US20110166471A1

Movatterモバイル変換

Info

Publication number: US20110166471A1
Application number: US13/050,659
Authority: US
Inventors: Touby A. Drew; Jonathon E. Giftakis; Nina M. Graves
Original assignee: Medtronic Inc
Current assignee: Medtronic Inc
Priority date: 2004-11-02
Filing date: 2011-03-17
Publication date: 2011-07-07
Also published as: US7917199B2; US20060235489A1

Abstract

Methods for receiving a patient mark in an implanted device. A patient may experience symptoms that indicative of a neurological event or are common precursors to more severe clinical symptoms. The patient may activate a patient activator. Activation of the patient activator can cause the patient activator to transmit the patient mark to an implanted device. The implanted device may adjust physiological data being stored in response to the patient mark, including data before and after the patient mark. The patient mark may be used to time tune the detection algorithms and otherwise tune the detection capability of the implanted device and may also be used as an aid in statistical analysis of the stored physiological data.

Description

RELATED APPLICATIONS

This patent application is a divisional of U.S. application Ser. No. 11/412,230, filed Apr. 26, 2006, published as U.S. Patent Publication No. 2006-0235489 on Oct. 19, 2006, now allowed, which is a continuation-in-part of U.S. application Ser. No. 11/264,974, filed Nov. 2, 2005, published as U.S. Patent Publication No. 2006-0094971 on May 4, 2006, currently pending, which claims priority to Provisional Application Ser. No. 60/624,232 filed Nov. 2, 2004, all of which are incorporated herein by reference in their entirety.

FIELD OF THE INVENTION

The invention relates to techniques for selecting, storing and reporting data associated with physiologic signals that may be further associated with a neurological event.

BACKGROUND OF THE INVENTION

Nervous system disorders affect millions of people, causing death and a degradation of life. Nervous system disorders include disorders of the central nervous system, peripheral nervous system, and mental health and psychiatric disorders. Such disorders include, for example without limitation, epilepsy, Parkinson's disease, essential tremor, dystonia, headache, and multiple sclerosis (MS). Additionally, nervous system disorders include mental health disorders and psychiatric disorders which also affect millions of individuals and include, but are not limited to, anxiety (such as general anxiety disorder, panic disorder, phobias, post traumatic stress disorder (PTSD), and obsessive compulsive disorder (OCD)), mood disorders (such as major depression, bipolar depression, and dysthymic disorder), sleep disorders (narcolepsy), obesity, and anorexia.

As an example, epilepsy is the most prevalent serious neurological disease across all ages. Epilepsy is a group of neurological conditions in which a person has or is predisposed to recurrent seizures. A seizure is a clinical manifestation resulting from excessive, hypersynchronous, abnormal electrical or neuronal activity in the brain. (A neurological event is an activity that is indicative of a nervous system disorder. A seizure is a type of a neurological event.) This electrical excitability of the brain may be likened to an intermittent electrical overload that manifests with sudden, recurrent, and transient changes of mental function, sensations, perceptions, and/or involuntary body movement. Because the seizures are unpredictable, epilepsy affects a person's employability, psychosocial life, and ability to operate vehicles or power equipment. It is a disorder that occurs in all age groups, socioeconomic classes, cultures, and countries, in developed countries, the age-adjusted incidence of recurrent unprovoked seizures ranges from 24/100,000 to 53/100,000 person-years and may be even higher in developing countries. In developed countries, age specific incidence is highest during the first few months of life and again after age 70. The age-adjusted prevalence of epilepsy is 5 to 8 per 1,000 (0.5% to 0.8%) in countries where statistics are available. In the United States alone, epilepsy and seizures affect 2.3 million Americans, with approximately 181,000 new cases occurring each year. It is estimated that 10% of Americans will experience a seizure in their lifetimes, and 3% will develop epilepsy by age 75.

There are various approaches in treating nervous system disorders. Treatment therapies can include any number of possible modalities alone or in combination including, for example, electrical stimulation, magnetic stimulation, drug infusion, and/or brain temperature control. Each of these treatment modalities can be operated using closed-loop feedback control. Such closed-loop feedback control techniques receive from a monitoring element a neurological signal that carries information about a symptom or a condition or a nervous system disorder. Such a neurological signal can include, for example, electrical signals (such as EEG, ECoG, and/or EKG), chemical signals, other biological signals (such as change in quantity of neurotransmitters), temperature signals, pressure signals (such as blood pressure, intracranial pressure or cardiac pressure), respiration signals, heart rate signals, pH-level signals, and peripheral nerve signals (cuff electrodes on a peripheral nerve). Monitoring elements can include, for example, recording electrodes or various types of sensors.

For example, U.S. Pat. No. 5,995,868 discloses a system for the prediction, rapid detection, warning, prevention, or control of changes in activity states in the brain of a patient. Use of such a closed-loop feed back system for treatment of a nervous system disorder may provide significant advantages in that treatment can be delivered before the onset of the symptoms of the nervous system disorder.

During the operation of a medical device system, the patient is likely to experience multiple detections of the nervous system disorder. For example, in the case of seizures, the patient may have thousands of seizures over the course of a time period, but only a few will have behavioral manifestations. The other seizure episodes that don't exhibit behavioral manifestations are considered sub-clinical or electrographic seizures. When the medical device system monitors for seizure occurrences, however, the medical device system will detect many seizure events although only some of these events will spread to other parts of the brain such that the patient will exhibit it (e.g., convulsions, unconsciousness, etc.).

In order to effectively provide treatment therapy, an implanted device may be required to record physiologic data that is related to the disorder. However, an implanted device is typically limited by memory capacity and by battery capacity. Thus, the implanted device is limited in the amount of data that can be stored and reported.

An implanted device often stores physiologic data in a data structure and manages memory allocation for the data structure. However, the memory allocation management supported by the implanted device may have deficiencies. For example, with a FIFO memory buffer if the amount of collected physiologic data exceeds the available memory space, the oldest physiologic data is lost regardless of the importance of the lost data.

It is therefore desirable to selectively store physiologic data in a limited memory space of an implanted device. The implanted device can report the most relevant data from the stored data so that the implanted device can be configured to provide efficacious treatment.

BRIEF SUMMARY

The following represents a simplified summary of some embodiments of the invention in order to provide a basic understanding of various aspects of the invention. This summary is not an extensive overview of the invention nor is it intended to identify key or critical elements of the invention or to delineate the scope of the invention. Its sole purpose is to present some embodiments of the invention in simplified form as a prelude to the more detailed description that is presented thereafter.

In accordance with an aspect of the invention, a device is configured to automatically record physiological signals. A user may provide an input, which will be referred to as a patient mark, to the implanted device via telemetry so as to indicate the onset of precursor symptoms such as an aura or chest pain or other clinical symptoms that precede more severe clinical onset. The patient mark may allow for improved detection sensitivity and may allow for a better recording of data before and after the mark. In addition, clusters that were not otherwise connected may be linked into a single cluster due to the patient mark. Furthermore, clusters may be started earlier or end later in response to the patient mark. The patient mark may also aid in statistical analysis and may help a physician to identify certain true positives and/or false negatives in a more precise manner.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a simplified schematic view of a thoracic cavity leadless medical device implanted in a patient that monitors cardiac and respiratory parameters in accordance with an embodiment of the invention.

FIG. 2 is a simplified block diagram of a core monitor as shown inFIG. 1.

FIG. 3 is a graphical representation of the signals sensed by core monitor as shown inFIG. 1 above.

FIG. 4 is a flow diagram showing operation of a core monitor as shown inFIG. 1 above.

FIG. 5 shows a first apparatus for collecting physiologic data that can be retained in accordance with an embodiment of the invention.

FIG. 6 is a simplified schematic view of an alternative embodiment of a thoracic and cranial leaded medical device implanted in a patient that monitors cardiac, respiratory and brain parameters in accordance with an embodiment of the invention.

FIG. 7 is a simplified block diagram of a full monitor as shown inFIG. 6 above.

FIG. 8 is a flow diagram showing operation of a full monitor as shown inFIG. 6 above.

FIG. 9 is a simplified schematic view of another embodiment of a cardiac, cranial and phrenic nerve leaded medical device implanted in a patient that monitors cardiac, respiratory and brain parameters and provides brain, respiration, and cardiac treatment thereof in accordance with an embodiment of the invention.

FIG. 10 is a simplified block diagram of a full monitor with brain, respiration and cardiac stimulation therapy as shown inFIG. 9 above.

FIG. 11 is a flow diagram showing operation of a full monitor with therapy (including brain, respiration and/or cardiac stimulation therapy) as shown inFIG. 9 above.

FIG. 12 shows apparatus that supports reporting neurological data in accordance with an embodiment of the invention;

FIG. 13 is a schematic diagram of a system utilizing any of the above-described embodiments and allowing remote monitoring and diagnostic evaluation of at risk patients.

FIG. 14 is a schematic diagram of an alternative system utilizing any of the above-described embodiments and allowing remote monitoring and diagnostic evaluation of at risk patients.

FIG. 15 shows a scenario for storing neurological data in accordance with an embodiment of the invention.

FIG. 16 shows a data structure for storing neurological data in accordance with an embodiment of the invention.

FIG. 17 shows a flow diagram for reporting neurological data in accordance with an embodiment of the invention.

FIG. 18 shows an exemplary data entry that contains physiologic data and that may be retained in memory of an implantable medical device.

FIG. 19 shows components of the exemplary data entry as shown inFIG. 18.

FIG. 20 shows the selection of channel data for retention in a data structure.

FIG. 21 shows a method of receiving input from a patient activator in accordance with an embodiment of the invention.

DETAILED DESCRIPTION OF THE INVENTION

The following description discloses techniques for selecting, storing and reporting data associated with physiologic signals that may be further associated with a neurological event. These techniques are suitable for use within any implantable medical device system including, but not limited to, a core monitoring device, a full monitoring device, and/or a combination monitoring and treatment device (e.g., involving brain, respiration, and/or cardiac physiologic signals, discussed below). For example, a core monitor system may consist of ECG and respiratory inputs, a lull monitor system may consist of ECG, respiratory and EEG inputs, a monitor/treatment system may include brain, cardiac inputs and phrenic nerve stimulation in various combinations.

In an embodiment, the invention is implemented within an implantable neurostimulator system, however, as already discussed, those skilled in the art will appreciate that the techniques disclosed herein may be implemented generally within any implantable medical device system having monitoring capabilities of physiological conditions of the patient including, but not limited to, implantable drug delivery systems, implantable systems providing stimulation and drug delivery, pacemaker systems, defibrillator systems, cochlear implant systems, and implantable diagnostic system for detecting bodily conditions, including those in organs like the brain and/or the heart. The implantable medical device may provide therapeutic treatment to neural tissue in any number of locations in the body including, for example, the brain (which includes the brain stem), the vagus nerve, the spinal cord, peripheral nerves, etc. The treatment therapies can include any number of possible modalities alone or in combination including, for example, electrical stimulation, magnetic stimulation, drug infusion, brain temperature control, and/or any combination thereof.

In addition, the invention may be embodied in various forms to analyze and treat nervous system and other disorders, namely disorders of the central nervous system, peripheral nervous system, and mental health and psychiatric disorders. Such disorders include, for example without limitation, epilepsy, Sudden Unexpected Death in Epilepsy Patients (SUDEP), Parkinson's disease, essential tremor, dystonia, multiple sclerosis (MS), anxiety (such as general anxiety, panic, phobias, post traumatic stress disorder (PTSD), and obsessive compulsive disorder (OCD)), mood disorders (such as major depression, bipolar depression, and dysthymic disorder), sleep disorders (narcolepsy), obesity, tinnitus, stroke, traumatic brain injury, Alzheimer's, and anorexia.

The physiologic signals that are selected, stored and reported in accordance with various aspects of the invention may include any number of sensed signals. Such physiological signals can include, for example, electrical signals (such as EEG, ECoG and/or EKG), chemical signals, biological signals (such as change in quantity of neurotransmitters), temperature signals, pressure signals (such as blood pressure, intracranial pressure or cardiac pressure), respiration signals, heart rate signals, pH-level signals, activity signals (e.g., detected by an accelerometer), and/or peripheral nerve signals (cuff electrodes on a peripheral nerve). Such physiological signals may be recorded using one or more monitoring elements such as monitoring electrodes or sensors. For example, U.S. Pat. No. 6,227,203 provides examples of various types of sensors that may be used to detect a symptom or a condition or a nervous system disorder and responsively generate a neurological signal. In addition, various types of physiologic activities may be sensing including, for example, brain, heart and/or respiration.

As discussed, the techniques disclosed herein are suitable for use within any implantable medical device system including, but not limited to, a core monitoring device, a full monitoring device, and/or a combination monitoring and treatment device. Each of these medical device implementations are discussed in further detail below. In general, however, each of these embodiments utilize one or more monitoring elements that receive signals associated with the physiological conditions being sensed, a memory component that contains multiple data structures (for example, at least one that is age-based and at least one other that is priority-based), and a processing component (logic or software) that stores data records in the data structures as disclosed herein (utilizing, for example, severity and priority determinations for the data).

Core Monitor

In an embodiment, a Core Monitor device monitors cardiac (ECG) and respiration signals and records these signals as discussed herein. Real-time analysis of the ECG signal evaluates rate disturbances (e.g., bradycardia; tachycardia; asystole) as well as any indications of cardiac ischemia (e.g., ST segment changes; T wave inversion, etc.). Real-time analysis of the respiration signal evaluates respiration disturbances (e.g., respiration rate, minute ventilation, apnea, prolonged pauses, etc.). Abnormalities detected during real-time analysis may lead to an immediate patient alert, which can be audible (beeps, buzzers, tones, spoken voice, etc.), light, tactile, or other means. Manual indication of a seizure may be achieved through an external activator device22. The patient (or caregiver) may push a button on the external device, while communicating with the implanted device. This will provide a marker and will initiate a recording, as discussed herein, of the sensed data (for example, in the event the patient is experiencing a neurological event).

In treating SUDEP, for example, prolonged ECG recordings may be possible (e.g., recording all data during sleep since the incidence of SUDEP is highest in patients during sleep). Post-processing of the signal can occur in the implanted device, the patient's external device and/or in the clinician external device. Intermittently (e.g., every morning, once/week, following a seizure), the patient may download data from the implantable device to the patient external device (as will be discussed further herein), which may then be analyzed by the external device (and/or sent through a network to the physician) to assess any ECG abnormalities. If an abnormality is detected, the device may notify the patient/caregiver. At that time, the patient/caregiver may inform the healthcare provider of the alert to allow a full assessment of the abnormality. The clinician external device may also be capable of obtaining the data from the implanted device and conducting an analysis of the stored signals. If a potentially life-threatening abnormality is detected, the appropriate medical treatment may be prescribed (cardiac abnormality: a pacemaker, an implantable defibrillator, or a heart resynchronization device may be indicated or respiration abnormality: CPAP, patient positioning, or stimulation of respiration may be indicated).

FIG. 1 is a simplified schematic view of acore monitor100 implanted in apatient10.Monitor100 continuously senses and monitors one or more physiological conditions of the patient via monitoring elements140 (in the embodiment, the physiological conditions are cardiac and respiration functions ofpatient10 and themonitoring elements140 are subcutaneous electrodes). Stored diagnostic data is uplinked and evaluated by an external computing device120 (e.g., a patient's physician utilizing programmer) via a 2-way telemetry link132. An externalpatient activator122 may optionally allow thepatient10, or other care provider (not shown), to manually activate the recording of diagnostic data.

In the embodiment, monitor100 encompasses monitoringelements140 that are several subcutaneous spiral electrodes and may be embedded individually into three or four recessed casings placed in a compliant surround that is attached to the perimeter of implantedmonitor100 as substantially described in U.S. Pat. Nos. 6,512,940 and 6,522,915. These electrodes are electrically connected to the circuitry of the implantedmonitor100 to allow the detection of cardiac depolarization waveforms (as substantially described in U.S. Pat. No. 6,505,067) that may be further processed to detect cardiac electrical characteristics (e.g., heart rate, heart rate variability, arrhythmias, cardiac arrest, sinus arrest and sinus tachycardia). Further processing of the cardiac signals signal amplitudes may be used to detect respiration characteristics (e.g., respiration rate, minute ventilation, and apnea). To aid in the implantation ofmonitor100 in a proper position and orientation, an implant aid may be used to allow the implanting physician to determine the proper location/orientation as substantially described in U.S. Pat. No. 6,496,715.

FIG. 2 depicts a block diagram of the electronic circuitry of the core monitor100 ofFIG. 11 in accordance with an embodiment of the invention.Monitor100 comprises a (primary control circuit220 and may be similar in design to that disclosed in U.S. Pat. No. 5,052,388.Primary control circuit220 includessense amplifier circuitry224, acrystal clock228, a random-access memory and read-only memory (RAM/ROM)unit230, a central processing unit (CPU)232, aMV Processor circuit238 and atelemetry circuit234, all of which are generally known in the art.

Monitor

100 may includeinternal telemetry circuit234 so that it is capable of being programmed by means of external programmer/control unit120 via a 2-way telemetry link132 (shown inFIG. 1). External programmer/control unit120 communicates via telemetry with themonitor100 so that the programmer can transmit control commands and operational parameter values to be received by the implanted device, and so that the implanted device can communicate diagnostic and operational data to theprogrammer120. For example,programmer120 may be Models 9790 and CareLink® programmers, commercially available from Medtronic, Inc., Minneapolis, Minn. Various telemetry systems for providing the necessary communications channels between an external programming unit and an implanted device have been developed and are well known in the art. Suitable telemetry systems are disclosed, for example, in U.S. Pat. Nos. 5,127,404; 4,374,382; and 4,556,063.

Typically, telemetry systems such as those described in the above referenced patents are employed in conjunction with an external programming/processing unit. Most commonly, telemetry systems for implantable medical devices employ a radio-frequency (RF) transmitter and receiver in the device, and a corresponding RF transmitter and receiver in the external programming unit. Within the implantable device, the transmitter and receiver utilize a wire coil as an antenna for receiving downlink telemetry signals and for radiating RF signals for uplink telemetry. The system is modeled as an air-core coupled transformer. An example of such a telemetry system is shown in U.S. Pat. No. 4,556,063.

In order to communicate digital data using RE telemetry, a digital encoding scheme such as is described in U.S. Pat. No. 5,127,404 can be used. In particular, a pulse interval modulation scheme may be employed for downlink telemetry, wherein the external programmer transmits a series of short RF “bursts” or pulses in which the interval between successive pulses (e.g., the interval from the trailing edge of one pulse to the trailing edge of the next) is modulated according to the data to be transmitted. For example, a shorter interval may encode a digital “0” bit while a longer interval encodes a digital “1” bit. For uplink telemetry, a pulse position modulation scheme may be employed to encode uplink telemetry data. For pulse position modulation, a plurality of time slots are defined in a data frame, and the presence or absence of pulses transmitted during each time slot encodes the data. For example, a sixteen-position data frame may be defined, wherein a pulse in one of the time slots represents a unique four-bit portion of data.

Programming units such as the above-referenced Medtronic Models 9790 and CareLink® programmers typically interface with the implanted device through the use of a programming head or programming paddle, a handheld unit adapted to be placed on the patient's body over the implant site of the patient's implanted device. A magnet in the programming head effects reed switch closure in the implanted device to initiate a telemetry session. Thereafter, uplink and downlink communication takes place between the implanted device's transmitter and receiver and a receiver and transmitter disposed within the programming head.

As previously noted,primary control circuit220 includescentral processing unit232 which may be an off-the-shelf programmable microprocessor or microcontroller, but in an embodiment of the invention is a custom integrated circuit. Although specific connections betweenCPU232 and other components ofprimary control circuit220 are not shown inFIG. 2, it will be apparent to those of ordinary skill in the art thatCPU232 functions to control the timed operation ofsense amplifier circuit224 under control of programming stored in RAM/ROM unit230. Those of ordinary skill in the art will be familiar with such an operative arrangement.

With continued reference toFIG. 2,crystal oscillator circuit228, in an embodiment, is a 32,768-Hz crystal controlled oscillator, provides main timing clock signals toprimary control circuit220. The various components ofmonitor100 are powered by means of a battery (not shown), which is contained within the hermetic enclosure ofmonitor100. For the sake of clarity in the figures, the battery and the connections between it and the other components ofmonitor100 are not shown.Sense amplifier224 is coupled tomonitoring elements140. Where cardiac intrinsic signals are sensed, they may sensed bysense amplifier224 as substantially described in U.S. Pat. No. 6,505,067.

Processing byCPU232 allows the detection of cardiac electrical characteristics and anomalies (e.g., heart rate, heart rate variability, arrhythmias, cardiac arrest, sinus arrest and sinus tachycardia). Further processing of the cardiac signals signal amplitudes may be used to detect respiration characteristics/anomalies (e.g., respiration rate, tidal volume, minute ventilation, and apnea) inMV Processor238.

FIG. 3 shows theintracardiac signals370 presented tosense amplifier224 from monitoringelements140. Note the amplitude variation of cardiac signals caused by the change in thoracic cavity pressure due to respiration (i.e., inspiration and expiration). By low pass filtering thecardiac signals370, a signal representing minute ventilation may be obtained as depicted inwaveform372. This respiration signal may further be examined to detect respiration rate and reduced or absence of inspiration and expiration (central apnea) byCPU232 and software resident in RAM/ROM230.

Upon detection of either a cardiac or respiration anomaly,CPU232, under control of firmware resident in RAM/ROM230, will initiate recording of the appropriate diagnostic information into RAM of RAM/ROM230 (discussed further herein), initiate a warning or alert to the patient, patient caregiver, or remote monitoring location.

FIG. 4 is a flow diagram400 showing operation of a core monitor sensing/monitoring cardiac and respiration parameters for the detection of a neurological event as shown and described inFIG. 1 above. Beginning atblock402, the interval between sensed cardiac signals are measured. Atblock404, a rate stability measurement is made on each cardiac interval utilizing a heart rate average fromblock406. Atblock408, a rate stable decision is made based upon preprogrammed parameters. If YES, the flow diagram returns to theHR Measurement block402. If NO, the rate stability information is provided to FormatDiagnostic Data block412.

Atblock416, thoracic impedance is continuously measured in a sampling operation. Atblock418, a MV and respiration rate calculation is made. Atblock422, a pulmonary apnea decision is made based upon preprogrammed criteria. If NO, the flow diagram returns toMV Measurement block416. If YES, the occurrence of apnea and MV information is provided to FormatDiagnostic Data block412. Format Diagnostic Data block412 formats the data from the cardiac and respiration monitoring channels, adds a time stamp (i.e., date and time) and provides the data to block414 where the data is stored in RAM, SRAM or MRAM memory for later retrieval by a clinician via, telemetry (using the techniques discussed herein). Optionally, block412 may add examples of intrinsic ECG and/or respiration signals recorded during a sensed neurological event. Additionally, optionally, block415 may initiate a warning or alert to the patient, patient caregiver, or remote monitoring location (as described in U.S. Pat. No. 5,752,976).

It will be appreciated that alternative embodiments of the core monitor device may also be utilized. As discussed above, core monitor100 may sense any number of physiologic conditions of the patient for purposes of detecting, and storing data relating to, any number of the neurological events. For example, cardiac lead(s) may be used to facilitate detection of a neurological event and the recording of data and signals pre and post event. Cardiac leads may consist of any typical lead configuration as is known in the art, such as, without limitation, right ventricular (RV) pacing or defibrillation leads, right atrial (RA) pacing or defibrillation leads, single pass RA/RV pacing or defibrillation leads, coronary sinus (CS) pacing or defibrillation leads, left ventricular pacing or defibrillation leads, pacing or defibrillation epicardial leads, subcutaneous defibrillation leads, unipolar or bipolar lead configurations, or any combinations of the above lead systems.

In another embodiment, an electrode located distally on a sensor stub may be used to facilitate detection of a neurological event and the recording of data and signals pre and post event. The sensor stub is inserted subcutaneously in a thoracic area of the patient. Themonitor100 may sense cardiac signals between an electrode on the distal end of the sensor stub and the monitor case as described in conjunction with the embodiment shown in FIG. 5 in U.S. Pat. No. 5,987,352. Themonitor100 may also sense respiration parameters such as respiration rate, minute ventilation and apnea via measuring and analyzing the impedance variations measured from the implantedmonitor100 case to the electrode located distally on the sensor stub lead as substantially described in U.S. Pat. Nos. 4,567,892 and 4,596,251.

In yet another embodiment, an external wearable device such as a wearable patch, a wristwatch, or a wearable computing device may be used may be used to continuously sense and monitor cardiac and respiration function ofpatient10. Optionally, a button on the external wearable device may be activated by the patient10 (or a caregiver) to manually activate data recording (for example, in the event the patient is experiencing a neurological event). The external wearable device may comprise an amplifier, memory, microprocessor, receiver, transmitter and other electronic components as substantially described in U.S. Pat. No. 6,200,265. In the embodiment of a wearable patch, the device may consist of a resilient substrate affixed to the patient's skin with the use of an adhesive. The substrate flexes in a complimentary manner in response to a patient's body movements providing patient comfort and wearability. The low profile patch is preferably similar in size and shape to a standard bandage, and may be attached to the patient's skin in an inconspicuous location.

FIG. 5 shows an apparatus500 for collecting physiologic data that can be retained in accordance with an embodiment of the invention. Apparatus comprises implanteddevice501 andexternal device503. In the embodiment, implanteddevice501 collects physiological data throughEEG lead505, which is coupled to an electrode. The embodiment may support a plurality of electrodes that may collect physiological data from desired locations in the brain or other neural tissue within the body. With respect to sensing of the brain, the embodiment may support electrodes that are positioned in the brain, positioned on the scalp, or a combination.EEG amplifier521 processes neurological signals fromEEG lead505, while stimulator511 forms appropriate electrical waveforms to simulate electrodes that are coupled to EEG lead505.CPU513 controls the circuitry ininternal device501 and is clocked bycrystal clock519.CPU513 obtains instructions from and stores neurological data to RAM/ROM515. Communication within internal device between components is transported on I/O bus523.

External device

503 collects other physiologic data through patch leads507aand507b, which are coupled to patch electrodes.External device503 is typically worn by the patient. In the embodiment, the patch electrodes monitor and stimulate the respiration of the patient in concert with monitoring neurological data. Respiratory data is collected from the patch electrodes throughlead interface525. Minute Ventilation (MV)module527 excites the patch electrodes (if necessary) while respiratory data (indicative of the minute ventilation) is obtained throughsense amplifier531, (Minute or total ventilation is the product of respiratory rate and tidal volume.) Additionally,impedance module529, MVlow pass filter533, andMV calibration module535 calibrateMV excitation module527. Data communication withinexternal device503 is transported on I/O bus539.

Internal device

501 andexternal device503 communicate with each other overtelemetry channel543 through

telemetry interfaces

517 and537 and

antennas

539 and541. However,internal device501 andexternal device503 may communicate using an alternative communications path. For example, I/O bus535 and I/O bus539 may be electrically connected through a wired channel.

While the embodiment shown inFIG. 5 collects and retains physiological data and respiratory data, other physiologic data may be collected and retained. For example, apparatus500 may collect cardiac data and motor coordination data. The embodiment may further correlate the neurological data with the physiological data.

Full Monitor

The full monitor is capable of monitoring cardiopulmonary parameters as in the core monitor described above, and additionally an EEG from an intracranially implanted lead system. This will allow the full monitor to collect cardiovascular, respiratory and neurologic signals in close proximity to detected neurologic events as well as notifying the patient/caregiver of a prolonged event (and/or status epilepticus). Like the core monitor, cardiovascular and respiratory monitoring may occur around a neurologic event (peri-ictal). The full monitor device may detect the neurological event and analyze the peri-ictal signals and initiate loop recording. The addition of the neurologic monitor allows shorter loop-recording durations since the beginning of the neurological event may be known with greater certainty. In addition, the EEG monitoring may be programmed to provide alerts when a neurologic event has exceeded a pre-determined duration or severity.

FIG. 6 is a simplified schematic view of afull monitor620 implanted in apatient10.Monitor620 continuously senses and monitors cardiac, brain and respiration function ofpatient10 via severalsubcutaneous electrodes616 and abrain lead618 to allow detection of neurological events and the recording of data and signals pre and post event. Stored diagnostic data is uplinked and evaluated by the patient'sphysician utilizing programmer612 via a 2-way telemetry link632. An externalpatient activator622 may optionally allow thepatient10, or other care provider (not shown), to manually activate the recording of diagnostic data. An implant aid may be used withmonitor620 to ensure a proper position and orientation during implant as described above in connection with the system ofFIG. 1.

FIG. 7 is a block diagram of the electronic circuitry that makes upfull monitor620 withbrain618 and cardiac616 leads ofFIG. 6 in accordance with the disclosed alternative embodiment of the invention.Monitor620 comprises aprimary control circuit720 andMV circuit722 that are described herein above in conjunction withFIG. 2. In addition, thefull Monitor620 also includes anamplifier725 to amplify and sense EEG signals from a cranially implantedlead618. TheCPU732, in conjunction with a software program resident in RAM/ROM730, integrates the information from the sensed cardiac, respiration and EEG signals, detects the onset of cerebral, cardiac and/or respiratory anomalies, formats and stores diagnostic data for later retrieval by the patient's clinician and, optionally, may warn or alert the patient, patient caregiver or remote monitoring location.

FIG. 8 is a flow diagram840 showing operation of a full monitor sensing and monitoring cardiac, respiration and electroencephalogram parameters for the detection of a neurological event as shown and described in embodiment inFIG. 6 above. Beginning atblock802, the interval between sensed cardiac signals are measured. Atblock804, a rate stability measurement is made on each cardiac interval utilizing a heart rate average fromblock806. Atblock808, a rate stable decision is made based upon preprogrammed parameters. If YES, the flow diagram returns to theHR Measurement block802. If NO, the rate stability information is provided to FormatDiagnostic Data block812.

Atblock816, thoracic impedance is continuously measured in a sampling operation. Atblock818, a MV and respiration rate calculation is made. Atblock822, a pulmonary apnea decision is made based upon preprogrammed criteria. If NO, the flow diagram returns to MV Measurement block816.1fYES, the occurrence of apnea and MV information is provided to FormatDiagnostic Data block812.

Atblock824, the electroencephalogram is sensed and measured. An EEG seizure determination is performed atblock826 as described in U.S. Patent Publication No. 2004/0138536. Atblock828, a seizure cluster episode is determined. If NO, the flow diagram returns toEEG Measurement block824. If YES, the occurrence of a seizure cluster is provided to FormatDiagnostic Data block812. Format Diagnostic Data block812 formats the data from the cardiac, respiration and EEG monitoring channels, adds a time stamp (i.e., date and time) and provides the data to block814 where the data is stored in RAM memory for later retrieval by a clinician via telemetry (using techniques discussed herein). Optionally, block812 may add examples of intrinsic ECG, respiration and/or EEG signals recorded during a sensed episode/seizure. Additionally, optionally, block815 may initiate a warning or alert to the patient, patient caregiver, or remote monitoring location (as described in U.S. Pat. No. 5,752,976).

Again, it will be appreciated that alternative embodiments of the full monitor device may also be utilized. For example, cardiac lead(s), a sensor stub, and/or a wearable patch may be used to facilitate detection of a neurologic event and the recording of data and signals pre and post event. An integrated electrode may also be used that senses ECG signals as described in U.S. Pat. No. 5,987,352 and respiration signals as described in U.S. Pat. Nos. 4,567,892 and 4,596,251. Optionally, the monitor may warn/alert the patient10 via buzzes, tones, beeps or spoken voice (as substantially described in U.S. Pat. No. 6,067,473) via a piezo-electric transducer incorporated in the housing of monitor and transmitting sound to the patient's inner ear.

In another embodiment, the monitor may be implanted cranially in thepatient10. In such an embodiment, the monitor may be constructed as substantially described in U.S. Pat. Nos. 5,782,891 and 6,427,086. EEG sensing may be accomplished by the use of integrated electrodes in the housing of the monitor, cranially implanted leads, and or leadless EEG sensing.

Alternatively, ECG rate and asystole may be inferred using a variety of technologies. For example, ECG rate and asystole may be inferred (along with a blood pressure signal) from a capacitive dynamic pressure signal (i.e., dP/dt) as substantially described in U.S. Pat. No. 4,485,813. Similarly, an acoustic signal (i.e., sound) may be used as substantially described in U.S. Pat. No. 5,554,177 (the sensed acoustic signal is low pass filtered to limit ECG signals to 0.5-3 Hz while filtering out speech, swallowing and chewing sounds). As another example, ECG rate and asystole may be inferred (along with a blood saturation measurement) by monitoring a reflectance oximetry signal (i.e., O₂sat) as substantially described in U.S. Pat. No. 4,903,701. As yet another example, a blood temperature signal (i.e., dT/dt) may be used as substantially described in U.S. Pat. No. 5,336,244. As still another example, ECG rate and asystole may be inferred (along with an arterial flow measurement) by monitoring a blood flow signal (from an adjacent vein via impedance plethysmography, piezoelectric sensor or Doppler ultrasound) as substantially described in U.S. Pat. No. 5,409,009. As another example, ECG rate and asystole may be inferred (along with a blood pressure measurement) by monitoring a blood pressure signal utilizing a strain gauge s substantially described in U.S. Pat. No. 5,168,759. As another example, ECG rate and asystole may be inferred by monitoring a blood parameter sensor (such as oxygen, pulse or flow) located on a V-shaped lead as substantially described in U.S. Pat. No. 5,354,318.

Monitor and Treat

As exemplified above, any number of implantable medical device systems are envisioned that may incorporate the recording and retention techniques discussed herein. For example, the monitoring may be achieved using any of the above techniques in conjunction with treatment by delivery of treatment therapy (e.g., electrical stimulation) to the brain, cardiac or respiration.

FIG. 9 is a simplified schematic view of a full Monitor/Brain, Respiration andCardiac Therapy unit980 implanted in apatient10. Monitor/Brain, Respiration andCardiac Therapy unit980 continuously senses and monitors cardiac, brain and/or respiration function ofpatient10 via, cardiac lead(s)916 and abrain lead918 to allow detection of epileptic neurological events, the recording of data and signals pre and post event, and the delivery of therapy viabrain lead918, cardiac lead(s)916 andphrenic nerve lead928. Stored diagnostic data is uplinked and evaluated by the patient'sphysician utilizing programmer912 via a 2-way telemetry link932. An externalpatient activator922 may optionally allow thepatient10, or other care provider (not shown), to manually activate the recording of diagnostic data and delivery of therapy. Optionally, lead928 may connect to the diaphragm to provide direct diaphragmatic stimulation. As discussed any combination of monitoring and any combination of treatment may be implemented.

FIG. 10 is a block diagram of the electronic circuitry that makes up full Monitor/Brain, Respiration andCardiac Therapy device980 ofFIG. 9. Monitor/Brain, Respiration andCardiac Therapy device980 comprises aprimary control circuit1020 andMV circuit1022 whose function was described herein above in conjunction withFIG. 2. In addition the Monitor/Brain, Respiration and Cardiac Therapy device ofFIG. 10 also includes anamplifier1025 to amplify and sense EEG signals from a cranially implantedlead918 and anoutput stimulator1029 to provide brain stimulation via cranially implantedlead918 and phrenic nerve stimulation via,respiration lead928. TheCPU1032, in conjunction with software program in RAM/ROM1030, integrates the information from the sensed cardiac, respiration and EEC signals, detects the onset of cerebral, cardiac and/or respiratory anomalies, provides preprogrammed stimulation therapy to the patient's brain vialead918, stimulation of the patient's phrenic nerve viarespiration lead928 and stimulation of the patient's heart via cardiac leads916, formats and stores diagnostic data for later retrieval by the patient's clinician and, optionally, may warn or alert the patient, patient caregiver or remote monitoring location. Optionally, lead928 may connect to the diaphragm to provide direct diaphragmatic stimulation.

FIG. 11 is a flow diagram1150 showing operation of a hill monitor/therapy sensing and monitoring cardiac, respiration and electroencephalogram parameters as shown and described inFIG. 9 above. Beginning atblock1102, the interval between sensed cardiac signals are measured. Atblock1104, a rate stability measurement is made on each cardiac interval utilizing a heart rate average fromblock1106. Atblock1108, a rate stable decision is made based upon preprogrammed parameters. If YES, the flow diagram returns to theHR Measurement block1102. If NO, the rate stability information is provided to Determine Therapy andDuration block1130.

Atblock1116, thoracic impedance is continuously measured in a sampling operation. Atblock1118, a MV and respiration rate calculation is made. Atblock1122, a pulmonary apnea decision is made based upon preprogrammed criteria if NC), the flow diagram returns toMV Measurement block1116. If YES, the occurrence of apnea and MV information is provided to Determine Therapy andDuration block1130.

Atblock1124, the electroencephalogram is sensed and measured. An EEG calculation is performed atblock1126. Atblock1128, a seizure episode is determined. If NO, the flow diagram returns toEEG Measurement block1124. If YES, the occurrence of a seizure is provided to Determine Therapy andDuration block1130.

Based upon the data presented to it, Determine Therapy andDuration block1130 determines the type of therapy and the duration to block1132, which controls the start of the therapy by evaluating the severity and ranking of each event (i.e., maximum ratio, duration of seizure detection, spread, number of clusters per unit time, number of detections per cluster, duration of an event cluster, duration of a detection, and inter-seizure interval).Block1134 monitors the completion of the determined therapy. If the therapy is not complete, block returns to block1134. If the therapy is determined to be complete, block1134 returns the flow diagram to blocks1102 (Measure HR),1116 (Measure Impedance) and1124 (Measure EEG) to continue the monitoring of cardiac, respiratory and brain signal parameters. Therapy may consist of neural stimulation, cardiac pacing, cardioversion/defibrillation, and drug delivery via, a pump or any combination of therapies.

Whenblock1130 determines that a therapy is to be initiated FormatDiagnostic Data block1112 formats the data from the cardiac, respiration and EEG monitoring channels, adds a time stamp (i.e., date and time), type and duration of therapy and provides the data to block1114 where the data is stored in RAM memory for later retrieval by a clinician via telemetry (using techniques discussed herein). Optionally,block1112 may add examples of intrinsic ECG, respiration and/or EEG signals recorded during a sensed episode/seizure.

The above embodiments illustrate that the disclosed techniques may be implemented within any number of medical device systems (drug delivery, electrical stimulation, pacemaking, defibrillating, cochlear implant, and/or diagnostic) but configured to retain sensed data records in accordance with the teachings disclosed herein. In general, the implanted medical component utilizes one or more monitoring elements (e.g., electrodes or other sensors); a memory component having a plurality of data structures (and/or data structure types), a processing component (such as a CPU) to process received data for storage in memory as disclosed herein, and a telemetry component.

Data Reporting and Remote Monitoring

FIG. 12

shows apparatus

1200 that supports reporting physiological data in accordance with an embodiment of the invention. Withapparatus1200, the implantedcomponent1205 of the medical device system communicates with the relayingmodule1215 via telemetry antenna1210. Similarly, theexternal component1225 communicates with the relayingmodule1215 viaantenna1220. In the embodiment, atelemetry link1221 between relayingmodule1215 andantenna1220 comprises a 3 MHz body wave telemetry link. To avoid interference, the relayingmodule1215 may communicate with the external and implanted components using differing communication schemes. In some embodiments, the reverse direction and the forward direction oftelemetry link1221 may be associated with different frequency spectra. The relayingmodule1215 thereby provides a greater range of communications between components of medical device system. For example, in the embodiment of an implanted system, an external programmer may communicate with an implanted device from a more remote location. The external programmer may be across the room and still be in communication via the relayingmodule1215. With the telemetry booster stage, the use of an implanted system is more convenient to the patient, in particular at night while sleeping or when taking a shower, eliminating the need for an external device to be worn on the body. In an embodiment, relatingmodule1215 may also have the features and functionality ofexternal device503 as shown inFIG. 5.

As shown inFIG. 13, in an embodiment, the system allows the residential, hospital or ambulatory monitoring of at-risk patients and their implanted medical devices at any time and anywhere in the world.Medical support staff1306 at a remotemedical support center1314 may interrogate and read telemetry from the implanted medical device and reprogram its operation while thepatient10 is at very remote or even unknown locations anywhere in the world. Two-way voice communications1310 via,satellite1304, viacellular link1332 orland lines1356 with thepatient10 and data/programming communications with the implantedmedical device1358 via a belt worntransponder1360 may be initiated by the patient10 or themedical support staff1306. The location of thepatient10 and the implantedmedical device1358 may be determined viaGPS1302 andlink1308 and communicated to the medical support network in an emergency. Emergency response teams can be dispatched to the determined patient location with the necessary information to prepare for treatment and provide support after arrival on the scene. See for example, U.S. Pat. No. 5,752,976.

An alternative or addition to the remote monitoring system as described above in conjunction withFIG. 13 is shown in thesystem1450 ofFIG. 14, which shows a patient10 sleeping with animplantable Monitor1458 and/or optional therapy device as described above in connection with the above-described systems. Theimplantable device1458, upon detection of a neurologic event may alert a remote monitoring location via local remote box1452 (as described in U.S. Pat. No. 5,752,976),telephone1454 andphone lines1456 or the patient's care provider via anRF link1432 to a pager-sizedremote monitor1460 placed in other locations in the house or carried (i.e., belt worn) by thecare provider1462. Theremote caregiver monitor1460 may include audible buzzes/tones/beeps, vocal, light and/or vibration to alert thecaregiver1462 of patient's monitor in an alarm/alert condition. The RF link may include RF portable phone frequencies, power line RF links, HomeRF, Bluetooth, ZigBee, WIFI, MICS band (medical implant communications service), or any other interconnect methods as appropriate.

Each of the above embodiments utilized graphical user interfaces that are suitable for displaying data records that have been retrieved from the implantable medical device.

Data Retention and Recording Techniques

Discussed herein are techniques for selecting and storing sensed physiological data in an implanted medical device for subsequent reporting to an external device. As used herein, the term data record encompasses the sensed physiological data, summary information, or simply a pointer that references a location in memory where the sensed physiological data is stored. Thus, the concept of storage of data records in first and second data structures envisions possibilities of storage of the sensed physiological data and the storage of their associated pointers. As an example, summary information data may be stored in the first and second data structures wherein the more detailed and more space consuming waveform data (pre-detection data, post-detection data, etc.) may be stored, and pointed to, in an associated memory (such as a loop record buffer).

Mapping from entries in the first and second data structures to the waveform physiological data that is stored in the associated memory may be achieved with pointers. Each entry in the event log may point to its corresponding waveform data, or each waveform data may point to its corresponding data in the event log. Alternatively, multi-directional pointers in an “allocation table” or “allocation data structure” may be pointed to by the priority structures. Thus, when a data record is overwritten or replaced as discussed herein, both the data record itself and its mapping to the event log may be changed/removed in the allocation structure.

FIG. 15

shows process

1500 for storing physiological data in accordance with an embodiment of the invention. Again, the techniques disclosed herein may be implemented in any of the implantable medical device systems discussed above and may be performed in conjunction with the flow diagrams ofFIGS. 4,8, and11 above. The process described therein may be implemented by computer executable instructions performed by a processor or dedicated hardware in the form of digital logic gates (also referred as hardware rather than software/firmware) in the implanted device. (Process1500 may be performed by a logic component that may include a processor and/or digital logic such as an ASIC.) The process described herein discusses the storage of data records within one or more data structures and the moving of those data records. It will be appreciated, however, that these processes may be implemented by storage and movement of pointers, wherein each pointer points to data that is stored in memory, and/or the storage and movement of the sensed physiological data.

The technique for storing data records may begin withstep1501, wherein an implanted device receives an instruction or configuration data from a clinician, typically via telemetry. For example, the clinician or patient may activate loop recording to store data collected for a specified neurological event between a specified time and having a minimum severity level (as will be discussed). Instep1503, an implanted device initiates a detection process for a neurological event. This detection process may implement an event detection algorithm to process the sensed physiological data and determine the possible onset of a neurological event. The neurological event may be associated with a nervous system disorder (e.g., an epileptic seizure) having a severity level and a time duration. The neurological event may also be a cardiac event. For example, a patient's heart rate may increase with an occurrence of an epileptic seizure. Accordingly, the implanted device may initiate loop recording upon determination of a neurological event by processing of sensed physiologic signals or upon determination of any number of other criteria.

By way of example only, the system may monitor one set of physiological signals (e.g., EEG) and trigger loop recording of those same signals upon detection of a neurological event. As another example, the system may monitor any combination of a plurality of physiological signals (including EEG, respiratory, and cardiac) and trigger loop recording of only a subset of those physiological signals based on predetermined criteria being satisfied with respect to one or more of the other physiological signals.

In an embodiment, the implantable medical device may have a set of monitoring elements sensing brain activity and another set of monitoring element that sense a physiological activity other than the brain (e.g., heart activity such as a heart arrhythmia and/or respiratory activity). The device may then implement a detection algorithm to determine the possible onset of a possible neurological event (e.g., a seizure) based on the sensed signals from either the first or second monitoring elements. Once a neurological event is detected, data records associated with the first and second monitoring elements may be stored in memory in accordance with the teachings herein.

Alternatively or additionally, the device may initiate loop recording upon indication to do so by the patient based, for example, on a patient detecting a neurological event. In the event the patient initiates loop recording based on detection of a neurological event (wherein, however, the detection process of the implanted device has not detected the neurological event), the priority index (discussed below) for such data may be set at a higher level such that the data is stored in a memory. In the situation where the patient experiences a neurological event but the medical device has not detected the event, the physiological sensed data may be particularly important for storage and subsequent evaluation. In an embodiment, once activated by a patient, loop recording may save the data for 30 seconds before the indicated seizure and 3 minutes after the seizure. However, to allow for the fact that the patient may not mark the seizure until the seizure has ended, the ECG loop recording may begin 3 to 5 minutes before the patient mark. This time period may be programmable. As discussed below, a subset or a composite of physiologic channels is selected from the available physiologic channels based on a selection criterion.

As can be appreciated, the patient marking of an event such as a seizure may occur while the loop recording is already in process. In such a situation, the data recorded may be given a higher priority and/or may be recorded in greater detail.

As noted above, clusters of data may be stored where each cluster represents a detected event. Algorithms for determining where a cluster starts and ends may vary but typically will include some threshold level that the signal must exceed for the start of the cluster and the same or a second threshold level that the signal must drop below for a period of time to end the cluster. It should be noted that while the signal used to determine the threshold level may be actual EEG signals, it is expected that the signal will be filtered and/or processed so that the signal being used is a derived signal based on the underlying EEG activity. The threshold level may also include time constraints so that spikes of EEG activity, such as sleep spindles, are not interpreted as a detected event. The determination of the cluster period can be aided with the patient mark. In an embodiment, the event mark providing by the patient may be used to link one cluster that took place right before the patient marking and a second cluster that occurs right after the marking provided by the patient. In an embodiment, clusters within 30 or 60 or 120 seconds of the mark could be joined with the data surrounding the mark. Furthermore, if the patient mark is provided near the end of a planned loop recording, the loop recording may be extended a predetermined period.

FIG. 21 illustrates an exemplary method of using the patient's input as a marker for improved analysis of the received signals, although numerous other methods will become apparent after reviewing this disclosure. First instep2110, a patient (or caregiver) may use the patient activator922 (FIG. 9) (also known as a patient therapy manager) or an equivalent device to indicate observance of a symptom that typically corresponds with and may typically precede more severe clinical symptoms. A signal from thepatient activator922 is received by the implanted medical device. In an embodiment, the observed symptoms may be an aura or chest pain indicative of an impending seizure but may also indicate other neurological or physiological events. Upon receipt of the patient marker, a check instep2115 is made to determine whether data from a sensor is currently being loop recorded. If it is not, then instep2120, the implanted device begins to loop record. If the implantable device uses a circular buffer, then the loop recording can begin prior to the patient mark, as discussed herein. Instep2125, a record indicating the time of input is stored. Once the loop recording is complete (or during the recording if a substantially real-time determination is desired), the relevance of the patient marking may be determined instep2130.

As can be appreciated, the patient mark of a seizure can be used to measure and/or improve the accuracy of the detection algorithm. For example, if a patient provides an indication of a seizure that the detection algorithm has also detected, the relevance of the patient mark is that it is a true positive. However, if the detection algorithm did not detect the event, the patient mark may indicate a false negative. As discussed above, marking a false negative may be helpful in improving the sensitivity of the detection algorithm and may help tune the implanted device. For example, the sensed data surrounding a false negative can be stored, downloaded and analyzed by a caregiver and the detection algorithm may be updated so as to better detect such events. The true positive also helps improve the statistical analysis of the data because certain clusters can be treated as known seizures instead of solely relying on the detection algorithm to determine whether the sensed signals are a seizure. Thus, the patient mark can be used to improve statistical analysis of detected events and to tune the detection process. For example, the additional data stored because of the patient mark can also be used to determine when and/or what was not detected. As can be appreciated, know what was not detected can also be useful in tuning the detection algorithm. Furthermore, the additional data may be used by an expert or recommender system to suggest both changes in settings (such as sensitivity thresholds including duration) and also whether additional processing would be helpful to detect the event. As discussed herein, the transmission of data to and from the implanted device can be accomplished via telemetry in a known manner as described elsewhere in the art.

In addition, it is expected that the patient may not observe or appreciate every event that is detected, because, for example, the event may not have noticeable clinical symptoms. Thus, the implanted system can automatically keep track of patient detected events versus non-patient detected events. The data associated with the patient mark can be calculated, summarized and automatically displayed. For example, but without limitation, the number of additional or linked clusters stored due to one or more patient marks could be displayed as a summary to provide an indication of the effectiveness of the patient's use of the patient activator.

Referring again toFIG. 15, if the detected neurological event meets the criterion provided by the clinician, the system triggers loop recording instep1505. Instep1507, the implanted device stores a data record associated with the sensed physiological data in a first data structure (in an embodiment, a circular buffer as will be discussed in greater detail inFIG. 16). Again, the data record may either be a pointer or the physiological data (which may be actual EEG data or may be derived data through some form of processing). An exemplary data entry is discussed below in relation toFIGS. 18 and 19. In an embodiment, the circular buffer stores chronologically sequenced data records in sequential memory locations. When the last location in the circular buffer is reached, the next data record is entered in the first memory location of the circular buffer. Although the embodiment ofFIG. 15 and the corresponding discussion herein is with respect to a circular buffer, it will be appreciated that the circular buffer may be any memory device or structure and the data records may be stored and replaced according to an associated priority index (discussed below). When the memory device is full, any new data may replace data in the memory device having the lowest associated priority index. Alternatively, as discussed below, the data may be automatically transferred out once the priority index is calculated.

With the storage of neurological data instep1507, the implanted device may also obtain and retain indicators of severity when the neurological event occurs instep1509. Examples of retained indicators include a maximum value of foreground ictal severity and heart rate during the neurological event. Indicators of severity are discussed further below and may be used to determine a priority index.

If the first data structure becomes full, a new data record replaces either the oldest data record or the existing data record having the lowest associated priority index. However, instep1511, if the associated priority index of the data record being replaced by the new data record is sufficiently large, the data record being replaced may be stored in a priority buffer instead of being permanently removed. It will be appreciated, that the data records may vary in size. Accordingly, for a new data record having a sufficiently large size, the system may require multiple old data records to be removed. Alternatively, only the least important portions of the data record would need to be removed from the first data structure.

The priority index of a data entry, discussed next, is determined in

steps

1515,1517, and1519 in concert with clinician input and information obtained by the implanted device. In the embodiment of the invention, steps1507,1509, and1511 are executed in a parallel manner. Instep1513, the recording of neurological and other physiological data in the circular buffer is completed. In an alternative embodiment, the data entry may be transferred from the first data structure (such as the circular buffer) to the second data structure (priority buffer) upon the priority index of the data entry being determined. In such an embodiment, the second data structure may begin to be populated with data records without requiring for the first data structure to become full and start replacing older data entries. Since the stored data entry is promptly removed to the second data structure, the circular buffer may vary in size.

One embodiment for determining the priority index of a data record is as follows. Instep1515, the most recent data record is classified. For example, the neurological event may be classified as an epileptic seizure. Instep1517, a priority index of the neurological event is determined. The priority index may be dependent on a severity level. In an embodiment, the severity level is a function of at least one characteristic of the neurological event, including the duration, spread, and measure of ictal content (e.g., peak ratio of algorithmically derived foreground to background ictal score) of the neurological event. Moreover, the embodiment may utilize other derivative indicators for determining the severity level. For example, a binary evidence count may be determined in which each new sample is compared to a criterion (e.g., a threshold determined by the product of a background value and a configured threshold value). If the criterion is met, a value of “1” is inserted into a data structure that contains Y previous Boolean values. Otherwise, if unmet, a value of “0” is inserted into the data structure. Upon insertion, a second value is removed from the data structure to ensure that the data structure retains a size of Y bits. The Boolean value of the first sample is then compared to the Boolean value of the second sample. The difference is used to maintain a separate counter value which reflects how many Boolean values in the data structure (e.g., Y-bit buffer) are “1”. In this way, the evidence count reflects the number of samples out of the last Y samples which met the threshold criterion, and therefore ranges from 0 to Y.

As another example, a slew-rate limited output of an associated physiologic channel may be determined by incrementing the slew-rate limited output by a small amount if the sample is greater than the current value of the slew-rate limited output. Otherwise, the slew-rate limited output is decremented by the small amount. The severity level may be determined from the maximum value of the evidence count or the slew-rate limited output associated with the set of physiologic channels contained in a data record.

In addition, the priority index of the neurological event may be dependent on an associated event, e.g., a cardiac event or a user signal in the form of telemetry, presence of a medical magnet, and so forth. In an embodiment, the priority index is stored in a data field in the data record. However, another embodiment of the invention may store the priority index in an associated data structure or may be calculated when it is needed.

Instep1519, a composite severity measure of a group of neurological events, which includes the current neurological event, is determined. The composite severity measure may be a function of determined severity levels of individual events and the relative chronological occurrence of events. For example, a group of epileptic seizures that occur within a short period time may be indicative of a more significant cumulative event than a single epileptic seizure.

While the embodiment illustrated inFIG. 15 shows a sequential ordering of steps, it will be appreciated (as discussed above) that the process may execute the steps in a different order or may execute some of the steps in a parallel fashion.

The priority index may be dependent on the severity level, which may be expressed as a function f(x₁, x₂, . . . , x_n), and/or on associated factors, which may be expressed as a function g(y₁, y₂, . . . y_m). Variable x_icorresponds to a characteristic of a physiologic event (e.g., a neurological event), and variable y_jis an associated factor. For example, characteristics of a neurological event may include a duration of a neurological cluster, a spread of a neurological cluster (which may be measured by the number of electrodes detecting the cluster), a statistic of an associated neurological signal, and an event class. A statistic may be a computed quantity characteristic of a physiologic signal or an ensemble of physiologic signals. An event class may include a most-severe seizure (potentially correlating to status epilepticus), a severe seizure, a sub-clinical seizure, a least severe seizure (potentially correlating to what is more likely to be an erroneous detection), a brady event, a tachy event, arrhythmia, and a detection on each electrode channel.

An associated event may include external event information and associated physiologic information that are correlated to a neurological event. For example, the associated physiologic information may be the age of the event. As another example, an external event may be a cardiac event, a chemical event, a scheduling event, a patient-initiated event (through a physical interface, e.g., a magnet interface or a telemetry interface), a caregiver-initiated event (e.g., through a telemetry channel), and a device-based event. Associated physiologic information may include heart rate, blood pressure, breathing rate, and glucose level. Moreover, an external event may not be associated with a neurological event. For example, a physician may instruct, through a programmer, that physiologic data for a designated time be stored and reported.

The priority index may be expressed as a mathematical combination of the severity level function f(x₁, x₂, . . . , x_n) and the associated factor function g(y₁, y₂, . . . y_m). For example, the priority level may be expressed as:

priority index=f(x₁,x₂, . . . ,x_n)+g(y₁,y₂, . . . y_m) (EQ. 1A)

Either f(x₁, x₂, . . . , x_n) or g(y₁, y₂, . . . y_m) may be a continuous function, a discrete-value function, a Boolean function, or a combination of the above function types. As another example, the priority level may be expressed as:

priority index=f(x₁,x₂, . . . ,x_n)·g(y₁,y₂, . . . y_m) (EQ. 1B)

The priority index may be more generally expressed as a function h(z₁, z₂), where

priority index=h(f(x₁,x₂, . . . ,x_n),g(y₁,y₂, . . . y_m)) (EQ. 1C)

FIG. 16 showsdata structure1600 for storing physiological data in accordance with an embodiment of the invention. In this embodiment,data structure1600 includescircular buffer1601 andpriority queue1603. (Another embodiment of the invention may implementcircular buffer1601 and/orpriority queue1603 as a different type of data structure.)Circular buffer1601 stores the most recent ten data records (DATA(−2) to DATA(−3) corresponding to data entries1607-1625). (An exemplary record is discussed withFIG. 18.) However, the embodiment may store a different number of data records, depending on the allocated memory forcircular buffer1601. In the example shown,circular buffer1601 currently stores data records that have been acquired between the previous nine time units (corresponding to data record DATA(−9)) and the current time (corresponding to data record DATA(0)).New data pointer1605 points to a data entry incircular buffer1601. At the current time,new data pointer1605 points to the data entry where new data is stored. Previous to storing data record DATA(0),data entry1611 stored data record DATA(−10), which was previously acquired 10 time units ago. The new data record may be associated with physiologic signals and may be further associated with a neurological event. With the next subsequent data acquisition,new data pointer1605 will point to the next data entry that currently stores data record (DATA(−9)). In the embodiment, the new data record is stored in the data entry that stores the oldest data record. When last data entry1625 (currently corresponding to data record DATA(−3)) ofcircular buffer1601 is reached, the subsequent data acquisition is stored indata entry1607 currently storing data record DATA(−2).

As discussed, the data record may be the sensed physiological data or a pointer that references a location in memory where the sensed physiological data is stored. Thus, the concept of storage of data records in first and second data structures envisions both possibilities of storage of the sensed physiological data and the storage of their associated pointers. Also as discussed,circular buffer1601 may be in the form of any other data structure. In such an alternative embodiment, physiological data may be stored in a memory device such that when a new data record is to replace an existing data record, the data record with the lowest priority index is replaced. Alternatively, the data entries being replaced may vary depending on the relative sizes of the new and old data entries.

When a data record has been overwritten by the new data record, a portion of the old data record (which may be the entire data record) is stored inpriority queue1603 if the corresponding priority index is sufficiently large or exceeding a threshold criterion. In one embodiment, the more significant the corresponding physiologic event, the larger the corresponding priority index. However, in another embodiment the priority level may decrease with the significance of the physiologic event.Priority queue1603 is capable of storing data records in data entries1627-1635; however, as shown in the example,priority queue1603 currently stores two data records. The remaining data entries1631-1635 are empty and may store subsequent data acquisition.Data entry1627 is currently storing data record DATA(−101), which was acquired 101 time units ago, anddata entry1629 is currently storing data record DATA(−24), which was acquired 24 time units ago. For example, the priority index must have a value of 3 or higher in order to be stored inpriority queue1603 when the corresponding data record is overwritten incircular buffer1601. In the event that thepriority queue1603 is also filled, the system may determine that a data record shall be retained inpriority queue1603 only if the priority index of the removed data record fromcircular buffer1601 has a larger priority index than any of the stored data entries inpriority queue1603. In other embodiments, the system may utilize a combination of the above retention strategies or other strategies.

In an embodiment of the invention, a data record (stored in data entries1607-1625) incircular buffer1601 may contain a different amount of data than a data record (stored in data entries1627-1635) inpriority queue1603. For example, data may be discarded to reduce the precision or to discard physiologic data that is not directly associated with a neurological event (e.g., cardiac data).

In an embodiment of the invention, stored data (either stored incircular buffer1601, orpriority queue1603, or both) may be selected in accordance with a retention policy in order to support intelligent data loss (data triage) or offloading (e.g., to an external device) and to conserve memory usage and reduce/bound the time required to transfer the data. For example, an implanted device may store data indata structure1600 for sensor channels fulfilling a predetermined criteria, may discard waveform data while retaining statistics, summary data, or burden information, enable or disable or prioritize sensor channels based on algorithm criteria, sensor location, or sensor type, or override the retention policy when a clinician provides instructions over a telemetry channel. Moreover, the embodiment may further use data compression if the data is compressible to compress the data content of the new or old record.

Reporting Physiological Data

In accordance with another aspect of the invention, in response to an instruction from a clinician, an implanted device organizes stored physiological data according to the associated priority index and reports a predetermined number of data records that are deemed as having a higher priority index than the other stored data records. Such structures for reporting stored physiological data are disclosed inFIGS. 12-14.

FIG. 17 shows flow diagram1700 for reporting physiological data in accordance with an embodiment of the invention.Step1701 corresponds to an implanted device receiving a request (e.g., through a telemetry communications channel). Instep1703, the data records (both incircular buffer1601 and priority queue1603) are sorted according to the associated priority index if the associated priority index is sufficiently large (e.g., three or higher).Step1705 determines if the number of reported data records is less than or equal to N. N can be, for example, the number of records that the user wishes to retrieve. If so, the physiological data associated with the data record is reported to a clinician's workstation over the telemetry communications channel and the number of reported data is incremented instep1707.Step1705 is then repeated. If the number of records is greater than N, the reporting of neurological data is terminated instep1709, andprocedure1700 is completed.

A physician may request physiological data associated with N data records that are stored incircular buffer1601 andpriority queue1603 and having a minimum priority index to be reported to, for example, a physician programmer over a telemetry link. The physician programmer may thereby display the requested physiological data or at least those where the priority index is above a threshold. When so instructed, implanted device sorts the physiological data associated with the data records incircular buffer1601 andpriority queue1603 and reports physiological data associated with the N data records having at least the minimum priority index. Once the data is reported, the implanted device may release the memory space previously allocated to the identified data records, mark that memory space for deletion, and/or delete the data records. If more than N data records have the minimum priority index, the implanted device may select N data records based on a temporal criterion (e.g., the most recent or the oldest), being greater than a predetermined priority index, or a priority index that is dependent on an external event. Moreover, system may select a data record so that each occurring event type is included in the reported data records. Because the invention may be implanted within any implantable medical device system, however, it will be appreciated that the external components may vary in configuration, components and/or features.

Data Format

FIG. 18 shows exemplaryphysiological data1800 that can be retained in implanted medical device. Thephysiological data1800 comprisessummary information data1801,pre-detection data1803,post-detection data1805, and event/etcinformation data1807.Pre-detection data1803 contains sampled data.1809-1823 that is collected before a detected physiologic event such as an epileptic seizure.Post-detection data1805 contains sampled data1825-1839 after the detected physiologic event. In the exemplary data shown inFIG. 18,pre-detection data1803 includes sampled

data

1809 and1817 that are obtained fromneurological channel1, sampled

data

1811 and1819 that are obtained fromneurological channel2, sampled

data

1813 and1821 that are obtained fromneurological channel3, and sampled

ECG data

1815 and1823 that are obtained from the ECG channel.Post-detection data1805 includes sampled

data

1825 and1833 that are obtained fromneurological channel1, sampled

data

1827 and1835 that are obtained fromneurological channel2, sampled

data

1829 and1837 that are obtained fromneurological channel3, and sampled

data

1831 and1839 that are obtained from the ECG channel. While the physiologic channels sampled before and after the physiologic event are same inexemplary record1800, different physiologic channels may be sampled before and after the physiologic event. Additionally,summary information data1801 contains summary information aboutrecord1800. The embodiment of the invention supports physiological data that has a fixed length or that has a variable length. Event/Etc information data1807 includes information about the associated physiologic event type and other associated information. Although depicted as data from different channels interleaved together, such data may also be stored in a non-interleaved fashion. For example, ECG may be stored separately from the other channels, etc.

FIG. 19 shows components of the exemplary physiological data as shown inFIG. 18.Summary information data1801 includes triggersource data1901, length of event/etcinformation data1903, size ofloop record1905, andrelevance information1907, Triggersource data1901 identifies what triggered recording (e.g., a neurological channel number). Length of event/etcinformation data1903 is indicative of the length of event/etcdata1807, size ofloop record1905 contains the length ofrecord1800, andrelevance data1907 indicates the relevance level of each sampled channel as will be later discussed. Event/Etc information1807 contains

entries

1909 and1911. Each entry includesevent type1913 that is indicative of an event associated withrecord1800, start offsetdata1915 includes the offset ofrecord1800 from the start of the loop recording, and typespecific field1917 includes specific data that is associated with the event. (However, in an embodiment of the invention, a data entry may store a pointer to physiologic data rather than store the physiologic data itself.)Record1800 may be associated with a plurality of event types, each being indicated in separate entries, e.g.,

entries

1909 and1911.

Data structure1600 (as shown inFIG. 16) typically can store 1 to 16 Mbytes of data.Record1800 may contain anywhere from approximately 5 Kbytes to approximately 1 Mbyte. Typically, a record contains 30-90 Kbytes per record. The embodiment supports both fixed size records (e.g., 60 Kbytes) or variable size records. With variable size records the retention scheme, as previously discussed, may require that a plurality of records or only portions of records be removed from circular buffer201 in order to store a new record. Also, different parts of a record may be stored in physically different memory components (e.g., summary/header information data601 in one place (a log) and sampled

data

1803 and1805 in another (waveform memory) with a pointer in the header/summary information indicating where the start of the data is and how much data is there).

As discussed, in an embodiment of the invention, a data record may be a pointer to physiologic data rather than store the physiologic data itself. Thus, for example,summary information data1801 and/or event/etcinformation data1807 may be stored in the first and second data structures wherein some or all of the entries may be mapped, via pointers, to more detailed and more space consuming waveform data (pre-detection data1803, post-detection data1805) in an associated memory (such as a loop record buffer).

Selective Storage of Data Channels

The embodiment supports the generation of records at different rates. The record rate will vary widely with an implanted device's configuration and with the patient. For example, a clinician may desire to record every detected seizure over a time period. There are some patients that have 1 seizure every 6 months, but there are some that have 500 seizures a day. A typical patient may have 2 unique events (seizure, button press, etc) per day.

WhileFIG. 18 shows an exemplary configuration with three neurological channels and an ECG channel, an embodiment of the invention supports a variable number of neurological channels so that the memory of an implanted device can store necessary information over a desired period of time. The channels may be adjusted either in quasi-static manner or in a dynamic manner. In an embodiment, the implanted medical device contains a channel selection component, which may be software or hardware logic for performing a process of selecting the one or more channels for data storage as described herein. An evaluation processing component may thereby obtain the selected channel data and process the channel data also in accordance with the techniques described herein.

For applications like the detection of epileptic seizures and the subsequent recording of data, it is advantageous, from a quality and resource stand-point, to be able to select a subset (which may include all physiologic channels) of the sensed input channels for detection processing and/or loop-recording. While this can be done manually prior to the availability of relevant data by defining the selected channels, manual configuration is less flexible and less desirable than allowing the closed-loop, real-time, quasi-real-time, periodic, or otherwise automated selection of physiologic channels.

With implantable detection algorithms and applications, there are typically a multitude of input sense channels. The degree of interest for each of these sense channels is often discernable while an implanted device is collecting data and storing the data in a memory component, e.g., a memory structure. As each sample is processed or stored as a short recording of samples, an application specific method may be employed to identify the most interesting (relevant) physiologic channels. Samples from selected physiologic channels may be utilized. Detection processing is typically resource intensive such that all physiologic channels may be utilized. Data recording in a loop-recording rapidly fills memory and then later prologs telemetry. In order to effectively process and store the samples, judicious selection of physiologic channels may be desired.

With a detection algorithm operating on physiological signals, it is likely that simply obtained information may reveal which channels to retain or continue processing. Often, physicians/users may be interested in examining, using, and retaining data from the physiologic channels in which an indication of an event is present. For example, if a seizure is detected on two of eight detection channels of data being recorded from the brain, only those two neurological channels need be retained. In this example, the savings in storage may be significant since only two channels are kept and six neurological channels are discarded, corresponding to a 4 to 1 sample compression.

Sample compression may be important if memory of an implanted device is limited in the above example, one needs to decide which two channels should be kept. The decision may be based upon a number of factors including, independently-calculated channel severity e.g., save the most severe), channel event onset time (e.g., save the first to detect), and seizure burden indication on the physiologic channel. In such a scenario, the flow diagram ofFIG. 15 may be modified such that a severity calculation is performed prior to actual storage of the data in the first data structure. Or stored data records may be automatically removed from those data channels that are determined that the data is not important. Similarly, this approach may be extended to different and mixed physiologic signals. For example, if the multiple EEG channels are being polled, one may keep only those EEG channels where the neurological event data is most important (e.g., based on a severity scoring). If both cardiac, intra-cranial pressure (ICP) and EEG channels are being polled, one may keep only the signal types that deviated from norm. If the intra-cranial pressure rises significantly, one may keep samples from the ICP channel. If the previous scenario occurs while there is a tachycardia event, one may further keep samples from the ECU channel. Furthermore, if a seizure occurs near the same time, one may keep samples form all three channels.

The above approach may be extended to include the retention of more than one channel from a channel list sorted by relevancy as determined by a function of various factors (e.g., onset time, presence and severity of an event) as previously discussed. One may keep the most relevant physiologic channels of the channel list. For example, one may keep the three most relevant (“interesting”) physiologic channels of five physiologic channels. Keeping the two or most relevant physiologic channels is referred as the “multi-max” of the channel list.

Alternatively, the system may retain a composite signal, which may be, for example, the sum of a plurality of channels. Other embodiments of the invention may utilize other derivatives. Furthermore, the channels themselves may be down-sampled or precision-reduced so that, for example, the least relevant samples are kept at 16 times less precision in amplitude and 2 times lower sampling rate resulting in about 4 times smaller data size.

Furthermore, with seizure detection and many other neurological applications the most interesting physiologic channels may be identified as those physiologic channels that are associated with the maximum energy value, normalized energy value, severity, or energy approximated value in the applications seizure frequency domain. For seizure detection, the seizure frequency domain may be 5-70 Hz (more narrowly this is typically 5-45 Hz) or (for fast ripple detections) at very high frequency (approximately 200 Hz or more). The energy value may often be sampled as the “rectified” (absolute value, square, etc.) of the samples taken directly from the filtered EEG signal. While this data may constantly be higher in some neurological locations, it is useful to normalize the data across the physiologic channels with manual or automatic gain control. Automatic gain control may be implemented as the daily comparison of running 1-pole filters and with the appropriate adjustment in response or in some other manner. For example, the gain of each physiologic channel may be adjusted so that the associated background signals are similar (to each other and/or to a target value), which may be referred to as background normalization. Because short signals may appear similar to interesting material over the short term, it is useful to smooth or reject outliers from the energy signals before comparison. Smoothing may be done, for example, using 1-pole filters, voting schemes, running or block percentile filters, percentile tracking technologies or slew-rate limiters to eliminate or mitigate, for example, the effect of epileptiform discharges (ED's) and sleep-spindles.

With an embodiment of the invention, the selection of physiologic channels may occur after filtering (e.g., bandpass, notch, FIR, and IIR) the physiologic channels. For example, an EEG signal may be filtered in the 10-60 Hz range to remove the bulk of the EEG energy content that may otherwise mask the ictal content. As another example, the physiologic channels may be filtered in the 180-250 Hz range in order to study “fast-ripple” events.

The above channel selection process is applicable both for records stored incircular buffer1601 andpriority queue1603. Moreover, the channel selection criterion forcircular buffer1601 may be different from the channel selection criterion forpriority queue1603.

FIG. 20 shows the selection of channel data for retention indata structure1600. In the embodiment,channel selection component2001 anddata structure1600 are supported by an implanted component.Channel selection component2001 obtains N physiologic channels from various sensors that may include neurological signals, EEG signals, and a mixture of physiologic signals.Channel selection component2001 selects a subset of physiologic channels from the N physiologic channels using a selection criterion as previously discussed. Data from the selected physiologic channels is included inrecord1800. Moreover, the selected physiologic channels may be processed by the implanted device and/or stored into memory. The embodiment, as shown inFIG. 20, stores record1800 intodata structure1600 as previously discussed. Similarly, the above approach may be extended to different and mixed physiological signals.

In the embodiment,channel selection component2001 includes a channel processor that selects a subset of physiological channels using the selection criterion. The channel processor may be implemented with a microprocessor, a signal processor, or other processor means.

As can be appreciated by one skilled in the art, a computer system with an associated computer-readable medium containing instructions for controlling the computer system can be utilized to implement the exemplary embodiments that are disclosed herein. The computer system may include at least one computer such as a microprocessor, digital signal processor, and associated peripheral electronic circuitry.

Thus, embodiments of the invention are disclosed. One skilled in the art will appreciate that the above disclosure is not limited to being practiced with the embodiments disclosed. The disclosed embodiments are presented for purposes of illustration and not limitation, and the inventions are limited only by the claims that follow.

Claims

1. A method of recording physiological signals detected from a patient, comprising:

sensing physiological data with a sensor of an implantable medical device;

transmitting a patient mark indicative of an onset of clinical symptoms related to a neurological event from a patient activator;

receiving the patient mark in the implantable medical device;

storing the patient mark in memory of the implantable medical device; and

initiating storing in memory of the implantable medical device the physiological data sensed before and after the patient mark based on the reception of the patient mark.

2. The method ofclaim 1, wherein the storing of physiological data comprises:

storing a first cluster from a first electrode; and

storing a second cluster from a second electrode, wherein the first and second electrode are positioned in different anatomical locations.

3. The method ofclaim 1, increasing a priority of the physiological data in memory based on reception of the patient mark relative to other stored physiological data.

4. The method ofclaim 1, further comprising displaying a summary of an impact of the patient mark to the stored data.

5. The method ofclaim 1, wherein storing the patient mark comprises:

creating a time stamped record indicating that the patient mark was received; and

saving the time stamped record in memory.

6. The method ofclaim 1, wherein the physiological data for which storage is initiated based on the reception of the patient mark comprises physiological data that is sensed 3 to 5 minutes before the patient mark.

7. The method ofclaim 1, wherein initiating storing of the physiological data sensed before the patient mark based on the reception of the patient mark comprises transferring the physiological data from a circular buffer to memory.

8. The method ofclaim 1, further comprising:

detecting a neurological event based on the physiological data;

comparing the patient mark to results of the neurological event detection; and

determining a relevance of the results of neurological event detection.

9. The method ofclaim 8, wherein the detection results do not include a cluster associated with a detected seizure.

10. The method ofclaim 8, wherein the relevance is selected from a list consisting of true positive and false negative.

11. The method ofclaim 8, further comprising adjusting a detection algorithm in response to the relevance.

12. An apparatus comprising:

a processor; and

a non-transitory computer-readable memory storing instructions that, when executed by the processor, are configured to cause an implantable medical device to perform operations comprising:

sensing physiological data;

receiving a patient mark in the implantable medical device from an external patient activator, the patient mark indicative of an onset of clinical symptoms related to a neurological event;

storing the patient mark in memory of the implantable medical device; and

13. The apparatus ofclaim 12, wherein the instructions, when executed by the processor, are configured to cause the apparatus to perform operations comprising:

storing a first cluster from a first electrode; and

14. The apparatus ofclaim 12, wherein the instructions, when executed by the processor, are configured to cause the apparatus to perform operations comprising: increasing a priority of the physiological data in memory based on reception of the patient mark relative to other stored physiological data.

15. The apparatus ofclaim 12, wherein the instructions, when executed by the processor, are configured to cause the apparatus to perform operations comprising: displaying a summary of an impact of the patient mark to the stored data.

16. The apparatus ofclaim 12, wherein the instructions, when executed by the processor, are configured to cause the apparatus to perform operations comprising:

saving the time stamped record in memory.

17. The apparatus ofclaim 12, wherein the physiological data for which storage is initiated based on the reception of the patient mark comprises physiological data that is sensed 3 to 5 minutes before the patient mark.

18. The apparatus ofclaim 12, wherein initiating storing of the physiological data sensed before the patient mark based on the reception of the patient mark comprises transferring the physiological data from a circular buffer to memory of the implantable medical device.

19. The apparatus ofclaim 12, wherein the instructions, when executed by the processor, are configured to cause the apparatus to perform operations comprising:

detecting a neurological event based on the sensed physiological data;

comparing the patient mark to results of the neurological event detection; and

determining a relevance of the results of neurological event detection.

20. The apparatus ofclaim 19, wherein the detection results do not include a cluster associated with a detected seizure.

21. The apparatus ofclaim 19, wherein the relevance is selected from a list consisting of true positive and false negative.

22. The apparatus ofclaim 19, wherein the instructions, when executed by the processor, are configured to cause the apparatus to perform operations comprising: adjusting a detection algorithm in response to the relevance.