System for monitoring physiological parameters

Abstract

An example system includes memory configured to store a machine learning (ML) model; processing circuitry coupled to the memory, the processing circuitry being configured to: receive a first optical signal sensed during a first period of time; receive a second optical signal via a first sensing device positioned at a location of a patient, the second optical signal being sensed by the first sensing device during the first period of time, the second optical signal being a lower-fidelity optical signal than the first optical signal; and apply the first optical signal and the second optical signal to the ML model to train the ML model to generate at least one of: a reconstructed third optical signal based on a third optical signal sensed via a second sensing device positioned at a location having a pulsatility below a pulsatility threshold, or a blood pressure value based on the third optical signal.

Description

Atty Ref. No. A0012750W001SYSTEM FOR MONITORING PHYSIOLOGICAL PARAMETERS

[0001] This application claims the benefit of U.S. Provisional Patent Application Serial No. 63 / 733,751, filed December 13, 2024, the entire content of which is incorporated herein by reference.TECHNICAL FIELD

[0002] The disclosure relates generally to medical systems and, more particularly, medical systems configured to monitor patient activity for changes in patient health.BACKGROUND

[0003] Some types of medical systems may monitor one or more physiological parameters of a patient. The medical system may detect signals associated with the one or more physiological parameters and may determine values for the physiological parameters based on the detected signals. The determined values may be used to detect changes in medical conditions, to evaluate efficacy of a medical therapy, or to generally evaluate patient health. In some examples, the medical system may include one or more of an implantable medical device or a wearable device to collect the data.SUMMARY

[0004] A medical system may sense one or more physiological parameters of a patient to generally evaluate health of the patient. The medical system may sense the one or more physiological parameters via one or more medical devices, e.g., one or more implantable medical devices (IMDs) and / or wearable devices. The medical system may sense signals from the patient via optical sensor(s) coupled to the one or more medical devices and determine values for physiological parameters based at least in part on the sensed optical signals. Physiological parameters may include, but are not limited to, blood pressure levels, blood oxygen saturation levels, blood glucose levels, cardiac signals (e.g., electrocardiogram (EGM) signals), or the like.

[0005] Monitoring blood pressure via an optical signal sensed by an optical sensor, such as a photoplethysmography (PPG) sensor, may require an ability to detect and resolve features of sensed optical signal, such as a dicrotic notch. Features, such as aAtty Ref. No. A0012750W001 dicrotic notch, of a sensed optical signal, may be more accurately identified in a high- fidelity optical signal that is sensed in high-fidelity pulsatility areas, such as where pulsatility is between about 1% and 5%. Detecting and identifying features, such as a dicrotic notch, accurately from optical signals sensed via an optical sensor of an IMD that is implanted in a low-pulsatility location, such as the chest where pulsatility is between about 0.1% and 0.5%, may be more difficult due to a dicrotic notch in a low-fidelity optical signal being less identifiable and / or due motion artifact, noise, and / or distortion of a low-fidelity optical signal sensed by an IMD positioned at a low-pulsatility location.

[0006] This disclosure describes devices, systems, and methods for training a machine learning (ML) model based on low-fidelity optical signals and high-fidelity optical signals sensed during a period of time to produce a trained ML model that is configured to more accurately reconstruct a low-fidelity optical signal sensed via a sensing device positioned at a low-pulsatility location to a reconstructed high-fidelity optical signal. Features, such as the dicrotic notch, may be more accurately identifiable and extractable from the reconstructed high-fidelity optical signal than the low-fidelity optical signal sensed via a sensing device. In some examples, training the ML model on the particular sensed low- fidelity optical signals via a first sensing device, such as an IMD, and high-fidelity optical signals, via the IMD operating in a high-fidelity mode or via a second sensing device, such as an external sensing device positioned over a fingertip of the patient, may provide a ML model that is able to more accurately reconstruct a high-fidelity optical signal from a low- fidelity optical signal that is sensed via a sensing device, such as an IMD, positioned at a low-pulsatility location and / or operating in a low-fidelity mode. In some examples, a particularly trained ML model being able to more accurately reconstruct a high-fidelity optical signal from a low-fidelity optical signal may enable a medical system and / or IMD to more accurately detect blood pressure values of a patient via an optical sensor of an IMD, such as an IMD positioned in a chest of the patient.

[0007] In some examples, this disclosure describes devices, systems, and methods for training a ML model based on low-fidelity optical signals and high-fidelity optical signals sensed during a period of time to produce a trained ML model that is configured to more accurately generate a blood pressure value of the patient based on optical signal sensed via a sensing device, such as an IMD, positioned at a low-pulsatility location. In some examples, a particularly trained ML model being able to more accurately generate a bloodAtty Ref. No. A0012750W001 pressure value of the patient based on a sensed low-fidelity optical signal may enable a medical system and / or IMD to more accurately detect blood pressure values of a patient via an optical sensor of an IMD, such as an IMD positioned in a chest of the patient.

[0008] In one example, this disclosure describes a system comprising: memory configured to store a machine learning (ML) model; processing circuitry coupled to the memory, the processing circuitry being configured to: receive a first optical signal sensed during a first period of time; receive a second optical signal via a first sensing device positioned at a location of a patient, the second optical signal being sensed by the first sensing device during the first period of time, the second optical signal being a lower- fidelity optical signal than the first optical signal; and apply the first optical signal and the second optical signal to the ML model to train the ML model to generate at least one of: a reconstructed third optical signal based on a third optical signal sensed via a second sensing device positioned at a location having a pulsatility below a pulsatility threshold, or a blood pressure value based on the third optical signal.

[0009] In another example, this disclosure describes an implantable medical device (IMD) comprising: an optical sensor configured to sense a first IMD optical signal of a patient during a first period of time and a second IMD optical signal of the patient during a second period of time, the second period of time being after the first period of time; and circuitry configured to: output the first IMD optical signal to a computing device; receive a trained machine learning (ML) model, the trained ML model being trained based on the first IMD optical signal and a third optical signal, the third optical signal sensed during the first period of time to generate at least one of: a reconstructed second IMD optical signal based on the second IMD optical signal sensed via the IMD, the IMD being positioned at a location having a pulsatility below a pulsatility threshold; or a blood pressure value based on the second IMD optical signal; and apply the second IMD optical signal to the trained ML model to generate at least one of the reconstructed second IMD optical signal or the blood pressure value of the patient based on the second IMD optical signal, the reconstructed second IMD optical signal having a higher-fidelity than the second IMD optical signal.

[0010] In another example, this disclosure describes a method for training a machine learning (ML) model, the method comprising: receiving a first optical signal during a first period of time; receiving a second optical signal via a first sensing device positioned at aAtty Ref. No. A0012750W001 location of a patient, the second optical signal being sensed by the first sensing device during the first period of time, the second optical signal being a lower-fidelity optical signal than the first optical signal; and applying the first optical signal and the second optical signal to the ML model to train the ML model to generate at least one of: a reconstructed third optical signal based on a third optical signal sensed via a second sensing device positioned at a location having a pulsatility below a pulsatility threshold, or a blood pressure value based on third optical signal.

[0011] The summary is intended to provide an overview of the subject matter described in this disclosure. It is not intended to provide an exclusive or exhaustive explanation of the systems, device, and methods described in detail within the accompanying drawings and description below. Further details of one or more examples of this disclosure are set forth in the accompanying drawings and in the description below. Other features, objects, and advantages will be apparent from the description and drawings, and from the claims.BRIEF DESCRIPTION OF THE DRAWINGS

[0012] FIG. 1 illustrates an example medical system in conjunction with a patient, in accordance with one or more examples of the present disclosure.

[0013] FIG. 2A is a perspective drawing illustrating an example configuration of an implantable medical device (IMD) of FIG. 1.

[0014] FIG. 2B is a perspective drawing illustrating another example configuration of the IMD of FIG. 1.

[0015] FIG. 3 A is a block diagram illustrating an example configuration of the example IMD of FIG. 1.

[0016] FIG. 3B is a graph drawing illustrating an example low-fidelity optical signal.

[0017] FIG. 3C is a graph drawing illustrating an example high-fidelity optical signal.

[0018] FIG. 3D is a block diagram illustrating an example configuration of the example external sensing device of FIG. 1.

[0019] FIG. 4 is a block diagram illustrating an example configuration of a computing device that operates in accordance with one or more techniques of the present disclosure.Atty Ref. No. A0012750W001

[0020] FIG. 5 is a block diagram illustrating an example configuration of a health monitoring system that operates in accordance with one or more techniques of the present disclosure.

[0021] FIG. 6A is a flowchart illustrating an example process of training a machine learning model to generate at least one of a reconstructed optical signal or a blood pressure value based on at least one sensed optical signal.

[0022] FIG. 6B is a flowchart illustrating an example process of training a machine learning model to generate at least one of a reconstructed optical signal or a blood pressure value based on at least one sensed optical signal.

[0023] FIG. 7 is a conceptual diagram illustrating an example training process for a machine learning model, in accordance with examples of the current disclosure.

[0024] FIG. 8 is a conceptual diagram illustrating an example training process for a machine learning model, in accordance with examples of the current disclosure.

[0025] Like reference characters denote like elements throughout the description and figures.DETAILED DESCRIPTION

[0026] A medical system may be configured to monitor one or more physiological parameters of a patient to monitor a condition of the patient. For example, a medical system may evaluate a blood pressure value of a patient based on optical signals sensed via an IMD.

[0027] In some examples, a system may train a ML model based on low-fidelity optical signals sensed via an IMD positioned at a low-pul satility location and high-fidelity optical signals sensed via an external sensing device positioned at a high-pul satility location sensed during a period of time to produce a trained ML model that is configured to more accurately reconstruct a sensed low-fidelity optical signal via a sensing device positioned at a low-pulsatility location to a reconstructed high-fidelity optical signal may improve the identifiability and extractability of features, such as a dicrotic notch, from the reconstructed high-fidelity optical signal compared to the identifiability and extractability of features, such as a dicrotic notch, from the low-fidelity optical signal sensed via a sensing device.Atty Ref. No. A0012750W001

[0028] In some examples, a system may train a ML model based on low-fidelity optical signals sensed via an IMD positioned at a low-pul satility location and high-fidelity optical signals sensed via the IMD operating in a high-fidelity mode or via an external sensing device positioned at a high-pul satility location sensed during a period of time to produce a trained ML model that is configured to more accurately generate a blood pressure value of the patient based on optical signal sensed via sensing device, such as an IMD, positioned at a low-pulsatility location. In some examples, a particularly trained ML model being able to more accurately generate a blood pressure value of the patient based on a sensed low-fidelity optical signal may enable a medical system and / or IMD to more accurately detect blood pressure values of a patient via an optical sensor of an IMD, such as an IMD positioned in a chest of the patient.

[0029] FIG. 1 illustrates an example medical system 100 (e.g., “system 100”) including an IMD 104 in conjunction with a patient 102, in accordance with one or more examples of the present disclosure. IMD 104 is configured for continuous, long-term monitoring of patient 102. IMD 104 is configured to sense signals corresponding to one or more physiological parameters from patient 102. Continuous monitoring and / or sensing by IMD 104 may include monitoring and / or sensing on a triggered or periodic basis, without requiring user or clinician intervention.

[0030] IMD 104 may include an implantable cardiac monitor (ICM), implantable pacemaker, implantable cardioverter, or other implantable monitoring device. In some examples, IMD 104 includes a Reveal LINQ™ or LINQ II™ ICM, available from Medtronic, Inc., of Minneapolis, Minnesota, which may be inserted subcutaneously. Such IMDs may facilitate relatively longer-term and continuous monitoring of patients during normal daily activities and may periodically transmit collected data to a remote patient monitoring system, such as the Medtronic Carelink™ Network.

[0031] While FIG. 1 illustrates system 100 with IMD 104, other examples of system 100 may include one or more other IMDs and / or one or more external medical devices (e.g., wearable devices) instead of or in addition to IMD 104. The other IMDs and / or external medical devices may be configured to perform the processes attributed to IMD 104 as described herein.

[0032] IMD 104 may include one or more optical sensors configured to sense optical signals from patient 102. In some examples, the one or more optical sensors may include aAtty Ref. No. A0012750W001 photoplethysmography (PPG) sensor. In some examples, IMD 104 may be configured to be positioned at a low pulastility location of patient 102. For example, IMD 104 may be configured to be positioned subcutaneously in a chest of the patient 102. In some examples, IMD 104 being positioned subcutaneously in a chest region of patient 102 may cause optical signals sensed by IMD 104 to be low pulastility signals. In some examples, a low pulsatility signal may be less than 1%. In some examples, a low pulsatility signal may be between 0.1% and 0.5%. In some examples, IMD 104 being positioned subcutaneously in a chest region of patient 102 may cause optical signals sensed by IMD 104 to be a low- fidelity optical signal. In some examples, a low-fidelity optical signal may have one or more of a lower amplitude, absence of a dicrotic notch, motion artifact, noise, or distortion compared to a high-fidelity optical signal. In some examples, IMD 104 may be configured to operate in a high-fidelity mode, such as a temporary high-fidelity mode, to sense optical signals that have a higher fidelity than the optical signals sensed by IMD 104 during a regular mode. In some examples, the IMD 104 operating in a high-fidelity mode may include the IMD 104 operating in a high-power mode may include operating the emitter of an optical sensor at a higher-emitting mode to obtain an optical signal having a higher fidelity. In some examples, IMD 104 may be configured to operate in a high-power mode for brief intervals to conserve energy as operating an emitter of an optical sensor at a higher-emitting mode may use more power of the IMD 104. In some examples IMD 104 may be configured to generate high-fidelity optical signals while operating in a high- fidelity mode and generate low-fidelity optical signals while operating in a normal lower fidelity mode.

[0033] In addition to IMD 104, system 100 may include external sensing device 105, and at least one patient computing device 106. External sensing device 105 may include optical sensors configured to sense physiological optical signals of patient 102 at high pulsatility location of patient 102 to generate high-fidelity optical signals of patient 102. In some examples, external sensing device 105 is positioned at a location having higher pulsatility than a location of IMD 104. In some examples, external sensing device 105 may generate higher-fidelity optical signals than IMD 104. In some examples, external sensing device 105 may be referred to as “first sensing device” in system 100 and IMD 104 may be referred to as the “second sensing device” in system 100.Atty Ref. No. A0012750W001

[0034] In some examples, external sensing device 105 may collect and store physiological parameter values based on the sensed physiological signals. In some examples, external sensing device 105 may be configured to be positioned on a fingertip of patient 102 to sense optical signals. In some examples, external sensing device 105 may be a wearable device, such as a smartwatch, configured to sense optical signals. In some examples, external sensing device 105 may include additional sensors, such as one or more of electrodes or accelerometers.

[0035] Patient computing device 106 is configured for wireless communication with IMD 104. Patient computing device 106 may retrieve physiological parameter data (e.g., physiological signals, physiological parameter values) from IMD 104. Patient computing device 106 may be personal computing devices of patient 102. In some examples, patient computing device 106 is worn by patient 102. Patient computing device 106 may be any computing device configured for wireless communication with IMD 104 and external sensing device 105, such as a desktop computer, a laptop computer, a tablet, a smartwatch, or a smartphone. Patient computing device 106 may communicate with IMD 104 and with external sensing device 105 according to a wireless communication protocol (e.g., according to Bluetooth® or Bluetooth® Low Energy (BLE) protocols).

[0036] Computing system 110 may include processing circuitry 112 and memory 114. Processing circuitry 112 may include fixed function circuitry and / or programmable processing circuitry. Processing circuitry 112 may include any one or more of a microprocessor, a controller, a digital signal processor (DSP), graphics processing unit (GPU), tensor processing unit (TPU), an application specific integrated circuitry (ASIC), a field-programmable gate array (FPGA), or equivalent discrete or analog logic circuitry. In some examples, processing circuitry 112 may include multiple components, such as any combination of one or more microprocessors, one or more controllers, one or more DSPs, GPUs, TPUs, one or more ASIC, or one or more FPGAs, as well as other discrete or integrated logic circuitry, which may be physically located in one or more devices in one or more physical locations. Computing system 110 may be configured as a cloud computing system.

[0037] Processing circuitry 112 may be capable of processing instructions stored in memory 114. In some examples, memory 114 includes a computer-readable medium that includes instructions that, when executed by processing circuitry 112, cause computingAtty Ref. No. A0012750W001 system 110 and processing circuitry 112 to perform various functions attributed to them herein. In some examples, computing system 110 implements a health monitoring system (HMS) 116. As will be described in greater detail below, HMS 116 may monitor reconstructed optical signals and / or blood pressure values generated by ML model 118 and determine one more health conditions statuses based on the reconstructed optical signals and / or blood pressure values generated by ML model 118. Memory 114 may include any volatile, non-volatile, magnetic, optical, or electrical media, such as random-access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), electrically- erasable programmable ROM (EEPROM), ferroelectric RAM (FRAM), dynamic randomaccess memory (DRAM), flash memory, or any other digital media.

[0038] Network 108 may include one or more computing devices, such as one or more non-edge switches, routers, hubs, gateways, security devices such as firewalls, intrusion detection, and / or intrusion prevention devices, servers, cellular base stations and nodes, wireless access points, bridges, cable modems, application accelerators, or other network drives. Network 108 may include one or more networks administered by service providers and may thus form part of a large-scale public network infrastructure, e.g., the Internet. Network 108 may provide computing devices and systems, such as those illustrated in FIG. 1, access to the Internet, and may provide a communication framework that allows the computing devices and systems illustrated in FIG. 1 to communicate with each other but isolates some of the data flows from devices external to the private network for security purposes. In some examples, the communications between the computing devices and systems illustrated in FIG. 1 are encrypted.

[0039] IMD 104 may be configured to transmit physiological parameters values and / or the notifications to wireless access point 124 and / or patient computing device 106. Wireless access point 124 and / or patient computing device 106 may then communicate the retrieved data to computing system 110 via network 108.

[0040] In some examples, computing system 110 may be configured to provide a secured storage site for data that has been collected from IMD 104 and / or external sensing device 105. Computing system 110 may include a database that stores medical-related and health-related data. For example, computing system 110 includes a cloud server or other remote server that stores data collected from IMD 104 and / or external sensing device 105. Computing system 110 may assemble data in webpages or other documents for viewing byAtty Ref. No. A0012750W001 trained clinicians, such as clinicians 130, via clinician computing devices 128. Clinicians 130 may include, but are not limited to, medical care providers and emergency medical services (EMS) providers. One or more aspects of the system 100 illustrated in FIG. 1 may be implemented with general network technology and functionality, which may be similar to that provided by the Medtronic CareLink® Network.

[0041] One or more of clinician computing devices 128 may be a tablet, smartphone, laptop computer, desktop computer, or other computing devices associated with one or more of clinicians 130, by which one or more clinicians 130 may program, receive notifications from, and / or interrogate IMD 104 and / or external sensing device 105. For example, one or more clinicians 130 may access data generated by a ML model (e.g., reconstructed optical signal or blood pressure values) through a clinician computing device 128, such as when clinician computing device 128 receives a notification from IMD 104, patient computing devices 106, and / or computing system 110. Clinician 130 may determine, based on the data accessed through clinician computing device 128, whether patient 102 may require medical aid. The data and notification transmitted to clinician computing device 128 may facilitate the provision of medical aid by one or more clinicians 130 to patient 102.

[0042] Local network 122 may facilitate communication between patient computing devices 106 and other computing devices and systems (e.g., computing system 110, HMS 116) connected to network 108. Local network 122 may be configured with wireless technology, such as IEEE 802.11 wireless networks, IEEE 802.15 ZigBee networks, an ultra-wideband protocol, near-field communications, or the like. Local network 122 may include one or more wireless access points 124 configured to provide support for wireless communications throughout an environment encompassing local network 122. In some examples, patient computing devices 106 may communicate with network 108 (e.g., with HMS 116) via a cellular base station and / or cellular network.

[0043] In some examples, an optical signal sensed by IMD 104 having a relatively lower-fidelity, such as shown in FIG. 3B, may make detecting features in the optical signal, such as a dicrotic notch, more difficult. In accordance with techniques described here, processing circuitry may train a machine learning (ML) model based on low-fidelity optical signals and high-fidelity optical signals sensed during a same period of time to produce a ML model that is configured to more accurately reconstruct a sensed low-Atty Ref. No. A0012750W001 fidelity optical signal to a high-fidelity optical signal which may improve detection of features from the low-fidelity optical signal, such as the dicrotic notch.

[0044] Although the techniques for storing a machine learning (ML) model 118 are described herein as primarily being performed by memory 114 of computing system 110, such techniques may be performed, in whole or part, by memory 422 of patient computing device 106 and / or memory 304 of IMD 104.

[0045] Although the techniques for applying the first optical signal sensed by external sensing device 105 and the second optical signal sensed by IMD 104 to the machine learning model to train the machine learning model to generate at least one of a reconstructed optical signal based on optical signals sensed via a sensing device positioned at a location having a pulsatility below a pulsatility threshold or blood pressure values based on optical signals sensed via a sensing device positioned at a location having a pulsatility below the pulsatility threshold are described herein primarily as being performed by processing circuitry 112 of computing system 110, such techniques may be performed, in whole or part, by processing circuitry of any one or more devices of system 100, such as processing circuitry 420 of patient computing device 106, and / or processing circuitry 302 of IMD 104.

[0046] Memory 114 may be configured to store a ML model 118. Processing circuitry 112 may be configured to receive the first optical signal sensed via external sensing device 105 or sensed via IMD 104 operating in a high-fidelity mode and receive the second optical signal sensed via IMD 104. Processing circuitry 112 may be configured to apply the first optical signal and the second optical signal to the machine learning model to train the machine learning model 118, such as a deep neural network, to generate a reconstructed optical signal based on optical signals sensed via a sensing device positioned at a location having a pulsatility below a pulsatility threshold and / or generate blood pressure values based on optical signals sensed via a sensing device positioned at a location having a pulsatility below the pulsatility threshold. In some examples, processing circuitry 112 may be configured to determine a blood pressure value of the patient based on the reconstructed optical signal. In some examples, a pulsatility threshold is up to 1.0%. In some examples, a pulsatility threshold is up to 0.5%. In some examples, a pulsatility threshold may be based on the pulsatility of the location of IMD 104. In some examples, a pulsatility threshold may be within + / - 10% of the pulsatility of the location ofAtty Ref. No. A0012750W001IMD 104. In some examples, a pulsatility threshold may be within + / - 20% of the pulsatility of the location of IMD 104.

[0047] In some examples, an optical signal sensed by IMD 104 having a low-fidelity may make detecting features in the optical signal, such as a dicrotic notch, more difficult.

[0048] In some examples, the sensing device positioned at a location having a pulsatility below a pulsatility threshold may be IMD 104. In some examples, the sensing device positioned at a location having a pulsatility below a pulsatility threshold may be a separate IMD than IMD 104 and configured to be implanted in patient 102 or configured to be implanted in a person other than patient 102.

[0049] In some examples, processing circuitry 112 may be configured to receive a third optical signal via IMD 104 during a second period of time, the second period of time being after the first period of time. The third optical signal having a similar fidelity as the second optical signal and being a lower-fidelity optical signal than the first optical signal. In some examples, processing circuitry 112 may be configured to apply the third optical signal to the trained ML model 118 to generate at least one of a reconstructed optical signal of the third optical signal or a blood pressure value of the patient based on the third optical signal. In some examples, the reconstructed optical signal of the third optical signal has a higher fidelity than the third optical signal.

[0050] In some examples, in response to training ML model 118, such as described above, computing system 110 may send the trained ML model 118 to one or more of IMD 104 or patient computing device 106. In some examples, in response to receiving the trained ML model 118, one or more of IMD 104 and / or patient computing device 106 may apply a sensed optical signal to the trained ML model 118 to generate at least one of a reconstructed optical signal based on optical signals sensed via a sensing device positioned at a location having a pulsatility below a pulsatility threshold and / or blood pressure values based on optical signals sensed via a sensing device positioned at a location having a pulsatility below the pulsatility threshold.

[0051] FIG. 2A is a perspective drawing illustrating an implantable medical device (IMD) 104 A, which may be an example configuration of IMD 104 of FIG. 1. In the example shown in FIG. 2 A, IMD 104 A may be embodied as a monitoring device having housing 202, proximal electrode 206A and distal electrode 206B. Housing 202 may further comprise a first major surface 204, a second major surface 208, a proximal endAtty Ref. No. A0012750W001210, and a distal end 212. Housing 202 encloses electronic circuitry located inside IMD 104 A and protects the electronic circuitry contained therein from body fluids. Housing 202 may be hermetically sealed and configured for subcutaneous implantation. Electrical feedthroughs provide electrical connection of electrodes 206A and 206B.

[0052] The electronic circuitry may contain components defining an optical sensor 224 of IMD 104 A. Optical sensor 224 may include one or more light emitters 226 and one or more light detectors 228. Optical sensor 224 may be entirely disposed within housing 202 and light emitter(s) 226 may transmit light signals out of housing 202 and light detector(s) 228 may detect light signals external to housing 202 via one or more transparent windows disposed within housing 202. The one or more transparent windows may hermetically seal housing 202 and may define at least a portion of first major surface 204 or second major surface 208. In some examples, IMD 104A may be configured to sense optical signal(s) via optical sensor 224. In some examples, IMD 104 may sense optical signals at a lower fidelity compared to an optical signal measured via external sensing device positioned at a location such as a fingertip since IMD 104 is positioned in a location having a pulsatility below a pulsatility threshold.

[0053] In the example shown in FIG. 2A, IMD 104A is defined by a length / ., a width W, and a thickness or depth D and is in the form of an elongated rectangular prism wherein the length L is much larger than the width W, which in turn is larger than the depth D. In one example, the geometry of IMD 104 A, in particular a width W greater than the depth D, is selected to allow IMD 104A to be inserted under the skin of the patient using a minimally invasive procedure and to remain in the desired orientation during insertion. For example, IMD 104 A as shown in FIG. 2 A includes radial asymmetries (notably, the rectangular shape) along the longitudinal axis that maintains the device in the proper orientation following insertion. For example, the spacing between proximal electrode 206A and distal electrode 206B may range from 5 millimeters (mm) to 55 mm, 30 mm to 55 mm, 35 mm to 55 mm, and from 40 mm to 55 mm and may be any range or individual spacing from 5 mm to 60 mm. In addition, IMD 104A may have a length L that ranges from 30 mm to about 70 mm. In other examples, the length L may range from 5 mm to 60 mm, 40 mm to 60 mm, 45 mm to 60 mm and may be any length or range of lengths between about 30 mm and about 70 mm. In addition, the width W of major surface 204 may range from 3 mm to 15, mm, from 3 mm to 10 mm, or from 5 mm to 15 mm, and may be any single or range of widthsAtty Ref. No. A0012750W001 between 3 mm and 15 mm. The thickness of depth D of IMD 104A may range from 2 mm to 15 mm, from 2 mm to 9 mm, from 2 mm to 5 mm, from 5 mm to 15 mm, and may be any single or range of depths between 2 mm and 15 mm. In addition, IMD 104A according to an example of the present disclosure is has a geometry and size designed for ease of implant and patient comfort. Examples of IMD 104A described in this disclosure may have a volume of three cubic centimeters (cm) or less, 1.5 cubic cm or less or any volume between three and 1.5 cubic centimeters.

[0054] In the example shown in FIG. 2A, once IMD 104A is inserted within patient 102, first major surface 204 faces outwards and towards the skin of patient 102 while second major surface 208 is located opposite first major surface 204. In addition, in the example shown in FIG. 2A, proximal end 210 and distal end 212 are rounded to reduce discomfort and irritation to surround tissue once IMD 104A is inserted under the skin of patient 102. IMD 104A and instruments and methods for inserting IMD 104A are described, for example, in U.S. Patent No. 11,311,312 filed on March 11, 2014, issued on April 26, 2022, and entitled “Subcutaneous Delivery Tool,” the entirety of which is herein incorporated by reference in its entirety.

[0055] Proximal electrode 206A is at or proximate to proximal end 210, and distal electrode 206B is at or proximate to distal end 212. Proximal electrode 206A and distal electrode 206B are used to sense cardiac signals, e.g., ECG signals, and measure interstitial impedance thoracically outside the ribcage, which may be sub-muscularly or subcutaneously. ECG signals and impedance measurements may be stored in a memory of IMD 104 A, and data may be transmitted via integrated antenna 216A to another device, which may be another implantable device or an external device, such as one or more of patient computing devices 106. In some examples, electrodes 206A, 206B are additionally or alternatively used for sensing any bio-potential or physiological signals of interest, which may be, for example, an electrogram (EGM), EEG, electromyogram (EMG), a nerve signal, or any other physiological signal, from any implanted location.

[0056] In the example shown in FIG. 2A, proximal electrode 206A is at or in close proximity to the proximal end 210 and distal electrode 206B is at or in close proximity to distal end 212. In this example, distal electrode 206B is not limited to a flattened, outward facing surface, but may extend from first major surface 204 around rounded edges and / or end surface 214 and onto second major surface 208 so that distal electrode 206B has aAtty Ref. No. A0012750W001 three-dimensional curved configuration. In some examples, distal electrode 206B is an uninsulated portion of a metallic, e.g., titanium, part of housing 202.

[0057] In the example shown in FIG. 2A, proximal electrode 206A is located on first major surface 204 and is substantially flat and outward facing. In some examples, proximal electrode 206A utilizes the three-dimensional curved configuration of distal electrode 206B and provide a three-dimensional proximal electrode (not shown in FIG. 2A). Similarly, in other examples distal electrode 206B may utilize a substantially flat, outward facing electrode located on first major surface 204 similar to that shown with respect to proximal electrode 206A.

[0058] The various electrode configurations allow for configurations in which proximal electrode 206A and distal electrode 206B are located on both first major surface 204 and second major surface 208. In other configurations, such as than shown in FIG. 2A, only one of proximal electrode 206A, distal electrode 206B, and optical sensor 224 is located on both major surfaces 204 and 208, and in still other configurations two or more of proximal electrode 206A, distal electrode 206B, and optical sensor 224 are located on one of first major surface 204 or second major surface 208. In some examples, IMD 104 A may include electrodes and optical sensors 224 on both major surfaces 204 and 208, such that a total of four electrodes and two optical sensors 224 are included on IMD 104A. Electrodes 206A and 206B may be formed of a plurality of different types of biocompatible conductive material, e.g., stainless steel, titanium, platinum, iridium, or alloys thereof, and may utilize one or more coatings such as titanium nitride or fractal titanium nitride.

[0059] In the example shown in FIG. 2 A, proximal end 210 includes a header assembly 218 that includes one or more of proximal electrode 206 A, integrated antenna 216A, anti-migration projections 221, and / or suture hole 220. Integrated antenna 216A allows IMD 104A to transmit and / or receive data. In some examples, integrated antenna 216A may be formed on the opposite major surface as proximal electrode 206 A or may be incorporated within housing 202 of IMD 104 A. In the example shown in FIG. 2 A, antimigration projections 221 are located adjacent to integrated antenna 216A and protrude away from first major surface 204 to inhibit longitudinal movement of IMD 104 A. In the example shown in FIG. 2A, anti-migration projections 221 include a plurality (e.g., nine) small bumps or protrusions extending away from first major surface 204. As discussedAtty Ref. No. A0012750W001 above, in other examples anti-migration projections 221 may be located on the opposite major surface as proximal electrode 206A and / or integrated antenna 216A. In addition, in the example shown in FIG. 2 A, header assembly 218 includes suture hole 220, which provides another means of securing IMD 104 A to patient 102 to inhibit movement of IMD 104A following insertion. In the example shown in FIG. 2A, suture hole 220 is located adjacent to proximal electrode 206A. In one example, header assembly 218 is a molded header assembly made from a polymeric or plastic material, which may be integrated or separable from the main portion of IMD 104 A.

[0060] Optical sensor 224 includes one or more light emitters 226 and one or more light detectors 228. In some examples, optical sensor 224 includes a proximal light detector 228 disposed at or around proximal end 210 and a distal light detector 228 disposed at or around distal end 212 (not shown in FIG. 2A). In some examples, the proximal light detector 228 is disposed at a first longitudinal location along housing 202 between light emitter(s) 226 and proximal end 210 and distal light detector 228 is disposed at a second longitudinal location between light emitter(s) 226 and distal end 212. In some examples, light emitter(s) 226 and light detector(s) 228 may be disposed on one of major surfaces 204, 208. In some examples, at least one light emitter 226 and one light detector 228 is disposed on each of major surfaces 204, 208. Each of light detectors 228 may be positioned at least a minimum longitudinal distance away from one or more light emitters 226 disposed on a same surface of major surfaces 204, 208. In some examples, optical sensor 224 is a PPG sensor.

[0061] Each of light emitter(s) 226 or light detector(s) 228 may include a transparent window defining at least a portion of an outer surface of IMD 104 (e.g., at least a portion of one or more of major surfaces 204, 208). The transparent windows may isolate the circuitry and other components of light emitter(s) 226 and light detector(s) 228 from body fluid of patient 102 and facilitate the maintenance of the hermetic seal by housing 202. Each transparent window may be formed from a glass or sapphire. In some examples, each of light emitter(s) 226 or light detector(s) 228 may be positioned beneath a portion of housing 202 formed from a transparent or translucent material (e.g., glass or sapphire).

[0062] Each light emitter 226 may include a light source, such as a light-emitting diode (LED), which may emit light at one or more wavelengths within the visible and / or near-infrared (NIR) spectra. For example, each light emitter 226 is configured to emit lightAtty Ref. No. A0012750W001 with wavelengths of about 680 nanometers (nm), 720 nm, 760 nm, 800 nm, or any other suitable wavelengths. IMD 104 may use the one or more wavelengths of light emitted by light emitter 226 to determine values for one or more physiological parameters, including, but are not limited to, tissue oxygen saturation (StCh), blood oxygen saturation (SpCh), tissue hemoglobin index (THI), blood pressure level, or the like.

[0063] Light emitter(s) 226 may emit light into a target site of patient 102. The target site generally includes blood within blood vessels in the interstitial space around IMD 104A when IMD 104A is implanted in patient 102. Light emitter(s) 226 may emit light directionally (e.g., to a side of IMD 104A). In some examples, the light is a cloud of light directly generally inwards (e.g., towards the musculature of patient 102 and away from the skin of patient 102). In some examples, the light is non-directi onal once emitted from light emitter(s) 226).

[0064] In some examples, light emitter(s) 226 emit visible (VIS) and / or NIR light at multiple wavelengths, either simultaneously or sequentially. Light emitter(s) 226 may emit light at one or more wavelengths capable of shallow penetration into tissue (e.g., tissue relatively close to IMD 104 A as well as at one or more wavelengths capable of deeper penetration into tissue (e.g., tissue relatively further away from IMD 104A). In such examples, using wavelengths capable of penetrating the tissue to different depths may enable the elimination of measurement errors, which may be caused by factors such as fibrous tissue encapsulation that may form around IMD 104A post implant. In some examples, using wavelengths capable of penetrating the tissue to different depths may allow IMD 104 A to identify and reduce the effects of common errors associated with the different light signals, such as, but are not limited to, tissue formation, adhesion, changes in tissue content.

[0065] Light emitter(s) 226 may include a single photodiode that is capable of emitting light at a range of wavelengths. In some examples, light emitter(s) 226 include multiple photodiodes, each photodiode being configured to emit light at one or more different wavelengths. Light detector(s) 228 may receive light from light emitter(s) 226 that is reflected by the tissue and generate electrical signals indicating intensities of the light detected by light detector(s) 228. IMD 104A may evaluate the electrical signals to determine values for one or more physiological parameters (e.g., oxygen saturation level, blood pressure level).Atty Ref. No. A0012750W001

[0066] The physiological parameters of patient 102 may affect an amount of light absorbed by blood within tissue adjacent to IMD 104A and / or an amount of light reflected by the tissue. IMD 104A may receive the reflected light via light detector(s) 228 and determine physiological parameter values based on the received light signals. For example, IMD 104A is configured to determine a blood pressure level of patient 102 based on the light signals received via light detector(s) 228. In some examples, IMD 104A is configured to determine the blood pressure level of patient 102 based on the received light signals via a PPG technique.

[0067] FIG. 2B is a perspective drawing illustrating another IMD 104B, which may be another example configuration of IMD 104 from FIG. 1. IMD 104B of FIG. 2B may be configured substantially similarly to IMD 104A of FIG. 2A, with differences between them discussed herein.

[0068] IMD 104B may include a leadless, subcutaneously implantable monitoring device, e.g., an ICM. IMD 104B includes housing having a base 223 and an insulative cover 222. Proximal electrode 206C and distal electrode 206D may be formed or placed on an outer surface of cover 222. Various circuitries and components of IMD 104B may be formed or placed on an inner surface of cover 222, or within base 223. In some examples, a battery or other power source of IMD 104B may be included within base 223. In the illustrated example, antenna 216B is formed or placed on the outer surface of cover222 but may be formed or placed on the inner surface in some examples. In some examples, insulative cover 222 may be positioned over an open base 223 such that base223 and cover 222 enclose the circuitries and other components and protect them from fluids such as body fluids. The housing including base 223 and insulative cover 222 may be hermetically sealed and configured for subcutaneous implantation. At least a portion of base 223 and / or cover 222 may be transparent or translucent, e.g., to facilitate the transmission of light out of IMD 104B by light emitter(s) 226 and the detection of lights signals by light detector(s) 228.

[0069] Circuitries and components may be formed on the inner side of insulative cover222, such as by using flip-chip technology. Insulative cover 222 may be flipped onto base223. When flipped and placed onto base 223, the components of IMD 104B formed on the inner side of insulative cover 222 may be positioned in a gap 225 defined by base 223. Electrodes 206C and 206D and antenna 216B may be electrically connected to circuitryAtty Ref. No. A0012750W001 formed on the inner side of insulative cover 222 through one or more paths (not shown) formed through insulative cover 222. Insulative cover 222 may be formed of sapphire (i.e., corundum), glass, perylene, and / or any other suitable insulating material. Insulative cover 222 may be at least partially or entirely transparent or translucent. Base 223 may be formed from titanium or any other suitable material (e.g., a biocompatible material). Electrodes 206C and 206D may be formed from any of stainless steel, titanium, platinum, iridium, or alloys thereof. In addition, electrodes 206C and 206D may be coated with a material such as titanium nitride or fractal titanium nitride, although other suitable materials and coatings for such electrodes may be used.

[0070] In the example shown in FIG. 2B, the housing of IMD 104B defines a length / ., a width W, and thickness or depth D, and is in the form of an elongated rectangular prism wherein the length L is much larger than the width W, which in turn is larger than the depth D, similar to IMD 104A of FIG. 2A. For example, the spacing between proximal electrode 206C and distal electrode 206D may range from 5 mm to 50 mm, from 30 mm to 50 mm, from 35 mm to 45 mm, and may be any single spacing or range of spacings from 5 mm to 50 mm, such as approximately 40 mm. In addition, IMD 104B may have a length L that ranges from 5 mm to about 70 mm. In other examples, the length L may range from 30 mm to 70 mm, 40 mm to 60 mm, 45 mm to 55 mm, and maybe any single length of range of lengths from 5 mm to 50 mm, such as approximately 45 mm. In addition, the width may range from 3 mm to 15 mm, 5 mm to 15 mm, 5 mm to 10 mm, and may be any single width or range of widths from 3 mm to 15 mm, such as approximately 8 mm. The thickness or depth D of IMD 104B may range from 2 mm to 15 mm, from 5 mm to 15 mm, or from 3 mm to 5 mm, and may be any single depth or range of depths between 2 mm and 15 mm, such as approximately 4 mm. IMD 104B may have a volume of three cubic centimeters (cm3) or less, or 1.5 cm3or less, such as approximately 1.4 cm3.

[0071] In the example shown in FIG. 2B, once IMD 104B is inserted subcutaneously within patient 102, the outer surface of cover 222 faces outwards and towards the skin of patient 102. In addition, as shown in FIG. 2B, proximal end 229A and distal end 229B are rounded to reduce discomfort and irritation to surround tissue once inserted under the skin of patient 102. In addition, edges of IMD 104B may be rounded.

[0072] FIG. 3 A is a block diagram illustrating an example configuration of IMD 104 of FIG. 1. As shown in FIG. 3 A, IMD 104 may include processing circuitry 302, memoryAtty Ref. No. A0012750W001304, sensing circuitry 306 coupled to electrodes 206 A and 206B (collectively referred to herein as “electrodes 206”), to optical sensor 224, and to one or more sensor(s) 308, and communications circuitry 310.

[0073] Processing circuitry 302 may include fixed function circuitry and / or programmable processing circuitry. Processing circuitry 302 may include any one or more of a microprocessor, a controller, a GPU, a TPU, a DSP, an ASIC, a FPGA, or equivalent discrete or analog logic circuitry. The functions attributed to processing circuitry 302 herein may be embodied as software, firmware, hardware, or any combination thereof. In some examples, memory 304 includes computer-readable instructions that, when executed by processing circuitry 302, cause IMD 104 and processing circuitry 302 to perform various functions attributed herein to IMD 104 and processing circuitry 302. Memory 304 may include any volatile, non-volatile, magnetic, optical, or electrical media, such as RAM, ROM, NVRAM, EEPROM, flash memory, or any other digital media.

[0074] Sensing circuitry 306 may sense an electrocardiogram (ECG) signal and measure impedance, e.g., of tissue proximate to IMD 104, via electrodes 206. The measured impedance may vary based on respiration, cardiac pulse, or flow, and a degree of perfusion or edema. Processing circuitry 302 may determine physiological parameter values relating to respiration, fluid retention, cardiac pulse or flow, perfusion, and / or edema based on the measured impedance. In some examples, processing circuitry 302 may identify features of the sensed ECG, such as heart rate, heart rate variability, T-wave alternans, intra-beat intervals (e.g., QT intervals), and / or ECG morphologic features. Processing circuitry 302 may determine that patient 102 has experienced an external impact based on changes in identified features of the sensed ECG (e.g., changes in the heart rate) in addition to changes in other physiological parameters (e.g., changes in acceleration experienced by patient 102, changes in posture of patient 102, changes in g- force on patient 102). The techniques described in this disclosure may be implemented, at least in part, in hardware, software, firmware, or any combination thereof. For example, various aspects of the techniques may be implemented within one or more microprocessors, DSPs, ASICs, FPGAs, or any other equivalent integrated or discrete logic QRS circuitry, as well as any combinations of such components, embodied in external devices, such as physician or patient programmers, stimulators, or other devices. The terms “processor” and “processing circuitry” may generally refer to any of theAtty Ref. No. A0012750W001 foregoing logic circuitry, alone or in combination with other logic circuitry, or any other equivalent circuitry, and alone or in combination with other digital or analog circuitry.

[0075] IMD 104 may include one or more sensors 308, such as one or more accelerometers, gyroscopes inertial measurement units (IMUs), microphones, oximeters, optical sensors 224 (e.g., including one or more light emitters 226 and one or more light detectors 228), temperature sensors, pressure sensors, and / or chemical sensors. Sensing circuitry 306 may include one or more filters and amplifiers for filtering and amplifying signals from one or more of electrodes 206 and / or sensor(s) 308. Sensing circuitry 306 and / or processing circuitry 302 may include a rectifier, filter and / or amplifier, a sense amplifier, comparator, and / or analog-to-digital converter. Processing circuitry 302 may determine physiological parameter data 320, e.g., values of physiological parameters of patient 102, based on signals from sensor(s) 308, which may be stored as data 316 in memory 304. Physiological parameter may include a posture of patient 102, acceleration of the body (e.g., of the torso) of patient 102, a g-force on patient 102, a heart rate of patient 102, a respiration rate of patient 102, a stress hormone level of patient 102, a body temperature of patient 102, glucose level of patient 102, blood pressure of patient 102, EEG signals of patient 102, change in electrocardiogram (ECG) morphology of patient 102, change in EGM morphology of patient 102, impedance between two or more locations on patient 102, an oxygen saturation of patient 102 or a rate of change of a physiological parameter.

[0076] Sensing circuitry 306 may continuously or periodically sensed physiological signals via electrodes 206 and / or sensor(s) 308. Processing circuitry 302 may receive an instruction to sense physiological signals (e.g., from computing system 110, patient computing devices 106) and cause sensing circuitry 306 to sense physiological signals from patient 102 in response to the instruction.

[0077] In addition to data 316, memory 304 may store application(s) 312 executable by processing circuitry 302. Application(s) 312 may include physiological parameter monitoring application 314 executable by processing circuitry 302 to determine physiological parameter values based on physiological signals sensed via sensing circuitry 306, electrodes 206, and / or sensor(s) 308. Processing circuitry 302 may execute physiological parameter monitoring application 314 to compare physiological parameter values (e.g., physiological parameter data 320) against threshold conditions 318 stored inAtty Ref. No. A0012750W001 data 316 and select a sensing mode of a plurality of sensing modes based on the comparison.

[0078] Processing circuitry 302 may execute physiological parameter monitoring application 314 to compare physiological parameter values of a single type (e.g., blood pressure levels only) or of two or more types (e.g., blood pressure level and heart rate) against corresponding threshold conditions 318 stored in data 316. In some examples, when processing circuitry 302 executes physiological parameter monitoring application 314, processing circuitry 302 selects a same sensing mode (e.g., a current sensing mode) based on satisfaction of threshold conditions 318 and selects a sensing mode different from a current sensing mode based on a failure to satisfy threshold conditions 318. In such examples, satisfaction of threshold conditions 318 may correspond to onset of a medical condition, changes in an existing medical condition, onset of the effects of a medical therapy, and / or changes in medical therapies. In some examples, when processing circuitry 302 executes physiological parameter monitoring application 314, processing circuitry 302 selects a different sensing mode (e.g., from a current sensing mode) based on a failure to satisfy threshold conditions 318 and selects the same sensing mode based on satisfaction of threshold conditions 318. In such examples, a failure to satisfy threshold conditions 318 may correspond to onset of a medical condition, changes in an existing medical condition, onset of the effects of a medical therapy, and / or changes in medical therapies. Processing circuitry 302 may receive an indication that patient 102 is receiving a new medical therapy and / or that there are changes to applied medical therapies from one or more other computing devices (e.g., from computing system 110 and / or clinician computing devices 128).

[0079] Communications circuitry 310 may include any suitable hardware, firmware, software, or any combination thereof for wireless communication with another device. Communications circuitry 310 may be configured to transmit and / or receive signals via inductive coupling, electromagnetic couple, Near Field Communication (NFC), Radio Frequency (RF) communication, Bluetooth®, WiFi, or other proprietary or nonproprietary wireless communications schemes.

[0080] In some examples, IMD 104 may include one or more optical sensors 224 configured to sense optical signals from patient 102. In some examples, the one or more optical sensors 224 may include a PPG sensor. In some examples, IMD 104 may beAtty Ref. No. A0012750W001 configured to be positioned at a low pulastility location of patient 102. In some examples, IMD 104 may be configured to sense a low-fidelity optical signal 370 of patient 102. For example, IMD 104 may be configured to be positioned subcutaneously in a chest of the patient 102. In some examples, IMD 104 being positioned subcutaneously in a chest region of patient 102 may cause optical signals sensed by IMD 104 to be a low-fidelity optical signal. In some examples, IMD 104 being positioned subcutaneously in a chest region of patient 102 may cause optical signals sensed by IMD 104 to be low pulastility signals. In some examples, a low pulsatility signal may be less than 1%. In some examples, a low pulsatility signal may be between 0.1% and 0.5%.

[0081] FIG. 3B is a graph showing an example low-fidelity optical signal 370. In some examples, low-fidelity optical signal 370 may be a low-fidelity PPG signal. FIG. 3C is a graph showing an example high-fidelity optical signal 380. In some examples, high- fidelity optical signal 380 may be a high-fidelity PPG signal. Low-fidelity optical signal 370 has a relatively lower fidelity than high-fidelity optical signal 380, and conversely high-fidelity optical signal 380 has a relatively higher fidelity than low-fidelity optical signal 370. Fidelity of an optical signal may refer to how accurately it reflects a time varying physiological parameter of interest, such pulsatile blood pressure.

[0082] In some examples, a high-fidelity optical signal 380 may include a dicrotic notch 382. In some examples, high-fidelity optical signal 380 may include high-fidelity signal amplitude 388. In some examples, low-fidelity optical signal 370 may include an absence of a dicrotic notch. In some examples, low-fidelity optical signal 370 may additionally or alternatively include one or more of a low-fidelity signal amplitude 378 that has a lower amplitude than an amplitude 388 of a high-fidelity optical signal 380 sensed during a same period of time, motion artifact or noise 374, and / or distortion 376. In some examples, as shown in FIGS. 3B-3C, high-fidelity signal amplitude 388 may be greater than low-fidelity signal amplitude 378. In some examples, high-fidelity optical signal 380 may be referred to as “first optical signal” that is sensed by external sensing device 105 or sensed by IMD 104 operating in a high-fidelity mode and low-fidelity optical signal 370 may be referred to as “second optical signal” that is sensed by IMD 104. In some examples, it may be more difficult for processing circuitry to detect and / or extract features, such as a dicrotic notch, from fidelity optical signal 370.Atty Ref. No. A0012750W001

[0083] In accordance with techniques described here, processing circuitry 112 being configured to train a ML model based on low-fidelity optical signal 370 and high-fidelity optical signal 380 sensed during a same period of time to produce a trained ML model that is configured to more accurately reconstruct a sensed low-fidelity optical signal, such as an optical signal similar to optical signal 370, to a high-fidelity optical signal, such as similar to optical signal 380, which may improve detection of features, such as a dicrotic notch 382, from a sensed low-fidelity optical signal.

[0084] FIG. 3D is a block diagram illustrating an example configuration of external sensing device 105 of FIG. 1. As shown in FIG. 3B, external sensing device 105 may include processing circuitry 332, memory 334, sensing circuitry 336 coupled to one or more sensor(s) 338, and communications circuitry 340. In some examples, sensor(s) 338 may include one or more optical sensors 339.

[0085] In some examples, external sensing device 105 may be configured to sense, via optical sensor(s) 339, physiological optical signals of patient 102 at high pulsatility location of patient 102 to generate high-fidelity optical signals 380 of patient 102 via optical sensor(s) 339. In some examples, the one or more optical sensors 339 may include a PPG sensor. In some examples, external sensing device 105 may collect and store physiological parameter values based on the sensed physiological signals. In some examples, external sensing device 105 may be configured to be positioned high pulsatility location of patient 102, such as over or around a fingertip of patient, 102 to sense optical signals. In some examples, external sensing device 105 being positioned at high pulsatility location of patient 102 may cause optical signals sensed by external sensing device 105 to be high pulastility signals. In some examples, a high pulsatility signal may be between 1% and 5%. In some examples, a high pulsatility signal may be greater than or equal to 2.5%. In some examples, a high pulsatility signal may be greater than or equal to 3%. In some examples, external sensing device 105 may be a wearable device, such as a smartwatch, configured to sense optical signals. In some examples, sensor(s) 338 may further include one or more of electrodes or accelerometers.

[0086] FIG. 4 is a block diagram illustrating an example configuration of a patient computing device 106. In some examples, patient computing device 106 includes a smartphone, a laptop, a table computer, a personal digital assistant (PDA), a smartwatch, or other wearable computing devices.Atty Ref. No. A0012750W001

[0087] As shown in the example illustrated in FIG. 4, patient computing device 106 may be logically divided into user space 402, kernel space 404, and hardware 406. Hardware 406 may include one or more hardware components that provide an operating environment for components executed in user space 402 and kernel space 404. User space 402 and kernel space 404 may represent different sections or segmentations of memory, where kernel space 404 provides higher privileges to processes and threshold than user space 402. For instance, kernel space 404 may include operating system 432, which operates with higher privileges than components executing in user space 402.

[0088] As shown in FIG. 4, hardware 406 includes process circuitry 420, memory 422, one or more input devices 424, one or more output devices 426, one or more sensors 428, and communications circuitry 429. Although shown in FIG. 4 as a stand-alone device for purposes of examples, patient computing device 106 may be any component or system that includes processing circuitry or other suitable computing environment for executing software instructions and, for example, need not necessarily include one or more elements shown in FIG. 4.

[0089] Processing circuitry 420 is configured to implement functionality and / or process instructions for execution within patient computing device 106. For example, processing circuitry 420 may be configured to receive and process instructions stored in memory 422 that provide functionality of components included in kernel space 404 and user space 402 to perform one or more operations in accordance with techniques of this disclosure. Examples of processing circuitry 420 may include, any one or more microprocessors, controllers, GPUs, TPUs, DSPs, ASICs, FPGAs, or equivalent discrete or integrated logic circuitry.

[0090] Memory 422 may be configured to store information within patient computing device 106, for processing during operation of patient computing device 106. Memory 422, in some examples, is described as a computer-readable storage medium. In some examples, memory 422 includes a temporary memory or a volatile memory. Examples of volatile memories include RAM, DRAM, SRAM, and other forms of volatile memories known in the art. Memory 422, in some examples, also includes one or more memories configured for long-term storage of information, e.g., including non-volatile storage elements. Examples of such non-volatile storage elements include magnetic hard discs, optical discs, floppy discs, flash memories, or forms of electrically programmableAtty Ref. No. A0012750W001 memories (EPROM) or electrically erasable and programmable (EEPROM) memories. In some examples, memory 422 includes cloud-associated storage.

[0091] One or more input devices 424 of patient computing device 106 may receive input, e.g., from patient 102, clinician 130, or another user. Examples of input are tactile, audio, kinetic, and optical input. Input devices 424 may include, as examples, a mouse, keyboard, voice responsive system, camera, buttons, control pad, microphone, presencesensitive or touch-sensitive component (e.g., screen), or any other device for detecting input from a user or a machine.

[0092] One or more output devices 426 of patient computing device 106 may generate output, e.g., to patient 102 or another user. Examples of output are tactile, haptic, audio, and visual output. Output devices 426 of patient computing device 106 may include a presence-sensitive screen, sound card, video graphics adapter card, speaker, cathode ray tube (CRT) monitor, liquid crystal display (LCD), light emitting diodes (LEDs), or any type of device for generating tactile, audio, and / or visual output.

[0093] One or more sensors 428 of computing devices 106 may sense physiological parameters or physiological signals of patient 102. Sensor(s) 428 may include electrodes, accelerometers (e.g., 3-axis accelerometers), IMUs, gyroscopes, optical sensors, impedance sensors, temperature sensors, pressure sensors, heart sound sensors (e.g., microphones or accelerometers), and other sensors.

[0094] Communication circuitry 429 of patient computing device 106 may communicate with other devices by transmitting and receiving data. Communication circuitry 429 may receive data from IMD 104, such as physiological signals and / or physiological parameter values, from communication circuitry 310 in IMD 104. Communication circuitry 429 may include a network interface card, such as an Ethernet card, an optical transceiver, a radio frequency transceiver, or any other type of device that can send and receive information. For example, communication circuitry 429 may include a radio transceiver configured for communication according to standards or protocols, such as 3G, 4G, 5G, WiFi (e.g., 802.11 or 802.15 ZigBee), Bluetooth®, or Bluetooth® Low Energy (BLE).

[0095] As shown in FIG. 4, health monitoring application 410 executes in user space 402 of patient computing device 106, health monitoring application 410 may be logically divided into presentation layer 412, application layer 414, and data layer 416. PresentationAtty Ref. No. A0012750W001 layer 412 may include a user interface (UI) component 430, which generates and renders user interfaces of health monitoring application 410.

[0096] Data layer 416 may include threshold condition data 450 and physiological parameter data 452, which may be received from IMD 104 via communications circuitry 429 and stored in memory 422 by processing circuitry 420. Threshold condition data 450 may contain threshold conditions corresponding to different physiological parameters, such as a pulsatility threshold. IMD 104 may receive threshold conditions from clinician 130, clinician computing device 128, computing system 110, or other computing devices or systems connected to network 108 (e.g., via input device(s) 424 and / or communications circuitry 429). Patient computing device 106 may determine or receive changes in physiological parameter values and store the changes in physiological parameter values in physiological parameter data 452. In some examples, patient computing device 106 may receive physiological parameter values from IMD 104. Application layer 414 may include, but is not limited to, physiological parameter analyzer 440. Physiological parameter analyzer 440 may monitor reconstructed optical signals and / or blood pressure values (e.g., generated by ML model 118) and determine a health condition status based on the reconstructed optical signals and / or blood pressure values. Physiological parameter analyzer 440 may output (e.g., via output device(s) 426) notifications to patient 102 and / or clinician 130 in response to a determination of a health condition status.

[0097] FIG. 5 is a block diagram illustrating an operating perspective of HMS 116. HMS 116 may be implemented in a computing system 110, which may include hardware components such as processing circuitry 112, memory 114, and communications circuitry, embodies in one or more physical devices. FIG. 5 provides an operating perspective of HMS 116 when hosted as a cloud-based platform. In the example of FIG. 5, components of HMS 116 are arranged according to multiple logical layers that implement the techniques of this disclosure. Each layer may be implemented by one or more modules comprised of hardware, software, or a combination of hardware and software.

[0098] Computing devices, such as patient computing devices 106 and / or client computing devices 128, operate as clients that communicate with HMS 116 via interface layer 500. The computing devices typically execute client software applications, such as desktop application(s), mobile application(s), and web application(s). Interface layer 500 represents a set of application programming interfaces (API) or protocol interfacesAtty Ref. No. A0012750W001 presented in and supported by HMS 116 for the client software applications. Interface layer 500 may be implemented with one or more web servers.

[0099] As shown in FIG. 5, HMS 116 also includes an application layer 502 that represents a collection of services 506 for implementing the functionality ascribed to HMS 116 herein. Application layer 502 receives information from client applications, e.g., data from patient computing devices 106, some or all of which may have been received from IMD 104, and further processes the information according to one or more of services 506 to respond to the information. Application layer 502 may be implemented as one or more discrete software services 506 executed on one or more application servers, e.g., physical or virtual machines. That is, the application servers provide runtime environments for execution of services 506. In some examples, the functionality interface layer 500 as described above and the functionality of application layer 502 may be implemented at the same server. Services 506 may communicate via a logical service bus 511. Service bus 511 generally represents a logical interconnection or set of interfaces that allows different services 506 to send messages to other services, such as by a publish / subscription communication model.

[0100] Data layer 504 of HMS 116 provides persistence for information in HMS 116 using one or more data repositories 508. A data repository 508, generally, may be any data structure or software that stores and / or manages data. Examples of data repositories 508 include, but are not limited to, relational databases, multi-dimensional databases, maps, and / or hash tables.

[0101] As shown in FIG. 5, each of services 510-514 is implemented in modular form within HMS 116. Although shown as separate modules for each service, in some examples the functionality of two or more services may be combined into a single module or component. Each of services 510-514 may be implemented in software, hardware, or a combination of hardware and software. Moreover, services 510-514 may be implemented as standalone devices, separate virtual machines or containers, processes, threads or software instructions generally for execution on one or more physical processors. Record management service 514 may store received data such as threshold condition data 520 and physiological parameter data 522 from client computing devices (e.g., from patient computing devices 106, IMD 104).Atty Ref. No. A0012750W001

[0102] Threshold condition detection service 510 may compare threshold conditions (e.g., stored in threshold condition data 520) against physiological parameters (e.g., stored in physiological parameter data 522) and determine whether the threshold conditions are satisfied. In some examples, threshold condition detection service 510 is configured to compare physiological parameter values from a prior time period against one or more threshold conditions to determine whether the one or more threshold conditions are satisfied. The one or more threshold conditions may correspond to a single type of physiological parameter or to two or more different types of physiological parameters. Physiological parameter monitoring service 512 may monitor physiological parameters of patient 102.

[0103] FIG. 6A is a flowchart illustrating an example process of training a ML model based on a first optical signal and a second optical signal to generate at least one of a reconstructed optical signal based on optical signals sensed via a sensing device positioned at a location having a pulsatility below a pulsatility threshold or a blood pressure value based on optical signals sensed via a sensing device positioned at a location having a pulsatility below the pulsatility threshold. While the technique of FIG. 6A is primarily described herein as being performed by processing circuitry 112 of computing system 110, such techniques may be performed, in whole or part, by processing circuitry of any one or more devices of system 100, such as processing circuitry 420 of patient computing device 106, and / or processing circuitry 302 of IMD 104.

[0104] Processing circuitry 112 may be configured to receive a first optical signal (602). In some examples, the first optical signal is sensed during a first period of time. In some examples, the first optical signal is a high-fidelity optical signal. In some examples, the first optical signal is sensed via an IMD 104 operating in a high-fidelity mode to generate a high-fidelity optical signal, such as due to an emitter temporarily provided with more power to generate a higher-fidelity optical signal. In some examples, the first optical signal is sensed via an external sensing device 105. Processing circuitry 112 may be configured to receive a second optical signal sensed via a sensing device, such as IMD 104 (604). In some examples, sensing device may be a plurality of sensing devices of the same type, such as a plurality of IMDs 104 and processing circuitry 112 may be configured to receive a plurality of second optical signals sensed via the respective plurality of sensing devices. In some examples, the second optical signal is sensed by theAtty Ref. No. A0012750W001 sensing device during the first period of time. In some examples, the second optical signal is a low-fidelity optical signal. In some examples, the second optical signal is a lower- fidelity optical signal than the first optical signal. In some examples, the first optical signal and the second optical signal are sensed via the same sensing device, such as IMD 104, but operating in respective different modes during sensing of the optical signals (e.g., IMD 104 operating in a high-fidelity mode to generate the first optical signal but operating in a normal lower-fidelity mode to generate the second optical signal).

[0105] Processing circuitry 112 may be configured to apply the first optical signal and the second optical signal to a ML model 118 to train the ML model 118 to generate at least one of a reconstructed optical signal or blood pressure value(s) based on sensed optical signals (606). In some examples, the sensed optical signals are sensed via a sensing device positioned at a location having a pulsatility below a pulsatility threshold. In some examples, the pulsatility threshold is within 0.5%. In some examples, the pulsatility threshold is within 1.0%. In some examples, the pulsatility threshold is based on the pulsatility of the second location.

[0106] Processing circuitry 112 may be configured to receive a third optical signal sensed via an additional sensing device (608). In some examples, the additional sensing device is the same as the sensing device (e.g., IMD 104), or of the same type of device as the sensing device. In some examples, the third optical signal is a lower-fidelity signal than the first optical signal. Processing circuitry 112 may be configured to apply the third optical signal to the trained ML model 118 to generate at least one of a reconstructed optical signal of the third optical signal or a blood pressure value of the patient based on the third optical signal (610). In some examples, the reconstructed optical signal of the third optical signal has a higher-fidelity than the third optical signal

[0107] FIG. 6B is a flowchart illustrating an example process of training a ML model based on a first optical signal and a second optical signal to generate at least one of a reconstructed optical signal based on optical signals sensed via a sensing device positioned at a location having a pulsatility below a pulsatility threshold or a blood pressure value based on optical signals sensed via a sensing device positioned at a location having a pulsatility below the pulsatility threshold. While the technique of FIG. 6 is primarily described herein as being performed by processing circuitry 112 of computing system 110, such techniques may be performed, in whole or part, by processing circuitry of any one orAtty Ref. No. A0012750W001 more devices of system 100, such as processing circuitry 420 of patient computing device 106, and / or processing circuitry 302 of IMD 104.

[0108] Processing circuitry 112 may be configured to receive a first optical signal sensed via a first sensing device, such as external sensing device 105 (612). In some examples, the first optical signal is sensed during a first period of time. In some examples, first sensing device may be positioned at a first location of a patient, the first position having a high pulsatility. In some examples, the first optical signal is a high-fidelity optical signal. Processing circuitry 112 may be configured to receive a second optical signal sensed via a second sensing device, such as IMD 104 (614). In some examples, second sensing device may be a plurality of sensing devices of the same type, such as a plurality of IMDs 104 and processing circuitry 112 may be configured to receive a plurality of second optical signals sensed via the respective plurality of second sensing devices. In some examples, the second optical signal is sensed by the second sensing device during the first period of time. In some examples, second sensing device is positioned at a second location of the patient. In some examples, the second location has a low pulsatility. In some examples, the second location has a lower pulsatility than the first location. In some examples, the second optical signal is a low-fidelity optical signal. In some examples, the second optical signal is a lower-fidelity optical signal than the first optical signal.

[0109] Processing circuitry 112 may be configured to apply the first optical signal and the second optical signal to a ML model 118 to train the ML model 118 to generate at least one of a reconstructed optical signal or blood pressure value(s) based on sensed optical signals (616). In some examples, the sensed optical signals are sensed via a sensing device positioned at a location having a pulsatility below a pulsatility threshold. In some examples, the pulsatility threshold is within 0.5%. In some examples, the pulsatility threshold is within 1.0%. In some examples, the pulsatility threshold is based on the pulsatility of the second location.

[0110] Processing circuitry 112 may be configured to receive a third optical signal sensed via a third sensing device (618). In some examples, the third sensing device is the same as the second sensing device (e.g., IMD 104), or of the same type of device as the second sensing device. In some examples, the third optical signal is a lower-fidelity signal than the first optical signal. Processing circuitry 112 may be configured to apply the thirdAtty Ref. No. A0012750W001 optical signal to the trained ML model 118 to generate at least one of a reconstructed optical signal of the third optical signal or a blood pressure value of the patient based on the third optical signal (620). In some examples, the reconstructed optical signal of the third optical signal has a higher-fidelity than the third optical signal.[oni] In some examples, training the ML model 118 based on the first optical signal and the second optical signal sensed during the first period of time may train ML model 118 to more accurately generate a reconstructed optical signal and / or generate a blood pressure value based on an optical signal that has a low-fidelity due to being sensed via a sensing positioned at a low pulsatility location, which may improve the detection of features from the third optical signal, such as a dicrotic notch of the third optical signal.

[0112] FIG. 7 is a conceptual diagram illustrating an example machine learning model 700 configured to generate one or more of a reconstructed optical signal and / or a blood pressure value based on low-fidelity optical signal(s) sensed via a sensing device positioned at a low pulsatility location, e.g., sensed by an IMD as described herein. In some examples, machine learning model 700 may correspond to ML model 118. Machine learning model 700 is an example of a deep learning model, or deep learning algorithm. One or more of IMD 104, patient computing device 106, or computing system 110 may train, store, and / or utilize machine learning model 700, but other devices may apply inputs associated with a particular patient to machine learning model 700 in other examples. Some non-limiting examples of machine learning techniques include Bayesian probability models, Support Vector Machines, K-Nearest Neighbor algorithms, and Multi-layer Perceptron.

[0113] As shown in the example of FIG. 7, machine learning model 700 may include three layers. These three layers include input layer 702, hidden layer 704, and output layer 706. Output layer 706 comprises the output from the transfer function 705 of output layer 706. Input layer 702 represents each of the input values XI through X4 provided to machine learning model 700. The number of inputs may be less than or greater than 4, including much greater than 4, e.g., hundreds or thousands. In some examples, the input values may include values of a sensed low-fidelity optical signal, and a sensed high- fidelity optical signal during a first period of time described herein.

[0114] Each of the input values for each node in the input layer 702 is provided to each node of hidden layer 704. In the example of FIG. 7, hidden layers 704 include twoAtty Ref. No. A0012750W001 layers, one layer having four nodes and the other layer having three nodes, but fewer or greater number of nodes may be used in other examples. Each input from input layer 702 is multiplied by a weight and then summed at each node of hidden layers 704. During training of machine learning model 700, the weights for each input are adjusted to establish the relationship between values of a sensed low-fidelity optical signal and values of a sensed high-fidelity optical signal during a first period of time and one or more output values indicative of a reconstructed high-fidelity optical signal and / or a blood pressure value of the patient. In some examples, one hidden layer may be incorporated into machine learning model 700, or three or more hidden layers may be incorporated into machine learning model 700, where each layer includes the same or different number of nodes.

[0115] The result of each node within hidden layers 704 is applied to the transfer function of output layer 706. The transfer function may be liner or non-linear, depending on the number of layers within machine learning model 700. Example non-linear transfer functions may be a sigmoid function or a rectifier function. The output 707 of the transfer function may be a value or values indicative of recovery of a reconstructed high-fidelity optical signal and / or a blood pressure value of the patient. By applying a sensed low- fidelity optical signal to a machine learning model trained based on a sensed low-fidelity optical signal and a sensed high-fidelity optical signal during a first period of time, such as machine learning model 700, processing circuitry of system 100 is able to generate a reconstructed high-fidelity optical signal and / or a blood pressure value of the patient with great accuracy, specificity, and sensitivity.

[0116] FIG. 8 is an example of a machine learning model 700 being trained using supervised and / or reinforcement learning techniques. Machine learning model 700 may be implemented using any number of models for supervised and / or reinforcement learning, such as but not limited to, an artificial neural network, a decision tree, naive Bayes network, support vector machine, or k-nearest neighbor model, to name only a few examples. In some examples, processing circuitry one or more of computing system 110, IMD 104, and / or patient computing device 106 initially trains the machine learning model 700 based on training set data 800 including a sensed low-fidelity optical signal and a sensed high-fidelity optical signal during a first period of time. An output of the machine learning model 700 may be compared 804 to the target output 803, e.g., as determinedAtty Ref. No. A0012750W001 based on the label. Based on an error signal representing the comparison, the processing circuitry implementing a leaming / training function 805 may send or apply a modification to weights of machine learning model 700 or otherwise modify / update the machine learning model 700. For example, one or more of computing system 110, IMD 104, and / or patient computing device 106, may, for each training instance in the training set 800, modify machine learning model 700 to change a reconstructed high-fidelity optical signal and / or a blood pressure value generated by the machine learning model 700 in response to data applied to the machine learning model 700.

[0117] For aspects implemented in software, at least some of the functionality ascribed to the systems and devices described in this disclosure may be embodied as instructions on a computer-readable storage medium such as RAM, DRAM, SRAM, magnetic discs, optical discs, flash memories, or forms of EPROM or EEPROM. The instructions may be executed to support one or more aspects of the functionality described in this disclosure.

[0118] In addition, in some aspects, the functionality described herein may be provided within dedicated hardware and / or software modules. Depiction of different features as modules or units is intended to highlight different functional aspects and does not necessarily imply that such modules or units must be realized by separate hardware or software components. Rather, functionality associated with one or more modules or units may be performed by separate hardware or software components, or integrated within common or separate hardware or software components. Also, the techniques could be fully implemented in one or more circuits or logic elements. The techniques of this disclosure may be implemented in a wide variety of devices or apparatuses, including an IMD, an external programmer, a combination of an IMD and external programmer, an integrated circuit (IC) or a set of ICs, and / or discrete electrical circuitry, residing in an IMD and / or external programmer.

[0119] The following is a non-limiting list of examples that are in accordance with one or more aspects of this disclosure.

[0120] Example 1 : A system includes memory configured to store a machine learning (ML) model; processing circuitry coupled to the memory, the processing circuitry being configured to: receive a first optical signal sensed during a first period of time; receive a second optical signal via a first sensing device positioned at a location of a patient, the second optical signal being sensed by the first sensing device during the first period ofAtty Ref. No. A0012750W001 time, the second optical signal being a lower-fidelity optical signal than the first optical signal; and apply the first optical signal and the second optical signal to the ML model to train the ML model to generate at least one of a reconstructed third optical signal based on a third optical signal sensed via a second sensing device positioned at a location having a pulsatility below a pulsatility threshold, or a blood pressure value based on the third optical signal.

[0121] Example 2: The system of example 1, wherein the first optical signal is received via an additional sensing device positioned at a different location of the patient, the location of the first sensing device having a lower pulsatility than the different location of the additional sensing device.

[0122] Example 3: The system of example 1, wherein the first optical signal is received via the first sensing device operating in a high-fidelity mode.

[0123] Example 4: The system of any of examples 1-3, wherein the pulsatility threshold is 1.0%.

[0124] Example 5: The system of example 2, wherein the pulsatility threshold is based on the pulsatility of the location of the additional sensing device.

[0125] Example 6: The system of any of examples 1-5, wherein the processing circuitry is further configured to: apply the first optical signal and the second optical signal to the ML model to train the ML model to generate a blood pressure value based on the third optical signal.

[0126] Example 7: The system of any of examples 1-5, wherein the processing circuitry is further configured to apply the first optical signal and the second optical signal to the ML model to train the ML model to generate a reconstructed third optical signal based on the third optical signal.

[0127] Example 8: The system of example 7, wherein the processing circuitry is further configured to determine a blood pressure value of the patient based on the reconstructed third optical signal.

[0128] Example 9: The system of any of examples 1-8 wherein the processing circuitry is further configured to: receive the third optical signal, the third optical signal sensed during a second period of time, the second period of time being after the first period of time, the third optical signal being a lower-fidelity optical signal than the first optical signal; and apply the third optical signal to the trained ML model to generate atAtty Ref. No. A0012750W001 least one of a reconstructed third optical signal or a blood pressure value of the patient based on the third optical signal, wherein the reconstructed third optical signal has a higher-fidelity than the third optical signal.

[0129] Example 10: The system of any of examples 1-9, wherein the first sensing device is an implantable cardiac monitor (ICM).

[0130] Example 11 : The system of any of examples 1-10, wherein the first sensing device and the second sensing device are the same device.

[0131] Example 12: An implantable medical device (IMD) includes an optical sensor configured to sense a first IMD optical signal of a patient during a first period of time and a second IMD optical signal of the patient during a second period of time, the second period of time being after the first period of time; and circuitry configured to: output the first IMD optical signal to a computing device; receive a trained machine learning (ML) model, the trained ML model being trained based on the first IMD optical signal and a third optical signal, the third optical signal sensed during the first period of time to generate at least one of: a reconstructed second IMD optical signal based on the second IMD optical signal sensed via the IMD, the IMD being positioned at a location having a pulsatility below a pulsatility threshold; or a blood pressure value based on the second IMD optical signal; and apply the second IMD optical signal to the trained ML model to generate at least one of the reconstructed second IMD optical signal or the blood pressure value of the patient based on the second IMD optical signal, the reconstructed second IMD optical signal having a higher-fidelity than the second IMD optical signal.

[0132] Example 13: The IMD of example 12, wherein the third optical signal is received via an additional sensing device positioned at a higher pulsatility location than the IMD, the location of the sensing device having a lower pulsatility than the different location of the additional sensing device.

[0133] Example 14: The IMD of example 12, wherein the third optical signal is received via the IMD operating in a high-fidelity mode.

[0134] Example 15: The IMD of any of examples 12-14, wherein the circuitry is further configured to: apply the second IMD optical signal to the trained ML model to generate the reconstructed second IMD optical signal based on the second IMD optical signal.Atty Ref. No. A0012750W001

[0135] Example 16: The IMD of any of examples 15, wherein the circuitry is further configured to: determine the blood pressure value of the patient based on the reconstructed second IMD optical signal.

[0136] Example 17: The IMD of any of examples 12-14, wherein the circuitry is further configured to: apply the second IMD optical signal to the trained ML model to generate the blood pressure value of the patient based on the second IMD optical signal.

[0137] Example 18: The IMD of any of examples 12-17, wherein the pulsatility threshold is 1.0%.

[0138] Example 19: The IMD of example 13, wherein the pulsatility threshold is based on the pulsatility of a location of the additional sensing device.

[0139] Example 20: A method for training a machine learning (ML) model includes receiving a first optical signal during a first period of time; receiving a second optical signal via a first sensing device positioned at a location of a patient, the second optical signal being sensed by the first sensing device during the first period of time, the second optical signal being a lower-fidelity optical signal than the first optical signal; and applying the first optical signal and the second optical signal to the ML model to train the ML model to generate at least one of: a reconstructed third optical signal based on a third optical signal sensed via a second sensing device positioned at a location having a pulsatility below a pulsatility threshold, or a blood pressure value based on third optical signal.

[0140] Example 21 : The method of example 20, wherein the first optical signal is received via an additional sensing device positioned at a different location of the patient, the location of the first sensing device having a lower pulsatility than the different location of the additional sensing device.

[0141] Example 22: The method of example 20, wherein the first optical signal is received via the first sensing device operating in a high-fidelity mode.

[0142] Example 23: The method of any of examples 20-22, wherein the pulsatility threshold is 1.0%.

[0143] Example 24: The method of example 21, wherein the pulsatility threshold is based on a pulsatility of the of location of the additional sensing device.Atty Ref. No. A0012750W001

[0144] Example 25: The method of any of examples 20-24, further includes applying the first optical signal and the second optical signal to the ML model to train the ML model to generate the blood pressure value based on the third optical signal.

[0145] Example 26: The method of any of examples 20-24, further includes applying the first optical signal and the second optical signal to the ML model to train the ML model to generate the reconstructed third optical signal based on the third optical signal.

[0146] Example 27: The method of example 26, further includes determining the blood pressure value of the patient based on the reconstructed third optical signal.

[0147] Example 28: The method of any of examples 20-27, further includes receiving the third optical signal, the third optical signal sensed during a second period of time, the second period of time being after the first period of time, the third optical signal being a lower-fidelity optical signal than the first optical signal; and applying the third optical signal to the trained ML model to generate at least one of the reconstructed third optical signal or a blood pressure value of the patient based on the third optical signal, wherein the reconstructed third optical signal has a higher-fidelity than the third optical signal.

[0148] Example 29: The method of any of examples 20-27, wherein the first sensing device is an implantable cardiac monitor (ICM).

[0149] Example 30: The method of any of examples 20-29, wherein the wherein the first sensing device and the second sensing device are the same device.

[0150] Example 31 : A computer-readable medium comprising instructions that, when executed, causes processing circuitry of a medical system to perform the method of any of examples 20-30.

[0151] Various aspects of the disclosure have been described. These and other aspects are within the scope of the following claims.

Claims

Atty Ref. No. A0012750W001WHAT IS CLAIMED IS:

1. A system comprising: memory configured to store a machine learning (ML) model; processing circuitry coupled to the memory, the processing circuitry being configured to: receive a first optical signal sensed during a first period of time; receive a second optical signal via a first sensing device positioned at a location of a patient, the second optical signal being sensed by the first sensing device during the first period of time, the second optical signal being a lower- fidelity optical signal than the first optical signal; and apply the first optical signal and the second optical signal to the ML model to train the ML model to generate at least one of: a reconstructed third optical signal based on a third optical signal sensed via a second sensing device positioned at a location having a pulsatility below a pulsatility threshold, or a blood pressure value based on the third optical signal.

2. The system of claim 1, wherein the first optical signal is received via an additional sensing device positioned at a different location of the patient, the location of the first sensing device having a lower pulsatility than the different location of the additional sensing device.

3. The system of claim 1, wherein the first optical signal is received via the first sensing device operating in a high-fidelity mode.

4. The system of any of claims 1-3, wherein the pulsatility threshold is 1.0%.

5. The system of claim 2, wherein the pulsatility threshold is based on the pulsatility of the location of the additional sensing device.Atty Ref. No. A0012750W0016. The system of any of claims 1-5, wherein the processing circuitry is further configured to: apply the first optical signal and the second optical signal to the ML model to train the ML model to generate a blood pressure value based on the third optical signal.

7. The system of any of claims 1-5, wherein the processing circuitry is further configured to apply the first optical signal and the second optical signal to the ML model to train the ML model to generate a reconstructed third optical signal based on the third optical signal.

8. The system of claim 7, wherein the processing circuitry is further configured to determine a blood pressure value of the patient based on the reconstructed third optical signal.

9. The system of any of claims 1-8 wherein the processing circuitry is further configured to: receive the third optical signal, the third optical signal sensed during a second period of time, the second period of time being after the first period of time, the third optical signal being a lower-fidelity optical signal than the first optical signal; and apply the third optical signal to the trained ML model to generate at least one of a reconstructed third optical signal or a blood pressure value of the patient based on the third optical signal, wherein the reconstructed third optical signal has a higher-fidelity than the third optical signal.

10. The system of any of claims 1-9, wherein the first sensing device is an implantable cardiac monitor (ICM).

11. The system of any of claims 1-10, wherein the first sensing device and the second sensing device are the same device.

12. A method for training a machine learning (ML) model, the method comprising:Atty Ref. No. A0012750W001 receiving a first optical signal during a first period of time; receiving a second optical signal via a first sensing device positioned at a location of a patient, the second optical signal being sensed by the first sensing device during the first period of time, the second optical signal being a lower-fidelity optical signal than the first optical signal; and applying the first optical signal and the second optical signal to the ML model to train the ML model to generate at least one of a reconstructed third optical signal based on a third optical signal sensed via a second sensing device positioned at a location having a pulsatility below a pulsatility threshold, or a blood pressure value based on third optical signal.

13. The method of claim 12, wherein the first optical signal is received via an additional sensing device positioned at a different location of the patient, the location of the first sensing device having a lower pulsatility than the different location of the additional sensing device.

14. The method of any of claims 12-13, further comprising: applying the first optical signal and the second optical signal to the ML model to train the ML model to generate the reconstructed third optical signal based on the third optical signal.

15. The method of any of claims 12-14, further comprising: receiving the third optical signal, the third optical signal sensed during a second period of time, the second period of time being after the first period of time, the third optical signal being a lower-fidelity optical signal than the first optical signal; and applying the third optical signal to the trained ML model to generate at least one of the reconstructed third optical signal or a blood pressure value of the patient based on the third optical signal, wherein the reconstructed third optical signal has a higher-fidelity than the third optical signal.

Images