Medical device for detecting compromised atrial mechanical function

Abstract

A medical device is configured to sense a cardiac mechanical signal such as a motion signal sensed by a motion sensor of the medical device. The medical device may determine a maximum amplitude of the motion signal from multiple cardiac cycles and detect a compromised atrial function episode based on the motion sensor signal maximum amplitudes meeting a threshold criterion. The medical device may generate an output in response to detecting the compromised atrial function episode and store the output in a memory of the medical device.

Description

MEDICAL DEVICE FOR DETECTING COMPROMISED ATRIAL MECHANICAL FUNCTIONREFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Patent Application Serial No. 63 / 678,743, filed August 2, 2024, the entire content of which is incorporated herein by reference.TECHNICAL FIELD

[0002] This disclosure relates to medical devices and methods for detecting compromised atrial mechanical function from a cardiac mechanical signal.BACKGROUND

[0003] During normal sinus rhythm (NSR), the heartbeat is regulated by electrical signals produced by the sino-atrial (SA) node located in the right atrial wall. Each depolarization signal produced by the SA node spreads across the atria, causing the depolarization and contraction of the atria, and arrives at the atrioventricular (AV) node. The AV node responds by propagating a depolarization signal through the bundle of His of the atrioventricular septum and thereafter to the left and right bundle branches in the interventricular septum and the Purkinje fibers of the right and left ventricles, sometimes referred to as the ' His- Purkinje system” and referred to herein as the "conduction system.”

[0004] Patients with poor SA node function, poor AV node conduction (sometimes referred to as AV block), or conduction system abnormalities of the His bundle or left and / or right bundle branches (sometimes referred to as bundle branch block) or other conduction system abnormalities may receive a pacemaker to restore a more normal heart rhythm and heart chamber synchrony. Atrial pacing may be performed to provide a regular atrial rate in a patient having SA node dysfunction. Ventricular pacing may be performed to promote a regular ventricular rate in a patient having AV conduction abnormalities. Implantable cardiac pacemakers can be placed in a subcutaneous pocket and coupled to one or more transvenous medical electrical leads carrying pacing and sensing electrodes positioned in the heart. A cardiac pacemaker implanted subcutaneously may be a single chamber pacemaker coupled to one transvenous medical lead for positioning electrodes in one heart chamber, atrial or ventricular, or a dual chamber pacemaker coupled to two transvenous, intracardiac leads for positioning electrodes in both an atrial and a ventricular chamber. Multi-chamber pacemakers are also available that may be coupled to three leads, for example, for positioning electrodes for pacing and sensing in one atrial chamber andboth the right and left ventricles to provide cardiac resynchronization therapy (CRT) to promote atrial and ventricular synchrony.

[0005] Leadless pacemakers have been proposed or introduced commercially that are implantable within a ventricular chamber of a patient's heart for delivering ventricular pacing pulses. Such a pacemaker may sense R-wave signals attendant to intrinsic ventricular depolarizations and deliver ventricular pacing pulses in the absence of sensed R-waves. While single chamber ventricular sensing and pacing by an intracardiac ventricular pacemaker may adequately address some heart rhythm conditions, some patients may benefit from atrial and ventricular (dual chamber) sensing for providing atrial-synchronized ventricular pacing by a leadless intracardiac ventricular pacemaker in order to maintain a more normal heart rhythm. Atrial synchronous ventricular pacing can be achieved when the leadless ventricular pacemaker is configured to sense atrial systolic events from an acceleration signal sensed by an accelerometer, for example.

[0006] Atrial tachyarrhythmias are atrial rhythms that may arise from a non-sinus node location and occur with a relatively high rate of incidence. Atrial fibrillation may be the most common form of arrhythmia. Non-sinus atrial tachycardia (AT) and atrial fibrillation (AF) can lead to serious and life-threatening complications, including blood clots, stroke, heart failure and more serious arrhythmias. Atrial tachyarrhythmias, while highly prevalent, tend to be underdiagnosed and undertreated.SUMMARY

[0007] The techniques of this disclosure generally relate to a medical device and method for detecting comprised atrial mechanical function, which in some instances may be due to atrial tachyarrhythmia. The techniques disclosed herein may be implemented in an implantable medical device (IMD) system that is capable of sensing a cardiac mechanical signal, e.g., a cardiac motion signal. The IMD may be a leadless pacemaker implantable in a ventricular chamber of the heart in some examples. The IMD may sense a signal from a motion sensor, such as an accelerometer, that produces a cardiac motion signal responsive to atrial mechanical contraction. An IMD operating according to the techniques disclosed herein may determine maximum peak amplitudes from the cardiac motion signal during cardiac cycles and perform an analysis of the maximum peak amplitudes for detecting compromised atrial mechanical function, which may be due to atrial tachyarrhythmia, e.g., AT or AF, and / or termination of the compromised atrial mechanical function, e.g.. termination of an atrial tachyarrhythmia episode.

[0008] In one example, the disclosure provides a medical device including a pulse generator configured to generate ventricular pacing pulses, a motion sensor configured to sense a motion signal and a control circuit. The control circuit may be configured to, for multiple cardiac cycles that may include ventricular pacing pulses delivered by the pulse generator, determine a maximum amplitude of the motion signal. The control circuit may determine a representative value from the maximum amplitudes, compare the representative value to a threshold criterion, and detect a compromised atrial mechanical function episode in response to at least the representative value meeting the threshold criterion. The control circuit may generate an output in response to detecting the compromised atrial mechanical function episode. The medical device may further include a memory configured to store the output generated by the control circuit. The medical device may further include a telemetry circuit configured to transmit the output.

[0009] In another example, the disclosure provides a method including sensing a motion signal, determining maximum amplitudes of the motion signal sensed during a plurality of cardiac cycles, which may include ventricular pacing cycles, and determining a representative value from the maximum amplitudes. The method may include comparing the representative value to a threshold criterion, detecting an episode of compromised atrial mechanical function in response to at least the representative value meeting the threshold criterion, generating an output in response to detecting the episode and transmitting the output.

[0010] In yet another example, the disclosure provides a non-transitory. computer-readable storage medium comprising a set of instructions which, when executed by a control circuit of a medical device, cause the medical device to sense a motion signal, determine maximum amplitudes of the motion signal sensed during a plurality of cardiac cycles that may include ventricular pacing cycles and determine a representative value from the maximum amplitudes. The instructions may further cause the medical device to compare the representative value to a threshold criterion, detect an episode of compromised atrial mechanical function in response to at least the representative value meeting the threshold criterion, generate an output in response to detecting the episode and transmit the output.

[0011] The details of one or more aspects of the disclosure are set forth in the accompanying drawings and the description below. Other features, objects, and advantages of the techniques described in this disclosure will be apparent from the description and drawings, and from the claims.BRIEF DESCRIPTION OF DRAWINGS

[0012] FIG. 1 is a diagram illustrating an implantable medical device (IMD) system that may be used to sense cardiac signals and detect heart rhythms.

[0013] FIG. 2 is a diagram of the IMD shown in FIG. 1.

[0014] FIG. 3 is a diagram of an example configuration of the IMD shown in FIG. 1.

[0015] FIG. 4 is a diagram of a cardiac electrical signal and a motion sensor signal that may be sensed by a medical device.

[0016] FIG. 5 is an example of motion sensor signals that may be sensed by a medical device over two different cardiac cycles.

[0017] FIG. 6 is a flow chart of a method for detecting compromised atrial mechanical function that may be performed by a medical device capable of sensing a cardiac mechanical signal according to some examples.

[0018] FIG. 7 is a flow' chart of a method for labeling individual cardiac cycles that occur during an atrial function analysis interval.

[0019] FIG. 8 is a flow chart of a method for classifying an interval for detecting compromised atrial mechanical function according to some examples.

[0020] FIG. 9 is a diagram of a histogram of atrial systolic event sensing window maximum amplitudes that may be determined by a medical device and stored in its memory over an atrial function analysis interval.

[0021] FIG. 10 is a diagram of a histogram of atrial systolic event sensing window maximum amplitudes that may be determined and stored by a medical device over another atrial function analysis interval.

[0022] FIG. 11 is a flow chart of a method that may be performed by a medical device for accepting or rejecting maximum amplitudes of a motion sensor signal for use in detecting compromised atrial function according to some examples.

[0023] FIG. 12 is a flow chart of a method for classifying atrial function analysis intervals based on motion sensor signal analysis according to some examples.

[0024] FIG. 13 is a diagram of a method for detecting an atrial tachyarrhythmia episode by a medical device according to some examples.

[0025] FIG. 14 is a diagram of another method for detecting an atrial tachyarrhythmia episode by a medical device according to some examples.

[0026] FIG. 15 is a flow chart of a method for classify ing an AF analysis interval by a medical device according to another example.

[0027] FIG. 16 is a flow chart of a method for establishing AF interval criteria thresholds by a medical device according to some examples.DETAILED DESCRIPTION

[0028] In general, this disclosure describes medical devices and techniques for detecting compromised atrial mechanical function. Compromised atrial mechanical function may be a loss or decline in the atrial mechanical contraction strength or “atrial kick” that normally occurs during the ventricular filling phase. The compromised atrial mechanical function may be due to atrial tachyarrhythmia or other atrial dysfunction, which may be electrical or mechanical dysfunction. In some examples, the medical device may detect the compromised atrial mechanical function as an episode of atrial tachyarrhythmia, e.g., an episode of AF. In the illustrative examples presented herein, a ventricular pacemaker is configured to sense atrial systolic events for synchronizing the ventricular pacing pulses to the atrial rate. As described below, the atrial systolic events may be sensed from a signal produced by a motion sensor that includes an atrial systolic event signal corresponding to atrial mechanical contraction and the active filling phase of the ventricle, sometimes referred to as the “atrial kick.” In other examples, atrial systolic event sensing may be performed using other techniques, such as sensing the atrial systolic event from another cardiac mechanical signal (e.g., a pressure signal, acoustical signal, impedance signal, etc.).

[0029] FIG. 1 is a diagram illustrating an implantable medical device (IMD) system 10 that may be used to sense cardiac signals and detect heart rhythms. IMD system 10 includes an IMD 14 configured to detect atrial tachyarrhythmia, or more generally compromised atrial mechanical function, according to the techniques disclosed herein. In some examples, IMD 10 can provide cardiac electrical stimulation therapy, e.g., cardiac pacing, to a patient’s heart 8. However, an IMD operating according to the techniques disclosed herein for detecting an episode of compromised atrial mechanical function does not necessarily have therapy delivery capabilities and may be provided for detecting and tracking episodes of compromised atrial mechanical function in a patient for use in managing cardiac rhythm therapies, nsk of stroke, etc.

[0030] In the example shown, IMD 14 is a leadless ventricular pacemaker, which may be configured as a transcatheter intracardiac pacemaker adapted for implantation wholly within a heart chamber, e.g., wholly within the right ventricle (RV) of heart 8. IMD 14 may be reduced in size compared to subcutaneously implanted pacemakers and may be generally cylindrical in shape to enable transvenous implantation via a delivery catheter.

[0031] In FIG. 1, IMD 14 is shown implanted along an endocardial wall of the RV. e.g., near the RV apex though other endocardial RV locations are possible, e.g., along the interventricular septum or the lateral free wall. IMD 14 may be positioned within or on the right ventricle or left ventricle to provide respective right ventricular or left ventricular pacing and cardiac signal sensing. It is to be understood that the techniques disclosed herein for sensing cardiac signals for detecting atrial tachyarrhythmia or compromised atrial mechanical function are not limited to a particular implant location of IMD 14 and other positions different than the implant position shown in FIG. 1 are possible. For example, IMD 14 may be implanted along the interventricular septum for sensing cardiac signals and, in some examples, delivering ventricular pacing via the septal myocardium and / or via conduction system pacing. For instance, IMD 14 may be implanted at a location along the interventricular septum 12 to operatively position a pacing electrode vector for delivering pacing pulses in the area of the His bundle, the left bundle branch, the right bundle branch or Purkinje fibers.

[0032] Furthermore, in some examples, IMD 14 may be implanted in an atrial chamber for sensing cardiac signals and optionally delivering atrial and / or ventricular pacing pulses from an atrial implant location. IMD 14 may be positioned within the right atrium (RA), for example, for providing ventricular pacing from an atrial implant location, which may include ventricular pacing of myocardial tissue and / or the His bundle via a RA approach. In this case, IMD 14 may be advanced into the RA with a tissue piercing tip electrode (not shown in FIG. 1 ) that is advanced into the interatrial septum tow ard the His bundle, e.g., at the inferior end of the interatrial septum. The tip electrode of IMD 14, which may be provided as an elongated or helical tissue piercing electrode, may be advanced toward the His bundle from a location beneath the AV node and near the tricuspid valve annulus, generally in the Triangle of Koch, to provide ventricular pacing via the conduction system.

[0033] In examples presented herein, IMD 14 is capable of generating electrical stimulation pulses, e.g., cardiac pacing pulses, delivered to heart 8 via one or more electrodes on the outer housing of IMD 14 (see FIG. 2). IMD 14 is configured to deliver ventricular pacing pulses and sense a cardiac electrical signal using housing-based electrodes for producing a ventricular electrogram (EGM) signal. The cardiac electrical signal may be sensed using the housing based electrodes that are also used to deliver pacing pulses to the ventricles in some examples. IMD 14 may be configured to control the delivery of ventricular pacing pulses in a manner that promotes synchrony between atrial activation and ventricular activation, e.g.. by delivering ventricular pacing pulses at an atrioventricular (AV) interval after sensed atrial events. That is, IMD 14 controls ventricular pacing pulse deli very to promote a desired AV interval between atrialcontractions corresponding to atrial systole and ventricular pacing pulses delivered to cause ventricular depolarization and ventricular systole.

[0034] According to the illustrative examples described herein, atrial systolic events producing the active ventricular filling phase of the cardiac cycle can be detected by IMD 14 from a motion sensor signal, such as an accelerometer signal. The motion sensor may be enclosed by the housing of IMD 14 when IMD 14 is implanted within or on a heart chamber. The motion signal produced by an accelerometer implanted within the RV, for example, includes motion signals caused by cardiac mechanical events, including ventricular and atrial mechanical events. Acceleration of blood flowing into the RV through the tricuspid valve 6 between the RA and RV caused by atrial systole and other cardiac motion can be detected by IMD 14 from an acceleration signal produced by an accelerometer included in IMD 14, for example. Other motion signals detected by IMD 14, such as motion caused by ventricular contraction, motion caused by ventricular relaxation, and motion caused by passive filling of the ventricles are described below, e.g., in conjunction with FIG. 4.

[0035] IMD 14 may be configured to sense an atrial event signal corresponding to atrial mechanical activation or atrial systole from the motion sensor signal. It is contemplated that other ty pes of sensors of cardiac mechanical or hemodynamic function may be used to produce a cardiac mechanical signal to enable IMD 14 to sense atrial systolic event signals from the cardiac mechanical signal. Such sensors may include impedance sensors (which produce a signal correlated to blood volume in the ventricle), pressure sensors, acoustical sensors or other sensors that produce a signal correlated to the mechanical contractions of the heart chambers.

[0036] Ventricular pacing pulses can be synchronized to the sensed atrial systolic event signals, which may be detected from the accelerometer signal as described below or any of the other cardiac mechanical signals listed above, by setting a programmable AV pacing interval that controls the timing of the ventricular pacing pulse relative to the detected atrial systolic event signal. The AV pacing interval may be set to about 10 to 100 ms, in some examples, to control IMD 14 to deliver a ventricular pacing pulse synchronized to the atrial systolic event signal sensed from the motion signal. As described below, the cardiac mechanical signal, e.g., a motion signal or any of the other cardiac mechanical signals listed as examples herein, may be processed and analyzed according to the techniques disclosed herein.

[0037] In some instances, the atrial systolic event signal may not be sensed by IMD 14. In addition to setting an AV pacing interval in response to sensing an atrial systolic event signal, IMD 14 may start a ventricular pacing interval, which may be referred to as a ventricular lower rate interval (LRI), in response to delivered ventricular pacing pulses and sensed ventricularevent signals, e.g., sensed R- waves. The ventricular LRI may correspond to a programmed lower rate, e.g., 40 to 60 beats per minute. If the ventricular LRI expires without sensing an atrial systolic event signal, IMD 14 may deliver a ventricular pacing pulse at the LRI. The LRI set in response to a ventricular pacing pulse or sensed R-wave may be set to a temporary LRI corresponding to a faster pacing rate than the programmed lower rate in some instances. For example, the LRI may be set to a temporary LRI to provide rate response pacing. In some instances, the LRI may be set to a rate smoothing interval that is determined based on the actual ventricular rate (paced and / or sensed), which may be faster than the programmed lower rate, to avoid abrupt changes in the ventricular rate when an atrial systolic event signal is not sensed. In other instances, an atrial systolic event signal may not be sensed during a ventricular cycle when the IMD 14 senses a ventricular event, e.g., a ventricular R-wave, from the cardiac electrical signal, before an atrial systolic event signal is sensed. An early sensed ventricular event signal may be a premature ventricular contraction (PVC) occurring at a relatively short ventricular interval or due to an increasing intrinsic heart rate. As further described below, IMD 14 may be configured to analyze the amplitude of the sensed cardiac motion signal in combination with the occurrence of delivered ventricular pacing pulses, sensed ventricular electrical event signals and / or sensed atrial systolic event signals for detecting compromised atrial mechanical function, which may include detecting an atrial tachyarrhythmia episode and termination of the detected atrial tachyarrhythmia episode.

[0038] Medical device system 10 is shown including an external medical device 50 for receiving data from IMD 14 and for transmitting programming commands to IMD 14, which may include various sensing and pacing control parameters used by IMD 14. External device 50 may receive data from IMD 14 which may include atrial and / or ventricular signal episodes, sensed cardiac event signal data, and therapy delivery data logged by IMD 14. External device 50 may receive data from IMD 14 pertaining to detected atrial tachyarrhythmia episodes (or more generally compromised atrial mechanical function episodes), which may include any of a ventricular EGM signal segment, cardiac motion signal segment, the time, date and duration of a detected episode of compromised atrial mechanical function, atrial tachyarrhythmia burden (e.g., the percentage of time that the patient is determined to have compromised atrial mechanical function), the ventricular rate and / or RR interval (RRI) variability' during a detected episode, as examples.

[0039] External device 50 may be embodied as a programmer used in a hospital, clinic or physician’s office to program IMD 14 and to acquire data from IMD 14. External device 50 may alternatively be a handheld device, such as a tablet or cell phone. In some examples, external device 50 is a home monitor configured to interrogate IMD 14 to receive signals or data fromIMD 14 and transmit data to IMD 14 via a wireless communication link 48. An example programmer that may be configured to program IMD 14 and be included in medical device system 10 configured to perform the techniques disclosed herein is the CARELINK® Programmer, commercially available from Medtronic, Inc., Dublin, Ireland.

[0040] External device 50 may include a processor 52, memory 53. display unit 54, user interface unit 56, and telemetry unit 58. Processor 52 is coupled to the other components and units of external device 50, e.g., via a data bus, for controlling the functions attributed to external device 50 herein. Processor 52 may execute instructions stored in memory' 53. Processor 52 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, processor 52 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 processor 52 herein may be embodied as software, firmware, hardware or any combination thereof. Memory 53 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 or analog media. Memory’53 may include non-transitory computer-readable media that may store instructions that, when executed by processor 52, cause medical device system 10 to perform various methods and functions attributed to medical device system 10 as disclosed herein.

[0041] User interface unit 56 may include a mouse, touch screen, keypad or the like to enable a user to interact with external device 50, e.g., to initiate and terminate an interrogation session for retrieving data from IMD 14, adjust settings of display unit 54, enter programming commands or selections or make other user requests. Display unit 54, may generate a display of cardiac electrical signals received from IMD 14 and / or data derived therefrom. Display unit 54 may be configured to generate a graphical user interface (GUI) including various windows, icons, user selectable menus, etc. to facilitate interaction by a user with the external device 50. Display unit54 may display various windows to a user, e.g., in a GUI, for enabling a clinician or other user to review cardiac event signal sensing, arrhythmia detections, compromised atrial mechanical function episodes, and therapy delivery related data retrieved from IMD 14.

[0042] Display unit 54 may function as an input and / or output device using technologies including liquid crystal displays (LCD), quantum dot display, dot matrix displays, light emitting diode (LED) displays, organic light-emitting diode (OLED) displays, cathode ray tube displays,e-ink, or monochrome, color, or any other type of display capable of generating tactile, audio, and / or visual output. In some examples, display unit 54 is a presence-sensitive display. Display unit 54 may serve as a user interface device that operates both as one or more input devices and one or more output devices.

[0043] External device 50 may receive data, via telemetry’ unit 58, from IMD 14 via the wireless communication link 48. Data received from IMD 14 may include cardiac signals, e.g., EGM signals sensed by IMD 14, marker channel data indicating the timing of pacing pulses delivered by IMD 14 and the timing of sensed cardiac event signals (e.g., sensed atrial mechanical event signals and sensed ventricular electrical event signals), and data relating to the pacing history.

[0044] Telemetry unit 58 includes a transceiver and antenna configured for bidirectional communication with a communication circuit included in an implantable IMD 14, e g., in response to user requests. Telemetry' unit 58 includes communication circuitry’ configured to operate in conjunction with processor 52 for sending and receiving data relating to pacemaker functions via a wireless communication link 48 with IMD 14. Communication link 48 may be established using a radio frequency (RF) link such as BLUETOOTH®, Wi-Fi, Medical Implant Communication Service (MICS) or other communication bandwidth or protocol. In some examples, external device 50 may include a programming head that is placed proximate IMD 14 to establish and maintain communication link 48, and in other examples external device 50 and IMD 14 may be configured to communicate using a distance telemetry algorithm and circuitry that does not require the use of a programming head and does not require user intervention to maintain a communication link 48.

[0045] It is contemplated that external device 50 may be in wired or wireless connection to a communications network via telemetry unit 58 that includes a transceiver and antenna or via a hardwired communication line for transferring data to a centralized database or computer to allow remote management of the patient. External device telemetry’ unit 58 may be coupled to a communication network / cloud (not show n in FIG. 1) for receiving and transmitting data to a remote computing device (not shown in FIG. 1), which may be a personal computer, personal mobile device or other computing device at a remote location from the patient to enable remote monitoring of data obtained from IMD 14 by a clinician or other user. The CARELINK™ network available from Medtronic, Inc., Dublin, Ireland, is an example of a remote patient monitoring system and database that may collect and display data retrieved from a patient's IMD for review by a clinician or other user.

[0046] FIG. 2 is a diagram of IMD 14 shown in FIG. 1. IMD 14 includes a housing 15 for enclosing internal components and circuitry’, e.g., as described below in conjunction with FIG. 3.Housing 15 may enclose, for example, control circuitry, sensing circuitry, pulse generating circuitry, a motion sensor, other optional physiological sensors, communication circuitry, and a power source, e.g., one or more batteries, to provide power to the circuitry and components of IMD 14 as needed. Housing 15 may have a generally cylindrical, longitudinal sidewall 17 that extends between a proximal end 20 and a distal end 22 of IMD housing 15. Housing 15 may be generally cylindrical in shape to facilitate advancement of IMD 14 through a catheter or other delivery tool to an implant site, e.g., transvenously to an intracardiac implant site. Distal end 22 is referred to as “distal” in that it is expected to be the leading end of IMD 14 as IMD 14 is advanced in a delivery tool to an implant site within or on the heart. Other shapes of housing 15 may be used, e.g., prismatic shapes, disk shapes, etc.

[0047] IMD 14 may include at least two electrodes 16 and 18 spaced apart along the housing 150 for sensing cardiac electrical signals and delivering pacing pulses. Electrode 16 is shown as a tip electrode on the distal end 22 of IMD 14. Electrode 18 is shown as a ring electrode along a mid-portion of housing 15, for example circumscribing longitudinal sidewall 17 and spaced proximally from distal tip electrode 16, e.g., adjacent proximal end 20. In other examples, the proximal electrode 18 may be on proximal end 20 or positioned relatively closer or on distal end 22 to provide a bipolar electrode pair.

[0048] Electrodes 16 and 18 form a cathode and anode pair for bipolar cardiac pacing and sensing. Electrodes 16 and 18 may be. without limitation, titanium, platinum, iridium or alloys thereof and may include a low polarizing coating, such as titanium nitride, iridium oxide, ruthenium oxide, platinum black, among others. Electrodes 16 and 18 may be positioned at locations along IMD 14 other than the locations shown. While distal electrode 16, also referred to herein as “tip electrode” 16. is shown as a button type electrode and proximal electrode 18, also referred to herein as “ring electrode” 18, is shown as a ring electrode, electrodes 16 and 18 may be provided as other types of electrodes, such as a tissue piercing electrode, helical electrode, hook electrode, hemispherical electrode, segmented electrode, short coil electrode, etc.

[0049] In some examples, the distal tip electrode 16 may be configured as a tissue piercing electrode that can be inserted into cardiac tissue to advance electrode 16 to a desired pacing site. For example, distal tip electrode 16 may be a straight, helical or other tissue piercing electrode that can be inserted into the inferior end of the interatrial septum, beneath the AV node and near the tricuspid valve annulus to position tip electrode 16 in, along or proximate to ventricular tissue, e.g.. near the His bundle, for delivering ventricular pacing pulses that capture at least a portion of the conduction system. In other examples, tip electrode 16 may be provided as tissue piercing electrode that can be advanced into the interventricular septum to deliver ventricularpacing in the area of the left bundle branch, right bundle branch, His bundle or Purkinje fibers for delivering ventricular pacing pulses that capture at least a portion of the conduction system and / or septal myocardial tissue.

[0050] Housing 15 is formed from a biocompatible material, such as a stainless steel or titanium alloy. In some examples, the housing 15 may include an insulating coating. Examples of insulating coatings include parylene, urethane, PEEK, or polyimide, among others. The entirety of the housing 15 may be insulated, but only electrodes 16 and 18 are uninsulated. Electrode 1 may serve as the cathode electrode and be coupled to internal circuitry, e.g., a pacing pulse generator and cardiac electrical signal sensing circuitry, enclosed by housing 15 via an electrical feedthrough crossing housing 15. Electrode 18 may be formed as a conductive portion of housing 15 defining a ring electrode that is electrically isolated from the other portions of the housing 15 as generally shown in FIG. 2. In other examples, the entire periphery of the housing 15 may function as an electrode that is electrically isolated from tip electrode 16, instead of providing a localized ring electrode such as anode electrode 18. Electrode 18 formed along an electrically conductive portion of housing 15 can serve as a return anode during pacing and sensing via cathode tip electrode 16.

[0051] IMD 14 may include a fixation member 26, e.g., one or more fixation tines, to secure IMD 14 to cardiac tissue, e.g., by actively engaging with the ventricular endocardium and / or interacting with the ventricular trabeculae. Fixation tines 26 are configured to anchor IMD 14 to position electrode 16 in operative proximity to a targeted tissue for delivering therapeutic electrical stimulation pulses and / or sensing cardiac electrical signals. Numerous types of active and / or passive fixation members, e.g., a helix, hook, curved tine, etc., may be employed for anchoring or stabilizing IMD 14 in an implant position.

[0052] IMD 14 may optionally include a delivery tool interface 24. Delivery tool interface 24 may be located on the proximal end 20 of IMD 14 and is configured to connect to a delivery device, such as a catheter, used to position IMD 14 at an implant location during an implantation procedure, for example within a heart chamber. Delivery tool interface 24 may be used in retaining IMD 14 within a delivery tool as IMD 14 is advanced to an implant site and subsequently deployed and released at the implant site. In some instances, a tether, wire, hook, plunger or other delivery' tool member may be configured to interact with delivery' tool interface 24 and / or proximal end 20 to advance and retract IMD 14 out of, and in some instances back into, a delivery tool lumen or receptacle to facilitate positioning, and repositioning if needed, of IMD 14 at a desired implant site.

[0053] Leadless IMD 14 shown in FIGs. 1 and 2 having housing-based electrodes and a motion sensor enclosed by the IMD housing is one example of a medical device that may be configured to perform the techniques disclosed herein. It is contemplated that the methods disclosed herein are not necessarily limited to being performed by leadless IMDs and may be implemented in an IMD coupled to one or more leads carrying a motion sensor and optionally one or more electrodes. A lead carrying a motion sensor may be advanced into a heart chamber for sensing a cardiac motion signal that is processed and analyzed according to the methods disclosed herein.

[0054] FIG. 3 is a diagram of IMD 14 according to some examples. The circuitry of IMD 14 may include a pulse generator 202, a cardiac electrical signal sensing circuit 204. a control circuit 206, memory 210, telemetry circuit 208, motion sensor 212 and a power source 214. The various circuits represented in FIG. 3 may be combined on one or more integrated circuit boards which include a specific integrated circuit (ASIC), an electronic circuit, a processor (shared, dedicated, or group) and memory that execute one or more software or firmware programs, a combinational logic circuit, state machine and / or other suitable components that provide the described functionality7.

[0055] Motion sensor 212 is implemented as an accelerometer in the examples described herein. Motion sensor 212 is not limited to being an accelerometer, however, and other motion sensors or cardiac mechanical signal sensors may be utilized successfully in IMD 14 for detecting cardiac mechanical signals for use in sensing atrial systolic event signals, controlling atrial synchronized ventricular pacing and for detecting compromised atrial mechanical function, such as during an atrial tachyarrhythmia episode, according to the techniques disclosed herein. Examples of motion sensors that may be implemented in motion sensor 212 include piezoelectric sensors and MEMS devices.

[0056] Motion sensor 212 may be a one dimensional sensor having a single sensor axis or a multi-axis sensor, e.g., a two-dimensional or three-dimensional sensor, with each axis providing a motion signal, e.g.. acceleration signal, that may be analyzed individually or in combination for detecting cardiac mechanical events and for monitoring for compromised atrial mechanical function according to the techniques disclosed herein. Motion sensor 212 produces an electrical signal correlated to motion or vibration of sensor 212 (and IMD 14), e.g., when subjected to acceleration forces due to flowing blood and cardiac motion. The motion sensor 212 may include filters, amplifiers, rectifiers, an ADC and / or other components for producing a motion signal that is passed to control circuit 206. For example, each axis signal produced by each individual axis of a multi-axis accelerometer may be filtered by a high pass filter, e.g., a 10 Hz high pass filter,and rectified for use by atrial event detector circuit 240 for detecting atrial systolic event signals. The high pass filter may be lowered (e.g., to 2.5 Hz or 5 Hz) if needed to detect atrial systolic event signals that have lower frequency content. In some examples, high pass filtering is performed with no low pass filtering. In other examples, each accelerometer axis signal is filtered by a low pass filter, e.g., a 30 Hz low pass filter, with or without high pass filtering. Furthermore, a patient posture signal may be sensed by control circuit 206 from the motion sensor signal as the static or DC acceleration signal.

[0057] When IMD 14 is capable of providing rate response pacing, control circuit 206 may be configured to adjust the pacing rate based on a patient physical activity metric. Control circuit 206 may determine the patient physical activity metric, correlated to the patient’s physical activity level or metabolic demand, from the motion sensor signal. In various examples, an acceleration signal received by control circuit 206 from motion sensor 212 may be filtered by a band pass or low pass filter, e.g., a 1-10 Hz bandpass filter or a 10 Hz low pass filter, digitized by an ADC and rectified for use by processor 244 for determining the patient physical activity metric, also referred to herein as the '‘activity metric.” Various activity metrics may be derived from the acceleration signal by control circuit 206. For instance, an activity metric derived from an acceleration signal received from motion sensor 212 may be obtained by integrating the absolute value of the acceleration signal received over a predetermined time duration (such as 2 seconds as an example). The amplitude of the sampled data points over a two-second time interval may be summed to obtain the activity metric. This activity metric may be referred to as an “activity count” and is correlated to the acceleration due to patient body motion imparted on IMD 14 during the predetermined time interval. The 2-second (or other time interval) activitycounts may be used by control circuit 206 for determining a sensor indicated pacing rate (SIR) for use in controlling rate response pacing. A patient activity metric or SIR may be used by control circuit 206 to set a target rate of ventricular pacing pulses. Pace timing circuit 242 may schedule ventricular pacing pulses by setting temporary LRIs according to the target rate to provide rate response pacing for ventricular rate support during periods of increased physical activity. The temporary LRIs set by pace timing circuit 242 are shortened from the LRI that corresponds to the programmed ventricular lower rate, which can also be referred to as a minimum or base ventricular rate.

[0058] The rate response pacing rate controlled by control circuit 206 based on the activity metric may be determined according to a transfer function relating the activity metric to the SIR. The transfer function may include one or more “set points,” which correspond to an SIR for a given level of the activity metric and a slope or rate of change of the SIR between the set pointsas the activity metric increases or decreases. For example, a resting set point may correspond to a resting level of the activity metric. When the activity metric is equal to or less than the resting level set point, the SIR may be equal to the programmed base ventricular rate. Pace timing circuit 242 may set ventricular pacing intervals to the corresponding LRI.

[0059] An activities of daily living (ADL) set point may correspond to non-resting, but non- strenuous, physical activity of the patient. The ADL may correspond to activities such as moving about the house, doing light chores or personal hygiene etc. Control circuit 206 may determine the SIR between the resting set point and the ADL set point according to a slope of the transfer function and set a target pacing rate based on the SIR. When the activity metric exceeds the ADL set point, control circuit 206 may determine the SIR according to a slope of the transfer function between the ADL set point and a maximum upper pacing rate, e.g., corresponding to strenuous or maximal exertion by the patient.

[0060] A “set point’' can be thought of as a threshold level of activity above which control circuit 206 determines an increasing SIR and associated rate response pacing rate according to the transfer function as the activity metric increases. A set point, e.g., the ADL set point, may be defined at a change in the slope of the transfer function, e.g., between the resting set point and a maximum upper pacing rate, that relates the patient physical activity7metric to the SIR such that the pacing rate may increase faster or slower above or below a given set point. One, none or more than one set point may be defined between the base pacing rate and the maximum upper pacing rate for use in adjusting the SIR as a function of the activity metric. In this way, control circuit 206 may control pace timing circuit 242 to schedule rate response pacing pulses for delivery7by pulse generator 202 based on the patient physical activity metric determined from the motion sensor signal.

[0061] Cardiac electrical signal sensing circuit 204, also referred to herein as “sensing circuit” 204, is configured to receive a cardiac electrical signal via electrodes 16 and 18 by a pre-filter and amplifier circuit 220. Pre-filter and amplifier circuit 220 may include a high pass filter to remove DC offset, e.g., a 2.5 to 5 Hz high pass filter, or a wideband filter having a passband of 2.5 Hz to 100 Hz to remove DC offset and high frequency noise. Pre-filter and amplifier circuit 220 may further include an amplifier to amplify the “raw” cardiac electrical signal passed to analog-to-digital converter (ADC) 226. ADC 226 may pass a multi-bit, digital EGM signal to control circuit 206 for performing morphological analysis of the cardiac electrical signal for detecting heart rhythms, passing an EGM signal to telemetry circuit 208 for transmission to external device 50 (see FIG. 1) or other cardiac signal processing and analysis purposes. The digital signal from ADC 226 may be passed to rectifier and amplifier circuit 222 of sensingcircuit 204. which may include a rectifier, bandpass filter, and amplifier for passing a filtered, amplified and rectified cardiac signal to cardiac event detector circuit 224.

[0062] Cardiac event detector circuit 224 may include a sense amplifier, comparator or other detection circuitry that compares the incoming rectified, cardiac electrical signal to a cardiac event signal sensing threshold, e.g., an R-wave sensing threshold, which may be an autoadjusting threshold. When the incoming signal crosses the R-wave sensing threshold, the cardiac event detector circuit 224 may produce a ventricular sensed event signal (Vsense) that can be passed to control circuit 206 for use in determining RR intervals (RRIs), determining the ventricular rate, and in controlling ventricular pacing pulses. In other examples, cardiac event detector circuit 224 may receive the digital output of ADC 226 for detecting R-waves by a comparator, morphological signal analysis of the digital EGM signal or other R-wave detection techniques. Vsense signals passed from cardiac event detector circuit 224 to control circuit 206 may be used for scheduling ventricular pacing pulses by pace timing circuit 242 and for use in identifying the timing of ventricular electrical events in algorithms performed by atrial event detector circuit 240 for applying atrial event windows and detecting atrial systolic event signals from a motion signal received from motion sensor 212, as further described below.

[0063] Control circuit 206 may include an atrial event detector circuit 240, pace timing circuit 242, and processor 244. Atrial event detector circuit 240 is configured to detect atrial systolic event signals from a signal received from motion sensor 212. In some examples, one or more ventricular mechanical events may be detected from the motion sensor signal in a given cardiac cycle to facilitate positive detection of the atrial systolic event signal from the motion sensor signal during the ventricular cycle.

[0064] Control circuit 206 may receive Vsense signals and / or digital cardiac electrical signals from cardiac electrical signal sensing circuit 204 for use in detecting and confirming cardiac events, determining the heart rhythm and controlling ventricular pacing. For example, Vsense signals may be passed to pace timing circuit 242 for inhibiting a scheduled ventricular pacing pulse and / or scheduling a next ventricular pacing pulse, depending on the pacing mode that is in effect at the time the Vsense signal is received. In response to a Vsense signal received from sensing circuit 204, control circuit 206 may start a ventricular pacing interval set to an LRI, a temporary rate response LRI, or a rate smoothing interval, as examples. Vsense signals may be passed to atrial event detector circuit 240 for use in setting time windows (as described below in conjunction with FIG. 5) used by control circuit 206 for detecting atrial systolic event signals from the motion sensor signal.

[0065] Atrial event detector circuit 240 receives a motion signal from motion sensor 212 and may start an atrial blanking period and / or an atrial refractory period in response to a ventricular electrical event, e.g., a Vsense signal from sensing circuit 204 or delivery' of a ventricular pacing pulse by pulse generator 202. In some examples, atrial event detector circuit 240 determines if the motion sensor signal crosses an atrial event sensing threshold outside the atrial refractory period. Atrial event detector circuit 240 may set time windows corresponding to the passive ventricular filling phase and the active ventricular filling phase of the cardiac cycle relative to the timing of a preceding ventricular electrical event, either a Vsense signal or a ventricular pacing pulse. A motion sensor signal crossing of an atrial event sensing threshold during either of these windows may be detected as the atrial systolic event signal. As described below, two different atrial event sensing thresholds may be applied during the respective passive filling phase window and active filling phase window for sensing an early atrial systolic event signal when the atrial contraction occurs relatively early, during the passive filling phase window, or for sensing a relatively later atrial systolic event signal when the atrial contraction occurs after the passive filling phase window.

[0066] Atrial event detector circuit 240 may pass a sensed atrial mechanical event (AMsense) signal to processor 244 and / or pace timing circuit 242 in response to sensing the atrial systolic event signal from the motion sensor signal. Pace timing circuit 242 (or processor 244) may additionally receive Vsense signals from cardiac event detector circuit 224 for use in controlling the timing of pacing pulses delivered by pulse generator 202. Processor 244 may include one or more clocks for generating clock signals that are used by pace timing circuit 242 to time out an AV pacing interval that is started upon receipt of an AMsense signal from atrial event detector circuit 240. Pace timing circuit 242 may include one or more pacing escape interval timers or counters that can be used to time out the AV pacing interval, a LRI, a rate response interval, a rate smoothing interval, etc. The AV pacing interval may be a programmable interval stored in memory' 210 and retrieved by processor 244 for use in setting the AV pacing interval used by pace timing circuit 242.

[0067] Pace timing circuit 242 may additionally include a ventricular pacing interval timer for scheduling a ventricular pacing pulse in response to a delivered ventricular pacing pulse (Vpace) or in response to a Vsense signal received from sensing circuit 204. The ventricular pacing interval can be set to the LRI that is timed out for providing ventricular rate support and preventing ventricular asystole if the intrinsic rate falls below the programmed base pacing rate. If an AMsense signal is not produced by7atrial event detector circuit 240 for triggering a Vpace at the programmed AV pacing interval, a Vpace may be delivered by pulse generator 202 uponexpiration of the ventricular pacing interval. In order to avoid abrupt changes in ventricular rate and promote atrial event signal sensing recovery when an atrial systolic event signal is not sensed during a cardiac cycle, control circuit 206 may be configured to set the ventricular pacing interval to a rate smoothing interval (RSI) during an atrial tracking ventricular pacing mode. The RSI may be determined based on one or more preceding paced ventricular event intervals. For example, the actual paced ventricular cycle lengths (VCLs) between consecutively delivered Vpaces (Vp-Vp cycle lengths) and / or Vsense signals and subsequent pacing pulses (Vs-Vp cycle lengths) may be determined. An RSI may be set based on the actual paced VCLs so that a Vpace that is delivered in the absence of an AMsense signal is delivered at a pacing interval that is within a predetermined interval of preceding paced VCLs, e.g., within 200 ms or within 100 ms or within 50 ms, to avoid an abrupt change in the ventricular rate. The RSI may be gradually increased toward the ventricular LRI (based on the programmed ventricular base pacing rate) for pacing at the ventricular lower rate or until AMsense signals are produced by atrial event detector circuit to restore atrial synchronous ventricular pacing at the AV pacing interval.

[0068] At times, control circuit 206 may control pulse generator 202 in anon-atrial tracking ventricular pacing mode (also referred to as “asynchronous ventricular pacing”). In this case, pace timing circuit 242 may schedule ventricular pacing pulses at the “permanent” ventricular LRI corresponding to the programmed ventricular lower (base) rate. If the intrinsic rate falls below the ventricular lower rate, ventricular pacing pulses are delivered at the LRI. When pacemaker 14 is programmed to provide rate response pacing, the ventricular pacing interval may be set to a temporary LRI that is shorter than the “permanent” ventricular LRI. The temporary LRI can be determined by control circuit 206 based on the patient physical activity metric determined from the motion sensor signal received from motion sensor 212. as generally described above. If control circuit 206 switches from an atrial-tracking ventricular pacing mode to a non-atrial tracking ventricular pacing mode, control circuit 206 may set RSIs to control a gradual adjustment of the ventricular pacing rate from the atrial tracking paced VCLs to the ventricular lower rate used to control the ventricular pacing rate during the asynchronous ventricular pacing mode, e.g., a VVI(R) or VDI(R) pacing mode.

[0069] Processor 244 may retrieve other programmable pacing control parameters from memory 210, such as pacing pulse amplitude and pacing pulse width, which are passed to pulse generator 202 for controlling pacing pulse delivery-. In addition to providing control signals to pace timing circuit 242 and pulse generator 202 for controlling pacing pulse delivery, processor 244 may provide sensing control signals to sensing circuit 204, e.g., R-wave sensing threshold controlparameters such as the ventricular sensitivity and / or various blanking and refractory intervals applied to the cardiac electrical signal.

[0070] Pulse generator 202 generates electrical pacing pulses that are delivered to pace the ventricles of the patient’s heart via cathode electrode 16 and return anode electrode 18. Pulse generator 202 may include charging circuit 230. switching circuit 232 and an output circuit 234. Charging circuit 230 may include a holding capacitor that may be charged to a pacing pulse amplitude by a multiple of the battery voltage signal of power source 214, e.g., under the control of a voltage regulator. The pacing pulse amplitude may be set based on a control signal from control circuit 206. Switching circuit 232 may control when the holding capacitor of charging circuit 230 is coupled to the output circuit 234 for delivering the pacing pulse. For example, switching circuit 232 may include a switch that is activated by a timing signal received from pace timing circuit 242 upon expiration of an AV pacing interval or a ventricular pacing interval and kept closed for a programmed pacing pulse width to enable discharging of the holding capacitor of charging circuit 230. The holding capacitor, previously charged to the pacing pulse voltage amplitude, is discharged across electrodes 16 and 18 through the output capacitor of output circuit 234 for the programmed pacing pulse duration.

[0071] Memoiy 210 may include computer-readable instructions that, when executed by control circuit 206. cause control circuit 206 to perform various functions attributed throughout this disclosure to IMD 14, including compromised atrial mechanical function detection. The computer-readable instructions may be encoded within memoiy' 210. Memory 210 may include any non-transitory, computer-readable storage media including 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 other digital media. Memory 210 may store timing intervals and other data used by control circuit 206 to control the delivery' of pacing pulses by pulse generator 202, e.g., by detecting an event signal by atrial event detector circuit 240 from the motion sensor signal and setting a pacing escape interval timer included in pace timing circuit 242.

[0072] Memory 210 may buffer a series of motion sensor signal amplitudes for use in detecting episodes of compromised atrial mechanical function as further described below. Control circuit 206 may be configured to detect compromised atrial mechanical function, which may be detected as an episode of AT or AF, based on motion sensor signal amplitudes acquired over time intervals that can be subsequently classified by control circuit 206 as being an atrial tachyarrhythmia interval, not an atrial tachyarrhythmia interval, or unclassified (or unknown), for example. Control circuit 206 may be configured to detect episodes of compromised atrialmechanical function and log episode data in memory 210 for transmission to external device 50 by telemetry circuit 208. Detection of compromised atrial mechanical function episodes, which may be evidence of atrial tachyarrhythmia episodes, and related data is important clinical information for a physician in managing the patient’s heart rhythm for preventing more serious arrhythmias and / or stroke.

[0073] Power source 214 provides power to each of the other circuits and components of IMD 14 as required. Power source 214 may include one or more energy storage devices, such as one or more rechargeable or non-rechargeable batteries. The connections between power source 214 and other IMD circuits and components are not shown in FIG. 3 for the sake of clarity but are to be understood from the general block diagram of FIG. 3. For example power source 214 may provide power to charging circuit 230 for charging a holding capacitor to a pacing voltage amplitude, current to switch 232 and other circuitry7included in pulse generator 202 as needed, power to transceiver 209, motion sensor 212, and ADC 226 and other circuitry of sensing circuit 204 as needed as well as memory 210.

[0074] Telemetry circuit 208 includes a transceiver 209 and antenna 211 for transferring and receiving data, e.g., via a radio frequency (RF) communication link. Telemetry circuit 208 may be capable of bi-directional communication with external device 50 (FIG. 1) as described above. Motion sensor signals, cardiac electrical signals (e.g., EGM signals) and / or data derived therefrom may be transmitted by telemetry circuit 208 to external device 50. Programmable control parameters and algorithms for performing atrial event signal sensing, ventricular event signal sensing, ventricular pacing control, and detecting compromised atrial mechanical function may be received by telemetry circuit 208 and stored in memory 210 for access by control circuit 206.

[0075] The functions attributed to IMD 14 herein may be embodied as one or more processors, controllers, hardware, firmware, softw are, or any combination thereof. Depiction of different features as specific circuitry is intended to highlight different functional aspects and does not necessarily imply that such functions must be realized by separate hardware, firmware or softw are components or by any particular circuit architecture. Rather, functionality associated with one or more circuits described herein may be performed by separate hardware, fir ware or software components, or integrated within common hardw are, firmware or software components. For example, compromised atrial mechanical function detection algorithms may be implemented in control circuit 206 executing instructions stored in memory 210 and relying on input from sensing circuit 204, timing of Vpaces delivered by pulse generator 202, and the motion signal from motion sensor 212.

[0076] The operation of circuitry included in IMD 14 as disclosed herein should not be construed as reflective of a specific form of hardware, firmware and software necessary to practice the techniques described. It is believed that the particular form of software, hardware and / or firmware will be determined primarily by the particular system architecture employed in the IMD 14 and by the particular sensing and therapy delivery’ circuitry employed by the IMD 14. Providing software, hardware, and / or firmware to accomplish the described functionality in the context of any modem medical device, given the disclosure herein, is within the abilities of one of skill in the art.

[0077] FIG. 4 is a diagram 250 of a cardiac electrical signal 251 and a motion sensor signal 253 that may be sensed by IMD 14. Vertical dashed lines 252 and 262 denote the timing of two consecutive ventricular events, Vpaces 280 and 282, marking the respective beginning and end of one ventricular cycle 251. Ventricular cycle 251 may be referred to as one cardiac cycle including one ventricular systolic phase and one ventricular diastolic phase. In other instances, one cardiac cycle could begin and / or end with an intrinsic ventricular depolarization sensed as an R-wave by sensing circuit 204 resulting in a Vsense signal passed to control circuit 206.

[0078] The motion sensor signal 253 is an unrectified signal that may be sensed by motion sensor 212 (shown in FIG. 3). Motion sensor 212 may pass the rectified motion sensor signal 255 to control circuit 206. Motion signal 253 and rectified motion signal 255 include an Al event signal 254. an A2 event signal 256, an A3 event signal 258 and an A4 event signal 260. The Al event signal 254 is an acceleration signal (in this example when motion sensor 212 is implemented as an accelerometer) that occurs during ventricular contraction and marks the approximate onset of ventricular mechanical systole. The Al event signal 254 occurs shortly after the Vpace 280 marking the beginning of ventricular cycle 251. The A2 event signal 256 is an acceleration signal that may occur with closure of the aortic and pulmonic valves, marking the approximate offset or end of ventricular mechanical systole.

[0079] The A3 event signal 258 is an acceleration signal that occurs during passive ventricular filling during ventricular mechanical diastole. The A4 event signal 260 is an acceleration signal that occurs during atrial contraction and active ventricular filling and marks atrial mechanical systole (normally during ventricular diastole). The A4 event signal 260 is also referred to herein as the “atrial systolic event signal” or merely the “atrial event signal.” The A4 event signal 260 is the atrial event signal that can be detected or sensed from motion sensor signal 253 (or rectified motion sensor signal 255) by atrial event detector circuit 240. resulting in an AMsense signal 284. In response to the AMsense signal 284, pace timing circuit 242 may trigger delivery of an atrial synchronous Vpace 282 by starting the AV pacing interval 283.

[0080] In response to a ventricular electrical event, e.g., Vpace 280, control circuit 206 may apply a post-ventricular atrial blanking (PVAB) period 264 to the rectified motion sensor signal 255, during which sensing cardiac event signals from the motion sensor signal 255 may be disabled or inhibited. In other examples, control circuit 206 may optionally be configured to detect one or more of the Al, A2, and A3 event signals 254. 256 and / or 258 from motion sensor signal 250, for at least some ventricular cardiac cycles, for use in positively detecting the A4 event signal 260 and / or for use in setting atrial event sensing control parameters.

[0081] Following the PVAB period 264, control circuit 206 may apply an A4 sensing threshold 274 to the rectified motion sensor signal 255 for sensing the atrial event signals (e.g., A4 event signal 260). As described below in conjunction with FIG. 5. the A4 sensing threshold 274 may be applied as a multi-level sensing threshold including an early high sensing threshold amplitude 276 that is applied during a time window 266. The time window 266 is also referred to herein as a “passive ventricular filling window’' or “A3 window’' because it is expected to encompass the early ventricular diastolic phase of passive ventricular filling and the A3 event signal 258.Control circuit 206 may schedule the time window 266, during which the high sensing threshold amplitude 276 is applied, to begin upon expiration of the PVAB period 264. Control circuit 206 may schedule the time window 266 to expire at an ending time VE 281, which may be automatically adjustable by control circuit 206.

[0082] If the high sensing threshold amplitude 276 is not crossed by the rectified motion sensor signal 255 prior to the expiration of the time window 266, control circuit 206 may decrease the A4 sensing threshold 274 to a low sensing threshold amplitude 278. The low sensing threshold amplitude 278 may be applied during an A4 window 270 that begins upon expiration 281 of the time window 266 and extends until the A4 signal 260 is sensed (e.g., when the rectified motion sensor signal 255 crosses the low sensing threshold amplitude 278), a Vpace is delivered by pulse generator 202 or a Vsense signal is received from sensing circuit 204, whichever occurs earliest.

[0083] FIG. 5 is an example of motion sensor signals 300 and 310 acquired over two different cardiac cycles. A Vpace is delivered at time 0.0 seconds for both cardiac cycles. The top motion sensor signal 300 is received over one cardiac cycle, and the bottom motion sensor signal 310 is received over a different cardiac cycle. The two signals 300 and 310 are aligned in time at 0.0 seconds, the time of the Vpace delivery. While motion signals 300 and 310 are shown as raw accelerometer signals, it is recognized that control circuit 206 may receive a digitized, filtered, amplified and rectified signal from motion sensor 212 for processing and analysis.

[0084] The Al event signals 302 and 312 of the respective motion sensor signals 300 and 310. which occur during ventricular contraction, are observed to be well -aligned in time following the ventricular pacing pulse at time 0.0 seconds. Similarly, the A2 events 304 and 314 (which may mark the end of ventricular systole) and the A3 events 306 and 316 (occurring during passive ventricular filling of the ventricular diastolic phase) are well-aligned in time. Since the Al. A2 and A3 event signals occur during ventricular contraction, at the end of ventricular systole and during passive ventricular filling, respectively, these cardiac mechanical event signals are expected to occur at relatively consistent intervals following a ventricular electrical event (a Vpace in this example or a Vsense in other instances) and relative to each other. The time relationship of the Al, A2 and A3 event signals may be different following a Vpace compared to following a Vsense; however, during a stable paced or intrinsic ventricular rhythm the relative timing of ventricular Al, A2 and A3 event signals to each other and the immediately preceding ventricular electrical event is expected to be consistent from beat-to-beat.

[0085] The A4 event signals 308 and 318 of the first and second motion sensor signals 300 and 310, respectively, are not aligned in time. The A4 event signal occurs during atrial systole and, as such, the time interval of the A4 event signal following the immediately preceding ventricular electrical event (sensed R-wave or ventricular pacing pulse) and the preceding Al through A3 event signals may vary between cardiac cycles, e.g., as the atrial rate changes and / or when asynchronous ventricular pacing is being delivered.

[0086] The consistency of the timing of the Al through A3 event signals relative to each other and the immediately preceding ventricular electrical event may be relied upon in applying the PVAB period 336 to the motion sensor signals 300 and 310 to avoid falsely sensing a ventricular mechanical event signals as the atrial systolic event signal, thereby increasing confidence in reliably sensing A4 event signals 308 and 318. The atrial systolic event signal is not detected by atrial event detector circuit 240 during the PVAB period 336, which may extend from the ventricular electrical event (at time 0.0) through an estimated onset of ventricular diastole so that the PVAB period 336 includes both the Al event signal 302 / 312 and A2 event signal 304 / 314 in some examples. An A3 window 324 may be applied to the motion sensor signal by atrial event detector circuit 240. The A3 window7324 may have a starting time 320 corresponding to the end of the PVAB period 336. The ending time 322 of A3 window7324 may also be considered a starting time of the A4 window 350, though A4 event signals may be detected by the atrial event detector circuit 240 during the A3 window 324 in some instances.

[0087] As seen by the lower motion sensor signal 310, the A4 event signal 318 may occur earlier after the A3 window7324. In some instances, as the atrial rate increases, the A4 event signal 318may occur within the A3 window 324. When this occurs, the A3 event 316 and the A4 event 318 may fuse as one relatively large acceleration signal, as passive and active ventricular filling occur concomitantly. The fused A3 / A4 event signal may have a relatively high amplitude, greater than the amplitude of either the A3 event signal 316 or the A4 event signal 318 when they occur as separate event signals spaced apart in time. As such, in some examples the first, higher A4 sensing threshold amplitude 346 may be applied to the motion sensor signal by control circuit 206 for detecting an early A4 event signal that is fused with the A3 event signal during the A3 window 324. The second, lower A4 sensing threshold amplitude 348 may be applied by control circuit 206 for detecting relatively later A4 event signals, after the ending time 322 of the A3 window 324. during an A4 window 350. The A4 window 350 extends from the ending time 322 of the A3 window 324 until the next ventricular electrical event, sensed or paced. The earliest crossing of the A4 sensing threshold 344 by the motion sensor signal after the starting time 320 of the A3 window (or after the expiration of the PVAB period 336) may be sensed as the atrial systolic event signal.

[0088] Atrial event detector circuit 240 may produce an AMsense signal (e.g., as shown in FIG. 4) in response to the earliest crossing time of the high A4 sensing threshold amplitude 346 or the low' A4 sensing threshold amplitude 348, whichever comes first. When the AMsense signal is produced before the expiration of the A3 window 324, pace timing circuit 242 schedules a Vpace at the AV pacing interval. Delivery of the Vpace marks the start of the next cardiac cycle such that an A4 window' 350 may not begin during a given cardiac cycle that ends with a Vpace triggered by an AMsense signal occurring during the A3 w indow 324.

[0089] In some examples, control circuit 206 may set a post-ventricular atrial refractory period (PVARP) 338. The PVARP 338 may extend from the ventricular electrical event (sensed R-wave or ventricular pacing pulse) for a time interval longer than the PVAB period 336. Depending on the ending time 322 of the A3 window' 324, the PVARP 338 may expire during the A3 window' 324. When the motion sensor signal crosses the high A4 sensing threshold amplitude 346 during the PVARP 338, but outside the PVAB period 336, a refractory AMsense signal may be produced by atrial event detector circuit 240. Pace timing circuit 242 may not start an AV pacing interval in response to a refractory AMsense signal, but control circuit 206 may use the refractory AMsense signal in adjusting A4 sensing control parameters, determining the intrinsic atrial rate or other purposes. When the motion sensor signal crosses the A4 sensing threshold 344 after the expiration of PVARP 338, atrial event detector circuit 240 senses the atrial systolic event signal, and pace timing circuit 242 starts the AV pacing interval (not shown in FIG. 5). Upon expirationof the AV pacing interval, pulse generator 202 generates a Vpace delivered to the ventricle to track the non-refractory sensed atrial event signal to promote AV synchrony.

[0090] FIG. 6 is a flow chart 400 of a method for detecting compromised atrial mechanical function that may be performed by a medical device capable of sensing a cardiac mechanical signal according to some examples. FIG. 6 and other flow charts and diagrams presented herein are described as being performed by IMD 14 of FIGs. 1 — 3 as an illustrative example. It is to be understood that the methods described herein can be implemented in other medical devices, including leadless IMDs or medical devices coupled to one or more leads carry ing at least cardiac mechanical signal sensor. For the sake of convenience, detection of compromised atrial mechanical function is described as detection of AF in the illustrative examples. It is to be understood that the methods described herein for detecting compromised atrial mechanical function are not necessarily limited to detecting the compromised atrial mechanical function as AF. Poor atrial mechanical function, resulting in a weak or absent A4 signal during the A4 windows, may occur during AF and other types of atrial tachyarrhythmia (e.g., non-sinus atrial tachycardia or atrial flutter), during other cardiac electrical function abnormalities (e.g., SA node dysfunction) or due to other atrial structural or mechanical function abnormalities (e.g., due to atrial fibrosis, atrial distension, etc.). The poor or compromised atrial mechanical function may be detected using the methods described herein.

[0091] At block 402, control circuit 206 of IMD 14 may start an atrial function analysis interval, also referred to herein as an “analysis interval" or an “AF (atrial fibrillation) analysis interval” for convenience even though compromised atrial mechanical function may not always be due to AF. During the AF analysis interval IMD 14 may acquire motion sensor signal amplitudes for use in detecting AF. The AF analysis interval may be a specified time interval, e.g.. 10 seconds to 15 minutes in duration, or 1 to 10 minutes in duration as examples. In an illustrative example, control circuit 206 starts a two-minute time interval at block 402.

[0092] In other examples, the AF analysis interval started at block 402 may be a time interval defined by a specified number of cardiac cycles that are counted by counting a series of ventricular electrical events, e.g., Vsense signals and / or delivered Vpaces. For example, control circuit 206 may start counting consecutive ventricular electrical events until a specified number is reached, e.g., a count of 10 to 1,000 cardiac cycles or between 100 and 200 cardiac cycles. In still other examples, the AF analysis interval started at block 402 may be a time interval that is defined by a minimum number of delivered ventricular pacing pulses or a minimum number of cardiac cycles that include an A4 window (e.g., cardiac cycles that are longer than the A3 window ending time described above in conjunction with FIGs. 4 and 5).

[0093] At block 404, control circuit 206 may determine the maximum (absolute) amplitude of the motion sensor signal sensed during the A4 window of cardiac cycles that are longer than the A3 window ending time during the AF analysis interval. The A4 window maximum amplitudes may be buffered in memory 210. It is recognized that in some instances, as described above, the A4 window may not be started in a cardiac cycle if the atrial event signal is sensed in the A3 window or a Vsense or Vpace occurs during the A3 window. If an A4 window does not occur in a given cardiac cycle, control circuit 206 does not determine an A4 window maximum amplitude for that cardiac cycle. As such, an A4 window maximum amplitude may not be stored for all cardiac cycles of the AF analysis interval started at block 402. In other examples, if an A4 window maximum amplitude cannot be determined, control circuit 206 may buffer a value of “0” in memory 210 for the given cardiac cycle.

[0094] Control circuit 206 may determine if the AF analysis interval is expired at block 406. If a specified time interval (or specified number of cardiac cycles, specified number of delivered Vpaces, or specified number of cardiac cycles having an A4 window) has elapsed, control circuit 206 may determine that the AF analysis interval is expired at block 406. In some examples, control circuit 206 may determine that the interv al is expired at block 406 when a specified number of A4 window maximum amplitudes have been stored in memory' 210.

[0095] If the AF analysis interval is not expired, control circuit 206 continues to determine and store A4 window maximum amplitudes (for cardiac cycles having an A4 window) at block 404 until the interval is determined to be expired at block 406. As further described below, it is possible that an AF analysis interval expires without any A4 window maximum amplitudes being stored if, for example, the A4 event signal is consistently sensed in the A3 window and / or Vsense or Vpace events occur during the A3 window.

[0096] At block 408, control circuit 206 may classify the AF analysis interval based on the A4 window maximum amplitudes acquired during the interval. In some examples, the AF analysis interv al may be classified as one of an AF interval (or more generally a compromised atrial function interval), a non-AF interval (or more generally a non-compromised atrial function interv al) or an unclassified interval. Methods for acquiring A4 window maximum amplitudes and labeling cardiac cycles that occur during the interval for use in classifying the interval at block 408 are further described below in conjunction with FIGs. 7 and 8.

[0097] If the AF analysis interval is classified as an AF interval (“yes” branch of block 410). control circuit 206 may determine if an AF detection threshold is met at block 412. The AF detection threshold may require that a threshold number of intervals be classified as AF in order to detect AF at block 414. In an illustrative example, if three consecutive 2-minute intervals areclassified as AF intervals (or compromised atrial function intervals), control circuit 206 may determine that the AF detection threshold is met at block 412 and detect AF (or more generally detect compromised atrial mechanical function) at block 414. In other examples, with no limitation intended, one to ten intervals may be required to be classified as AF and may be required to be consecutive intervals.

[0098] In still other examples, the threshold number of intervals required to be classified as AF intervals for detecting AF may not be required to be consecutive intervals. In some examples, if a specified portion or percentage of most recent AF analysis intervals (e.g., at least X out of Y) are classified as AF intervals, control circuit 206 may determine that the AF detection threshold is met at block 412. When the threshold number of X AF intervals out of a total number of Y consecutive intervals is reached for detecting AF, control circuit 206 may detect AF at block 414 if all of the Y intervals that are not classified as AF intervals are unclassified intervals. In still other examples, when the threshold number of X AF intervals out of a total number of Y consecutive intervals is reached for detecting AF, control circuit 206 may detect AF at block 414 when the intervals (out of the Y intervals) that are not classified as AF intervals include non-AF intervals and / or unclassified intervals.

[0099] After detecting AF at block 414, control circuit 206 may perform a response to detecting the compromised atrial mechanical function at block 415. The response, referred to here as an “AF detection response,’7may include generating a detection output. The AF response may include flagging the AF detection in memory 210 as an AF episode with a date and time stamp and optionally other AF episode related data. For example, the AF episode data may be subsequently updated with an AF episode duration if and when AF episode termination is detected as further described below. Memory 210 may store a history of AF detections including the time of onset, duration, updated AF burden (e.g., percentage of time in AF over a 24 hour time period, since implant, or other time period), ventricular rate during the detected AF episode, percentage of paced ventricular cycles during the AF episode, percentage of sensed ventricular cycles during the AF episode, and / or ventricular rate variability during the AF episode, as examples. As such, when AF is detected, control circuit 206 may analyze data stored for the AF analysis interval(s) classified as AF for determining the ventricular rate, percentages of paced and sensed ventricular cycles, and / or ventricular rate variability7. Additionally or alternatively, control circuit 206 may analyze subsequent AF analysis interval(s) started after the time of AF detection for determining ventricular rate and rhythm data, e.g.. based on RRIs, during the AF episode, until AF termination is detected as further described below. The AF response performedat block 415 may include transmitting data to external device 50 for presenting AF episode data to a clinician or other user, e.g., in a graphical user interface on display unit 54.

[0100] Additionally or alternatively, the AF detection response performed at block 415 may include adjusting a therapy delivery control parameter used to control the generation and delivery of electrical pulses by pulse generator 202. Control circuit 206 may switch the pacing mode in response to detecting AF. For instance, the pacing mode may be switched from an atrial tracking (atrial synchronous) ventricular pacing mode to a non-atrial tracking (atrial asynchronous) ventricular pacing mode. Control circuit 206 may adjust ventricular pacing intervals to regulate the ventricular rate during AF if the AF is being conducted at an irregular rate that causes variable RRIs. Control circuit 206 may increase the ventricular pacing rate to control the ventricular rate during AF. In still other examples, control circuit 206 may maintain the pacing mode and ventricular pacing control parameters without adjustment in response to AF detection.

[0101] When IMD 14 is positioned in the right atrium for delivering ventricular pacing from a right atrial approach, IMD 14 may be capable of delivering atrial electrical stimulation therapies, e.g., when IMD 14 is provided with a second, distal ring electrode or other second cathode electrode positioned in operative proximity to atrial tissue for pacing the atria, e.g., in combination with the proximal ring electrode 18. In this case. IMD 14 may deliver electrical stimulation therapy in response to detecting the AF that can include anti-tachycardia or overdrive pacing of the atria in an attempt to terminate the AF. In still other examples, IMD 14 may be coimplanted with an implantable cardioverter defibrillator (ICD) capable of delivering a cardioversion / defibrillation (CV / DF) shock to the patient's heart for terminating the AF. IMD 14 may perform an AF response at block 415 that includes transmitting an AF detection notification. Another co-implanted device may operate to deliver a cardiac electrical stimulation, e.g., atrial anti-tachycardia pacing or CV / DF shock, in an attempt to terminate the AF in response to receiving a transmitted AF detection notification. In other examples, another co-implanted device may withhold delivery of ventricular anti-tachyarrhythmia therapy and / or a CV / DF shock delivery based on the AF detection notification.

[0102] In still other examples, the response performed at block 415 may include storing current values of atrial sensing control parameters, such as the current value of the A3 window ending time, the high A4 sensing threshold (applied in the A3 window), and the low A4 sensing threshold (applied in the A4 window). A4 event signal sensing may be suspended in response to detecting AF. Any updating of the A4 sensing control parameters that may normally be performed as A4 event signals are being sensed during non-compromised atrial mechanicalfunction may be suspended during the detected AF episode. If control circuit 206 detects termination of the AF episode, control circuit 206 may restart A4 event signal sensing using the stored atrial sensing control parameters.

[0103] In yet another example, control circuit 206 may increase the high A4 sensing threshold applied during the A3 window in response to detecting AF. During AF, an increase in the A3 event signal may occur that could lead to false A4 event signal sensing due to the true A3 event signal crossing the A4 sensing threshold. Control circuit 206 may increase the high A4 sensing threshold at block 415 as an AF detection response. Control circuit 206 may decrease the high A4 sensing threshold at block 422 if AF termination is detected, as described below.

[0104] After an AF detection, control circuit 206 may start the next AF analysis interval by returning to block 402. If the next interval is also classified as an AF interval (block 410) after an AF detection (and before an AF termination detection), the AF detection threshold continues to be met (block 412). Control circuit 206 continues to detect the AF episode (block 414) and maytrack the number of intervals classified as AF for use in determining the AF episode duration as part of the AF response at block 415. Control circuit 206 may continue to determine other AF episode data as described above, such as the ventricular rate, percentage of paced and sensed ventricular cycles, and / or RRI variability-, as part of the AF response at block 415.

[0105] When an interval is not classified as AF (“no” branch of block 410), and an AF episode is not in progress (AF detection not previously made at block 414), control circuit 206 returns to block 402 (from ’‘no” branch of block 416) to start the next AF analysis interval. However, if an interv al is not classified as AF (“no’' branch of block 410) and an AF episode is in progress (AF detection previously made at block 414), control circuit 206 may determine if an AF termination threshold is met at block 418 (“yes” branch of block 416).

[0106] Control circuit 206 may determine at block 418 if a threshold number of AF analysis intervals are not classified as AF (e.g., classified as a non-AF interval or unclassified). In some examples, control circuit 206 may determine that the AF termination threshold is met at block 418 in response to one interval that is classified as a non-AF interval. In other examples, a threshold number of two or more intervals may be required to be classified as non-AF intervals in order to detect AF termination. The threshold number of intervals for detecting AF episode termination may or may not be required to be consecutive. For example, if at least M of P intervals are not classified as AF, control circuit 206 may determine that the AF termination threshold is met at block 418. The at least M intervals out of P consecutive intervals may be required to all be classified as non-AF intervals in some examples. In other examples, one or more of the M intervals may be unclassified intervals, and the AF termination threshold may bemet at block 418. In an illustrative example, an AF termination threshold is met when three consecutive two-minute intervals are classified as non-AF intervals. In other examples, one or more unclassified intervals may occur between the threshold number of non-AF intervals, and the AF episode termination threshold may be met at block 418. In still other examples, control circuit 206 may detect AF episode termination at block 420 when a threshold number of AF analysis intervals are not classified as AF, e.g., classified as unclassified and / or non-AF. For instance, control circuit 206 may detect AF episode termination in response to a single AF analysis interval that is classified as unclassified or non-AF.

[0107] When the AF termination threshold is met at block 418, control circuit 206 may detect AF episode termination at block 420. Control circuit 206 may perform an AF termination response at block 422. The AF termination response performed at block 422 may include completing any pending AF episode data determinations initiated in the AF detection response (at block 415), such as determining the AF episode duration, AF burden, ventricular rate during the AF episode, percentage of paced and sensed ventricular cycles during the AF episode, and / or RRI variability during the AF episode duration.

[0108] In some examples, the AF termination response may include adjusting a therapy delivery control parameter. For example, if the pacing mode is switched at block 415 in the AF detection response, the pacing mode may be switched back to a pre-AF detection pacing mode. If a ventricular pacing rate or ventricular pacing intervals were adjusted to promote a regular ventricular rate during AF, the ventricular pacing interval may be adjusted back toward a ventricular LRI, e.g., using one or more rate smoothing intervals. When IMD 14 is configured to transmit an AF detection notification, e.g.. to alert a co-implanted device of the detected AF, control circuit 206 may control telemetry circuit 208 to transmit an AF episode termination detection notification at block 422.

[0109] In some examples, the AF termination response may include redetermining atrial sensing control parameters, such as the low atrial event sensing threshold applied during the A4 window, the high atrial event sensing threshold applied during the A3 window and / or the A3 window ending time. Redetermining atrial sensing control parameters may be based on motion sensor signal amplitudes and / or the amplitude and / or timing of atrial systolic event signals sensed by atrial event detector circuit 240 prior to detecting the AF episode and / or post-AF termination, for instance. As described above, control circuit 206 may store the atrial sensing control parameters in effect at the time of AF detection and reset the atrial sensing control parameters to the stored values upon detecting termination of the AF episode. In other examples, if the high A4 sensing threshold applied during the A3 window was increased at block 415 in response to detecting AF,control circuit 206 may decrease the high A4 sensing threshold at block 422 in response to detecting AF termination. After detecting AF termination, control circuit 206 may return to block 402 to start the next AF analysis interval for continuing to monitor for another AF episode.

[0110] FIG. 7 is a flow chart 500 of a method for labeling individual cardiac cycles that occur during an AF analysis interval. As described above in conjunction with FIG. 6, one or more AF analysis intervals may be classified for detecting compromised atrial mechanical function. The AF analysis interval classification may be determined by control circuit 206, at least in part, by labeling cardiac cycles that occur during the AF analysis interval, e.g., as being possible AF cycles or notAF cycles.

[0111] At block 502, the AF analysis interval is started for acquiring A4 window maximum amplitudes, which may be accumulated in a buffer of memory 210. For the sake of illustration, control circuit 206 may start a 2-minute time interval at block 502. As described above, control circuit 206 may start the AF analysis interval to run for a specified time interval, a specified number of cardiac cycles (e.g.. based on a count ventricular electrical events), a specified number of Vpaces, or until a specified number of A4 window maximum amplitudes are acquired, in various examples. Generally the interval includes multiple cardiac cycles in order to classify the rhythm during the interval as likely being AF (or more generally compromised atrial mechanical function), non-AF (or more generally non-compromised atrial mechanical function) or, in some cases, unclassified due to data that is indeterminate for confidently classifying the interval as AF or non-AF.

[0112] At block 504, control circuit 206 identifies a ventricular event, which may be an electrical event, e.g., a Vsense signal received from sensing circuit 204 or a Vpace delivered by pulse generator 202. Methods disclosed herein are not necessarily limited to identifying the ventricular event by identifying an electrical event, however. In other examples, the ventricular event identified at block 504 could be a ventricular mechanical event, e.g., by detecting the Al signal from the acceleration signal received from motion sensor 212.

[0113] At block 506, control circuit 206 starts the A3 window having a specified starting time and ending time following the ventricular event. Control circuit 206 applies the high A4 sensing threshold amplitude during the A3 window for sensing the A4 signal that may be fused with the A3 signal of the acceleration signal received from motion sensor 212, e.g., as described above in conjunction with FIG. 5. If the motion sensor signal crosses the high A4 sensing threshold amplitude during the A3 window (as determined at block 508), control circuit 206 senses the atrial event signal (“yes” branch). Control circuit 206 may produce the AMsense signal to trigger a Vpace at the AV interval. Control circuit 206 may label the cardiac cycle having an AMsenseduring the A3 window as not AF at block 510. The cardiac cycle may be labeled as anot AF cycle in some examples because the relatively high amplitude acceleration signal sensed as the atrial event signal (the A3 event fused with the A4 event) is evidence of a relatively strong or normal atrial contraction (“atrial kick’') that is more likely to be a sinus atrial systolic event than AF and is not evidence of compromised atrial mechanical function. Control circuit 206 may buffer the not AF label for the cardiac cycle in memory 210. When the atrial event signal is sensed in the A3 window, an A4 window is not started for the cardiac cycle. As such, an A4 window maximum amplitude is not available for buffering in memory7210 for the not AF cycle. If the AF analysis interval is not expired (“no"’ branch of block 526), control circuit 206 returns to block 504 to wait for the next ventricular event, which may be a Vpace delivered at an AV pacing interval from the atrial event signal sensed at block 508.

[0114] If an atrial event signal is not sensed during the A3 window (“no” branch of block 508), control circuit 206 may determine if a Vsense signal is received (block 512) before the A3 window expires or if a Vpace is delivered (block 516) before the A3 window expires (as determined at block 520). In some instances, a Vsense signal may be received by control circuit 206 from sensing circuit 204 during the A3 window, e.g., due to a premature ventricular contraction or other non-sinus ventricular beat (or oversensing due to noise). If a Vsense signal is received during the A3 window (block 512), control circuit 206 may label the current cardiac cycle as an indeterminate cycle (block 514) because the early Vsense signal terminating the current cardiac cycle may not provide clear evidence of the cycle being either an AF cycle or a not AF cycle. When a Vsense signal is received during the A3 window, an A4 window is not started so that an A4 window- maximum amplitude is not determined for storing in memory- 210 for the associated cardiac cycle. A buffer in memory 210 may store the indeterminate label for the cunent cardiac cycle without storing an A4 window maximum amplitude. In some examples, a zero value may be stored for the A4 window- maximum amplitude to indicate that no A4 window- was started for the current cardiac cycle.

[0115] In other examples, the Vsense signal received during the A3 window without an atrial event signal being sensed could be a conducted atrial fibrillation wave (that is undersensed) and could be evidence of possible AF. As such, in some examples, a Vsense signal received during the A3 window (without an AMsense signal prior to the Vsense signal) may be labeled a possible AF cycle in other examples. The possible AF label may be buffered in memory 210 but without an A4 window maximum amplitude (or with a zero value stored) because the A4 window is not started when the early Vsense signal ends the cardiac cycle before the A3 window ending time.

[0116] In other instances, a Vpace may be delivered during the A3 window if a ventricular pacing interval expires during the A3 window. A Vpace may be delivered by pulse generator 202 during the A3 window at a rate smoothing interval or a rate response pacing interval, for instance. When a Vpace occurs in the A3 window (“yes” branch of block 516), without an AMsense signal preceding the Vpace (“no” branch of block 508). control circuit 206 may label the cardiac cycle as a not AF cycle at block 518. An A4 window maximum amplitude is not available for buffering in memory 210 (or a zero value may be stored), and the not AF cycle label can be buffered in memory' 2f 0 at block 518 for the cardiac cycle ending with the Vpace. In other examples, a cardiac cycle ending on a Vpace delivered during the A3 window without a preceding AMsense may be labeled as an indeterminate cycle because the A4 window maximum amplitude cannot be determined for the cycle that ends on the early Vpace.

[0117] If the A3 window' expires, as determined at block 520, without an AMsense, Vsense or Vpace, the A4 window begins (block 522) at the A3 window ending time (as described above in conjunction with FIG. 5). Control circuit 206 may determine the maximum (absolute) amplitude of the motion sensor signal during the A4 window which expires with the next ventricular event, e.g., the next Vpace or Vsense. The A4 window' maximum amplitude is buffered in memory 210. The associated cardiac cycle may be labeled as possible AF by control circuit 206. The possible AF label may be buffered with the A4 window maximum amplitude at block 524.

[0118] The A4 w indow maximum amplitude can be the maximum amplitude of the motion sensor signal whether or not an AMsense occurs during the A4 window- (e g., whether or not the motion sensor signal crosses the low A4 sensing threshold amplitude applied to the motion sensor signal during the A4 window as shown in FIGs. 4 and 5). As such, the A4 windowmaximum amplitude buffered in memory 210 could be less than the low A4 sensing threshold amplitude in some cardiac cycles. In other cardiac cycles, when the low A4 sensing threshold is crossed, the maximum amplitude may be the peak amplitude of the A4 signal sensed as the atrial event signal. The cardiac cycle may be labeled as a possible AF cycle at block 524 because an analysis of the A4 window maximum amplitudes can reveal compromised atrial mechanical function, e.g., due to AF. The A4 window maximum amplitudes stored for possible AF cycles can be evaluated upon expiration of the 2-minute time interval for classifying the AF analysis interv al, as further described below' in conjunction with FIG. 8.

[0119] In some examples, control circuit 206 may store the A4 window maximum amplitude time for the cardiac cycle labeled as a possible AF cycle at block 524. The time to the A4 window maximum amplitude from the Vpace or Vsense at the start of the cardiac cycle or from the A3 window' ending time may be buffered in memory 210 as the A4 window maximumamplitude time. A4 window maximum amplitudes may occur at more variable times during the A4 window during AF. Accordingly, the A4 window maximum amplitude times may be accumulated during the AF analysis interval for use in classifying the AF analysis interval as further described below.

[0120] If the interval started at block 502 has not expired (decision block 526). control circuit 206 may return to block 504 upon the expiration of the A4 window when the next ventricular event occurs. The process of labeling cardiac cycles and storing maximum A4 window amplitudes for the cardiac cycles that do include an A4 window can be repeated until the interval, e.g., the 2-minute time interval, expires at block 526. Upon expiration of the AF analysis interval (“yes” branch of block 526), control circuit 206 may classify the interval based on the cardiac cycle labels and / or the buffered A4 window maximum amplitudes at block 528, as described below in conjunction with FIG. 8. In other examples, as further described below, if a threshold number of cardiac cycles that are labeled as not AF and / or indeterminate cycles is reached during the AF analysis interval, the AF analysis interval could be terminated early and a new AF analysis interval could be started by control circuit 206, either immediately or after a delay.

[0121] FIG. 8 is a flow chart 600 of a method for classifying an interval for detecting AF according to some examples. At block 602, control circuit 206 determines that a previously started AF analysis interval, e.g.. a 2-minute time interval, is expired. At block 604, control circuit 206 may determine if the A4 window was started in at least a threshold percentage of the cardiac cycles spanning the expired interval. The cardiac cycles for which an A4 window maximum amplitude is stored in memory 210 may be labeled as “possible AF’" cycles as described above, in which case an A4 window maximum amplitude has been buffered for use in classifying the AF analysis interval. If the percentage of possible AF cycles during the AF analysis interval is at least a threshold percentage of all of the cardiac cycles of the expired AF analysis interval, control circuit 206 may evaluate the stored A4 window maximum amplitudes at block 605.

[0122] For example, when at least 30%, 40%, 50%, 60% or other specified threshold percentage of the cardiac cycles of the AF analysis interval are “possible AF” cycles, e.g., cycles for which an A4 window maximum amplitude is stored in memory' 210, control circuit 206 may evaluate the A4 window maximum amplitudes. If a zero value is stored for a cardiac cycle (e.g., a not AF cycle or an indeterminate cycle in which no A4 window is started), any zero values buffered as A4 window maximum amplitudes may be removed prior to evaluating the A4 window maximum amplitudes. In one example, if at least 50% of the cycles are possible AF cycles at block 604,control circuit 206 may evaluate the A4 window maximum amplitudes at block 605 to determine if the criteria for classifying the interval as an AF interval are met at block 606. The AF interval criteria may include one or more thresholds, ranges or other values that are compared to the A4 window maximum amplitudes and / or metrics determined therefrom and are indicative of a likelihood of AF being present during the AF analysis interv al.

[0123] In some examples, control circuit 206 may determine a representative A4 window maximum amplitude from the stored A4 window maximum amplitudes and compare the representative A4 window maximum amplitude to an AF threshold amplitude. For example, control circuit 206 may determine a mean, median, mode, or specified percentile of the A4 window maximum amplitudes stored for the AF analysis interval as a representative A4 window maximum amplitude. In an illustrative example, control circuit 206 may determine the 75thpercentile of the stored A4 window maximum amplitudes.

[0124] At block 606, control circuit 206 may compare the representative A4 window maximum amplitude to the AF threshold amplitude. The threshold amplitude may be a specified amplitude, above which the representative A4 window maximum amplitude may be evidence of coordinated, relatively strong atrial contractions during the AF analysis interval, which is evidence against AF (and against compromised atrial mechanical function). When the representative A4 window maximum amplitude is less than the threshold amplitude, the relatively weak or low amplitude atrial event signals or absence of atrial event signals can be evidence for AF (or more generally compromised atrial mechanical function). The AF threshold amplitude included in the AF interval criteria may be a specified, programmable value that is stored in memory 210, such as 0.5 meters / second / second (m / s2), 0.75 m / s2, 1 m / s2, 1.25 m / s2, 1.5 m / s2, 1.75 m / s2or 2.0 m / s2, as examples with no limitation intended. In an example, if the 75thpercentile of all buffered A4 window maximum amplitudes (after removing any zeros stored for indeterminate or not AF labeled cycles) is less than 1 m / s2, control circuit 206 may classify7the AF analysis interval as an AF interval at block 608.

[0125] In other examples, the AF threshold amplitude included in the AF interval criteria may be determined by control circuit 206 based on stored A4 window maximum amplitudes. For example, the AF threshold amplitude may be set as a percentage or offset less than the maximum peak amplitudes of atrial sensed event signals, e.g., the maximum peak amplitudes of the motion sensor signal following the low A4 sensing threshold crossings that result in AMsense signals being produced by atrial event detector circuit 240. Control circuit 206 may be configured to determine the peak amplitudes of sensed atrial event signals during time intervals that are classified as non-AF intervals, when atrial mechanical function is not compromised or when aclinician manually confirms that the patient is not experiencing an AF episode. Control circuit 206 may determine the AF threshold amplitude as a percentage, e.g., 50%, 60%, 70% or 80% of the average or median peak amplitude of sensed atrial event signals that resulted in AMsense signals produced in A4 windows, for example. In other examples, AF threshold amplitude may be determined by control circuit 206 as the mean or median peak amplitude of sensed atrial event signals less a specified offset, e.g., less 0.2, 0.3, 0.4, 0.5 or 1.0 m / s2

[0126] In still other examples, the mean or median (or other representative value) of the peak amplitudes of the sensed atrial event signals (that result in an AMsense signal produced during the A4 window) may be compared to a specified threshold amplitude, e.g., 1.0 m / s2. If the representative value of peak amplitudes of sensed atrial event signals (corresponding to AMsense signals produced by atrial event detector circuit 240 in A4 windows) is less than or equal to the specified threshold amplitude, control circuit 206 may determine the AF threshold amplitude as being an offset less than the representative peak amplitude of sensed atrial event signals. For instance, the AF threshold amplitude may be determined to be the lesser one of the specified threshold amplitude or an offset less than the average peak amplitude of A4 window sensed atrial event signals. When the average peak amplitude of atrial event signals sensed during A4 windows is less than or equal to the specified threshold amplitude (e.g., 1.0 m / s2), the AF threshold amplitude may be 0.2 m / s2(or other specified offset) less than the average peak amplitude of the atrial event signals. Otherwise, the AF threshold amplitude may be set equal to the specified threshold amplitude, or 1.0 m / s2in this illustrative example.

[0127] Referring again to block 606, if the representative A4 window maximum amplitude does not meet the specified or determined AF threshold amplitude, control circuit 206 may determine that the criteria for classifying the interval as an AF interval are met (“yes” branch of block 606). Control circuit 206 may store the interval classification in memory 210 as being an AF interval. Depending on the threshold number of intervals required to detect AF, e.g., as described above in conjunction with FIG. 6, control circuit 206 may or may not detect AF upon classifying the AF analysis interval as an AF interval at block 608.

[0128] If the representative A4 window maximum amplitude does meet the AF threshold amplitude, control circuit 206 may determine that the criteria for classifying the interval as an AF interval are not met (“no” branch of block 606). Control circuit 206 may store a non-AF interval classification in memory 210 at block 610. Depending on the threshold number of non-AF intervals required for detecting AF episode termination (if an AF episode has been detected), control circuit 206 may or may not detect an AF episode termination upon classifying the interval as a non-AF interval at block 10.

[0129] In other examples, instead of or in addition to determining a representative A4 window maximum amplitude that is compared to an AF threshold amplitude at block 606, control circuit 206 may determine a measure or metric of spread of the A4 window maximum amplitudes. For example, control the range of the stored A4 window maximum amplitudes, the interquartile range of the stored A4 window maximum amplitudes, the lower quartile range, the upper quartile range, a standard deviation or other metric of the spread of the A4 window maximum amplitudes. During AF, the range or spread of A4 window maximum amplitudes may be less than the range or spread of the A4 window maximum amplitudes during normal sinus rhythm or non-AF rhythms. A metric of the spread of the A4 window maximum amplitudes may be compared to a threshold value included in the AF interval criteria. When a metric or measure of spread of the A4 window maximum amplitudes is less than the threshold value ("yes" branch of block 606), the AF analysis interval may be classified as an AF interval (block 608). Otherwise (“no” branch of block 606), the AF analysis interval may be classified as a non-AF interval (block 610).

[0130] In some examples, control circuit 206 may determine a variability7of the A4 window maximum amplitude time. Control circuit 206 may determine a mean of the A4 window maximum amplitude times. Control circuit 206 may determine a variability of the A4 window maximum amplitude time as an average of the absolute differences between the buffered A4 window maximum amplitude times and the average. In other examples, control circuit 206 may determine the difference between the minimum and the maximum A4 window maximum amplitude times as a variability7metric of the A4 window maximum amplitude times. In still other examples, control circuit 206 may determine a standard deviation of the A4 window maximum amplitude times for assessing the variability of the A4 window maximum amplitude times. Greater variability in the A4 window maximum amplitude times may be evidence of AF. As such, control circuit 206 may determine a variability metric of the A4 window maximum amplitude times at block 606 and compare the variability metric to a corresponding variability threshold. The AF interval criteria may be met at block 606 when the variability metric is greater than a respective variability threshold.

[0131] In various examples, one or more representative values of the A4 window maximum amplitudes (e.g., mean, median, mode, specified percentile, maximum, etc.) and / or one or more metrics of A4 window maximum amplitude spread and / or one or more metrics of the variability of the A4 window maximum amplitude times may be determined at block 605 and compared to respective thresholds, ranges or values at block 606 by control circuit 206. When at least one or arequired combination of the AF interval criteria thresholds, ranges or values are met. control circuit 206 may classify the AF analysis interval as an AF interval at block 608.

[0132] Referring again to block 604, control circuit 206 may determine that the percentage of possible AF cycles (for which an A4 window maximum amplitude is stored in memory 210) does not meet the threshold percentage required for evaluating the A4 window maximum amplitudes (‘’no” branch of block 604). At block 612, control circuit 206 may determine if the percentage of cycles labeled as indeterminate cycles is greater than the percentage of cycles labeled as not AF cycles. If the percentage of cycles labeled as indeterminate cycles is greater than the percentage of cycles labeled not AF (“yes” branch of block 612), control circuit 206 may classify the AF analysis interval as being “unclassified” at block 614. The data gathered from the AF analysis interval is generally inconclusive as being evidence for detecting AF or for not detecting AF or for detecting AF termination or not detecting AF termination. Classification of the AF analysis interval as “unclassified,” therefore, may result in no detection of AF or no detection of AF termination (if an AF episode is being detected). In some examples, however, control circuit 206 may detect AF termination in response to a threshold number of unclassified AF interv als (or a combination of unclassified and non-AF interv als).

[0133] If the percentage of indeterminate cycles is not greater than the percentage of not AF cycles (“no” branch of block 612). control circuit 206 may classify the AF analysis interval as a non-AF interval. Classification of the interval as being non-AF may result in no detection of AF. If an AF episode is being detected, classification of the interval as being non-AF may or may not result in AF episode termination detection, depending on the threshold number of non-AF interv als required to detect termination.

[0134] As described above in conjunction with FIG. 7, in some examples cardiac cycles may be labeled using only two labels instead of using three labels (possible AF, indeterminate and not AF) as described here. For example, the cardiac cycles may be labeled as either possible AF or indeterminate (without using the not AF label). In other examples, the cardiac cycles may be labeled as possible AF or not AF (without using the indeterminate label). In these examples, when the percentage of possible AF cycles does not meet the threshold percentage at block 604, the AF analysis interval may be classified according to the second label. For example, when the cardiac cycles are labeled as either possible AF or indeterminate, and the percentage of possible AF cycles does not meet the threshold percentage at block 604, the AF analysis interval may be classified as unclassified. When the cardiac cycles are labeled as either possible AF or not AF, and the percentage of possible AF cycles does not meet the threshold percentage at block 604, the AF analysis interval may be classified as a non-AF interval by control circuit 206.

[0135] FIG. 9 is a diagram 650 of a histogram of A4 window maximum amplitudes (on x-axis) that may be determined by IMD 14 and stored in memory 210 over an AF analysis interval. The A4 window maximum amplitudes may be acquired for cardiac cycles that occur in the AF analysis interval and for which an A4 window is started. As described above, an A4 window may not be started in a given cardiac cycle if the high A4 sensing threshold is crossed in the A3 window (resulting in an AMsense signal produced in the A3 window) or if a Vsense or Vpace occur in the A3 window marking the end of the cardiac cycle and the start of the next cardiac cycle before the A4 window can start.

[0136] In some examples, the A4 window maximum amplitude may be stored for a given cardiac cycle only if other conditions are met that avoid storing A4 window maximum amplitudes that could be associated with a ventricular mechanical event. For instance, the A4 window maximum amplitudes buffered in memory 210 may be acquired for cardiac cycles that occur in the AF analysis interval for which an A4 window is started and reaches at least a minimum length or duration before a Vpace or Vsense terminates the cardiac cycle. In other examples, the A4 window maximum amplitude determined from a cardiac cycle may be rejected if the maximum amplitude occurs within a threshold time interval from a Vsense signal (that ends the A4 window and current cardiac cycle). Example conditions for accepting and buffering an A4 window maximum amplitude and for rejecting or discarding an A4 window maximum amplitude for use in detecting AF are further described below in conjunction with FIG. 11.

[0137] In FIG. 9, the number of cardiac cycles (y-axis) for which A4 window maximum amplitudes fall into a given histogram bin range of amplitudes is plotted. In this example, the histogram bin widths are 0.5 m / s2; however A4 window maximum amplitudes may be stored in memory 210 according to wider or narrow bin widths in other examples. When A4 window maximum amplitudes are stored in histogram bins in memory 210, memory 210 may include cycle label counters that are incremented one at a time as each cardiac cycle is labeled, e.g., as possible AF, indeterminate, or not AF. It is to be understood, however, that the A4 window maximum amplitudes may or may not be stored in memory 210 according to histogram bin widths. The A4 window maximum amplitudes may each be stored as individual values in a memory buffer over the AF analysis interval along with corresponding cardiac cycle labels.

[0138] The 75th percentile 652 of the A4 window maximum amplitudes stored for one AF analysis interval may be determined by control circuit 206 as a representative A4 window maximum amplitude for classifying the AF analysis interval according to some examples. Control circuit 206 may compare the determined percentile, which may be a different specified percentile than the 75thpercentile in other examples, to an AF threshold amplitude 654, show n as1 m / s2in this illustrative example. When the determined percentile 652 is less than the AF threshold amplitude 654, control circuit 206 may classify the AF analysis interval as an AF interval. As described above, a different representative A4 window maximum amplitude may be determined in other examples, such as a mean, median, mode, maximum, center value of the highest (furthest right) occupied bin. center value of the lowest (furthest left) occupied bin, the most frequently occupied bin, etc.

[0139] FIG. 10 is a diagram 670 of a histogram of A4 window maximum amplitudes (along the x-axis) that may be stored in memory 210 over another AF analysis interval. In FIG. 10, the number of cardiac cycles (y-axis) is plotted for which A4 window maximum amplitudes fall into each histogram bin range (bin widths of 0.5 m / s2in the example shown). Control circuit 206 may determine a representative A4 window maximum amplitude, shown here as the 75th percentile 752 of the A4 window maximum amplitudes. Control circuit 206 may compare the representative A4 window maximum amplitude to the AF threshold amplitude 674, shown as 1 m / s2in this example. When the representative A4 window maximum amplitude 672 is greater than the AF threshold amplitude 674, as is the case of FIG. 10, control circuit 206 may classify’ the AF analysis interval as a non-AF interval.

[0140] As shown by the example histograms of FIG. 9 and FIG. 10, the A4 w indow7maximum amplitudes buffered in memory 210 during AF or during compromised atrial mechanical function (see FIG. 9) can be expected to have a distribution that is overall lower in amplitude and have a narrower spread or range. The A4 window maximum amplitudes buffered in memory 210 when AF or compromised atrial mechanical function is not occurring (see FIG. 10) can have a distribution that is overall higher in amplitude and wider in range. Accordingly, control circuit 206 may be configured to determine a representative value of the A4 window maximum amplitudes, which may be a specified percentile, a measure of center (e.g., a mean, median, mode, or mid-range) or other representative value. The representative value may be determined with or without removing outliers. The representative value may be compared to a specified or determined AF threshold amplitude. The AF threshold amplitude may be determined by control circuit 206 from time intervals that are classified as non-AF (and may be confirmed as being non-AF intervals by a clinician). The AF threshold amplitude may be determined by control circuit 206 from peak amplitudes of atrial event signals sensed during A4 w indow s. For example, the amplitude threshold 674 may be determined from previous non-AF intervals as a specified percentile, e.g.. 20th, 25th, or 30thpercentile, of the A4 window maximum amplitudes buffered for one or more intervals that are classified as non-AF intervals. In this way, the amplitude threshold 674 may be tailored to a given patient.

[0141] In other examples, in addition to or alternatively to determining a percentile, a measure of center, or other representative value of the A4 window maximum amplitudes, control circuit 206 may be configured to determine a measure of spread of the A4 window maximum amplitudes. For example, the range between the maximum and the minimum A4 window maximum amplitudes, the interquartile range, the standard deviation, the number of occupied bins of the histogram bins storing values of the A4 window maximum amplitudes (each having a defined bin width), difference between a measure of center and a maximum or minimum A4 window maximum amplitude, or other metric that is representative of the spread of the distribution of the A4 window maximum amplitudes may be determined by control circuit 206. Control circuit 206 may compare the determined metric of spread to a respective threshold for discriminating between intervals when AF is present (e g., having relatively narrower spread or range) and intervals of non-AF (e.g., having wider spread or range). In still other examples, control circuit 206 may determine a metric of spread or variability of the A4 window maximum amplitude times for discriminating between intervals when AF is likely present (having greater variability’ in the A4 window maximum amplitude times) vs. when AF is not present (having more consistent, less variable A4 window maximum amplitude times). In various examples, control circuit 206 may classify the AF analysis interval as AF or non-AF based on the A4 window maximum amplitudes buffered in memory 210 by determining one or more measure(s) of center, specified percentile(s). measure(s) of spread, or any combination thereof and comparing each determined metric of the A4 w indow maximum amplitudes and / or A4 window maximum amplitude times to a respective threshold or range of values.

[0142] FIG. 11 is a flow chart 700 of a method that may be performed by IMD 14 for accepting or rejecting A4 window maximum amplitudes for use in detecting AF according to some examples. A relatively high A4 window maximum amplitude can be evidence of a relatively strong atrial mechanical contraction that is not expected to occur during AF. However, in some instances, the maximum amplitude of the motion sensor signal during the A4 w indow' could be associated with a ventricular mechanical event that occurs late in the A3 window or early in the A4 window. In other instances, if the maximum amplitude of the motion sensor signal occurs very close in time to a Vsense signal, the maximum amplitude could be due to an Al signal associated with an intrinsic ventricular contraction corresponding to the Vsense signal. As such, control circuit 206 may apply conditions to the A4 window and / or the A4 indow maximum amplitude time in order to accept or reject the A4 window maximum amplitude for a given cardiac cycle.

[0143] At block 702, the A4 window of a given cardiac cycle is started. As described above in conjunction with FIGs. 4 and 5, the A4 window begins at the ending time of the A3 window when the A4 event signal is not sensed in the A3 window and no Vsense or Vpace terminate the cardiac cycle prior to A3 window ending time. Control circuit 206 may track the amplitude of the motion sensor signal during the A4 window to determine the A4 window maximum amplitude at block 704.

[0144] At block 706, control circuit 206 may determine if the A4 window is greater than a threshold time duration (X ms). For example, if the A4 window ends upon a Vsense or a Vpace within 30 ms, 40 ms. 50 ms, 60 ms, 80 ms, or 100 ms of the A3 window ending time (as examples with no limitation intended), control circuit 206 may label the cardiac cycle as indeterminate at block 710. In some examples, the A4 window maximum amplitude, if determined, may be rejected or discarded and not retained in memory' 210 for use in classifying the current AF analysis interval. A Vsense or Vpace event that occurs within a short time, e.g., 50 ms. of the A3 window ending time, for example, may render the A4 window too short for evaluating the motion sensor signal for the purposes of AF detection. A maximum amplitude of the motion sensor signal during the relatively short A4 window' could be associated with a ventricular event, e.g., the A3 signal extending into the A4 window or an Al signal associated with a Vsense signal and occurring early in the A4 window or late in the A3 window. As such, control circuit 206 may label the cardiac cycle as an indeterminate cardiac cycle at block 710 because the motion sensor signal amplitude evidence for labeling the cardiac cycle as a possible AF cycle or as a not AF cycle is inconclusive when the A4 window is shorter than the threshold duration applied at block 706. In some examples, the cardiac cycle may be labeled as indeterminate at block 710 and the A4 window maximum amplitude may be stored in memory 210. The A4 window- maximum amplitude that is stored for an indeterminate cardiac cycle that ends w ithin the threshold X ms from the A3 window ending time may be evaluated by control circuit 206 with other A4 window maximum amplitudes (e.g., at block 605 of FIG. 8) that are stored for possible AF cycles, e.g.. when the percentage of cardiac cycles in the AF analysis interval that are labeled possible AF cycles meets a threshold percentage (see block 604 of FIG. 8).

[0145] In FIG. 11, block 706 for comparing the A4 window length to a threshold duration is shown after block 704 for determining the maximum amplitude in the A4 window. The order of blocks 704 and 706 could be reversed in that control circuit 206 may not determine the A4 window maximum amplitude until after the A4 window' length has reached at least the threshold duration. When the A4 window length does not reach the threshold duration, control circuit 206may not determine the A4 window maximum amplitude at all (or discard any buffered value) and only store the indeterminate label of the cardiac cycle at block 710.

[0146] When the A4 window meets the threshold duration (“yes” branch of block 706), control circuit 206 may determine if the A4 window maximum amplitude is at least a threshold time interval before a ventricular electrical event at block 708. For instance, control circuit 206 may determine if the A4 window maximum amplitude time is at least a threshold time interval (e.g., Y ms) earlier than a Vsense signal. The current A4 window may end on aVpace or aVsense signal. AVpace may be triggered by an AMsense and delivered at the AV pacing interval or at the expiration of a ventricular pacing interval when no AMsense or Vsense occurs. The ventricular myocardial contraction is expected to occur at a delay after the Vpace such that an A4 window maximum amplitude that is not associated with a ventricular contraction caused by the delivered Vpace is generally expected to occur at least the threshold time interval prior to the Vpace. AVsense, however, may occur prior to an AMsense or Vpace and terminate the A4 window. The ventricular contraction and associated Al event signal corresponding to the Vsense could occur close in time to (or even start earlier than) the Vsense because of inherent delays in sensing and processing the cardiac electrical signal or other factors. An A4 window maximum amplitude of the motion sensor signal that is close in time, e.g., within 10 ms, of a Vsense signal could be caused by the Al event signal (associated with ventricular myocardial contraction). In this case, the A4 window maximum amplitude may not be reliable for use in monitoring for AF or compromised atrial mechanical function.

[0147] If the time of the A4 window maximum amplitude is within the specified time interval threshold of a ventricular electrical event (block 708), therefore, control circuit 206 may label the cardiac cycle as indeterminate (block 710). The A4 window maximum amplitude may be ignored or discarded and not retained in memory 210 for analysis with other A4 window maximum amplitudes acquired during the current AF analysis interval. If the A4 window maximum amplitude time is more than the threshold time interval from a ventricular electrical event, the A4 window maximum amplitude is less likely to be caused by a ventricular mechanical event. Control circuit 206 may classify the cardiac cycle as a possible AF cycle and buffer the A4 window maximum amplitude (and time) in memory 210 (at block 712). In this way, control circuit 206 may select the A4 window maximum amplitudes and times determined during the AF analysis interval that are unlikely to be a sample point of a ventricular mechanical event signal, e.g., an Al event signal or an A3 event signal.

[0148] FIG. 12 is a flow chart 800 of a method for classifying AF analysis intervals based on motion sensor signal analysis according to some examples. When patient physical activity ishigh, motion sensor signal amplitudes may be relatively high and more variable than when the patient is at rest. As such, the motion sensor signal amplitude during A4 windows may be relatively high and or more variable due to patient physical activity, whether or not the patient is experiencing compromised atrial mechanical function, such as during an AF episode. The motion sensor signal may be somewhat less reliable for detecting compromised atrial mechanical function, e.g., for detecting an AF episode (or detecting termination of an AF episode) during increased patient physical activity. Accordingly, control circuit 206 may be configured to classify AF analysis intervals in a process that takes into account changes in the motion sensor signal due to patient physical activity in some examples.

[0149] At block 801, control circuit 206 may start an AF analysis interval, e.g., a 2-minute time interval or other interval according to any of the examples described above. At block 802, control circuit 206 may determine patient physical activity metrics (also referred to herein as “activity metrics'’) according to an activity monitoring protocol. As described above in conjunction with FIG. 3. control circuit 206 may determine patient physical activity metrics for use in determining an SIR and controlling rate responsive pacing. In an example, control circuit 206 may be configured to determine an activity count from the motion sensor signal sensed over each activity count interval that occurs during the AF analysis interval. The activity count interval may be 1 second, 2 seconds, 3 seconds or other selected time interval. For instance, control circuit 206 may determine an activity count from at least one axis signal of the motion sensor every 2 seconds, e g., by summing sample point amplitudes of the rectified axis signal. Other methods may be used for determining a patient physical activity metric in other examples. The techniques disclosed here for classifying an AF analysis interval, e.g., as AF, non-AF or indeterminate, for use in compromised atrial mechanical function detection is not limited to a particular method for determining a level or metric of patient physical activity. In some examples, the patient physical activity metric could be determined from a different sensor signal than the motion sensor signal. For instance, a patient physical activity metric can be determined from a temperature signal, a respiration signal such as a thoracic impedance signal, an oxygen sensor signal or other sensor signal that varies with patient physical activity or combinations thereof.

[0150] At block 804, control circuit 206 may determine if one or more rate response pacing pulses are delivered during the current AF analysis interval. When rate response pacing is enabled, one or more ventricular pacing pulses may be delivered at a temporary LRI that is adjusted (e g., shortened) from the programmed ventricular LRI based on an SIR determined from the patient physical activity metric. When patient physical activity' is increased resulting inrate response pacing, the motion sensor signal is expected to include non-cardiac acceleration signals caused by the increased patient activity. Non-cardiac acceleration signals can occur any time during the cardiac cycles of the AF analysis interval, including A4 windows. As such, analysis the A4 window maximum amplitudes may be less reliable for monitoring for during rate response pacing.

[0151] If one or more rate response pacing pulses are delivered during the AF analysis interval (‘■yes” branch of block 804), control circuit 206 may classify the AF analysis interval as unclassified at block 806. In some examples, at least a threshold number of rate response pacing pulses greater than one may be required during the AF analysis interval to classify the interval as unclassified. For instance, at least 2. 3, 4 or other specified number or percentage of cardiac cycles during the AF analysis interval may be required to be rate response pacing cycles for control circuit 206 to classify the entire AF analysis interval as being unclassified at block 806. By classifying the AF analysis interval as unclassified at block 806. classification of the AF analysis interval as non-AF is avoided when higher amplitude acceleration signals due to patient physical activity could cause a false negative AF detection. If an AF episode is already being detected (because the AF detection criteria were met prior to the current AF analysis interval), classifying the current AF analysis interval as “unclassified” at block 806 due to rate response pacing can avoid false detection of AF episode termination due to acceleration signals caused by the patient activity resulting in a non-AF interval classification (as opposed to true A4 event signals that return when the AF is terminated).

[0152] If no rate response pacing pulse (or less than a specified threshold number or percentage of rate response pacing pulses) are delivered during the AF analysis interval, control circuit 206 may evaluate the patient physical activity metrics determined during the AF analysis interval (at block 808) for use in classify ing the interval. Control circuit 206 may determine if the activity metrics are above a resting activity level for at least a threshold percentage of the AF analysis interval. Control circuit 206 may apply a non-resting activity level threshold to each of the activity metrics determined during the AF analysis interval. A non-resting activity level may be any activity metric greater than a resting level, e.g., greater than the resting set point of the transfer function used by control circuit 206 to convert activity counts to an SIR as described above in conjunction with FIG. 3. Anon-resting activity level may be any activity metric greater than an ADL level, e g., greater than the ADL set point of the transfer function used by control circuit 206 to convert activity counts to an SIR as described above in conjunction with FIG. 3. In still other examples, the non-resting activity level may be an activity count that is greater than anaverage of the resting set point and the ADL set point, a weighted combination of the resting set point and the ADL set point, or other specified activity count threshold.

[0153] When less than a threshold percentage of activity7metrics determined during the AF analysis interval are greater than the specified non-resting threshold (“no” branch of block 808), control circuit 206 may keep the AF interval criteria applied to the A4 window maximum amplitudes unchanged at block 810. For example, if the activity7metrics that are greater than the non-resting threshold are no more than 10% or other specified percentage of all the activity metrics determined during the AF analysis interval, control circuit 206 may advance to block 810, where no adjustment to the AF interval criteria is made. The threshold percentage applied at block 808 may be 10%, 15%, 20%, 25%, 30%, 40%, or 50% as examples. For instance, if the patient physical activity metrics are at a resting level for a majority of the AF analysis interval, control circuit 206 may advance to block 810 where no AF interval criteria adjustment is made.

[0154] At block 816, control circuit 206 may determine if the A4 window maximum amplitudes buffered during the AF analysis interval meet the AF interval criteria. As described above, a specified percentile, one or more measures of center or other representative A4 window maximum amplitude and / or one or more measures of spread of the A4 window maximum amplitudes and / or A4 window maximum amplitude times may be compared to respective thresholds or ranges of the AF interval criteria at block 816 to determine if AF analysis interval should be classified as an AF interval or a non-AF interval. If the AF interval criteria are met, control circuit 206 may classify the AF analysis interval as an AF interval at block 818. If the AF interval criteria are not met, control circuit 206 may classify the AF analysis interval as a non-AF interval at block 820. Control circuit 206 may return to block 801 to start the next AF analysis interval. Control circuit 206 may determine if AF detection criteria or AF episode termination criteria are met upon classification of the current AF analysis interval, e.g., as described above in conjunction with FIG. 6 and below in conjunction with FIGs. 13 — 14.

[0155] Referring again to block 808, if the percentage of activity metrics that are greater than the non-resting threshold exceeds the threshold percentage, (“yes” branch of block 808). control circuit 206 may evaluate the activity metrics at block 812 to determine whether to pause evaluation of the motion sensor signal for AF detection due to high patient activity. At block 812, control circuit 206 may determine if non-resting activity metrics, e.g., relatively high activity metrics, are dominant during the AF analysis interval compared to resting or relatively low activity metrics. For example, control circuit 206 may determine if the percentage of activity metrics that are greater than the non-resting threshold is greater than (or equal to) the percentage of activity7metrics that are less than the non-resting threshold.

[0156] Control circuit 206 may determine that high activity is dominant during the AF analysis interval when the percentage of activity metrics exceeding the non-resting threshold is greater than the percentage of activity metrics that do not exceed the non-resting threshold. If high activity is not dominant (“no” branch of block 812), control circuit 206 may optionally adjust the AF interval criteria at block 814. For example, an AF threshold amplitude applied to a representative A4 window maximum amplitude may be increased at block 814. A threshold range applied to the range of A4 window maximum amplitudes (or another metric of spread of the A4 window maximum amplitudes) may be increased when high activity is not dominant (“no” branch of block 812) but the activity metrics that are greater than the non-resting threshold account for at least the threshold percentage of all activity metrics determined for the AF analysis interval (“yes” branch of block 808). There may be some patient activity' causing greater variation and higher amplitudes in the motion sensor signal and greater variability in the A4 window maximum amplitude times but the increased variation and amplitude may be accounted for in elassilying the AF analysis interval by adjusting the AF interval criteria at block 814. In an illustrative example, if the non-resting activity’ metrics account for more than 10% but less than 50% of the AF analysis interval, control circuit 206 may classify the AF analysis interval (block 816) after adjusting the AF interval criteria (block 814).

[0157] At block 816, control circuit 206 may apply the adjusted AF criteria to the A4 window maximum amplitudes buffered in memory 210 for the AF analysis interval. Using the example shown in FIGs. 9 and 10, control circuit 206 may determine the 75thpercentile of the stored A4 window maximum amplitudes as a representative A4 window maximum amplitude for comparison to an adjusted AF threshold amplitude . The adjusted AF threshold amplitude may be increased at block 814 from the non-adjusted value by a selected increment or percentage, e.g., from 1.0 m / s2to 1.3, 1.5, 1.8, or 2.0 m / s2.

[0158] In some examples, the adjustment to a threshold or range of the AF interval criteria that is applied to a representative value of the A4 w indow maximum amplitudes or to a metric of spread of the A4 w indow maximum amplitudes may be determined by control circuit 206 based on the activity metrics determined during the AF analysis interval. The adjustment of the AF interval criteria may be scaled according to the level of the activity metrics. For example, if the predominant (e.g., highest percentage of) activity’ metrics fall into a first range of activity' levels greater than the non-resting threshold, control circuit 206 may increase an AF interval criteria threshold by a first increment. Using the examples of FIGs. 9 and 10. the AF threshold amplitude applied to the specified percentile of the A4 w indow maximum amplitudes may be increased from 1.0 m / s2to 1.3 m / s2. If the predominant activity metrics fall into a second range of activitylevels higher than the first range, control circuit 206 may increase the AF amplitude threshold by a second increment greater than the first increment. Again using the examples of FIGs. 9 and 10, the AF amplitude threshold applied to a specified percentile of the A4 window maximum amplitudes may be increased from 1.0 m / s2to 1.6 m / s2at block 814.

[0159] Control circuit 206 may classify the AF analysis interval as a non-AF interval (block 820) if the adjusted AF interval criteria are not met at block 816. Control circuit 206 may classify the AF analysis interval as an AF interval (block 818) if the adjusted AF interval criteria are met at block 816. In this way, acceleration forces imparted on the motion sensor due to increased patient physical activity or other increased non-cardiac acceleration forces may be accounted for by adjusting the AF interval criteria used for classifying the AF time interval when at least some activity metrics determined during the AF analysis interval are higher than the non-resting threshold.

[0160] Referring again to block 812, when high activity is dominant (“yes'’ branch of block 812), control circuit 206 may pause analysis of the motion sensor signal for AF detection for at least the current AF analysis interval. No determination of whether AF detection criteria (or AF episode termination criteria) may be performed by control circuit 206 based on the current AF analysis interval because the predominant, relatively high patient physical activity during the AF analysis interval may confound the analysis of the motion sensor signal for detecting relatively low and / or narrow distribution of A4 window maximum amplitudes associated with AF or compromised atrial mechanical function. Control circuit 206 may ignore or discard the A4 window maximum amplitudes and cardiac cycle labels accumulated during the current AF analysis interval and skip classifying the AF analysis interval as being one of unclassified, AF or non-AF. No AF detection or AF episode termination detection is made based on the current AF analysis interval. Control circuit 206 may return from block 822 to block 801 to start the next AF analysis interval.

[0161] In other examples, as shown in the example of FIG. 12, when evaluation of the motion sensor signal for AF detection is paused (at block 822) in response to predominantly high activity metrics (“yes branch of block 812), control circuit 206 may wait for a threshold number of activity metrics to be determined that are less than the non-resting threshold at block 824. For instance, if the activity metric is determined every' two seconds, control circuit 206 may wait for at least 2, 4, 6, 8, 10, 12, 15, 30 or other selected number of activity metrics to be less than the non-resting threshold prior to starting the next AF analysis interval at block 801. In some examples, control circuit 206 may wait for at least 50% of the most recent Y activity metrics, for example at least 50% of the most recent 20 activity metrics or other specified X out ofY activity'metrics, to be less than the non-resting threshold. Generally, control circuit 206 may wait at block 824 until the patient physical activity level has decreased from the predominantly high (non-resting) activity detected at block 812 to predominately lower activity (e.g., more than 50% resting level activity metrics) before starting the next AF analysis interval at block 801.

[0162] In other examples, control circuit 206 may advance from block 822 to block 801 without waiting for a change in patient physical activity level. Control circuit 206 may accumulate A4 window maximum amplitudes and cardiac cycle labels in memory 210 and evaluate the activity level metrics obtained in the next AF analysis window. Based on an evaluation of the activity level metrics according to the methods of flow chart 800, control circuit 206 may determine whether to continue pausing the evaluation of A4 window maximum amplitudes (block 822) or re-start evaluating the A4 window maximum amplitudes by classifying the next AF analysis interval as an unclassified interval (block 806), an AF interval (block 818) or a non- AF interval (block 820). In this way, false negative AF detection and false positive AF episode termination detection can be avoided due to confounding acceleration signals caused by patient physical activity or other non-cardiac motion.

[0163] It is to be understood that the methods described in conjunction with FIG. 12 can be combined with the cardiac cycle labeling methods and AF analysis interval classification methods described above in conjunction with FIGs. 7, 8 and 11. In FIG. 12, it is assumed that a threshold percentage of cardiac cycles are labeled as possible AF cycles so that the AF interval criteria are applied to the A4 window maximum amplitudes buffered for the possible AF cycles. It is to be understood that some cardiac cycles may be labeled as indeterminate and / or not AF during the process of FIG. 12 such that some AF analysis intervals may be classified as unclassified even when no rate response pacing pulses are delivered. Some AF analysis intervals may be classified as non-AF intervals without necessarily applying the AF interval criteria (nonadjusted or adjusted) due to a high percentage of cardiac cycles being labeled as not AF (e.g., see block 616 of FIG. 8).

[0164] FIG. 13 is a diagram 850 of a method for detecting an AF episode by an IMD according to some examples. The AF episode can be detected in real time in the ambulatory patient to allow recordation of cardiac signals representative of compromised atrial mechanical function, determination and storage of other AF episode data, determination of AF burden, and / or other responses performed upon detection of the AF episode by the IMD. Such real time detection of AF during an office visit with a clinician is not always practical and sometimes impossible to achieve because the occurrence of AF can be unpredictable and sporadic. By providing an IMD capable of detecting an AF episode as it is occurring, data collection and / or adjustments to thesensing and / or therapy delivery functions of the IMD can be performed. Such real time AT / AF detection and / or real time compromised atrial mechanical function detection and data collection can provide vital information to the clinician in managing the patient care for arrhythmia prevention, stroke prevention and / or other cardiac rhythm management therapies. Continuous collection of cardiac signals for transmission to an external device for post-processing and analysis for detecting atrial tachyarrhythmia can be limited at best due to power and processing constraints of the medical device and may be impractical or non-feasible due to the large amount of data storage required, power requirements for data transmission, etc. Furthermore, manual and / or automated post-processing of cardiac signals for detecting atrial tachyarrhythmias or compromised atrial mechanical function could delay critical patient care and precludes the ability of a medical device operating according to the presently disclosed techniques to respond in real time to compromised atrial mechanical function detection, e.g., AF detection, and / or termination detection, e.g., by adjusting operating parameters of the IMD in controlling sensing and / or therapy delivery’ functions, transmitting a detection notification to a co-implanted medical device, transmitting a termination notification to a co-implanted medical device, etc.

[0165] Accordingly, the techniques set forth herein provide specific improvements in the field of medical devices that have practical applications. By providing a medical device system configured to detect atrial tachyarrhythmia episodes in real time, or more generally detect compromised atrial mechanical function in real time, using a cardiac mechanical signal enables atrial tachyarrhythmia episode detection, even when ventricular pacing is being delivered such that intrinsic ventricular depolarization signals are not occurring for monitoring intrinsic RRIs for irregular or rapidly conducted atrial tachyarrhythmia. The disclosed techniques that utilize the cardiac mechanical signal can be implemented in a ventricular device that can provide ventricular pacing rate support without requiring P-wave sensing or RRI monitoring for detecting the atrial tachyarrhythmia. P-wave sensing used for detecting AF can be unreliable because of the small amplitude P-waves which can be undersensed or difficult to sense, even during normal sinus rhythm. The reliability of real time atrial tachyarrhythmia detection in a patient is improved by the cardiac mechanical signal analysis techniques disclosed herein, particularly during ventricular pacing (though the presently disclosed techniques are not necessarily limited to detecting atrial tachyarrhythmia during periods of ventricular pacing). Improved processor-based methods for detecting atrial tachyarrhythmia and compromised atrial mechanical function as disclosed herein reduces the likelihood of atrial tachyarrhythmia being undetected and untreated in a patient. A clinician may have no other conclusive evidence that the patient is experiencing episodes of atrial tachyarrhythmia or compromised atrial mechanicalfunction, which may require treatment or warrant stroke prevention therapies, without the real time detection and data collection provided by the techniques disclosed herein.

[0166] In the example of FIG. 13, a threshold number of consecutive AF analysis intervals may be required to be classified as AF in order for IMD 14 to detect an AF episode. As an illustrative example, with continued reference to IMD 14 of FIG. 3, control circuit 206 may detect AF (arrow 852) in response to three consecutive AF analysis intervals being classified as AF. In other examples, one, two, or more than three consecutive AF analysis intervals may be applied as the threshold number of AF intervals required to detect AF by control circuit 206.

[0167] A threshold number of consecutive AF analysis intervals may be required to be classified as non-AF in order for IMD 14 to detect termination of an AF episode. As an illustrative example, control circuit 206 may detect AF episode termination (arrow 854) in response to three consecutive AF analysis intervals being classified as non-AF (NAF). In other examples, one, two, or more than three consecutive AF analysis intervals may be applied as the threshold number of non-AF intervals required to detect AF episode termination by control circuit 206.

[0168] Control circuit 206 may determine the AF episode duration 856 as the time interval from AF detection 852 to AF episode termination detection 854. Any intervening AF analysis intervals may be included in the AF episode duration 856. The intervening AF analysis intervals may include one or more unclassified (U) and / or non-AF (NAF) intervals in some instances as shown. In other examples, the AF episode duration 856’ may be determined by control circuit 206 as the time interval from the AF detection 852 to the time of AF episode termination detection 854 less the total time of the threshold number of consecutive non-AF intervals leading up to AF episode termination detection 854. In still other examples, control circuit 206 may determine the AF episode duration 856” to begin from the start of the first AF analysis interval of the threshold number of consecutive AF analysis intervals classified as AF that resulted in the AF detection 852. The AF episode duration 856” may end upon AF episode termination detection 854 or at the start of the first non-AF interval that led to the AF episode termination detection criteria being met (as shown).

[0169] The AF episode duration (856, 856’ or 856”) may be stored in memory 210, along with a date and time stamp of AF detection 852, and may be used to update an AF burden as the percentage of time the patient is experiencing AF (or compromised atrial mechanical function), e.g., over a 24-hour period. The episode detection and duration may be stored in memory 210 with other episode related data determined by control circuit 206. such as intervening AF analysis interval classifications occurring during the AF episode, the representative A4 window maximum amplitudes determined from one or more AF analysis intervals, representativemeasure of spread of the A4 window maximum amplitudes and / or times determined from one or more of the AF analysis intervals, ventricular rate information, RRI variability, percentage of Vpaces, percentage of Vsenses, a segment of the cardiac EGM signal, a segment of the motion sensor signal, marker channel data indicating timing of AMsenses, Vsenses and Vpaces, or any combination thereof.

[0170] FIG. 14 is a diagram 860 of another method for detecting an AF episode by IMD 14. Control circuit 206 may detect AF (arrow 862) in response to a threshold number X out the most recent Y AF analysis intervals being classified as AF intervals. In the illustrative example, when at least 4 out of the most recent 5 AF analysis intervals are classified as AF (e.g., when at least 80% of the AF analysis intervals are AF intervals), control circuit 206 may detect AF (at arrow 862). In some examples, any AF analysis intervals that are not classified as AF may be required to be classified as “unclassified” in order to detect AF when the threshold number X AF intervals out of the most recent Y AF analysis intervals is reached. If one or more of the most recent Y AF analysis intervals is classified as non-AF, the AF detection criteria may remain unmet. When the most recent Y AF analysis intervals include only AF intervals or a combination of only AF intervals and unclassified intervals with at least X AF intervals, control circuit 206 may detect AF. In other examples, the most recent Y AF analysis intervals may include any combination of AF. non-AF and unclassified intervals, and control circuit 206 may detect AF if the most recent Y AF analysis intervals include at least X AF intervals. Control circuit 206 may evaluate the most recent 1, 2, 3, 5, 8, 12, 20 or other specified number of AF analysis intervals and detect AF when at least 50, 60, 66, 70, 75, 80, 90 or 100% of the AF analysis intervals are classified as AF interv als in various examples, with no limitation intended.

[0171] After detecting AF (862), control circuit 206 may evaluate the AF analysis intervals for detecting AF episode termination (at arrow 864). Control circuit 206 may detect AF episode termination 862 in response to a threshold number M of the P most recent AF analysis intervals being classified as non-AF intervals and / or unclassified. In the illustrative example shown, control circuit 206 may detect AF episode termination 864 in response to at least 4 out of 5 AF analysis intervals being classified as non-AF (NAF). One or more of the most recent Y AF analysis intervals may be unclassified (U) as shown. Control circuit 206 may not detect AF episode termination if any one of the most recent P AF analysis interv als is classified as AF. In other examples, if one or more of the most recent P AF analysis intervals is classified as an AF interv al, control circuit 206 may still detect AF episode termination 864 if at least M out of P interv als are classified as non-AF and / or unclassified. A threshold percentage of non-AF and / orunclassified intervals out of the most recent P AF analysis intervals may be required to detect termination of the AF episode.

[0172] FIG. 15 is a flow chart 900 of a method for classifying an AF analysis interval by IMD 14 according to another example. As described above in conjunction with FIG. 7, each cardiac cycle during the AF analysis interval may be labeled as being possible AF or labeled as not AF or indeterminate. The AF analysis interval may be subsequently classified as an AF interval, a non- AF interval or an unclassified interval, e.g., as described in conjunction with FIG. 8. In the example of FIG. 15, A3 window maximum amplitudes may be used in addition to A4 window maximum amplitudes for classifying the AF analysis interval.

[0173] At block 902, control circuit 206 may start the AF analysis interval for acquiring A4 window maximum amplitudes and A3 window maximum amplitudes, which may be accumulated in a buffer of memory 210. For the sake of illustration, control circuit 206 may start a 2-minute time interval at block 902. As described above, control circuit 206 may start the AF analysis interval to run for a specified time interval, a specified number of cardiac cycles (e.g.. based on a count ventricular electrical events), a specified number of Vpaces, or until a specified number of A4 window maximum amplitudes are acquired, in various examples. Generally the interv al includes multiple cardiac cycles in order to classify the rhythm during the interval as being AF (or more generally compromised atrial mechanical function), non-AF (or more generally non-compromised atrial mechanical function) or, in some cases, unclassified due to data that is indeterminate for confidently classifying the interval as AF or non-AF.

[0174] At block 904, control circuit 206 identifies a ventricular event, which may be an electrical event, e.g., a Vsense signal received from sensing circuit 204 or a Vpace delivered by pulse generator 202. At block 906. control circuit 206 starts the A3 window having a specified starting time and ending time following the ventricular event. At block 908, control circuit 206 determines the maximum (absolute) amplitude of the motion sensor signal sensed during the A3 window. This A3 window maximum amplitude may be buffered in memory 210.

[0175] As described above in conjunction with FIG. 5, during the A3 window, control circuit 206 applies the high A4 sensing threshold amplitude for sensing the A4 signal that may be fused with the A3 signal. The A3 window maximum amplitude determined at block 908 may be the maximum amplitude during the A3 window when the high A4 sensing threshold amplitude is not crossed during the A3 window. If the motion sensor signal crosses the high A4 sensing threshold amplitude during the A3 window, no A4 window is started C‘no?’ branch of block 910) and the cardiac cycle may be labeled as not AF at block 912. Control circuit 206 may buffer the not AF label for the cardiac cycle in memory 210. In other instances, a Vsense or Vpace may occurduring the A3 window such that an A4 window is not started. In this case, control circuit 206 may label the cardiac cycle as being indeterminate or not AF at block 912 according to any of the examples described above, e.g., in conjunction with FIG. 7. It is noted that if the A4 window is not started due to an AMsense. Vsense or Vpace occurring during the A3 window, the A3 window maximum amplitude determined for the A3 window at block 908 may be discarded. After labeling the cardiac cycle as not AF or indeterminate at block 912, control circuit 206 may return to block 904 to wait for the next ventricular event if the AF analysis interval is not expired (“no” branch of block 918).

[0176] If the A4 window is started (“yes” branch of block 910), control circuit 206 may determine the maximum (absolute) amplitude of the motion sensor signal during the A4 window. The associated cardiac cycle may be labeled as possible AF at block 914 by control circuit 206. The possible AF label may be buffered with the A4 window7maximum amplitude in memory' 210 at block 914.

[0177] The A4 w indow maximum amplitude can be the maximum amplitude of the motion sensor signal whether or not an AMsense occurs during the A4 window As such, the A4 window' maximum amplitude buffered in memory 210 could be less than the low A4 sensing threshold amplitude in some cardiac cycles. In other cardiac cycles, when the low' A4 sensing threshold is crossed, the maximum amplitude may be the peak amplitude of the A4 signal sensed as the atrial event signal. The cardiac cycle may be labeled as a possible AF cycle at block 914 because an analysis of the A4 window' maximum amplitudes can reveal compromised atrial mechanical function, e.g., due to AF. The A4 window' maximum amplitudes and the A3 window maximum amplitudes stored for possible AF cycles can be evaluated upon expiration of the AF analysis interval for classifying the AF analysis interval, as further described below. In some examples, a ratio of the A3 window' maximum amplitude and the A4 window maximum amplitude may be determined at block 916 and stored in memory 210 or the individual A3 window and A4 window maximum amplitudes may be determined and stored for subsequent analysis for classifying the AF analysis interval.

[0178] As described above, control circuit 206 may store the A4 window maximum amplitude time for the cardiac cycle labeled as a possible AF cycle at block 914. A4 window maximum amplitude times may be more variable during AF. Accordingly, the A4 window maximum amplitude times may be accumulated during the AF analysis interval for use in classifying the AF analysis interval in some examples. In some examples, the time of an A4 sensing threshold crossing during the A4 window or the time of the maximum peak following the A4 sensing threshold crossing may be determined by control circuit as the time of the A4 window maximumamplitude. The time of the A4 event signal may become more variable during an episode of AF and a metric of spread of the time of the A4 event signal may be determined from cardiac cycles having an A4 window in which the motion signal crosses the A4 sensing threshold. In other examples, the time of the A4 window maximum amplitude may be determined whether or not the A4 sensing threshold is crossed during the A4 window.

[0179] If the interval started at block 902 has not expired (decision block 918), control circuit 206 may return to block 904 upon the expiration of the A4 window when the next ventricular event occurs. The process of labeling cardiac cycles and storing maximum A3 window amplitudes and maximum A4 window amplitudes and optionally A4 window maximum amplitude times for the cardiac cycles that do include an A4 window can be repeated until the AF analysis interval expires at block 918.

[0180] Upon expiration of the AF analysis interval (“yes” branch of block 918), control circuit 206 may determine if a threshold percentage of the cardiac cycles during the AF analysis interval are labeled as possible AF cycles (block 920). If less than the threshold percentage of the cycles during the AF analysis interval are labeled possible AF cycles, control circuit 206 may classify the AF analysis interval as being an unclassified interval or a non-AF interval at block 922 according to any of the examples given above in conjunction with FIG. 8. If the number of cardiac cycles labeled as possible AF cycles meets the threshold percentage (“yes” branch of block 920). control circuit 206 may analyze the maximum A3 window amplitudes and the maximum A4 window amplitudes (and optionally A4 window maximum amplitude times) at block 924 for determining if AF interval criteria are met.

[0181] When atrial function is compromised, the A3 event signal may become larger, e.g., the A3 event signal may have an increased maximum absolute peak amplitude compared to the A3 event signal during periods of normal atrial function. An increase in the maximum A3 window' amplitude and / or an increase in the ratio of the maximum A3 window' amplitude to the maximum A4 window' amplitude (A3 / A4 ratio) may be detected by control circuit 206 as corroborating evidence of compromised atrial mechanical function. As such, at block 924, control circuit 206 may apply criteria to the maximum A4 window amplitudes and the maximum A3 window amplitudes for determining if AF interval criteria are met at block 924.

[0182] In an illustrative example, if a representative maximum A4 window amplitude (e.g., a specified percentile as described above in conjunction with FIGs. 9 and 10) is greater than the AF threshold amplitude, control circuit 206 may classify the AF analysis interval as a non-AF interval at block 928. However, if the representative maximum A4 window' amplitude is notgreater than the AF threshold amplitude, control circuit 206 may apply AF interval criteria to the maximum A3 window amplitudes and / or the A3 / A4 ratios buffered for the AF analysis interval.

[0183] For example, at block 924, control circuit 206 may determine a representative maximum A3 window amplitude and compare it to a respective threshold amplitude to detect a concomitant increase in the A3 event signal peak amplitude that accompanies the low maximum A4 window amplitudes during AF. The representative maximum A3 window amplitude may be a specified percentile, average, median, maximum or minimum of the stored maximum A3 window amplitudes. Additionally or alternatively, control circuit 206 may determine a representative A3 / A4 ratio from the stored A3 / A4 ratios. A3 / A4 ratios may be determined for each possible AF cycle as indicated at block 916, and a representative A3 / A4 ratio may be determined as a specified percentile, average, median, minimum or maximum of the stored A3 / A4 ratios. In other examples, control circuit 206 may determine the representative A3 / A4 ratio as the ratio of the representative maximum A3 window amplitude and the representative maximum A4 window amplitude.

[0184] If the representative maximum A3 window amplitude is greater than a respective threshold amplitude and / or the representative A3 / A4 ratio is greater than a respective threshold ratio, control circuit 206 may classify the AF analysis interval as an AF interval at block 926. In some examples, control circuit 206 may classify the interval as an AF interval (block 926) in response to the representative maximum A4 window amplitude being less than a respective threshold amplitude and the A3 / A4 ratio being greater than the respective threshold ratio and / or the representative maximum A3 window amplitude being greater than the respective A3 window threshold amplitude (“yes'’ branch of block 924). If the representative A3 window amplitude is not greater than the respective threshold amplitude and / or the representative A3 / A4 ratio is not greater than the threshold ratio, control circuit 206 may classify the AF analysis interval as a non-AF interval at block 928. In other examples, the AF analysis interval may be classified as an unclassified interval when the representative maximum A4 window amplitude is less than the AF threshold but the representative maximum A3 window amplitude and / or representative A3 / A4 ratio do(es) not meet the respective amplitude threshold or ratio threshold. Compromised atrial mechanical function may be present but may not be confirmed without an associated increase in the A3 event signal peak amplitude.

[0185] The threshold amplitude applied to the representative maximum A3 window amplitude and / or the threshold ratio applied to the A3 / A4 ratio may be stored in memory 210 as programmable values selected by a clinician or established by control circuit 206 based on maximum A3 window amplitudes and maximum A4 window amplitudes acquired during aperiod of normal atrial function, e.g., from AF analysis intervals classified as non-AF intervals or confirmed to be normal atrial function by a clinician as further described below in conjunction with FIG. 16.

[0186] In some examples, when the AF analysis interval is classified as a non-AF interval at block 928 based on the representative maximum A4 window amplitude being greater than the AF threshold, the maximum A3 window amplitudes and the maximum A4 window amplitudes accumulated for the possible AF cycles (those cycles in which an A4 window is started) may be used at block 930 for updating thresholds used in applying the AF interval criteria. The AF threshold applied to the representative maximum A4 window amplitude may be updated, for example, as a weighted combination of the current AF threshold and the representative maximum A4 window amplitude determined for the non-AF interval. The ratio threshold applied to the A3 / A4 ratio may be updated to a weighted combination of the current ratio threshold and the representative A3 / A4 ratio determined for the non-AF interval. When the representative maximum A3 window amplitude is being compared to a threshold, control circuit 206 may determine an updated respective threshold amplitude at block 930 that is a weighted combination of the current threshold amplitude and the representative maximum A3 window amplitude. In other examples, block 930 is optional and may be omitted. It is noted that when AF interval criteria thresholds are being updated using data accumulated during a non-AF interval, the non- AF interval may be required to be classified as non-AF based on at least the representative maximum A4 window amplitudes being greater than the current AF threshold.

[0187] FIG. 16 is a flow chart 950 of a method for establishing AF interval criteria thresholds by IMD 14 according to some examples. In some examples, control circuit 206 may receive an AF detection sensitivity level as a programmable value that may be entered by a clinician or other user into external device 50 (see FIGs. 1 and 3). Control circuit 206 may receive the AF detection sensitivity level from external device 50 via telemetry circuit 208 and store the AF detection sensitivity in memory' 210. The AF detection sensitivity level may be a value on a scale, e.g., one to five, one to eight or one to ten, that control circuit 206 may use to establish the AF interv al criteria threshold(s). In this way, a user may program IMD 14 to be relatively more sensitive or less sensitive to detecting AF. In an illustrative example, the AF detection sensitivity level may be programmed by a clinician to be a least sensitive level or a most sensitive level or to one or more intermediate levels between the least and most sensitive levels. Depending on the history, frequency and / or severity of symptoms associated with AF episodes in a given patient, the clinician may program the AF detection sensitivity level to be relatively more or less sensitive, tailored to individual patient need.

[0188] At block 952, control circuit 206 may confirm that the current cardiac rhythm is a non- AF rhythm or more generally a period of normal (n on-compromised) atrial mechanical function. The process of flow chart 950 may be performed, for example, after implantation of IMD 14, e.g., during the first 24 hours, when the patient is not expected to be experiencing AF or compromised atrial mechanical function. In some examples, control circuit 206 may receive a communication signal via telemetry circuit 208 from external device 50 (see FIGs. 1 and 3) that initiates the process of flow chart 950 and is a confirmation of normal atrial mechanical function at the time that the clinician or other user initiates the transmission of the communication signal. In other examples, control circuit 206 may confirm non-AF rhythm at block 952 based on classification of a threshold number of AF analysis intervals as being non-AF intervals due to representative maximum A4 window amplitudes being greater than a default or starting AF threshold. For instance, a starting AF threshold may be programmed into memory 210 that is based on data from a population of patients or programmed by a clinician. A non-AF rhythm may be confirmed based on an initial, default AF threshold.

[0189] At block 954, control circuit 206 may start an analysis interval. Control circuit 206 may accumulate amplitude data from the motion sensor signal that is used for classifying AF analysis intervals. For instance, the maximum amplitude during all A4 windows during the analysis interval may be determined and stored in memory 210. Optionally, the maximum amplitude during all A3 windows during the analysis interval for cardiac cycles that include an A4 window may be determined and stored in memory 210. In some examples, the A4 window maximum amplitude times may be determined and stored in memory 210. The analysis interval started at block 954 for accumulating amplitude data for establishing the AF interval criteria threshold(s) may be the same duration as the analysis intervals used by control circuit 206 when monitoring for AF episodes, but may be longer or shorter in some examples, e.g., several minutes or several hours. In still other examples, multiple analysis intervals may be started during a confirmed non- AF rhythm, e.g., in the first 24 hours after implanting IMD 14, for establishing the AF interval criteria threshold(s).

[0190] At block 958, control circuit 206 may establish the AF interval criteria threshold(s) based on the accumulated amplitude data. Control circuit 206 may determine a representative non-AF maximum A4 window amplitude from the accumulated amplitude data. The representative non- AF maximum A4 window amplitude determined from the motion sensor signal sensed during A4 windows during the confirmed non-AF rhythm corresponds to the approximate maximum A4 window amplitude that can be expected during non-AF intervals. Control circuit 206 may establish an AF threshold that can be applied to representative maximum A4 window amplitudesdetermined from subsequent AF analysis intervals by determining a percentage of the representative non-AF maximum A4 window amplitude. The percentage may be selected based on the programmed AF detection sensitivity level. A relatively lower percentage of the representative non-AF maximum A4 window amplitude may be selected for a relatively lower AF detection sensitivity level. A relatively higher percentage of the representative non-AF maximum A4 window amplitude may be determined as the AF threshold when the programmed AF detection sensitivity level is relatively high. In this way, when the AF threshold is set to a relatively high percentage of the representative non-AF maximum A4 window amplitude, IMD 14 will be more likely to detect AF (higher AF detection sensitivity) than when the AF threshold is set to a relatively low percentage of the representative non-AF maximum A4 window amplitude (lower AF detection sensitivity). The AF threshold may be scaled to the representative non-AF maximum A4 window7amplitude according to the programmed AF detection sensitivity level.

[0191] Any other thresholds included in the AF interval criteria may be similarly determined from the amplitude data accumulated during the confirmed non-AF rhythm, e.g., as a percentage of representative amplitude data w here the percentage may be scaled according to the programmed AF detection sensitivity level. For instance, if an A3 / A4 ratio is determined and used by control circuit 206 for classifying an AF analysis interval as being an AF or non-AF interval, as described above in conjunction with FIG. 15, control circuit 206 may determine a ratio threshold as a percentage of the A3 / A4 ratio determined from amplitude data (e g., A3 window maximum amplitudes and A4 window maximum amplitudes of cardiac cycles that include an A4 window) determined from the motion signal sensed during the confirmed non-AF rhythm. The percentage may be scaled according to the programmed AF detection sensitivity level. For instance, a relatively higher percentage of the representative non-AF A3 / A4 ratio may be determined when the AF detection sensitivity7level is programmed to a relatively low level and a relatively low er percentage of the representative non-AF A3 / A4 ratio may be determined when the AF detection sensitivity level is programmed to a relatively high level.

[0192] Additionally or alternatively, the percentile of the A4 window maximum amplitudes determined as the representative A4 window7maximum amplitude for an AF analysis interval may be scaled according to the programmed AF detection sensitivity level. Thus instead of or in addition to establishing a threshold amplitude applied to the representative A4 window maximum amplitude, the percentile used to determine the representative A4 window maximum amplitude may be scaled according to the programmed AF detection sensitivity7level. For example, w ith reference to FIGs. 9 and 10, the percentile of the A4 window maximumamplitudes accumulated over the AF analysis interval may be a lower percentile when the programmed AF sensitivity level is relatively high; a lower percentile of the A4 window maximum amplitudes will be more likely to be less than a given AF threshold making AF episode detection more likely. When the AF detection sensitivity is relatively low, a higher percentile of the A4 window maximum amplitudes may be determined as the representative A4 window maximum amplitude, making an AF analysis interval less likely to be detected for a given AF threshold.

[0193] As such, control circuit 206 may establish AF interval criteria at block 958 by determining one or more thresholds from amplitude data accumulated during a confirmed non- AF rhythm in a given patient. The one or more thresholds may include one or more of an AF threshold applied to a representative maximum A4 window amplitude, a second amplitude threshold applied to a representative maximum A3 window amplitude, and / or a ratio threshold applied to a representative A3 / A4 ratio. The threshold(s) detemiined from the non-AF amplitude data may be established as specified percentages of or offsets from non-AF representative values of the amplitude data. The specified percentages or offsets may be scaled according to the programmed AF detection sensitivity level and may be stored in memoy 210. Additionally or alternatively control circuit 206 may establish AF interval criteria at block 958 by selecting a percentile used for determining representative values of the amplitude data (e.g., A4 window maximum amplitudes. A3 window maximum amplitudes, and / or A3 / A4 ratios). The percentile may be scaled according to the received AF detection sensitivity level. However, in some examples, the percentile (or other method) for determining representative values of amplitude data accumulated during AF analysis intervals may be fixed or programmable, e.g., stored in memory 210. and only thresholds applied to the representative values may be determined at block 958, e.g., based on non-AF amplitude data and optionally scaled to a programmed AF detection sensitivity level.

[0194] Further disclosed herein is the subject matter of the following examples:

[0195] Example 1. A medical device including a motion sensor configured to sense a motion signal and a control circuit configured to, from a plurality of cardiac cycles, determine maximum amplitudes of the motion signal. The control circuit may determine a representative value from the maximum amplitudes, compare the representative value to threshold criteria, detect an episode of compromised atrial mechanical function in response to at least the representative value meeting the threshold criteria and generate an output in response to detecting the episode. The medical device may include a memory configured to store the output generated by the control circuit.

[0196] Example 2. The medical device of example 1 wherein the control circuit is further configured to start an analysis interval comprising the plurality of cardiac cycles and determine the maximum amplitudes of the motion signal from the plurality7of cardiac cycles of the analysis interval.

[0197] Example 3. The medical device of any one of examples 1 — 2 wherein the control circuit is further configured to determine each of the maximum amplitudes of the motion signal by: during a time window of a cardiac cycle of the plurality of cardiac cycles, applying a first atrial event sensing threshold to the motion signal; determining that the motion signal does not cross the first atrial event sensing threshold during the time window, and determining the maximum amplitude of the motion signal sensed by the motion sensor after the time window.

[0198] Example 4. The medical device of example 3 wherein the control circuit is further configured to label the cardiac cycle as a possible atrial tachyarrhythmia cycle in response to at least determining that the motion signal does not cross the first atrial event sensing threshold during the time window and store the label in the memory.

[0199] Example 5. The medical device of example 4 wherein the control circuit is further configured to start an analysis interval that comprises the plurality of cardiac cycles and determine the representative value of maximum amplitudes only from cardiac cycles of the analysis interval that are labeled possible atrial tachyarrhythmia cycles.

[0200] Example 6. The medical device of any one of examples 3 — 5 wherein the control circuit is further configured to determine if the cardiac cycle ends less than a threshold time interval after the time window and label the cardiac cycle as one of: an indeterminate cycle when the cardiac cycle ends less than the threshold time interval after the time window; or a possible atrial tachyarrhythmia cycle when the cardiac cycle does not end less than the threshold time interval after the time window.

[0201] Example 7. The medical device of any one of examples 3 — 6 wherein the control circuit is further configured to determine if the maximum amplitude of the motion signal after the time window of the cardiac cycle is at least a threshold time interval earlier than an ending time of the cardiac cycle and discard the maximum amplitude of the motion signal if the maximum amplitude is not at least the threshold time interval earlier than the ending time of the cardiac cycle.

[0202] Example 8. The medical device of any one of examples 4 — 7 wherein the control circuit is further configured to determine a percentage of cardiac cycles labeled as possible atrial tachyarrhythmia cycles, determine that the percentage of cardiac cycles labeled as possible atrialtachyarrhythmia cycles meets a threshold percentage, and determine the representative value of the maximum amplitudes in response to the percentage meeting the threshold percentage.

[0203] Example 9. The medical device of any one of examples 1 — 8 wherein the control circuit is further configured to, during a time window of a cardiac cycle of the plurality of cardiac cycles, apply a first atrial event sensing threshold to the motion signal and determine that the motion signal crosses the first atrial event sensing threshold during the time window. The control circuit may label the cardiac cycle as a not atrial tachyarrhythmia cardiac cycle without determining a maximum amplitude of the motion signal for the not atrial tachyarrhythmia cardiac cycle.

[0204] Example 10. The medical device of any one of examples 1 — 9 wherein the control circuit is further configured to, during a time window of a cardiac cycle of the plurality7of cardiac cycles, apply a first atrial event sensing threshold to the motion signal and determine that the cardiac cycle ends before the time window expires. The control circuit may label the cardiac cycle as an indeterminate cardiac cycle without determining a maximum amplitude of the motion signal for the indeterminate cardiac cycle.

[0205] Example 11. The medical device of any one of examples 1 — 10 further including a sensing circuit configured to sense a cardiac signal and sense a ventricular event signal from the sensed cardiac signal. The control circuit may be further configured to, during a time window of a cardiac cycle of the plurality of cardiac cycles, apply a first atrial event sensing threshold to the motion signal and determine that the ventricular event signal is sensed by the sensing circuit during the time window. The control circuit may label the cardiac cycle as an indeterminate cardiac cycle without determining a maximum amplitude of the motion signal for the indeterminate cardiac cycle.

[0206] Example 12. The medical device of any one of examples 1 — 11 further including a pulse generator configured to deliver pacing pulses. The control circuit may be further configured to determine a patient physical activity metric from the motion signal, determine that the patient physical activity metric is greater than anon-resting threshold level, control the pulse generator to deliver pacing pulses at a rate response rate according to the patient physical activity' metric, and pause determining the maximum amplitudes from the motion signal in response to at least one pacing pulse being delivered by the pulse generator at the rate response rate.

[0207] Example 13. The medical device of any one of examples 1 — 12 wherein the control circuit is further configured to determine a patient physical activity metric from the motion signal, determine that the patient physical activity metric is greater than a first threshold leveland adjust the threshold criteria in response to the patient physical activity metric being greater than the first threshold level.

[0208] Example 14. The medical device of any one of examples 1 — 13 wherein the control circuit is further configured to determine a patient physical activity metric from the motion signal, determine that the patient physical activity metric is greater than a second threshold level and pause determining the maximum amplitudes in response to the patient physical activity metric being greater than the second threshold level.

[0209] Example 15. The medical device of any one of examples 1 — 14 wherein the control circuit is further configured to determine the representative value from the maximum amplitudes by determining at least one of a measure of center of the maximum amplitudes or a percentile of the maximum amplitudes.

[0210] Example 16. The medical device of any one of examples 1 — 15 wherein the control circuit is further configured to determine the representative value from the maximum amplitudes by determining a metric of a spread of at least one of the maximum amplitudes or times of the maximum amplitudes.

[0211] Example 17. The medical device of any one of examples 1 — 16 wherein the control circuit is further configured to determine a second representative value from maximum amplitudes of the motion signal sensed after detecting the compromised atrial mechanical function episode and detect termination of the compromised atrial mechanical function episode based on at least the second representative value.

[0212] Example 18. The medical device of any one of examples 1 — 17 wherein the control circuit is further configured to generate the output by determining at least one of: a duration of the episode; an atrial tachyarrhythmia burden; a ventricular rate; a percentage of ventricular pacing cardiac cycles; or a ventricular interval variability.

[0213] Example 19. The medical device of any one of examples 1 — 18 wherein the control circuit is further configured to sense atrial systolic event signals from the motion signal and adjust a sensing control parameter used to sense the atrial systolic event signals from the motion signal after detecting the compromised atrial mechanical function episode.

[0214] Example 20. The medical device of any one of examples 1 — 19 further comprising a pulse generator configured to deliver ventricular pacing pulses. The control circuit may be further configured to sense atrial systolic event signals from the motion signal, schedule ventricular pacing pulses to be delivered by the pulse generator at atrioventricular intervals from the sensed atrial systolic event signals and adjust a pacing control parameter used to control thepulse generator in delivering ventricular pacing pulses in response to detecting the compromised atrial mechanical function episode.

[0215] Example 21. The medical device of any one of examples 1 — 20 further comprising a telemetry circuit configured to transmit the output. The medical device may be included in a medical device system comprising a display unit configured to receive the output transmitted by the telemetry circuit and generate a display of the output.

[0216] Example 22. The medical device of any one of examples 1 — 21 wherein the control circuit is further configured to detect the episode of compromised atrial mechanical function as an episode of atrial tachyarrhythmia.

[0217] Example 23. The medical device of any one of examples 1 — 22 further comprising a pulse generator configured to deliver ventricular pacing pulses. The control circuit may be further configured to determine the maximum amplitudes of the motion signal from the plurality of cardiac cycles that include one or more ventricular pacing pulses delivered by the pulse generator.

[0218] Example 24. The medical device of any one of examples 2 — 23 wherein the control circuit is further configured to, for each of the cardiac cycles of the analysis interval that the motion signal does not cross the first atrial event sensing threshold, determine a second maximum amplitude from the motion signal sensed during the time window. The control circuit may be further configured to detect the episode of compromised atrial mechanical function based on the second maximum amplitudes and the representative value meeting the threshold criteria. The control circuit may be further configured to determine at least one of a second representative value of the second maximum amplitudes or a representative ratio of the second maximum amplitudes to the maximum amplitudes determined from the motion signal sensed after the time window. The control circuit may compare at least one of the second representative value or the representative ratio to a respective second amplitude threshold or ratio threshold. The control circuit may detect the episode of compromised atrial mechanical function in response to at least the representative value being less than a first threshold of the threshold criteria and at least one of the second representative value or the representative ratio being greater than a respective second amplitude threshold or ratio threshold.

[0219] Example 25. The medical device of any one of claims 1 — 24 wherein the control circuit may be further configured to identify a non-compromised atrial mechanical function interval, determine maximum amplitude data from the motion signal sensed during the non-compromised atrial mechanical function interval, and establish the threshold criteria from the maximum amplitude data sensed during the non-compromised atrial mechanical function interval.

[0220] Example 26. The medical device of any one of examples 1-25 wherein the control circuit is further configured to receive a detection sensitivity level that may be stored in a memory of the medical device. The control circuit may be configured to identify a non-compromised atrial mechanical function time interval, determine maximum amplitude data from the motion signal sensed during the non-compromised atrial function time interval and establish the threshold criteria by determining from the maximum amplitude data at least a first threshold that is scaled according to the detection sensitivity level. The control circuit may be configured to determine the representative value from the maximum amplitudes by determining a percentile of the maximum amplitudes. The control circuit may be configured to scale the percentile of the determined maximum amplitudes according to the detection sensitivity level.

[0221] Example 27. A method including sensing a motion signal, determining maximum amplitudes of the motion signal sensed during a plurality of cardiac cycles, determining a representative value from the maximum amplitudes and comparing the representative value to threshold criteria. The method may include detecting an episode of compromised atrial mechanical function in response to at least the representative value meeting the threshold criteria and generating an output in response to detecting the episode.

[0222] Example 28. The method of example 27 further including starting an analy sis interval including the plurality of cardiac cycles and determining the maximum amplitudes of the motion signal from the plurality of cardiac cycles of the analysis interval.

[0223] Example 29. The method of any one of examples 27 — 28 wherein determining each of the maximum amplitudes of the motion signal includes: during a time window of a cardiac cycle of the plurality of cardiac cycles, applying a first atrial event sensing threshold to the motion signal, determining that the motion signal does not cross the first atrial event sensing threshold during the time window, and determining the maximum amplitude of the motion signal sensed by the motion sensor after the time window.

[0224] Example 30. The method of example 29 further including labeling the cardiac cycle as a possible atrial tachyarrhythmia cycle in response to at least determining that the motion signal does not cross the first atrial event sensing threshold during the time window and storing the label in a memory.

[0225] Example 31. The method of example 30 further including starting an analysis interval that includes the plurality of cardiac cycles and determining the representative value of maximum amplitudes only from cardiac cycles of the analysis interval that are labeled possible atrial tachyarrhythmia cycles.

[0226] Example 32. The method of any one of examples 29 — 31 further including determining if the cardiac cycle ends less than a threshold time interval after the time window and labeling the cardiac cycle as one of: an indeterminate cycle when the cardiac cycle ends less than the threshold time interval after the time window; or a possible atrial tachyarrhythmia cycle when the cardiac cycle does not end less than the threshold time interval after the time window.

[0227] Example 33. The method of any one of examples 29 — 32 further including determining if the maximum amplitude of the motion signal after the time window of the cardiac cycle is at least a threshold time interval earlier than an ending time of the cardiac cycle and discarding the maximum amplitude of the motion signal if the maximum amplitude is not at least the threshold time interval earlier than the ending time of the cardiac cycle.

[0228] Example 34. The method of any one of examples 30 — 33 further including determining a percentage of cardiac cycles labeled as possible atrial tachyarrhythmia cycles, determining that the percentage of cardiac cycles labeled as possible atrial tachyarrhythmia cycles meets a threshold percentage and determining the representative value of the maximum amplitudes in response to the percentage meeting the threshold percentage.

[0229] Example 35. The method of any one of examples 27 — 34 further including, during a time window of a cardiac cycle of the plurality' of cardiac cycles, applying a first atrial event sensing threshold to the motion signal and determining that the motion signal crosses the first atrial event sensing threshold during the time window. The method may include labeling the cardiac cycle as a not atrial tachyarrhythmia cardiac cycle without determining a maximum amplitude of the motion signal for the not atrial tachyarrhythmia cardiac cycle.

[0230] Example 36. The method of any one of examples 27 — 35 further including, during a time window of a cardiac cycle of the plurality of cardiac cycles, applying a first atrial event sensing threshold to the motion signal and determining that the cardiac cycle ends before the time window expires. The method may further include labeling the cardiac cycle as an indeterminate cardiac cycle without determining a maximum amplitude of the motion signal for the indeterminate cardiac cycle.

[0231] Example 37. The method of any one of examples 28 — 36 further including sensing a cardiac signal, sensing a ventricular event signal from the sensed cardiac signal and, during a time window of a cardiac cycle of the plurality' of cardiac cycles, applying a first atrial event sensing threshold to the motion signal. The method may include determining that the ventricular event signal is sensed during the time window and labeling the cardiac cycle as an indeterminate cardiac cycle yvithout determining a maximum amplitude of the motion signal for the indeterminate cardiac cycle.

[0232] Example 38. The method of any one of examples 27 — 37 further including determining a patient physical activity metric from the motion signal, determining that the patient physical activity metric is greater than a non-resting threshold level, delivering pacing pulses at a rate response rate according to the patient physical activity' metric and pausing determining the maximum amplitudes from the motion signal in response to at least one pacing pulse being delivered at the rate response rate.

[0233] Example 39. The method of any one of examples 27 — 38 further including determining a patient physical activity' metric from the motion signal, determining that the patient physical activity metric is greater than a first threshold level and adjusting the threshold criteria in response to the patient physical activity metric being greater than the first threshold level.

[0234] Example 40. The method of any one of examples 27 — 39 further including determining a patient physical activity metric from the motion signal, determining that the patient physical activity metric is greater than a second threshold level, and pausing determining the maximum amplitudes in response to the patient physical activity metric being greater than the second threshold level.

[0235] Example 41. The method of any one of examples 27 — 40 further including determining the representative value from the maximum amplitudes by determining at least one of a measure of center of the maximum amplitudes or a percentile of the maximum amplitudes.

[0236] Example 42. The method of any one of examples 27 — 41 further including determining the representative value from the maximum amplitudes by determining a metric of a spread of at least one of the maximum amplitudes or times of the maximum amplitudes.

[0237] Example 43. The method of any one of examples 27 — 42 further including determining a second representative value from maximum amplitudes of the motion signal sensed after detecting the compromised atrial mechanical function episode and detecting termination of the compromised atrial mechanical function episode based on at least the second representative value.

[0238] Example 44. The method of any one of examples 27 — 43 wherein generating the output includes determining at least one of a duration of the episode; an atrial tachyarrhythmia burden; a ventricular rate; a percentage of ventricular pacing cardiac cycles; or a ventricular interval variability'.

[0239] Example 45. The method of any one of examples 27 — 44 further including sensing atrial systolic event signals from the motion signal and adjusting a sensing control parameter used to sense the atrial systolic event signals from the motion signal after detecting the compromised atrial mechanical function episode.

[0240] Example 46. The method of any one of examples 27 — 45 further including sensing atrial systolic event signals from the motion signal, scheduling ventricular pacing pulses at atrioventricular intervals from the sensed atrial systolic event signals and adjusting a pacing control parameter used to control delivering ventricular pacing pulses in response to detecting the compromised atrial mechanical function episode.

[0241] Example 47. The method of any one of examples 27 — 46 further including transmitting the output. The method may include generating a display of the output.

[0242] Example 48. The method of any one of examples 27 — 47 further including detecting the episode of compromised atrial mechanical function as an episode of atrial tachyarrhythmia.

[0243] Example 49. The method of any one of examples 27 — 48 further including delivering ventricular pacing pulses and determining the maximum amplitudes of the motion signal from the plurality7of cardiac cycles that include one or more delivered ventricular pacing pulses.

[0244] Example 50. The method of any one of examples 28 — 49 further comprising, for each of the cardiac cycles of the analysis interval that the motion signal does not cross the first atrial event sensing threshold, determine a second maximum amplitude from the motion signal sensed during the time window. The method may further include detecting the episode of compromised atrial mechanical function based on the second maximum amplitudes and the representative value meeting the threshold criteria. The method may include determining at least one of a second representative value of the second maximum amplitudes or a representative ratio of the second maximum amplitudes to the maximum amplitudes determined from the motion signal sensed after the time window. The method may include comparing at least one of the second representative value or the representative ratio to a respective second amplitude threshold or ratio threshold. The method may include detecting the episode of compromised atrial mechanical function in response to at least the representative value being less than a first threshold of the threshold criteria and at least one of the second representative value or the representative ratio being greater than a respective second amplitude threshold or ratio threshold.

[0245] Example 51. The method of any one of claims 27 — 50 further comprising identifying a non-compromised atrial mechanical function interval, determining maximum amplitude data from the motion signal sensed during the non-compromised atrial mechanical function interval, and establishing the threshold criteria from the maximum amplitude data sensed during the noncompromised atrial mechanical function interval.

[0246] Example 52. The method of any one of examples 27-51 further comprising storing a detection sensitivity level in a memory. The method may include identifying a non-compromised atrial mechanical function time interval, determining maximum amplitude data from the motionsignal sensed during the non-compromised atrial function time interval and establishing the threshold criteria by determining from the maximum amplitude data at least a first threshold that is scaled according to the detection sensitivity level. The method may include determining the representative value from the maximum amplitudes by determining a percentile of the maximum amplitudes. The method may include scaling the percentile of the determined maximum amplitudes according to the detection sensitivity- level.

[0247] Example 53. A non-transitory, computer-readable storage medium comprising a set of instructions which, when executed by a control circuit of a medical, cause the medical device to sense a motion signal, determine maximum amplitudes of the motion signal sensed during a plurality of cardiac cycles, determine a representative value from the maximum amplitudes, compare the representative value to threshold criteria and detect an episode of compromised atrial mechanical function in response to at least the representative value meeting the threshold criteria, and generate an output in response to detecting the episode.

[0248] Example 54. In any one of examples 1 — 53, wherein the threshold criteria include one or more threshold criteria.

[0249] Example 55. In example 54, wherein the episode of compromised atrial mechanical function is detected in response to at least the first representative value meeting at least one of the threshold criteria.

[0250] It should be understood that, depending on the example, certain acts or events of any of the methods described herein can be performed in a different sequence, may be added, merged, or left out altogether (e.g., not all described acts or events are necessary- for the practice of the method). Moreover, in certain examples, acts or events may be performed concurrently, e.g., through multi-threaded processing, interrupt processing, or multiple processors, rather than sequentially. In addition, while certain aspects of this disclosure are described as being performed by a single circuit for purposes of clarity, it should be understood that the techniques of this disclosure may be performed by a combination of circuits or components associated with, for example, a medical device.

[0251] In one or more examples, the functions described may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, the functions maybe stored as one or more instructions or code on a computer-readable medium and executed by a hardware-based processing unit. Computer-readable media may include computer-readable storage media, which corresponds to a tangible medium such as data storage media (e.g.. RAM, ROM, EEPROM, flash memory, or any other medium that can be used to store desired program code in the form of instructions or data structures and that can be accessed by a computer).

[0252] Instructions may be executed by one or more processors, such as one or more digital signal processors (DSPs), general purpose microprocessors, application specific integrated circuits (ASICs), field programmable logic arrays (FPGAs), or other equivalent integrated or discrete logic circuitry. Accordingly, the term “processor,"’ as used herein may refer to any of the foregoing structure or any other structure suitable for implementation of the techniques described herein. Also, the techniques could be fully implemented in one or more circuits or logic elements.

[0253] Thus, a medical device has been presented in the foregoing description with reference to specific examples. It is to be understood that various aspects disclosed herein may be combined in different combinations than the specific combinations presented in the accompanying drawings. It is appreciated that various modifications to the referenced examples may be made without departing from the scope of the disclosure and the following claims.

Claims

WHAT IS CLAIMED IS:1 . A medical device system comprising: a motion sensor configured to sense a motion signal; a control circuit configured to: from a plurality of cardiac cycles, determine first maximum amplitudes of the motion signal; determine a first representative value from the first maximum amplitudes; compare the first representative value to one or more threshold criteria; detect an episode of compromised atrial mechanical function in response to at least the first representative value meeting the one or more threshold criteria; and generate an output in response to detecting the episode; and a memory configured to store the output generated by the control circuit.

2. The medical device system of claim 1 wherein the control circuit is further configured to determine each of the first maximum amplitudes of the motion signal by: during a time window of a cardiac cycle of the plurality of cardiac cycles, applying a first atrial event sensing threshold to the motion signal; determining that the motion signal does not cross the first atrial event sensing threshold during the time window; and determining the first maximum amplitude of the motion signal sensed by the motion sensor after the time window.

3. The medical device system of claim 2 wherein the control circuit is further configured to: determine if a cardiac cycle of the plurality of cardiac cycles ends less than a threshold time interval after the time window; and label the cardiac cycle as a possible atrial tachyarrhythmia cycle when the cardiac cycle does not end less than the threshold time interval after the time window.

4. The medical device system of any one of claims 1 — 3 wherein the control circuit is further configured to: determine if the first maximum amplitude of the motion signal after the time window of the cardiac cycle is at least a threshold time interval earlier than an ending time of the cardiac cycle; anddiscard the first maximum amplitude of the motion signal if the first maximum amplitude is not at least the threshold time interval earlier than the ending time of the cardiac cycle.

5. The medical device system of any one of claims 3 — 4 wherein the control circuit is further configured to: determine a percentage of cardiac cycles labeled as possible atrial tachyarrhythmia cycles; determine that the percentage of cardiac cycles labeled as possible atrial tachyarrhythmia cycles meets a threshold percentage; and determine the representative value of the first maximum amplitudes in response to the percentage meeting the threshold percentage.

6. The medical device system of any one of claims 1 — 5 wherein the control circuit is further configured to: during a time window of a cardiac cycle of the plurality of cardiac cycles, apply a first atrial event sensing threshold to the motion signal; determine that the cardiac cycle ends before the time window expires without determining a first maximum amplitude of the motion signal for the cardiac cycle.

7. The medical device system of any one of claims 1 — 6 further comprising a pulse generator configured to deliver pacing pulses; and wherein the control circuit is further configured to: determine a patient physical activity metric from the motion signal; determine that the patient physical activity metric is greater than a non-resting threshold level; control the pulse generator to deliver pacing pulses at a rate response rate according to the patient physical activity metric; and pause determining the first maximum amplitudes from the motion signal in response to at least one pacing pulse being delivered by the pulse generator at the rate response rate.

8. The medical device system of any one of claims 1 — 7 wherein the control circuit is further configured to: determine a patient physical activity metric from the motion signal;determine that the patient physical activity metric is greater than a first threshold level; and adjust the one or more threshold criteria in response to the patient physical activity metric being greater than the first threshold level.

9. The medical device system of any one of claims 1 — 8 wherein the control circuit is further configured to: determine a patient physical activity7metric from the motion signal; determine that the patient physical activity metric is greater than a second threshold level; and pause determining the first maximum amplitudes in response to the patient physical activity metric being greater than the second threshold level.

10. The medical device system of any one of claims 1 — 9 wherein the control circuit is further configured to determine the first representative value from the first maximum amplitudes by determining at least one of: a measure of center of the first maximum amplitudes; a percentile of the first maximum amplitudes; a metric of a spread of the first maximum amplitudes; or a metric of times of the first maximum amplitudes.

11. The medical device system of any one of claims 1 — 10 wherein the control circuit is further configured to generate the output by determining at least one of: a duration of the episode; an atrial tachyarrhythmia burden; a ventricular rate; a percentage of ventricular pacing cardiac cycles; or a ventricular interval variability.

12. The medical device system of any one of claims 1 — 11 further comprising: a pulse generator configured to deliver ventricular pacing pulses; and wherein the control circuit is further configured to: sense atrial systolic event signals from the motion signal; andschedule ventricular pacing pulses to be delivered by the pulse generator at atrioventricular intervals from the sensed atrial systolic event signals; determine the first maximum amplitudes of the motion signal from the plurality of cardiac cycles that include one or more ventricular pacing pulses delivered by the pulse generator; after detecting the compromised atrial mechanical function episode, adjust at least one of: a sensing control parameter used to sense the atrial systolic event signals from the motion signal; or a pacing control parameter used to control the pulse generator in delivering ventricular pacing pulses.

13. The medical device system of any one of claims 1 — 12 further comprising: a telemetry circuit configured to transmit the output; and a display unit configured to generate a display of the output.

14. The medical device system of any one of claims 2-13 wherein the control circuit is further configured to: for each of the cardiac cycles of the plurality of cardiac cycles that the motion signal does not cross the first atrial event sensing threshold during the time window, determine a second maximum amplitude of the motion signal sensed during the time window; and detect the episode of compromised atrial mechanical function in response to at least the first representative value and the second maximum amplitudes meeting the one or more threshold criteria.

15. The medical device system of any one of claims 1-14 wherein: the memoy is further configured to store a detection sensitivity level; and the control circuit is further configured to: identify a non-compromised atrial mechanical function time interval; determine maximum amplitude data from the motion signal sensed during the non-compromised atrial function time interval; establish the one or more threshold criteria by determining from the maximum amplitude data at least one threshold that is scaled according to the detection sensitivity level.

Images