EP4633448A1 - Methods and systems for automated calibration of local field potential sensing tools for deep brain stimulation - Google Patents

Abstract

A system for automated calibration of Local Field Potential (LFP) sensing for deep brain stimulation can comprise one or more electrodes and control circuitry. The control circuitry is configured to, prior to delivery of electrical stimulation via the one or more electrodes, sense, using the one or more electrodes, LFPs of the brain, record a snapshot of LFP activity across a wide frequency band, determine a peak frequency of the LFP activity from the snapshot, and update a Frequency of Interest (FOI) based on the peak frequency.

Description

METHODSAND SYSTEMS FOR AUTOMATED CALIBRATION OF LOCAL FIELD POTENTIAL SENSING TOOLS FOR DEEP BRAIN STIMULATION

[0001] This Application claims priority from U.S. Provisional Patent Application 63/433,331, filed 16 December 2022, the entire content of which is incorporated herein by reference.

FIELD

[0002] The present technology is generally related to automated calibration of Local Field Potential (LFP) sensing tools, aiding in the set up and application of LFP sensing in the practice of deep brain stimulation (DBS).

BACKGROUND

[0003] Implantable medical devices, such as electrical stimulators or therapeutic agent delivery devices, have been proposed for use in different therapeutic applications, such as deep brain stimulation. In some therapy systems, an implantable electrical stimulator delivers electrical therapy to a target tissue site within a patient with the aid of one or more electrodes, which may be deployed by medical leads and/or on a housing of the electrical stimulator, or both. In some therapy systems, therapy may be delivered via particular combinations of the electrodes carried by leads and/or by the housing of the electrical stimulator.

[0004] During programming sessions, which may occur during implant of the medical device, during a trial session, or during an in-clinic or remote follow-up session after the medical device is implanted in the patient, a clinician may generate one or more therapy programs (also referred to as therapy parameter sets) that are found to provide efficacious therapy to the patient, where each therapy program may define values for a set of therapy parameters.

[0005] In the case of DBS, DBS systems are capable of recording Local Field Potentials (LFPs) from the brain but require manual calibration by the user. In particular, a Frequency of Interest (FOI) or a set of FOIs (one unique FOI per lead) must be correctly chosen in order for the various sensing tools to function properly. This presents a persistent challenge to the user, as the FOI may or may not be visible to select at the time of setup.

[0006] Conventionally LFPs are manually calibrated by a clinician in follow-up sessions by interpreting LFP snapshots that were previously, and manually triggered by the patient, or by observing the correlation of the FOI power being tracked with patient clinical symptom severity. This presents a persistent challenge to the user as the identification of appropriate LFP calibration often require several follow-up clinician visits to determine the efficacy of previously set FOI. Accordingly, one of greatest barriers to recording valuable LFP data for DBS is the burden of the setup process.

[0007] Accordingly, there is a need to improve the efficiency of calibrating LFP sensing tools while reducing user burden.

SUMMARY

[0008] The techniques of this disclosure generally relate to automated calibration of LFP sensing tools prior to delivery of DBS therapy. This calibration occurs during the period between neurostimulator implant and initial programming and enables detection of the patient's FOI and automatic adjustment of the system's tracking to that frequency, improving the efficiency of the initial programming visit and reducing the patient burden.

[0009] In one aspect, the present disclosure provides a system for automated calibration of Local Field Potential (LFP) sensing for deep brain stimulation. The system can comprise one or more electrodes and control circuitry. The control circuitry is configured to, prior to delivery of electrical stimulation via the one or more electrodes, sense, using the one or more electrodes, LFPs of the brain at a Frequency of Interest (FOI), capture a snapshot of LFP activity across a wide frequency band, determine a frequency of peak LFP activity from the snapshot, and update the FOI based on the frequency of peak activity.

[0010] In another aspect, the disclosure provides a method for automated calibration of LFP sensing for deep brain stimulation. The method comprises, prior to delivery of electrical stimulation, sensing, using one or more electrodes, LFPs of the brain at an FOI, capturing a snapshot of LFP activity at a wide frequency band, determining a frequency of peak LFP activity from the snapshot, and updating the FOI based on the frequency of peak activity.

[0011] In yet another aspect, the disclosure provides for a method of calibrating a series of electrodes. Different electrode combinations can be selected from the series of electrodes as sensing electrodes. The sensing electrodes can then be automatically updated based on which electrodes had the largest LFP activity in the FOI. Thus, it should be appreciated that the described operations can be repeated across several electrodes such that various electrode combinations can be evaluated sequentially or simultaneously.

[0012] The details of one or more aspects of the disclosure are set forth in the accompanying drawings and the description below. Other features, objects, and advantages of the techniques described in this disclosure will be apparent from the description and drawings, and from the claims.

BRIEF DESCRIPTION OF DRAWINGS

[0013] FIG. l is a conceptual diagram illustrating an example system that provides deep brain stimulation, according to an embodiment.

[0014] FIG. 2 is a conceptual diagram illustrating an example system that provides robust adaptive brain stimulation, according to an embodiment.

[0015] FIG. 3A is a schematic illustrating an example deep brain stimulation (DBS) system configured to deliver electrical stimulation therapy to a tissue site within a brain of a patient, according to an embodiment.

[0016] FIG. 3B is a block diagram illustrating components of the system of FIG. 3 A, according to an embodiment.

[0017] FIG. 4 is a LFP snapshot of a DBS system, according to an embodiment.

[0018] FIG. 5 is a flowchart of a method for calibrating LFP sensing tools, according to an embodiment.

[0019] While various embodiments are amenable to various modifications and alternative forms, specifics thereof have been shown by way of example in the drawings and will be described in detail. It should be understood, however, that the intention is not to limit the claimed inventions to the particular embodiments described. On the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the subject matter as defined by the claims.

DETAILED DESCRIPTION

[0020] Embodiments of the present disclosure enable automated calibration of Local Field Potential (LFP) sensing tools in the practice of Deep Brain Stimulation (DBS) by leveraging full spectrum LFP snapshot data collected during the timeframe between implant of a neurostimulator and a patient's initial programming. During this discovery phase before initial programming, the neurostimulator, or the patient, can trigger full-spectrum LFP snapshots, revealing the correct Frequency of Interest (FOI) data that should be tracked. This tracked FOI data in addition to further full spectrum LFP snapshot data can then be used to automatically update the FOI to complete the setup process. The increased detail in the data acquired during this discovery phase could also inform further outputs at the initial programming visit, including circadian rhythm detection, threshold suggestions, and artifact detection or mitigation.

[0021] LFP Snapshots reveal activity across a wide spectrum of frequencies, which can be used to reveal which frequencies the longitudinal tools (e.g., Timeline, Streaming) should track (FOI). The longitudinal tracking tools, that are based on the FOI, plot the power of the FOI over time. Accordingly, an LFP snapshot is a more acute (30s) assessment of the many frequencies that could be present at that time that can be leveraged in embodiments of the present disclosure to automatically adjust the FOI without a clinician visit, speeding up and simplifying the calibration process for the clinician and patient.

[0022] Used herein, LFP snapshots refer to a wide band LFP power spectral density recording and an LFP timeline is integrated power in a defined frequency band around the peak FOI. One advantage of the wide-band LFP snapshot data is that it is agnostic of FOI, and captures all potential FOI at that time. LFP snapshots can be recorded via sensing electrodes and an implanted medical device as contemplated herein.

[0023] Referring to FIG. 1, a conceptual diagram illustrating an example system 100 that provides deep brain stimulation and can be configured to automatically calibrate LFP of a sensing engine is depicted, according to an embodiment. In an embodiment, system 100 comprises an implantable medical device (IMD) 102 for a patient 104.

[0024] IMD 102 delivers neurostimulation therapy to patient 104 at a target tissue site. For example, one or more leads may extend from IMD 102 to the brain of patient 104, and IMD 102 may deliver deep brain stimulation (DBS) therapy to patient 104 to, for example, treat neurodegenerative diseases or traumas.

[0025] IMD 102 is configured to deliver therapy according to one or more programs or control policies. A control policy includes one or more parameters that define an aspect of the therapy delivered by the medical device according to that control policy. For example, a control policy that controls delivery of stimulation by IMD 102 in the form of pulses may define a voltage or current pulse amplitude, a pulse width, a pulse rate, an electrode configuration or field shape, or a schedule or pattern for stimulation pulses delivered by IMD 102 according to that control policy. Further, each of the leads includes electrodes, and the parameters for a control policy that controls delivery of stimulation therapy by IMD 102 can include information identifying which electrodes have been selected for delivery of pulses according to the program, and the polarities of the selected electrodes, i.e., the electrode configuration for the control policy. Control policies that control delivery of other therapies by IMD 102 can include other parameters, as required by the particular policy. A selected control policy of IMD 102 is implemented based on LFP sensing and, in particular, FOI. Accordingly, optimal DBS relies on calibration of an LFP sensing engine of IMD 102. [0026] In embodiments, such as that depicted, system 100 further comprises an external device 106. External device 106 is communicatively coupled to IMD 102 via wireless communication. External device 106 can be a device for inputting information relating to patient 104, programming IMD 102, receiving information from IMD 102, and updating IMD 102 such as updating FOI.

[0027] Referring also to FIG. 2, a conceptual diagram illustrating an example system that provides robust adaptive brain stimulation is depicted, according to an embodiment. As depicted, external devices 106a and 106b are further illustrated as a handheld computing device and a laptop computing device but can further be a key fob or a wristwatch, smart phone, computer workstation, or networked computing device, and the like. Embodiments of systems can further include a server 108 and/or a database 110, as illustrated.

[0028] Server 108 can include one or more servers in a cloud computing environment. Server 108 can be configured to communicate with external devices 106a and/or 106b, and/or IMD 102 via wireless communication. In embodiments, server 108 can be co-located with external devices 106a and/or 106b, or located elsewhere, such as in a cloud computing data center or medical clinic.

[0029] Database 110 is configured to store data related to system 100, including IMD 102 and/or external device 106 data. In embodiments, database 110 can be integrated as part of server 108, or be standalone such that server 110 and/or external device 106a and 106b can be communicatively coupled to database 110. Database 110 can be a general- purpose database management storage system (DBMS) or relational DBMS as implemented by, for example, Oracle, IBM DB2, Microsoft SQL Server, PostgreSQL, MySQL, SQLite, Linux, or Unix solutions.

[0030] One purpose of database 110 is to store LFP snapshot data and programmer event tracking that can be used to update FOI or verify FOI thresholds, as necessary. LFP activity communicated to server 108 can be an effective way to assess efficacy of settings and train a machine learning algorithm (MLA) to auto-extract noise from an LFP timeline.

[0031] In embodiments database 110 stores LFP snapshots, LFP timeline data, and stimulation timeline data. This data could them be used to train and automatically set LFP thresholds and stimulation limits. For example, upper and lower stimulation limits for adaptive DBS could be automatically set based on how the patient uses and adapts their stimulation.

[0032] Referring to FIG. 3 A, an alternate illustration of system 100 is depicted. In the example shown in FIG. 3 A, therapy system 100 includes a medical device programmer as external device 106 and a neurostimulator as implantable medical device (IMD) 102, lead extension 112, and one or more leads 114 with respective sets of electrodes 116. IMD 102 includes a stimulation generator configured to generate and deliver electrical stimulation therapy to one or more regions of brain of patient 104 via one or more electrodes 116 of one or more leads 114, respectively.

[0033] DBS can be used to treat or manage various patient conditions, such as, but not limited to, seizure disorders (e.g., epilepsy), pain, migraine headaches, psychiatric disorders (e.g., major depressive disorder (MDD), bipolar disorder, anxiety disorders, post-traumatic stress disorder, dysthymic disorder, and obsessive compulsive disorder (OCD)), behavior disorders, mood disorders, memory disorders, mentation disorders, movement disorders (e.g., essential tremor or Parkinson's disease), Huntington's disease, Alzheimer's disease, or other neurological or psychiatric disorders and impairment of patient 104.

[0034] Leads 114 can be positioned to deliver electrical stimulation therapy to one or more target tissue sites within the brain to manage patient symptoms associated with a disorder of patient 104. Leads 114 may be implanted to position electrodes 116 at desired locations of the brain via any suitable technique, such as through respective burr holes in the skull of patient 104 or through a common burr hole in the cranium. Leads 114 can be placed at any location within the brain such that electrodes 116 are capable of providing electrical stimulation to target therapy delivery sites within the brain during treatment. Different neurological, motor, or psychiatric disorders can be associated with activity in one or more of regions of the brain, which may differ between patients. Accordingly, the target therapy delivery site for electrical stimulation therapy delivered by leads 114 may be selected based on the patient condition. For example, a suitable target therapy delivery site within the brain for controlling a movement disorder of patient 104 may include one or more of the pedunculopontine nucleus (PPN), thalamus, basal ganglia structures (e.g., globus pallidus, substantia nigra or subthalamic nucleus (STN)), zona inserta, fiber tracts, lenticular fasciculus (and branches thereof), ansa lenticularis, or the Field of Forel (thalamic fasciculus). The PPN may also be referred to as the pedunculopontine tegmental nucleus. [0035] As previously described, IMD 102 can deliver electrical stimulation therapy to the brain of patient 104 according to one or more control policies. A control policy may define one or more electrical stimulation parameter values for therapy generated by a stimulation generator of IMD 102 and delivered from IMD 102 to a target therapy delivery site within patient 104 via one or more electrodes 116. The electrical stimulation parameters may define an aspect of the electrical stimulation therapy, and may include, for example, voltage or current amplitude of an electrical stimulation signal, a charge level of an electrical stimulation, a frequency of the electrical stimulation signal, waveform shape, on/off cycling state (e.g., if cycling is “off,” stimulation is always on, and if cycling is “on,” stimulation is cycled on and off) and, in the case of electrical stimulation pulses, pulse rate, pulse width, and other appropriate parameters such as duration or duty cycle. In addition, if different electrodes are available for delivery of stimulation, a therapy parameter of a therapy program can be further characterized by an electrode combination, which can define selected electrodes 116 and their respective polarities. In some examples, stimulation may be delivered using a continuous waveform and the stimulation parameters can define this waveform.

[0036] In addition to being configured to deliver therapy to manage a disorder of patient 104, therapy system 100 can be configured to sense bioelectrical brain signals or another physiological parameter of patient 104. For example, IMD 102 can include a sensing engine that is configured to sense bioelectrical brain signals within one or more regions of the brain via electrodes 116. Accordingly, in some examples, electrodes 116 can be used to deliver electrical stimulation to target sites within the brain as well as sense brain signals. However, IMD 102 can also use a separate set of sensing electrodes to sense the bioelectrical brain signals. In some examples, the sensing engine of IMD 102 can sense bioelectrical brain signals via one or more of the electrodes 116 that are also used to deliver electrical stimulation to the brain. In other examples, one or more of electrodes 116 can be used to sense bioelectrical brain signals while one or more different electrodes 116 can be used to deliver electrical stimulation. The sensing engine can be used to record LFP snapshots.

[0037] Notably, during the time between implant of the neurostimulator and the patient's initial programming, stimulation is not yet turned on for the patient and the user has increased flexibility in selecting a sensing configuration. Optimal sensing electrodes can be selected with no regard for the location of the inactive stimulation electrodes and an arbitrary FOI can be selected for tracking. With the patient at home during this discovery phase, they can trigger full-spectrum LFP snapshots, revealing the correct FOI that should be tracked. At the initial programming, the clinician can confirm the auto-detected FOI and enable stimulation therapy.

[0038] External medical device programmer 106 is configured to wirelessly communicate with IMD 102 as needed to provide or retrieve therapy information. Programmer 106 is an external computing device that the user, e.g., the clinician and/or patient 104, can use to communicate with IMD 102. For example, programmer 106 can be a clinician programmer that the clinician uses to communicate with IMD 102 and program one or more therapy programs for IMD 102. In addition, or instead, programmer 106 can be a patient programmer that allows patient 104 to select programs and/or view and modify therapy parameter values. The clinician programmer can include more programming features than the patient programmer. In other words, more complex or sensitive tasks may only be allowed by the clinician programmer to programmer event an untrained patient from making undesired changes to IMD 102.

[0039] Programmer 106 can be a hand-held computing device with a display viewable by the user and an interface for providing input (i.e., a user input mechanism). For example, programmer 106 can include a small display screen (e.g., a liquid crystal display (LCD) or a light emitting diode (LED) display) that presents information to the user. In addition, programmer 106 can include a touch screen display, keypad, buttons, a peripheral pointing device, voice activation, or another input mechanism that allows the user to navigate through the user interface of programmer 106 and provide input. If programmer 106 includes buttons and a keypad, the buttons may be dedicated to performing a certain function, e.g., a power button, the buttons and the keypad may be soft keys that change in function depending upon the section of the user interface currently viewed by the user, or any combination thereof.

[0040] In other examples, programmer 106 can be a larger workstation or a separate application within another multi -function device, rather than a dedicated computing device. For example, the multi -function device may be a notebook computer, tablet computer, workstation, one or more servers, cellular phone, personal digital assistant, or another computing device that may run an application that enables the computing device to operate as a secure medical device programmer 106. A wireless adapter coupled to the computing device can enable secure communication between the computing device and IMD 102.

[0041] When programmer 106 is configured for use by the clinician, programmer 106 may be used to transmit programming information to IMD 102. Programming information can include, for example, hardware information, such as the type of leads 114, the arrangement of electrodes 116 on leads 114, the position of leads 114 within the brain, one or more therapy programs defining therapy parameter values, therapeutic windows for one or more electrodes 116, and any other information that may be useful for programming into IMD 102. Programmer 106 can also be capable of completing functional tests (e.g., measuring the impedance of electrodes 116 of leads 114) and confirming or adjusting autodetecting FOI.

[0042] The clinician can also generate and store therapy programs within IMD 102 with the aid of programmer 106. Programmer 106 can assist the clinician in the creation/identification of therapy programs by providing a system for identifying potentially beneficial therapy parameter values. For example, during a programming session, programmer 106 may automatically select a combination of electrodes for delivery to therapy to the patient.

[0043] Programmer 106 can also be configured for use by patient 104. When configured as a patient programmer, programmer 106 can have limited functionality (compared to a clinician programmer) in order to prevent event patient 104 from altering critical functions of IMD 102 or applications that may be detrimental to patient 104.

[0044] Whether programmer 106 is configured for clinician or patient use, programmer 106 is configured to communicate with IMD 102 and, optionally, another computing device, via wireless communication. Programmer 106, for example, may communicate via wireless communication with IMD 102 using radio frequency (RF) and/or inductive telemetry techniques known in the art, which can comprise techniques for proximal, mid-range, or longer-range communication. Programmer 106 can also communicate with another programmer or computing device via a wired or wireless connection using any of a variety of local wireless communication techniques, such as RF communication according to 802.11 or Bluetooth specification sets, infrared (IR) communication, or other standard or proprietary telemetry protocols. Programmer 106 can also communicate with other programming or computing devices via exchange of removable media, such as magnetic or optical disks, memory cards, or memory sticks. Further, programmer 106 can communicate with IMD 102 and another programmer via remote telemetry techniques known in the art, communicating via a personal area network (PAN), a local area network (LAN), wide area network (WAN), public switched telephone network (PSTN), or cellular telephone network, for example.

[0045] Referring to FIG. 3B, a block diagram illustrating components of IMD 102 of FIG. 3 A is depicted, according to an embodiment. In an embodiment, IMD 102 generally includes a processor 118, memory 120, a stimulation generator 122, a sensing engine 124, a power source 126, and a telemetry engine 128.

[0046] Processor 118 can include one or more digital signal processors (DSPs), general purpose microprocessors, application specific integrated circuits (ASICs), field programmable logic arrays (FPGAs), or other equivalent integrated or discrete logic circuitry, or combinations thereof. The functions attributed to processors described herein may be provided by a hardware device and embodied as software, firmware, hardware, or any combination thereof. Processor 118 is configured to control stimulation generator 122 according to therapy programs stored by memory 120 to apply particular stimulation parameter values specified by one or more programs, such as amplitude, pulse width, and pulse rate.

[0047] Memory 120 can be operably coupled to processor 118 and can include any volatile or non-volatile media, such as a random access memory (RAM), read only memory (ROM), non-volatile RAM (NVRAM), electrically erasable programmable ROM (EEPROM), flash memory, and the like. Memory 120 can store computer-readable instructions that, when executed by processor 118, cause IMD 102 to perform various functions described herein. [0048] In an embodiment, memory 120 can store therapy programs, operating instructions, and the like. Each stored therapy program defines a particular program of therapy in terms of respective values for electrical stimulation parameters, such as an electrode combination, current or voltage amplitude, and, if stimulation generator 122 generates and delivers stimulation pulses, the therapy programs can define values for a pulse width, and pulse rate of a stimulation signal. Each stored therapy program can also be referred to as a set of stimulation parameter values. Operating instructions guide general operation of IMD 102 under control of processor 118 and can include instructions for monitoring brain signals within one or more brain regions via electrodes 116 and delivering electrical stimulation therapy to patient 104.

[0049] Stimulation generator 122, under the control of processor 118, is configured to generate stimulation signals for delivery to patient 104 via selected combinations of electrodes 116.

[0050] Sensing engine 124, under the control of processor 118, is configured to sense bioelectrical brain signals of patient 104 via electrodes 116. In some examples, the bioelectrical brain signals can reflect changes in electrical current produced by the sum of electrical potential differences across brain tissue. Examples of neurological brain signals include, but are not limited to, electrical signals generated from LFP sensed within one or more regions of the brain, such as an electroencephalogram (EEG) signal, or an electrocorti cogram (ECoG) signal. LFP, however, may include a broader genus of electrical signals within the brain of patient 104.

[0051] Sensing engine 124 is configured to conduct LFP snapshots, such as that depicted in FIG. 4. LFP snapshots represent a wide spectrum of frequency powers for a sensed channel using one or more electrodes. Detection of peak frequencies over time can enable system 100 to identify and confirm patient-specific FOI.

[0052] Power source 126 delivers operating power to various components of IMD 102. Power source 126 can include a small rechargeable or non-rechargeable battery and a power generation circuit to produce the operating power. Recharging can be accomplished through proximal inductive interaction between an external charger and an inductive charging coil within IMD 102. In some examples, power requirements may be small enough to allow IMD 102 to utilize patient motion and implement a kinetic energy-scavenging device to trickle charge a rechargeable battery. In other examples, traditional batteries may be used for a limited period of time.

[0053] Telemetry engine 128 is configured to support wireless communication between IMD 102 and an external programmer 106 or another computing device under the control of processor 118. Processor 118 can receive, as updates to programs, values for various stimulation parameters such as amplitude and electrode combination, from programmer 106 via telemetry engine 128.

[0054] As depicted, system 100 can further comprise electrodes 116 of lead 114 that includes electrodes 116A-116D. Processor 118 can apply the stimulation signals generated by stimulation generator 122 to a selected combination of electrodes 116A-116D.

[0055] Embodiments of the present disclosure address shortcomings of conventional LFP sensing of DBS systems that are limited to user entered programmer event tracking prior to the initial programming visit. In conventional systems a default or arbitrary FOI is set for the patient following implantation. The clinicians then rely on patients to manually record one to several programmer events (e.g., eating, taking medication) during which an LFP snapshot is captured (i.e. recorded). This process carries a high patient burden during a period when the patient is not receiving therapy and is recovering from surgery. Further, since observation of captured LFP snapshots and modification of the FOI can only occur during a clinician session, the clinician must often update the FOI across several patient follow-up visits until an optimal FOI is determined.

[0056] In contrast, system 100 is configured to conduct automated discovery of FOI by generating default programmer events, taking timer-driven automatic snapshots of LFP activity, auto-adjusting FOI based on optimal LFP peak frequency. In embodiments, system 100 can run electrocardiograms (ECGs) periodically (e.g., daily) to learn and extract noise from the timeline. Patients with poor circadian patterns can be flagged for further analysis according to embodiments. Accordingly, system 100 enables clinicians to assess efficacy of settings sooner and requires less patient follow-up visits to interpret timeline and programmer event data and verify FOI. System 100 additionally enables ongoing automated honing of sensing based on LFP triggered snapshots (e.g. Circadian Wake-up) after clinician visits.

[0057] System 100 can dynamically trigger IMD 102 to record LFP snapshots based on one or more of a timer, time of day, accelerometer, and observance of changes within FOI. Observance of changes within FOI can indicate medication wash-in and/or out, the patient is asleep or awake, or the patient is engaging in physical activity.

[0058] In embodiments, determining patient-specific FOI prior to initial programming can be accomplished by applying a machine learning algorithm (MLA) to system 100. Embodiments of the present disclosure are operable to detect and classify FOI peaks associated with LFP activity without relying on conventional means of programmer event detection such as instructing users to manually record programmer events.

[0059] The inventors of the present disclosure have discovered that FOI can be automatically extracted from periodic LFP snapshots by machine learning approaches such as, for example, neural networks, to refine the spectrum of frequencies captured by each LFP snapshot prior to initial programming by a clinician. This adjustment can better identify FOI prior to an initial programming session and confirmation by a clinician.

[0060] In embodiments an MLA can be trained to recognize LFP snapshot triggers, particularly for changes within FOI that indicate patient events (e.g. sleeping, exercising). Such recognition can be accomplished by computing similarity metrics for these patient events using correlation or machine learning regression algorithms. For example, if the similarity of detected changes within FOI across LFP snapshots to training data of a programmer event, such as medication washi-in, is above a certain threshold, (e.g., 75%, 90%, 95% or 99% similarity) a matching process can determine that the detected change represents the programmer event. These detected programmer events can be presented to the patient for confirmation and/or flagged for observation by the clinician during a programming session.

[0061] Programmer event detection can be implemented by observing patient-specific FOI changes. For example, if a patient manually enters a “took medication” programmer event the observed changes in FOI can be compared to future changes in order to suggest or identify future “took medication” programmer events that are not be identified by the patient.

[0062] Similarly, programmer event detection can be applied across a database of patients. For example, the MLA can suggest an approach to therapy that has been historically successful for other patients to a clinician when a particular trend in LFP snapshots is observed. Additionally, observed trends across patients can be used to understand features of LFP snapshots that indicate disease progression. Updates to stimulation can be automatically applied based on what has been seen in other patients.

[0063] In embodiments the MLA can extract characteristics from frequency bands to better identify and confirm frequency peaks.

[0064] MLA techniques can be applied to labelled (supervised) or unlabeled (unsupervised) LFP snapshot data. Further, a classifier can take in parameters such as model of IMD and type of patient programmer event. For example, frequency peaks can be separated based on patient programmer events and frequency peaks within programmer events can each be compared using image recognition. Classifying the image recognition analysis can improve accuracy in some circumstances by limiting the influence of outlying data and differences between patients.

[0065] In operation, the sensed electric signals can be processed by the MLA determine contextual information for a detected frequency peak. In other words, the MLA can allow for processing of patient programmer events by maintaining state information over time and prompting or automatically categorizing patient programmer events based on historical patterns, time of day, or other factors. In one example, the MLA can detect the patient sleeping based on low signal and capture a snapshot. Similarly, if multiple LFP snapshots are being produced over a short window of time, they can be contributed to a common programmer event. Thus, the context surrounding an LFP snapshot can contribute to personalized Al insights that accounts for wide variation in FOI across patients.

[0066] In operation, programmer event prompts can be delivered to an external device 106. Programmer event prompts can ask the patient to verify a detected programmer event (e.g., eating or heightened physical activity) and can be predefined according to embodiments. Verified classification of detected programmer events can streamline clinician review and simplify LFP timeline analysis.

[0067] Referring to FIG. 5, a flowchart of a method 200 for automatically calibrating LFP sensing tools in conjunction with DBS is depicted according to an embodiment. Method 200 can collect baseline brain data and implement patient-specific FOI without user input. In embodiments method 200 can be implemented via a DBS system such as system 100. In such embodiments, method 200 can be ran locally via processor 118 or remotely via server 108. [0068] At 202 a machine learning model is trained to interpret LFP snapshots and/or an LFP timeline as described herein. In embodiments the machine learning model can be trained to interpret frequency peaks based on one or more of severity, number of occurrences within a particular frequency band, time of day, patient programmer event, and clinician or patient entered information.

[0069] At 204 LFP sensing is enabled. In embodiments, LFP sensing is enabled in a post anesthesia care unit (PACU) post implant. In such embodiments a default patient programmer event can be used to capture a first LFP snapshot and categorize the patient’s LFP peak. For example, this default event can be labeled “took medications.”

[0070] At 206 LFP activity is monitored by automatically capturing LFP snapshots over time. In embodiments LFP snapshots can be timer driven such that a snapshot is taken periodically (e.g., every minute, every 15 minutes, every hour, every five hours). LFP snapshots can further be captured (i.e. recorded) for each detected, or patient-entered, patient programmer event.

[0071] In embodiments, the FOI power overtime can be longitudinally tracked to create a timeline. The LFP snapshots can provide a glimpse as to which frequency should be tracked, and then the timeline is configured to plot the power of that particular frequency as it changes. The LFP snapshots that are captured are then also layered onto this timeline.

[0072] A power fluctuation threshold can be used in conjunction with the timeline to determine if fluctuations in power of the FOI that has been automatically selected (e.g., by the peak detection described herein) is significant enough such that it is the frequency peak should be tracked. Put differently, a power fluctuation threshold can act as a verification mechanism to ensure the proper FOI is selected. The timeline of FOI power over time can thus be utilized by an MLA (or be presented to a clinician) to confirm a suggested FOI.

[0073] At 208, new and/or collective data recorded by these wide-spectrum LFP snapshots allows the timeline FOI to be automatically adjusted, such that the power of the FOI is being tracked passively and accurately while the patient is still at home.

[0074] At 210, upon initial programming the clinician has the opportunity to confirm and utilize the automatically determined FOI (as long as the tracked timeline data for that FOI looks satisfactory). Stimulation can then be applied based on the confirmed FOI.

[0075] Optionally, at 212 the machine learning model can be updated based on the FOI selected by a clinician and trends observed in LFP snapshot data. [0076] It should be understood that the individual operations used in the methods of the present teachings may be performed in any order and/or simultaneously, as long as the teaching remains operable. Furthermore, it should be understood that the apparatus and methods of the present teachings can include any number, or all, of the described embodiments, as long as the teaching remains operable.

[0077] Embodiments of the present disclosure accordingly provide earlier collection of LFP data based on current Frequency of Interest (FOI). This LFP data is made available to the clinician and patient at the initial programming visit and can inform further outputs, including poor circadian rhythm detection for follow up by the clinician and DBS therapy threshold adjustment.

[0078] Thus, embodiments described herein reduce the technical barriers to entry for patients and clinicians and provide them with more immediate and effortless clinical value. Collection and adjustment of LFP data leverages the sensing capabilities of neurostimulators to aid in the setup, application, and adoption of that LFP data.

[0079] The ongoing collection of LFP data and refinement of FOI further aid in future adjustment of DBS and LFP thresholds. In embodiments, if a certain portion of LFP limits is outside of a threshold difference (e.g., more than 25% of the time) the clinician can adapt control policies accordingly.

[0080] In embodiments, ECG can be ran periodically during the discovery phase prior to the initial programming. In such embodiments, ECG can be ran daily. Over time the IMD processor can learn ECG patterns and automatically extract noise from the timeline. Additional ECG data collected prior to initial programming can aid the clinician in detecting and mitigating ECG artifacts.

[0081] It should be appreciated that while described with respect to determining FOI, embodiments of the present disclosure can be similarly applied to automatically determine the LFP threshold and the upper and lower stimulation limits based on the timeline data recorded during the continuous DBS phase.

[0082] Embodiments of the present disclosure can be implemented over different electrode combinations to automatically update the sensing electrodes based on which electrodes had the largest LFP activity in the FOI. The sensing electrodes can be updated as more LFP activity is recorded. Thus, it should be appreciated that the described operations can be repeated across several electrodes such that various electrode combinations can be evaluated sequentially or simultaneously.

[0083] It should be understood that various aspects disclosed herein may be combined in different combinations than the combinations specifically presented in the description and accompanying drawings. It should also be understood that, depending on the example, certain acts or programmer events of any of the processes or methods described herein may be performed in a different sequence, may be added, merged, or left out altogether (e.g., all described acts or programmer events may not be necessary to carry out the techniques). In addition, while certain aspects of this disclosure are described as being performed by a single engine or unit for purposes of clarity, it should be understood that the techniques of this disclosure may be performed by a combination of units or engines associated with, for example, a medical device.

[0084] In one or more examples, the described techniques may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, the functions may be stored as one or more instructions or code on a computer-readable medium and executed by a hardware-based processing unit. Computer-readable media may include non-transitory computer-readable media, which corresponds to a tangible medium such as data storage media (e.g., RAM, ROM, EEPROM, flash memory, or any other medium that can be used to store desired program code in the form of instructions or data structures and that can be accessed by a computer).

[0085] Instructions may be executed by one or more processors, such as one or more digital signal processors (DSPs), general purpose microprocessors, application specific integrated circuits (ASICs), field programmable logic arrays (FPGAs), or other equivalent integrated or discrete logic circuitry. Accordingly, the term “processor” as used herein may refer to any of the foregoing structure or any other physical structure suitable for implementation of the described techniques. Also, the techniques could be fully implemented in one or more circuits or logic elements.

[0086] The following are illustrative of the techniques described herein.

[0087] Example 1 : A system for automated calibration of Local Field Potential (LFP) sensing for deep brain stimulation comprising: one or more electrodes; control circuitry configured to, prior to delivery of electrical stimulation via the one or more electrodes: sense, using the one or more electrodes, LFPs of the brain; record a snapshot of LFP wide- band activity; determine a peak frequency of the LFP wide-band activity from the snapshot; and update a Frequency of Interest (FOI) based on the peak frequency.

[0088] Example 2: The system of Example 1, wherein the control circuitry is configured to record the snapshot periodically.

[0089] Example 3 : The system of Example 1, wherein the control circuitry is configured to record the snapshot based on a comparison of the sensed FOI to a training data.

[0090] Example 4: The system of Example 3, wherein the control circuitry is further configured to determine the sensed FOI represents one or more programmer events based on the comparison.

[0091] Example 5: The system of Example 4, wherein at least one of the one or more programmer events are predefined user activities corresponding to heightened or reduced brain activity.

[0092] Example 6: The system of Example 3, further comprising an accelerometer communicatively coupled to the control circuitry, wherein the control circuitry is further configured to determine the sensed FOI represents one or more programmer events based on accelerometer data received from the accelerometer.

[0093] Example 7: The system of Example 1, further comprising a user interface communicatively coupled to the control circuitry, wherein the control circuitry is configured to record the snapshot based on receiving an indication of a programmer event from the user interface.

[0094] Example 8: The system of Example 1, wherein the control circuitry is further configured to generate a timeline of sensed power of the FOI over time.

[0095] Example 9: The system of Example 8, wherein the control circuitry is further configured to determine one or more of an LFP threshold, an upper stimulation limit, and a lower stimulation limit based on the timeline.

[0096] Example 10: The system of Example 1, wherein the control circuitry is further configured to run an electrocardiogram (ECG).

[0097] Example 11 : A method for automated calibration of Local Field Potential (LFP) sensing for deep brain stimulation comprising, prior to delivery of electrical stimulation: sensing, using one or more electrodes, LFPs of the brain; recording a snapshot of LFP wideband activity; determining a peak frequency of the LFP wide-band activity from the snapshot; and updating a Frequency of Interest (FOI) based on the peak frequency. [0098] Example 12: The method of Example 11, wherein the snapshot is recorded periodically.

[0099] Example 13 : The method of Example 11, wherein the snapshot is recorded based on a comparison of the sensed FOI to training data.

[0100] Example 14: The method of Example 13, further comprising: determining the sensed FOI represents one or more programmer events based on the comparison.

[0101] Example 15: The method of Example 14, wherein at least one of the one or more programmer events are predefined user activities corresponding to heightened or reduced brain activity.

[0102] Example 16: The method of Example 11, wherein the snapshot is recorded based on an identification of a programmer event by a user.

[0103] Example 17: The method of Example 11, further comprising generating a timeline of sensed power of the FOI over time.

[0104] Example 18: The method of Example 17, wherein the FOI is updated based on the timeline.

[0105] Example 19: The method of Example 17, further comprising determining one or more of an LFP threshold, an upper stimulation limit, and a lower stimulation limit based on the timeline.

[0106] Example 20: The method of Example 11, further comprising running an electrocardiogram (ECG).

Claims

WHAT IS CLAIMED IS:

1. A system for automated calibration of Local Field Potential (LFP) sensing for deep brain stimulation comprising: one or more electrodes; control circuitry configured to, prior to delivery of electrical stimulation via the one or more electrodes: sense, using the one or more electrodes, LFPs of the brain; record a snapshot of LFP wide-band activity; determine a peak frequency of the LFP wide-band activity from the snapshot; and update a Frequency of Interest (FOI) based on the peak frequency.

2. The system of claim 1, wherein the control circuitry is configured to record the snapshot periodically.

3. The system of claim 1, wherein the control circuitry is configured to record the snapshot based on a comparison of the sensed FOI and a training data.

4. The system of claim 3, wherein the control circuitry is further configured to determine the sensed FOI represents one or more programmer events based on the comparison.

5. The system of claim 4, wherein at least one of the one or more programmer events are predefined user activities corresponding to heightened or reduced brain activity.

6. The system of claim 3, further comprising an accelerometer communicatively coupled to the control circuitry, wherein the control circuitry is further configured to determine the sensed FOI represents one or more programmer events based on accelerometer data received from the accelerometer.

7. The system of claim 1, further comprising a user interface communicatively coupled to the control circuitry, wherein the control circuitry is configured to record the snapshot based on receiving an indication of a programmer event from the user interface.

8. The system of claim 1, wherein the control circuitry is further configured to generate a timeline of sensed power of the FOI over time.

9. The system of claim 8, wherein the control circuitry is further configured to determine one or more of an LFP threshold, an upper stimulation limit, and a lower stimulation limit based on the timeline.

10. The system of claim 1, wherein the control circuitry is further configured to run an electrocardiogram (ECG).

11. A method for automated calibration of Local Field Potential (LFP) sensing for deep brain stimulation comprising, prior to delivery of electrical stimulation: sensing, using one or more electrodes, LFPs of the brain; recording a snapshot of LFP wide-band activity; determining a peak frequency of the LFP wide-band activity from the snapshot; and updating a Frequency of Interest (FOI) based on the peak frequency.

12. The method of claim 11, wherein the snapshot is recorded periodically.

13. The method of claim 11, wherein the snapshot is recorded based on a comparison of the sensed FOI to training data.

14. The method of claim 13, further comprising: determining the sensed FOI represents one or more programmer events based on the comparison.

15. The method of claim 14, wherein at least one of the one or more programmer events are predefined user activities corresponding to heightened or reduced brain activity.