Predictive QRS detection and RR timing system and method - Patent Application 20070122997

Abstract

To provide systems and methods built for predictively timing the inflation and / or deflation of an intra-aortic balloon pump.SOLUTION: A controller operates in three states: (1) initialization state, (2) learning state, and (3) peak detection state. The controller decomposes a patient's electrocardiogram signal to a power signal. It then learns characteristics of the patient's electrocardiogram signal during the learning state and computes adaptive threshold parameter values. During the peak detection state, the controller applies the learnt threshold parameter values on a current electrocardiogram signal to identify occurrence and timings of R peaks in the electrocardiogram signal. The R-to-R peak timings are then used to trigger inflation of an intra-aortic balloon pump.SELECTED DRAWING: Figure 4A

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. Patent Application No. 16 / 401,368, filed May 2, 2019, the disclosure of which is incorporated herein by reference in its entirety. [Background technology]

[0002] background The present invention relates generally to medical devices, and more particularly to systems and methods related to the operation of medical devices, such as intra-aortic balloon pumps, based on physiological measurements. According to the Centers for Disease Control and Prevention website, approximately 5.7 million adults in the United States suffer from heart failure. Each year, approximately 100,000 people nationwide are diagnosed with advanced heart failure and require some form of medical support, such as an intra-aortic balloon pump. A balloon pump is positioned inside the aorta, typically in the proximal descending aorta. The balloon pump (typically 40-50 milliliters in volume) is inflated and deflated in response to the contraction of the left ventricle. During diastole, the balloon is inflated, thereby transporting blood from the ascending aorta and aortic arch to the coronary arteries to oxygenate the myocardium. During systole, as the left ventricle contracts, the balloon is deflated to reduce afterload. This technique is called "counterpulsation." Summary of the Invention

[0003] [Brief explanation of the drawings]

[0004] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate several aspects and, together with the description, serve to explain the disclosed principles. [Figure 1A] 1 illustrates an intra-aortic balloon pump, according to some embodiments. [Figure 1B] 1 illustrates an intra-aortic balloon pump, according to some embodiments. [Figure 2] FIG. 2 is a block diagram illustrating some of the components typically incorporated into a controller, according to some aspects. [Figure 3A] 1 illustrates components of a controller, according to some aspects. [Figure 3B] 1 illustrates a state diagram of a controller, according to some aspects. [Figure 4A] 1 illustrates a QRS complex in an electrocardiogram signal, according to some embodiments. [Figure 4B] 1 is a flow diagram illustrating a method performed by a power signal module of a controller, in accordance with some aspects. [Figure 4C] 1 illustrates a received and processed electrocardiogram signal, in accordance with some aspects. [Figure 4D] 1 illustrates a received and processed electrocardiogram signal, in accordance with some aspects. [Figure 4E] 1 illustrates a conversion signal, according to some aspects. [Figure 4F] 1 illustrates a decomposed electrocardiogram signal, in accordance with some aspects. [Figure 4G] 1 illustrates a decomposed electrocardiogram signal, in accordance with some aspects. [Figure 4H] 1 illustrates an electrocardiogram signal and a corresponding decomposed power signal, in accordance with some aspects. [Figure 4I] 1 illustrates an electrocardiogram signal and a corresponding decomposed power signal, in accordance with some aspects. [Figure 5A] 1 is a flow diagram illustrating a method performed by a QRS detection module of a controller, in accordance with some aspects. [Figure 5B] 1 illustrates the selection of h-peak values ​​for calculating thresholds, according to some embodiments. [Figure 6] 1 illustrates the results of applying an adaptive threshold to detect the occurrence of an R peak, according to some embodiments. [Figure 7] 1 illustrates the results of applying an adaptive threshold to detect the occurrence of an R peak, according to some embodiments. DETAILED DESCRIPTION OF THE INVENTION

[0005] Detailed Description Intra-aortic balloon pump devices typically process ECG signals for efficient and optimal operation. One of the most important parts of ECG signal processing and intra-aortic balloon pump operation is interpreting and characterizing the QRS complex. The QRS complex is the name for the combination of three graphical deflections seen on a typical electrocardiogram. In adults, the QRS complex typically lasts 0.06–0.10 seconds; in children and during physical activity, the QRS complex may be shorter. The Q, R, and S waves occur in rapid succession, do not appear in all leads, and are usually considered together because they represent a single event. The Q wave is any downward deflection immediately following the P wave. The R wave follows as an upward deflection, and the S wave is any downward deflection after the R wave. The T wave follows the S wave, and sometimes an additional U wave follows the T wave.

[0006] The R wave is one of the most important parts of the complex, playing an essential role in diagnosing cardiac rhythm irregularities and determining heart rate variability (HRV). Conventional systems for detecting the QRS complex include differentiation methods, digital filters, neural networks, filter banks, hidden Markov models, genetic algorithms, and maximum a posteriori (MAP) estimators. These methods are highly sensitive to noise and typically fail to detect the exact timing of the R wave in an ECG signal. As a result, conventional intra-aortic balloon pump systems and methods suffer from an inability to predict the timing of balloon pump inflation and / or deflation.

[0007] In view of the deficiencies of conventional techniques for accurately detecting the occurrence and timing of R waves, the present inventors recognized that a new technique for more accurate, reliable, and efficient prediction of R wave and R peak timing would be of significant utility.

[0008] A software, hardware, and / or firmware facility ("Facility") is described that predicts the timing of future R-peak occurrences to provide a solution for efficiently inflating and / or deflating an intra-aortic balloon pump. In some embodiments, the facility operates as a state machine including three states: (1) an initialization state; (2) a learning state; and (3) a peak detection state. By implementing some or all of the methods discussed below, the facility improves prediction of future R-peaks so that the corresponding intra-aortic balloon pump is inflated and / or deflated in an efficient and reliable manner. This results in an improved patient therapy experience that keeps the patient's heart beating in a rhythm closer to its natural state. The facilities for QRS detection and / or RR timing described herein may be used in a variety of devices, including, but not limited to, pacemakers, cardiac monitors, defibrillators, heart rate monitors, smartwatches, athletic accessories, and the like.

[0009] DESCRIPTION OF THE DRAWINGS The following description provides certain specific details of illustrative examples. However, those skilled in the art will understand that aspects can be practiced without many of these details. Likewise, those skilled in the art will also understand that the present disclosure may include many other features not described in detail herein. Furthermore, some well-known structures or functions may not be shown or described in detail below to avoid unnecessarily obscuring the relevant description of the various examples.

[0010] Some implementations are discussed in more detail below with reference to the drawings. Referring now to the drawings, FIGS. 1A and 1B illustrate an example of an intra-aortic balloon pump facility 100. The intra-aortic balloon pump 100 may be dimensioned and configured to hang inside a patient's aorta. The intra-aortic balloon pump 100 may include a balloon 101 configurably connected to an internal driveline 103 that connects to a skin interface device 105. As shown in FIG. 1B, an external driveline 107 may connect the skin interface device 105 to a driver 109.

[0011] At its proximal end, balloon 101 is connected to the distal end of an internal driveline 103. A skin interface device 105 connects the proximal end of internal driveline 103 to the distal end of an external driveline 107. The proximal end of external driveline 107 is connected to a driver 109. Driver 109 may include or be connected to a controller 111. An arterial interface 113 may be sized and configured for passing internal driveline 103 through an arterial wall.

[0012] In some embodiments, the intra-aortic balloon pump 100 may include a controller 111 configured in accordance with the systems and methods described herein. The controller 111 may control the operation of the valves and bellows (not shown) of the driver 109 to control the flow of pumping medium (e.g., air) and the inflation and deflation of the balloon 101. The controller 111 may receive one or more signals from the balloon 101 and the surrounding area. The signals may include those received from electrodes, pressure sensors, etc. (e.g., electrocardiogram signals). In some embodiments, the controller 111 receives, for example, from the skin interface device 105, one or more of the following data values ​​about the patient: a patient identifier (e.g., a unique patient identification alphanumeric string); the patient's first and last name; the patient's age; the patient's average heart rate; the patient's maximum heart rate; the patient's minimum heart rate; the patient's pulse; the patient's average R-R time interval; the patient's maximum R-R time interval; the patient's minimum R-R time interval; the patient's average R-peak; the patient's maximum R-peak; the patient's minimum R-peak, etc. In some embodiments, the controller 111 retrieves and / or calculates one or more of the above-listed data values ​​based on a subset of the received information about the patient. For example, based on a patient identifier, the controller 111 retrieves (e.g., from a communicatively coupled memory) the following data values: the patient's first and last name; the patient's average heart rate; the patient's pulse; and the patient's average RR time interval.

[0013] 2 is a block diagram illustrating some of the components typically incorporated in the controller 111. In various embodiments, the controller 111 includes the following: a central processing unit (“CPU”) 101 for executing computer programs; computer memory 102 for storing programs and data while they are in use, including facilities and associated data, an operating system including a kernel, and device drivers; persistent storage 103, such as a hard drive or flash drive, for persistently storing programs and data (e.g., patient-related data); a computer-readable media drive 104, such as a floppy, CD-ROM, or DVD drive, for reading programs and data stored on a computer-readable medium; and a network connection 105 for connecting the computer system to other computer systems to send and / or receive data, such as via the Internet or another network, and one or more of each of its networking hardware, e.g., switches, routers, repeaters, electrical and optical cables, light emitters and receivers, wireless transmitters and receivers, etc. For example, via the network connection 105, the controller 111 can retrieve patient-related data from a remote data storage location (e.g., a cloud data storage location). Although a computer system configured as described above is typically used to support the operation of the controller 111, those skilled in the art will appreciate that the facility may be implemented using various types of devices and configurations and with various components.

[0014] 3A illustrates components of a controller according to some embodiments. In some embodiments, the controller may be configured on a drive unit for an intra-aortic balloon pump. The controller 111 may include a signal receiver module 307, a power signal module 309, a QRS detection module 311, an RR timing module 313, and a drive unit moderator module 315. In some embodiments, the controller operates as a state machine including three states depicted in FIG. 3B: (1) initialization state 325; (2) learning state 330; and (3) peak detection state 335.

[0015] Initialization state The signal receiver module 307 of the controller may be configured to receive a signal (e.g., an analog electrocardiogram signal). For example, the controller receives an electrocardiogram signal from a remote device (e.g., one or more skin interface devices (SIDs) implanted within the patient's body). Applicant's U.S. Pat. Nos. 9,265,871 and 10,137,230 provide further details regarding suitable SIDs and are incorporated herein by reference in their entireties. In some embodiments, the signal receiver module 307 digitizes the received electrocardiogram signal and / or applies one or more filters to obtain a filtered electrocardiogram signal suitable for further processing by the facility. Exemplary filters may include, but are not limited to, a bandpass filter, a differential filter, a square-law filter, and / or applying a moving window integration. FIG. 4A illustrates a portion of an electrocardiogram signal 400 (including a QRS complex 403) according to some embodiments. The QRS complex 403 is a feature of an electrocardiogram signal indicative of cardiac depolarization. The R-peak 401 is an element within the QRS complex 403. Real-time detection of these peaks requires an appropriate threshold that can distinguish between prominent peaks that typically represent QRS complexes and minor peaks such as T or P waves that may be associated with noise. The facilities described herein are configured to accurately detect and predict QRS complexes and R-peaks. During the initialization state 325, the controller receives electrocardiogram signal data and initializes one or more variables, buffers, memories, etc. For example, the controller receives a patient identifier (e.g., a unique patient identification alphanumeric string) and initializes values ​​for one or more of the following variables: the patient's first and last name; the patient's age; the patient's average heart rate; the patient's maximum heart rate; the patient's minimum heart rate; the patient's pulse rate; the patient's average R-R time interval; the patient's maximum R-R time interval; the patient's minimum R-R time interval; the patient's average R-peak; the patient's maximum R-peak; the patient's minimum R-peak, etc.

[0016] Learning status After completing the initialization state 325, the controller enters a learning state 330. During the learning state 330, the controller generates a first power spectrum signal by decomposing a received first electrocardiogram signal (e.g., using historical electrocardiogram signals for the patient). A power signal module 309 within the controller is configured to generate the power signal based on the digitized and / or filtered electrocardiogram signal. FIG. 4B is a flow diagram illustrating a method performed by the power signal module 309 of the controller 111, according to some aspects. The method includes an act of receiving an electrocardiogram (ECG) signal 401 (act 402). FIGS. 4C and 4D illustrate portions of received and / or processed electrocardiogram signals 410 and 415, according to some aspects. In act 403, the power signal module 309 may decompose the received electrocardiogram signal into one or more subbands or frequencies of interest. Thereafter, in operation 405, the power signal module 309 may aggregate one or more subbands of interest to generate a power signal. The one or more subbands of interest may be selected based on characteristics of the received signal, such as the number of information vectors received, the number of electrode leads used to acquire the electrocardiogram signal, the positions of the electrode leads used to acquire the electrocardiogram signal, etc.

[0017] As an illustrative example, the facility collects electrocardiogram signal data using three electrode leads (each corresponding to three vectors) and selects data from one of the electrode leads (corresponding to one of the three vectors) for monitoring at 1000 Hz sampling. The power signal module 309 applies a transform signal (e.g., Haar wavelet 420 illustrated in FIG. 4E) to decompose the electrocardiogram signal from the selected electrode lead into discrete frequency signals. For example, FIGS. 4F and 4G illustrate decomposed electrocardiogram signals 430 and 440 according to some embodiments. The power signal module 309 then selects the following subbands of interest: a first portion in the d5 frequency range (15.625 Hz to 31.25 Hz) and a second portion in the d6 frequency range (7.8125 Hz to 15.625 Hz). The power signal module 309 aligns the selected portions in the d5 and d6 frequency ranges to calculate the power signal ("h signal") = |d5.d6|.

[0018] In some embodiments, the facility collects electrocardiogram signal data using three (or more) electrode leads (each corresponding to three (or more) vectors) and selects data from all three electrode leads (all three vectors) for monitoring at 500 Hz sampling. After applying a Haar wavelet transform to the electrocardiogram data, the power signal module 309 selects the following subbands of interest from the vector with the best signal characteristics: a first portion in the d4 frequency range (31.25 Hz to 62.5 Hz) and a second portion in the d5 frequency range (15.625 Hz to 31.25 Hz). The delay in detecting and processing electrocardiogram data is a linear function of the number of electrodes (or vectors) for which data is utilized (delay = n / 2, where n is a function of the number of higher frequency bands). For example, when data from a single electrode (and a single vector) is used by the power signal module 309, such that the selected higher frequency band is d6, the delay = 32 seconds. Similarly, when data from three electrodes (and three vectors) are used by the power signal module 309, such that the selected higher frequency band is 5, the delay amount is 16 seconds. The file is TIFF0007776836000002.tif8128.

[0019] In some embodiments, the power signal includes two peaks corresponding to the rise from Q to R and the drop from R to S in the QRS complex. For example, Figure 4H illustrates a power signal 450 including two peaks 450a and 450b corresponding to the rise from Q to R and the drop from R to S in the QRS complex of the corresponding electrocardiogram signal 445. Figure 41 is another diagram depicting an electrocardiogram signal 460 and its corresponding decomposed signal 465.

[0020] In some embodiments, after generating a power signal corresponding to a received electrocardiogram signal, the QRS detection module 311 in the controller may detect future R-peaks based on features derived from the generated power signal. Figure 5A is a flow chart illustrating a method performed by the QRS detection module 311 of the controller according to some embodiments. In operation 501, the QRS detection module 311 collects information about the calculated power signal and separates values ​​corresponding to non-QRS and QRS signals. In operation 503, the QRS detection module 311 fits one or more models to the distributions of the non-QRS and QRS signals and determines initial parameters and / or one or more thresholds. For example, during a learning state of the controller, the QRS detection module 311 of the controller selects a first portion of the generated power spectrum signal and divides it into one or more subsets containing sub-portions of the selected portion of the generated power spectrum signal.

[0021] As shown in the example of FIG. 5B, the controller selects a 7.5-second portion of the power spectrum signal 530 corresponding to the electrocardiogram signal 520. The controller divides the power signal 530 into five subsets (530a, 530b, 530c, 530d, and 530e), each of which spans 1.5 seconds. For each of these subsets, the controller calculates an h-peak value for the subportion. In some embodiments, the controller stores the calculated h-peak values ​​of the selected portion of the power spectrum signal within the subset (i.e., the maximum h-signal value per subset). For example, the controller stores the h-peak values ​​of the subsets (535a, 535b, 535c, 535d, and 535e) in the first threshold buffer 540. The controller then calculates a first adaptive threshold R-peak value based on the calculated peak values ​​stored in the first threshold buffer. For example, the controller may calculate a median (or average, weighted average, weighted median, etc.) for the h-peak values ​​in the threshold buffer 540 and then calculate a percentage of the median (e.g., a configurable percentage of 40%) as the first adaptive threshold R-peak value for the patient. The configurable percentage may be a value within a predetermined range (e.g., 20% to 27%). The configurable percentage value may be configured (or reconfigured) automatically or manually (e.g., by a physician). For example, the configurable percentage value may be configured (or reconfigured) automatically using the signal-to-noise ratio (SNR) of the ECG signal. As the SNR increases, the configurable percentage value may be reduced, which may lead to faster detection. On the other hand, when the patient's noise level rises to the point where it nearly overwhelms the R-waves, the configurable percentage value may be increased to avoid inducing noise pumping. As another example, the configurable percentage value may be automatically configured (or reconfigured) using measurements of R-wave amplitude variability. When a patient's R-wave peak amplitude changes significantly or repeatedly, as may be the case in a heart failure patient with premature ventricular contractions (PVCs), bigeminy or trigeminality, or other cardiac conditions, the controller can adjust the configurable percentage value so that it can detect both low-amplitude and high-amplitude R-waves.An example of such a measurement might be the standard deviation or variance of several sets of maximum h signal peak values ​​measured during the learning phase, or occasionally during a detection phase, or even during another "test" period of a particular duration. The controller can periodically update the values ​​in the first threshold buffer 540 based on the updated electrocardiogram signal values ​​and then recalculate the first adaptive threshold R peak value for the patient.

[0022] In some embodiments, the controller implements an outlier detection routine that removes outliers from the calculation of the h-peak value. In this manner, the facility addresses the problem of conventional systems that are known to be sensitive to outliers by implementing outlier detection at various stages of the process performed by the QRS detection module 311 in which the h-peak value is calculated. In some embodiments, an outlier may be defined as being three standard deviations from the median (or mean, or weighted average, or weighted median, etc.) value. In some embodiments, the controller uses an approximation of the outlier for computational simplicity and to provide a more accurate estimate of the standard deviation. In some embodiments, the outlier is determined by γ * It may be defined as a value greater than the mean (peak), where γ > 1. In some embodiments, the QRS detection module 311 may only consider right-sided outliers (i.e., values ​​significantly greater than the median).

[0023] Peak detection status After calculating the first adaptive threshold R peak value for the patient, the controller enters a peak detection state 335. As depicted in FIG. 5A at operation 505, the controller's QRS detection module 311 receives the patient's current / most recent electrocardiogram signal (second electrocardiogram signal). As in the learning state 330, the controller in the peak detection state 335 generates a second power spectral signal by decomposing the received second electrocardiogram signal. For example, the controller selects a first portion of the generated second power spectral signal and divides the selected first portion of the generated second power spectral signal into one or more second subsets including subportions of the selected first portion of the generated second power spectral signal. For example, the controller selects a portion of the second power spectral signal spanning 7.5 seconds and divides it into five subsets, each spanning 1.5 seconds. For each of these subsets, the controller calculates the peak value of the subportion of the selected first portion of the second power spectral signal that is within the subset (i.e., the peak h signal value for each subset). In some embodiments, the controller stores the calculated peak values ​​of the subset in a second threshold buffer. The controller then calculates a second adaptive threshold R peak value based on the calculated peak values ​​stored in the second threshold buffer. For example, the controller calculates a median value for the h signal values ​​in the second threshold buffer and then calculates a percentage of the median value (e.g., a configurable percentage value of 40%) as the second adaptive threshold R peak value for the patient.

[0024] In operation 507, the controller applies the model learned during the learning state (e.g., in operation 503) to the second power spectrum signal. In some aspects, the controller determines whether the currently received electrocardiogram signal includes a QRS complex by applying a first adaptive threshold R-peak value, a second adaptive threshold R-peak value, or both to the power signal corresponding to the currently received electrocardiogram signal. For example, in some aspects, the presence of a QRS complex is determined when the power signal is greater than a first adaptive threshold R-peak value, a second adaptive threshold R-peak value, or both. The absence of a QRS complex is determined when the power signal is less than the threshold. Because electrocardiogram signals are non-stationary signals whose statistical characteristics change over time, in some aspects, techniques have been developed to make this threshold adaptive to change with the changing characteristics of the signal. Additionally, the facility may include other mechanisms for improving detection consistency, such as including advance information about the location of the QRS, limiting how closely adjacent QRS complexes can be in time, and improving future decisions based on past errors in detection.

[0025] 6 and 7 illustrate the results of applying an adaptive threshold to detect the occurrence of R-peaks. For example, as illustrated in FIG. 6, a power signal 601 may be generated, which reflects an electrocardiogram signal 603. An adaptive threshold 605 may be applied to the power signal 601 to detect a QRS complex 607 and estimate an R-peak 609. FIG. 7 similarly illustrates the results of applying an adaptive threshold to detect the occurrence of R-peaks. As shown, a power spectrum signal 701 may be generated, which reflects the electrocardiogram signal 703. As shown, an R-peak detection 705 may be overlaid on the power spectrum 701.

[0026] When the received electrocardiogram signal is determined to include a QRS complex, the QRS signal, QRS detection, R peak, and other data from the QRS detection module 311 may be passed to the RR timing module 313. The RR timing module 313 calculates an RR time interval value indicating the timing between two consecutive R peaks in consecutive QRS complexes in the patient's received second electrocardiogram signal. The RR time interval value may be determined by averaging a buffer of the seven most recent RR intervals. In some embodiments, the RR time interval value is calculated as the median of the five most recent RR intervals. Using the median value may result in outlier rejection above the median (vs. the mean value) as well as a faster response of the system to rapidly changing heart rates. The controller may determine the patient's heart rate by taking the reciprocal of the calculated RR interval and multiplying by 60 to provide the heart rate in beats per minute (BPM). The controller may also generate a prediction of the timing of one or more future R-peaks based on the calculated R-R time interval value indicating the timing between two successive R-peaks in successive QRS complexes. In some embodiments, the controller triggers inflation of the at least one intra-aortic balloon pump based on the generated prediction of the timing of the future R-peaks.

[0027] Based on the prediction of the future R element occurrence, the drive unit moderator 315 of the control device 111 may be used to control the drive unit 305 of an intra-aortic balloon pump or other medical device. For example, balloon inflation may be timed to the predicted occurrence of the future R element in the QRS complex.

[0028] The controller periodically updates the values ​​in the first and / or second threshold buffers based on the updated electrocardiogram signal values ​​and then recalculates the second adaptive threshold R-peak value for the patient. In some embodiments, the controller updates the first adaptive threshold R-peak value based on the second adaptive threshold R-peak value, and / or vice versa (FIG. 5A, operation 509).

[0029] In some embodiments, at least a portion of the contents of the second threshold buffer are replaced by those of the first threshold buffer (effectively resetting the adaptive threshold calculation based on the learning state value). For example, after every reconfigurable time interval (e.g., 30 seconds), the contents of the second threshold buffer (stored in the peak detection state) are replaced by the contents of the first threshold buffer (stored in the learning state). In some embodiments, if the controller does not detect an R wave within a certain time interval (e.g., 10 seconds, 30 seconds, etc.), the controller declares a coded warning indicating a missing R wave. The missing R wave warning can be transmitted and / or communicated (e.g., via a visual indicator (blinking light, flashing message, etc.), an audio indicator (horn, message readout, etc.), or any other means). When the controller declares a missing R wave warning, it may reset to the initialization state and restart the peak detection process.

[0030] In some embodiments, the QRS detection module 311 includes one or more correction processes. In some embodiments, the correction process may adapt the median metric in the presence of false negatives. In particular, the correction process may learn from errors in detection. In some embodiments, the correction process may include an adaptive search-back, which attempts to identify possible QRS events that were incorrectly missed (false negatives) and adjust the median metric accordingly. Assuming the heart rate does not drop dramatically from QRS event to QRS event, the lack of QRS event detection can be an indicator of possible missed events (i.e., a false negative).

[0031] The correction process may look at one or more windows of the power signal to determine whether a QRS event was detected in the window. If a QRS event was not detected, the correction process may look at a finite window and identify a maximum value of the power signal over that range. The correction process may assume that the maximum value corresponded to a QRS event that was missed by the QRS detection module 311 because the threshold used for detection was too high. Taking this missed detection into account, the correction process may incorporate the peak value associated with the missed QRS event and reduce the resulting median value by a predefined percentage (e.g., 15%). This incorporation of the missed value into the median value and further percentage reduction allows the QRS detection module 311 to adapt quickly in the presence of a reduction in the peak value associated with a QRS event, resulting in fewer missed beats.

[0032] A similar correction process may be used in the event of a false positive. In other words, the median metric used by the QRS detection module 311 may also be adapted in the presence of false positives. In some embodiments, the correction process may consider false positives resulting from inaccurate detections made directly before or after a true QRS event. Assuming that peak values ​​in the power signal associated with a true QRS event will have larger peak values ​​than those belonging to a pre- or post-QRS false event, peaks that may correspond to false detections are separated into two categories: pre-QRS false detections and post-QRS false detections. A pre-QRS false detection occurs when an event reaches a time window (e.g., 400 ms) before a true QRS event triggers the detection logic. The 400 ms time window reflects an assumed maximum heart rate of 150 bpm. Under this scenario, the QRS detection module 311 triggers a pre-QRS event because the threshold is too low, since both the peak value associated with the pre-QRS event and the peak value associated with the QRS event would be incorporated into the median metric.

[0033] Continual incorporation of peak values ​​associated with pre-QRS false events into the median metric would perpetuate this error. Thus, using a correction process, if two peaks are identified within a time window (e.g., 400 ms) of each other, where the second peak (i.e., a true QRS event) is higher than the first peak (i.e., a pre-QRS event), the value of the first peak is removed from the median metric and the resulting median value is increased by a certain percentage (e.g., 15%). This increase in the threshold attempts to ensure that similar false positives will not occur with subsequent events. In this way, the correction process may correct for pre-QRS false events.

[0034] A late QRS false positive can occur when an event triggers the QRS detection module 311 within a time window (e.g., 400 ms) after a true QRS event. Under this scenario, the first peak (i.e., the true QRS event) will be higher than the second peak (i.e., the late QRS event). The QRS detection module 311 triggering the second event indicates that the threshold is too low, resulting in a false positive. Given that incorporating the peak value associated with this late QRS false event would likely lead to a continued false positive, the value of the second peak is not incorporated into the median metric. The median metric is again increased by a certain percentage (e.g., 15%). This increase in the threshold attempts to ensure that similar false positives will not occur for subsequent events.

[0035] The techniques described herein have several advantages over conventional methods because they initialize quickly by approximating, rather than directly constructing, the characteristics of the QRS and non-QRS distributions. Furthermore, by approximating, rather than directly calculating, the QRS and non-QRS distributions, they require less memory and processing than conventional systems. Furthermore, the techniques described for the QRS detection module 311 are faster to adapt to sudden changes in the power signal.

[0036] example Several aspects of the present technology will become apparent from the following examples, which are numbered for convenience (1, 2, 3, etc.), and are provided as examples and are not intended to limit the subject technology. 1. A system for operating an intra-aortic balloon pump, comprising: at least one intra-aortic balloon pump positioned adjacent the patient's heart; at least one controller configurably connected to the at least one intra-aortic balloon pump; Including, the at least one controller includes at least one processor and at least one non-transitory memory; The at least one control device initializing values ​​of one or more parameters based on the patient's identity; receiving a first electrocardiogram signal from the patient; generating a first power spectrum signal by decomposing the received first electrocardiogram signal; selecting a first portion of the generated first power spectrum signal; dividing the selected first portion of the generated first power spectrum signal into one or more subsets comprising sub-portions of the selected first portion of the generated first power spectrum signal; For each subset in the one or more subsets of the selected first portion of the generated first power spectrum signal: calculating a peak value of the subportion of the selected first portion of the first power spectral signal that is within the subset; storing the calculated peak values ​​of the subset in a first threshold buffer; Calculating a first adaptive threshold R peak value based on the calculated peak value stored in the first threshold buffer. It is configured as follows: system. 2. At least one control device comprises: receiving a second electrocardiogram signal of the patient; generating a second power spectrum signal by decomposing the received second electrocardiogram signal; selecting a first portion of the generated second power spectrum signal; dividing the selected first portion of the generated second power spectrum signal into one or more second subsets comprising sub-portions of the selected first portion of the generated second power spectrum signal; For each subset in the one or more second subsets of the selected first portion of the generated second power spectrum signal: calculating a peak value of the subportion of the selected first portion of the second power spectral signal that is within the subset; storing the calculated peak values ​​of the subset in a second threshold buffer; Calculating a second adaptive threshold R peak value based on the calculated peak value stored in the second threshold buffer. 10. The system of Example 1, further comprising: 3. The system of Example 2, wherein the at least one controller is further configured to update the first adaptive threshold Rpeak value based on the second adaptive threshold Rpeak value. 4. At least one control device comprises: determining whether the received second electrocardiogram signal includes a QRS complex by applying the calculated first adaptive threshold R-peak value, or the calculated second adaptive threshold R-peak value, or both, to the generated second power spectrum signal; When the received electrocardiogram signal is determined to include a QRS complex, calculating an R-R time interval value indicative of timing between two successive R-peaks in successive QRS complexes in the patient's received second electrocardiogram signal based on the calculated first adaptive threshold R-peak value, or the calculated second adaptive threshold R-peak value, or both. The system of Example 2, further configured as follows: 5. The system of Example 4, wherein the at least one controller is further configured to generate a prediction of the timing of future R peaks based on the calculated R-R time interval value indicating the timing between two consecutive R peaks in consecutive QRS complexes. 6. The system of Example 5, wherein the at least one controller is further configured to trigger inflation of the at least one intra-aortic balloon pump based on the generated prediction of the timing of the future R-peak. 7. At least one skin interface device configurably connected to at least one intra-aortic balloon pump and configurably connected to at least one controller. further comprising the at least one control device is further configured to receive a patient identity and a set of patient-specific information from the at least one skin interface device. 10. The system of any of the preceding examples. 8. The first power spectrum signal is applying at least one transformation signal to decompose the received first electrocardiogram signal into discrete frequency signals; selecting one or more portions of the discrete frequency signal; and calculating the first power spectrum signal by aligning one or more portions of the selected discrete frequency signal; 2. The system of any of the preceding examples, 9. The system of example 8, wherein at least one transformed signal is based on a Haar wavelet. 10. The calculated power signal is based on a first selected portion of the discrete frequency signal and a second selected portion of the discrete frequency signal; the first selected portion of the discrete frequency signal is in a first frequency range of 15.625 Hz to 31.25 Hz; the second selected portion of the discrete frequency signal is in a second frequency range of 7.8125 Hz to 15.625 Hz; The system in Example 8. 11. The system of any of the preceding examples, wherein the first portion of the generated first power spectrum signal spans 7.5 seconds. 12. The system of Example 2, where the first portion of the generated second power spectrum signal spans 7.5 seconds. 13. The system of any of the preceding examples, wherein each subset in the one or more subsets of the selected first portion of the generated first power spectrum signal spans 1.5 seconds. 14. The system of Example 2, wherein each subset in the one or more second subsets of the selected first portion of the generated second power spectrum signal spans 1.5 seconds. 15. One or more parameters are Patient's first and last name, the patient's age, the patient's average heart rate, the patient's maximum heart rate, the patient's minimum heart rate, The patient's pulse, The patient's mean RR time interval, The patient's maximum RR time interval, The patient's minimum RR time interval, The patient's mean R-peak, The patient's maximum R peak, The patient's minimum R peak, or any combination of these 2. The system of any of the preceding examples, 16. A computer-implemented method for operating an intra-aortic balloon pump, comprising: initializing values ​​of one or more parameters based on the patient's identity; receiving a first electrocardiogram signal of the patient; generating a first power spectrum signal by decomposing the received first electrocardiogram signal; selecting a first portion of the generated first power spectrum signal; dividing the selected first portion of the generated first power spectrum signal into one or more subsets comprising sub-portions of the selected first portion of the generated first power spectrum signal; For each subset in the one or more subsets of the selected first portion of the generated first power spectrum signal: calculating a peak value of the sub-portion of the selected first portion of the first power spectral signal that is within the subset; and storing the calculated peak values ​​of the subset in a first threshold buffer; and calculating a first adaptive threshold R peak value based on the calculated peak value stored in the first threshold buffer; 11. A computer-implemented method comprising: 17. Receiving a second electrocardiogram signal of the patient; generating a second power spectrum signal by decomposing the received second electrocardiogram signal; selecting a first portion of the generated second power spectrum signal; dividing the selected first portion of the generated second power spectrum signal into one or more second subsets comprising sub-portions of the selected first portion of the generated second power spectrum signal; For each subset in the one or more second subsets of the selected first portion of the generated second power spectrum signal: calculating a peak value of the subportion of the selected first portion of the second power spectral signal that is within the subset; and storing the calculated peak values ​​of the subset in a second threshold buffer; and calculating a second adaptive threshold R peak value based on the calculated peak value stored in the second threshold buffer; 17. The computer-implemented method of Example 16, further comprising: 18. Updating the first adaptive threshold R peak value based on the second adaptive threshold R peak value. 18. The computer-implemented method of Example 17, further comprising: 19. determining whether the received second electrocardiogram signal includes a QRS complex by applying the calculated first adaptive threshold R-peak value, or the calculated second adaptive threshold R-peak value, or both, to the generated second power spectrum signal; and calculating an R-R time interval value indicative of timing between two successive R-peaks in successive QRS complexes in the patient's received second electrocardiogram signal based on the calculated first adaptive threshold R-peak value, or the calculated second adaptive threshold R-peak value, or both, when the received electrocardiogram signal is determined to include a QRS complex; 18. The computer-implemented method of Example 17, further comprising: 20. A computer-implemented method for operating an intra-aortic balloon pump, comprising: initializing the value of one or more parameters based on the patient's identity; receiving a first electrocardiogram signal of the patient; generating a first power spectrum signal by decomposing the received first electrocardiogram signal; selecting a first portion of the generated first power spectrum signal; dividing the selected first portion of the generated first power spectrum signal into one or more subsets comprising sub-portions of the selected first portion of the generated first power spectrum signal; For each subset in the one or more subsets of the selected first portion of the generated first power spectrum signal: calculating a peak value of the sub-portion of the selected first portion of the first power spectral signal that is within the subset; and storing the calculated peak values ​​of the subset in a first threshold buffer; calculating a first adaptive threshold R peak value based on the calculated peak value stored in the first threshold buffer; receiving a second electrocardiogram signal from the patient; generating a second power spectrum signal by decomposing the received second electrocardiogram signal; selecting a first portion of the generated second power spectrum signal; dividing the selected first portion of the generated second power spectrum signal into one or more second subsets comprising sub-portions of the selected first portion of the generated second power spectrum signal; For each subset in the one or more second subsets of the selected first portion of the generated second power spectrum signal: calculating a peak value of the subportion of the selected first portion of the second power spectral signal that is within the subset; and storing the calculated peak values ​​of the subset in a second threshold buffer; calculating a second adaptive threshold R peak value based on the calculated peak value stored in the second threshold buffer; calculating an R-R time interval value indicative of the timing between two successive R-peaks in successive QRS complexes in the patient's received second electrocardiogram signal based on the calculated first adaptive threshold R-peak value, or the calculated second adaptive threshold R-peak value, or both; generating a prediction of the timing of future R peaks based on the calculated R-R time interval value indicating the timing between two consecutive R peaks in consecutive QRS complexes; and triggering inflation of at least one intra-aortic balloon pump based on the generated prediction of the timing of the future R-peak. 11. A computer-implemented method comprising:

[0037] conclusion While illustrative embodiments are described herein, the scope includes any and all embodiments having equivalent elements, modifications, omissions, combinations (e.g., of aspects across various embodiments), adaptations, and / or variations that would be appreciated by one of ordinary skill in the art based on this disclosure. For example, the numbering and orientation of components shown in the exemplary systems may be modified.

[0038] Those skilled in the art will appreciate that the processes illustrated in Figures 4B, 5A, 5B, and 8 may be modified in various ways, for example, the order of operations may be rearranged, some operations may be performed in parallel, illustrated operations may be omitted or other operations may be included, illustrated operations may be divided into sub-operations, or multiple illustrated operations may be combined into a single operation, etc.

[0039] It will be appreciated by those skilled in the art that the above-described facilities may be adapted as is or extended in various ways. While the foregoing description refers to particular embodiments, the scope of the invention is defined only by the appended claims and the elements defined therein.

Claims

1. 1. A system for operating an intra-aortic balloon pump, comprising: at least one intra-aortic balloon pump configured to be positioned adjacent to the patient's heart; at least one controller configurably connected to the at least one intra-aortic balloon pump; the at least one controller comprises at least one processor and at least one non-transitory memory; The at least one control device detecting successive QRS complexes in the patient's electrocardiogram signal by generating a power signal corresponding to the patient's electrocardiogram signal, detecting when the value of the power signal is greater than a threshold R-peak value, and associating one QRS complex with when the value of the power signal is greater than the threshold R-peak value; using the detected successive QRS complexes to predict future occurrences of R waves in an electrocardiogram signal; controlling the intra-aortic balloon pump based on the predicted occurrence of the future R wave. system.

2. 2. The system of claim 1, wherein the at least one controller is configured to inflate the intra-aortic balloon pump based on the predicted occurrence of the future R-wave.

3. 2. The system of claim 1, wherein the at least one controller is configured to deflate the intra-aortic balloon pump based on the predicted occurrence of the future R-wave.

4. The system of claim 1 , wherein the threshold R-peak value is based on the patient's historical electrocardiogram signals.

5. 5. The system of claim 4, wherein the at least one controller is configured to generate a historical power signal based on the historical electrocardiogram signal, and the threshold R-peak value is a function of two or more peak values ​​associated with the historical power signal.

6. 2. The system of claim 1, wherein the predicted occurrence of the future R-wave represents an amount of time from a previously detected QRS complex in the electrocardiogram signal, the amount of time being a function of the time interval between the detected successive QRS complexes.

7. The system of claim 1 , wherein the at least one controller generates the power signal by applying a conversion signal to the electrocardiogram signal.

8. The system of claim 7 , wherein the transformed signal is a wavelet transform.

9. 10. The system of claim 1, wherein the at least one controller is further configured to determine if a QRS complex is not detected within a time frame in which a QRS complex is expected to occur, and to adjust the threshold R-peak value in response to the determination.

10. 10. The system of claim 1, wherein the at least one controller is further configured to determine whether a detected QRS complex is a false positive and to adjust the threshold R-peak value in response to the determination.

11. 1. A method for operating an intra-aortic balloon pump, comprising: a computer generating an electrical power signal corresponding to the patient's electrocardiogram signal, detecting when the value of the electrical power signal is greater than a threshold R-peak value, and detecting successive QRS complexes in the patient's electrocardiogram signal by associating an instance of the electrical power signal greater than the threshold R-peak value with a QRS complex; the computer using the detected successive QRS complexes to predict future occurrences of R waves in the electrocardiogram signal; the computer controlling an intra-aortic balloon pump based on the predicted occurrence of the future R-wave; A method comprising:

12. the computer inflating the intra-aortic balloon pump based on the predicted occurrence of the future R-wave; the computer deflating the intra-aortic balloon pump based on the predicted occurrence of the future R wave. The method of claim 11 , comprising one or more.

13. The method of claim 11 , wherein the threshold R-peak value is based on the patient's historical electrocardiogram signals.

14. 14. The method of claim 13, further comprising the computer generating a historical power signal based on the historical electrocardiogram signal, wherein the threshold R-peak value is a function of two or more peak values ​​associated with the historical power signal.

15. 12. The method of claim 11, wherein the predicted occurrence of the future R-wave represents an amount of time from a previously detected QRS complex in the electrocardiogram signal, the amount of time being a function of the time interval between the detected successive QRS complexes.

16. The method of claim 11 , further comprising the computer generating the power signal by applying a conversion signal to the electrocardiogram signal.

17. The method of claim 16, wherein the transformed signal is a wavelet transform.

18. The computer No QRS complex was detected within the time frame in which it was expected to occur; and The detected QRS complex is a false positive. determining one or more the computer adjusting the threshold R peak value in response to the determination. The method of claim 11 , comprising:

Images