and provide improved specificity and sensitivity of the personalized ML model.

[0072] In accordance with techniques of this disclosure, particular health events, such as AMI, may be detected and/or predicted earlier without a patient needing to go a hospital. Accordingly, medical intervention and/or treatment due to particular health events, such as AMI, may be applied sooner which may lower the risk of long-term complications of heart failure patients and may reduce mortality and/or morbidity.

[0073] FIG. 10 is an example of a machine learning model 1002 being trained using supervised and/or reinforcement learning techniques. Machine learning model 1002 may correspond to any machine learning model described herein, e.g., machine learning model discussed with respect to FIGS. 6-9D. The machine learning model 1002 may be implemented using any number of models for supervised and/or reinforcement learning, such as but not limited to, an artificial neural network, a convolutional neural network, such as a Residual Network (ResNet), a decision tree, naive Bayes network, support vector machine, or k-nearest neighbor model, to name only a few examples. In some examples, one or more of IMD 10, external device 12, server 94, and/or computing device(s) 100 initially trains the machine learning model 1002 based on a training set of metrics and corresponding to a particular health event, such as myocardial infarction or arrhythmia, and/or classification of a particular health event, such as myocardial infarction or arrhythmia. The training set 1000 may include a set of feature vectors, where each feature in the feature vector represents a value for a particular metric. One or more of IMD 10, external device 12, server 94, and/or computing device(s) 100 may select a training set comprising a set of training instances, each training instance comprising an association between one or more respective electrical signals, such as ECG signals or a plurality of ECG segments of an ECG signal, and a particular health event, such as myocardial infarction or arrhythmia and/or a classification of a particular health event, such as myocardial infarction or arrhythmia. A prediction or classification by the machine learning model 1002 may be compared 1004 to the target output 1003, and an error signal and/or machine learning model weights modification may sent/applied to the machine learning model 1002 based on the comparison to learn/train 1005 the machine learning model to modify/update the machine learning model 1002. For example, one or more of IMD 10, external device 12, server 94, and/or computing device(s) 100 may, for each training instance in the training set, modify, based on the respective electrical signals, such as ECG signals or a plurality of ECG segments of an ECG signal, a determination of particular health event, such as myocardial infarction or arrhythmia and/or a classification of a particular health event, such as myocardial infarction or arrhythmia of the training instance, the machine learning model 1002 to change a score generated by the machine learning model 1002 in response to subsequent electrical signals, such as ECG signals or a plurality of ECG segments of an ECG signal, applied to the machine learning model 1002. [0074] FIG. 11 is a conceptual diagram illustrating an example machine learning model 1100, such as described with respect to FIGS. 6-9D, configured to generate one or more values indicative of a particular health event or of a risk of a particular health event, e.g., myocardial infarction, AMI, based on physiological parameter values, e.g., sensed by an IMD and/or other devices as described herein. Machine learning model 1100 is an example of a deep learning model, or deep learning algorithm, such as a Residual Network. One or more of IMD 10, external device 12, or sever 94 may train, store, and/or utilize machine learning model 1100, but other devices may apply inputs associated with a particular patient to machine learning model 1100 in other examples. Some non-limiting examples of machine learning techniques include Bayesian probability models, Support Vector Machines, K-Nearest Neighbor algorithms, and Multi-layer Perceptron.

[0075] As shown in the example of FIG. 11, machine learning model 1100 may include three layers. These three layers include input layer 1102, hidden layer 1104, and output layer 1106. Output layer 1106 comprises the output from the transfer function 1105 of output layer 1106. Input layer 1102 represents each of the input values XI through X4 provided to machine learning model 1100. The number of inputs may be less than or greater than 4, including much greater than 4, e.g., hundreds or thousands. In some examples, the input values may be any of the of physiological or other patient parameter values described herein. In some examples, the input values may include an electrical signal, such as an ECG signal or an EEG signal, or a plurality of ECG segments of an ECG signal.

[0076] Each of the input values for each node in the input layer 1102 is provided to each node of hidden layer 1104. In the example of FIG. 11, hidden layers 1104 include two layers, one layer having four nodes and the other layer having three nodes, but fewer or greater number of nodes may be used in other examples. Each input from input layer 1102 is multiplied by a weight and then summed at each node of hidden layers 1104. During training of machine learning model 1100, the weights for each input are adjusted to establish the relationship between input values and one or more output values indicative of a particular health event or of a risk of a particular health event. In some examples, one hidden layer may be incorporated into machine learning model 1100, or three or more hidden layers may be incorporated into machine learning model 1100, where each layer includes the same or different number of nodes.

[0077] The result of each node within hidden layers 1104 is applied to the transfer function of output layer 1106. The transfer function may be liner or non-linear, depending on the number of layers within machine learning model 1100. Example non-linear transfer functions may be a sigmoid function or a rectifier function. The output 1307 of the transfer function may be a value or values indicative of a particular health event or of a risk of a particular health event, e.g., myocardial infarction, AMI. By applying the plurality of ECG segments to a machine learning model, such as machine learning model 1100, processing circuitry of system 2 is able to determine an indication of myocardial infarction or a classification of myocardial infarction with great accuracy, specificity, and sensitivity. By applying the electrical signals to a machine learning model, such as machine learning model 1100, processing circuitry of system 2 is able to determine a particular morphology of the electrical signal is indicative of a particular health event and/or determine whether the particular health event occurred or predict the particular health event will occur based on the determined morphology of the electrical signal with great accuracy, specificity, and sensitivity.

[0078] The techniques described in this disclosure may be implemented, at least in part, in hardware, software, firmware, or any combination thereof. For example, various aspects of the techniques may be implemented within one or more microprocessors, DSPs, ASICs, FPGAs, or any other equivalent integrated or discrete logic QRS circuitry, as well as any combinations of such components, embodied in external devices, such as physician or patient programmers, stimulators, or other devices. The terms “processor” and “processing circuitry” may generally refer to any of the foregoing logic circuitry, alone or in combination with other logic circuitry, or any other equivalent circuitry, and alone or in combination with other digital or analog circuitry.

[0079] For aspects implemented in software, at least some of the functionality ascribed to the systems and devices described in this disclosure may be embodied as instructions on a computer-readable storage medium such as RAM, DRAM, SRAM, magnetic discs, optical discs, flash memories, or forms of EPROM or EEPROM. The instructions may be executed to support one or more aspects of the functionality described in this disclosure. [0080] In addition, in some aspects, the functionality described herein may be provided within dedicated hardware and/or software modules. Depiction of different features as modules or units is intended to highlight different functional aspects and does not necessarily imply that such modules or units must be realized by separate hardware or software components. Rather, functionality associated with one or more modules or units may be performed by separate hardware or software components, or integrated within common or separate hardware or software components. Also, the techniques could be fully implemented in one or more circuits or logic elements. The techniques of this disclosure may be implemented in a wide variety of devices or apparatuses, including an IMD, an external programmer, a combination of an IMD and external programmer, an integrated circuit (IC) or a set of ICs, and/or discrete electrical circuitry, residing in an IMD and/or external programmer.

[0081] The following examples are illustrative of the techniques described herein. [0082] Example 1 : A system includes an implantable medical device configured to sense an electrocardiogram (ECG) signal of a patient via a single lead; and processing circuitry configured to: separate the ECG signal into a plurality of ECG segments corresponding to a plurality of heartbeat time periods; apply the plurality of ECG segments to a machine learning (ML) model; and determine whether one or more of the plurality of ECG segments indicate myocardial infarction, the model being trained on a plurality of training ECG segments corresponding to the plurality of heartbeat time periods.

[0083] Example 2: The system of example 1, wherein a respective heartbeat time period of the plurality of time periods begins at a particular R-wave of the ECG signal and ends a period of time after a following R-wave, the following R-wave being a next R- wave immediately after the particular R-wave.

[0084] Example 3 : The system of example 2, wherein the period of time after the following R-wave is within 50 milliseconds after the following R-wave.

[0085] Example 4: The system of example 2, wherein the period of time after the following R-wave is within 20 sample points of the ECG signal after the following R- wave. [0086] Example 5: The system of any of examples 1-4, wherein the processing circuitry is further configured to determine whether an amount of the plurality of ECG segments that indicate myocardial infarction satisfies a myocardial infarction threshold. [0087] Example 6: The system of example 5, wherein the myocardial infarction threshold is at least 30% of the plurality of ECG segments.

[0088] Example 7: The system of any of examples 5-6, wherein the processing circuitry is further configured to determine the ECG signal indicates myocardial infarction in response to the amount of the plurality of ECG segments that indicate myocardial infarction satisfying the myocardial infarction threshold.

[0089] Example 8: The system of any of examples 1-7, wherein the ML model is trained on arrhythmia classification.

[0090] Example 9: The system of any of examples 1-8, wherein the ML model is a deep learning model.

[0091] Example 10: The system of any of examples 1-9, wherein the processing circuitry is further configured to output an indication that the one or more of the plurality of ECG segments indicate myocardial infarction.

[0092] Example 11 : The system of any of examples 1-10, wherein the processing circuitry is positioned in the implantable medical device.

[0093] Example 12: The system of any of examples 1-10, wherein the processing circuitry is positioned in an external computing device configured to be in wireless communication with the implantable medical device.

[0094] Example 13: The system of any of examples 1-10, wherein the processing circuitry is positioned in one or more of a cloud computing device or server.

[0095] Example 14: A system includes an implantable medical device configured to sense an electrical signal of a patient via a single lead; and processing circuitry configured to: apply the electrical signal of the patient to a machine learning (ML) model to generate one or more predictions of whether the electrical signal indicates a particular health event; apply an integrated gradient to the one or more predictions; determine a particular morphology of the electrical signal that is indicative of the particular health event based on the application of the integrated gradient to the one or more predictions; and output an indication of the particular morphology of the electrical signal that is indicative of the particular health event. [0096] Example 15: The system of example 14, wherein the processing circuitry is further configured to determine whether the particular health event occurred or predict the particular health event will occur based on the determined morphology of the electrical signal.

[0097] Example 16: The system of any of examples 14-15, wherein the processing circuitry is further configured to re-train the ML model based on the determined particular morphology of the electrical signal that is indicative of the particular health event.

[0098] Example 17: The system of any of examples 14-16, wherein the particular health event is a myocardial infarction.

[0099] Example 18: The system of any of examples 14-17, wherein the electrical signal is an electrocardiogram (ECG) signal.

[0100] Example 19: The system of any of examples 14-18, wherein the particular morphology of the electrical signal includes at least one of ST elevation, ST depression, deep Q waves, or deep S waves.

[0101] Example 20: The system of any of examples 14-16, wherein the electrical signal is an electroencephalogram (EEG) signal.

[0102] Example 21 : The system of any of examples 14-20, wherein the processing circuitry is positioned in the implantable medical device.

[0103] Example 22: The system of any of examples 14-20, wherein the processing circuitry is positioned in an external computing device configured to be in wireless communication with the implantable medical device.

[0104] Example 23: The system of any of examples 14-20, wherein the processing circuitry is positioned in one or more of a cloud computing device or server.

[0105] Example 24: The system of any of examples 14-23, wherein the implantable medical device is configured to continuously sense the electrical signal of the patient via the single lead.

[0106] Example 25: A method includes receiving, via an implantable medical device, an electrocardiogram (ECG) signal of a patient sensed via a single lead of the implantable medical device; separating the ECG signal into a plurality of ECG segments corresponding to a plurality of heartbeat time periods; applying the plurality of ECG segments to a machine learning (ML) model; and determining whether one or more of the plurality of ECG segments indicate myocardial infarction, the model being trained on a plurality of training ECG segments corresponding to the plurality of heartbeat time periods.

[0107] Example 26: The method of example 25, wherein a respective heartbeat time period of the plurality of time periods begins at a particular R-wave of the ECG signal and ends a period of time after a following R-wave, the following R-wave being immediately after the particular R-peak.

[0108] Example 27: The method of example 26, wherein the period of time after the following R-wave is within 50 milliseconds after the following R-wave.

[0109] Example 28: The method of example 26, wherein the period of time after the following R-wave is within 20 sample points of the ECG signal after the following R- wave.

[0110] Example 29: The method of any of examples 25-28, wherein the method further comprises: determining whether an amount of the plurality of ECG segments that indicate myocardial infarction satisfies a myocardial infarction threshold.

[0111] Example 30: The method of example 29, wherein the myocardial infarction threshold is at least 30% of the plurality of ECG segments indicate myocardial infarction.

[0112] Example 31 : The method of any of examples 29-30, wherein the method further comprises: determining the ECG signal indicates myocardial infarction in response to the amount of the plurality of ECG segments that indicate myocardial infarction satisfying the myocardial infarction threshold.

[0113] Example 32: The method of any of examples 25-31, wherein the ML model is trained on arrhythmia classification.

[0114] Example 33: The method of any of examples 25-32, wherein the ML model is a deep learning model.

[0115] Example 34: The method of any of examples 25-33, wherein the method further comprises: outputting an indication that the one or more of the plurality of ECG segments indicate myocardial infarction.

[0116] Example 35: A method includes receiving, via an implantable medical device, an electrical signal of a patient sensed via a single lead; applying the electrical signal of the patient to a machine learning (ML) model to generate one or more predictions of whether the electrical signal indicates a particular health event; applying an integrated gradient to the one or more predictions; determining a particular morphology of the electrical signal that is indicative of the particular health event based on the application of the integrated gradient to the one or more predictions; and outputting an indication of the particular morphology of the electrical signal that is indicative of the particular health event.

[0117] Example 36: The method of example 35, wherein the method further comprises: determining whether the particular health event occurred or predict the particular health event will occur based on the determined morphology of the electrical signal.

[0118] Example 37: The method of any of examples 35-36, wherein the method further comprises: re-training the ML model based on the determined particular morphology of the electrical signal that is indicative of the particular health event.

[0119] Example 38: The method of any of examples 35-37, wherein the particular health event is a myocardial infarction.

[0120] Example 39: The method of any of examples 35-38, wherein the electrical signal is an electrocardiogram (ECG) signal.

[0121] Example 40: The method of any of examples 35-39, wherein the particular morphology of the electrical signal includes at least one of ST elevation, ST depression, deep Q waves, or deep S waves.

[0122] Example 41 : The method of any of examples 35-37, wherein the electrical signal is an electroencephalogram (EEG) signal.

[0123] Example 42: The method of any of examples 35-41, wherein receiving, via the implantable medical device, the electrical signal of the patient sensed via the single lead comprises continuously sensing the electrical signal of the patient.

[0124] Various examples have been described. These and other examples are within the scope of the following claims.

Claims

CLAIMS What is claimed is:

1. A system comprising: an implantable medical device configured to sense an electrocardiogram (ECG) signal of a patient via a single lead; and processing circuitry configured to: separate the ECG signal into a plurality of ECG segments corresponding to a plurality of heartbeat time periods; apply the plurality of ECG segments to a machine learning (ML) model; and determine whether one or more of the plurality of ECG segments indicate myocardial infarction, the model being trained on a plurality of training ECG segments corresponding to the plurality of heartbeat time periods.

2. The system of claim 1, wherein a respective heartbeat time period of the plurality of time periods begins at a particular R-wave of the ECG signal and ends a period of time after a following R-wave, the following R-wave being a next R-wave immediately after the particular R-wave.

3. The system of claim 2, wherein the period of time after the following R-wave is within 50 milliseconds after the following R-wave.

4. The system of claim 2, wherein the period of time after the following R-wave is within 20 sample points of the ECG signal after the following R-wave.

5. The system of any of claims 1-4, wherein the processing circuitry is further configured to determine whether an amount of the plurality of ECG segments that indicate myocardial infarction satisfies a myocardial infarction threshold.

6. The system of claim 5, wherein the myocardial infarction threshold is at least 30% of the plurality of ECG segments.

7. The system of any of claims 5-6, wherein the processing circuitry is further configured to determine the ECG signal indicates myocardial infarction in response to the amount of the plurality of ECG segments that indicate myocardial infarction satisfying the myocardial infarction threshold.

8. The system of any of claims 1-7, wherein the ML model is trained on arrhythmia classification.

9. The system of any of claims 1-8, wherein the processing circuitry is further configured to output an indication that the one or more of the plurality of ECG segments indicate myocardial infarction.

10. The system of any of claims 1-9, wherein the processing circuitry is positioned in the implantable medical device.

11. A system comprising: an implantable medical device configured to sense an electrical signal of a patient via a single lead; and processing circuitry configured to: apply the electrical signal of the patient to a machine learning (ML) model to generate one or more predictions of whether the electrical signal indicates a particular health event; apply an integrated gradient to the one or more predictions; determine a particular morphology of the electrical signal that is indicative of the particular health event based on the application of the integrated gradient to the one or more predictions; and output an indication of the particular morphology of the electrical signal that is indicative of the particular health event.

12. The system of claim 11, wherein the processing circuitry is further configured to determine whether the particular health event occurred or predict the particular health event will occur based on the determined morphology of the electrical signal.

13. The system of any of claims 11-12, wherein the processing circuitry is further configured to re-train the ML model based on the determined particular morphology of the electrical signal that is indicative of the particular health event.

14. The system of any of claims 11-13, wherein the electrical signal is an electrocardiogram (ECG) signal.

15. The system of any of claims 11-14, wherein the particular morphology of the electrical signal includes at least one of ST elevation, ST depression, deep Q waves, or deep S waves.