Novel electrocardiogram waveform parameterization for cardiac health

Abstract

In some embodiments, there is provided system for processing electrocardiogram waveforms. The system may include at least one processor and at least one memory including instructions, which when executed by the at least one processor causes operations including receiving a segmented electrocardiogram waveform, wherein the segmented electrocardiogram waveform is for a heartbeat epoch; determining, for the segmented electrocardiogram waveform, initial estimates for mean, height, and standard deviation for a Gaussian curve; determining, for the initial estimates, upper bounds and lower bounds; fitting, using the initial estimates, the upper bounds, and the lower bounds, the Gaussian curve to the segmented electrocardiogram waveform; and determining, using the fitted Gaussian curve, one or more parameters and / or waveform shape parameters to enable application of a treatment. Related systems, methods, and articles of manufacture are also disclosed.

Description

[0001] PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0002] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0003] Filing Date: August 15, 2025 Customer No.: 39564

[0004] NOVEL ELECTROCARDIOGRAM WAVEFORM PARAMETERIZATION FOR CARDIAC HEALTH

[0005] CROSS-REFERENCE TO RELATED APPLICATION

[0006] [1] This application claims priority to US Provisional Application No. 63 / 683,873 filed August 16, 2024, entitled “NOVEL ELECTROCARDIOGRAM WAVEFORM parameterization FOR CARDIAC HEALTH.” The disclosure of which is incorporated herein by reference in their entirety.

[0007] BACKGROUND

[0008] [2] The term “electrocardiogram” (which is abbreviated as ECG or EKG) represents a heart's electrical activity through one or more cardiac cycles, such as heartbeats. Using electrodes placed on the skin of a patient, electrical activity of the heart may be obtained. As the heart depolarizes and the repolarizes during each heartbeat, the electrodes detect the electrical changes caused by depolarization and repolarization. The ECG may be obtained using for example a 12-lead ECG, but other quantities of leads and / or other types of devices (e.g., 2 leads, a smartwatch, a Holter monitor, and / or the like) may be used to record the heart’ s electrical activity.

[0009] [3] The ECG signal (also referred to as a waveform) may comprise components, such as a P-wave, Q-wave, R-wave, S-wave, and T-wave. The P-wave is a component in the ECG that may be used to, for example, represent atrial depolarization of the atria (e.g., before contraction). The Q-wave, R-wave, and S-wave (also referred to as the QRS complex) may be used to, for example, represent ventricular depolarization. Ventricular depolarization is an electrical activation of the ventricles before contraction. And, the T-wave may be used to, for example, represent ventricular repolarization. Ventricular repolarization is an electrical recovery of the ventricles as the ventricles prepare for a beat.

[0010] SUMMARY

[0011] [4] In some example embodiments, there may be provided parametrization for electrocardiogram waveform analysis.

[0012] [5] In some embodiments, there is provided system for processing electrocardiogram waveforms. The system may include at least one processor and at least one memory including instructions, which when executed by the at least one processor causes operations including PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0013] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0014] Filing Date: August 15, 2025 Customer No.: 39564 receiving a segmented electrocardiogram waveform, wherein the segmented electrocardiogram waveform is for a heartbeat epoch; determining, for the segmented electrocardiogram waveform, initial estimates for mean, height, and standard deviation for a Gaussian curve; determining, for the initial estimates, upper bounds and lower bounds; fitting, using the initial estimates, the upper bounds, and the lower bounds, the Gaussian curve to the segmented electrocardiogram waveform; and determining, using the fitted Gaussian curve, one or more parameters and / or waveform shape parameters to enable application of a treatment.

[0015] [6] In some variations, the fitting includes fitting 5 Gaussian curves to the P-wave, Q- wave, R-wave, S-wave, and T wave of the segmented electrocardiogram waveform. The receiving includes receiving a plurality of segmented electrocardiogram waveforms, wherein each of the plurality of segmented electrocardiogram waveforms is for a corresponding heartbeat epoch. An electrocardiogram waveform may be received and / or the electrocardiogram waveform may be filtered using at least one of a high-pass filter and / or a notch filter. One or more R-peaks may be detected from the electrocardiogram waveform based on one or more upward deflections in the electrocardiogram waveform. The electrocardiogram waveform maybe segmented to form one or more segmented electrocardiogram waveforms including the received segmented electrocardiogram waveform, wherein the segmenting is based on the detected one or more R-peaks in the electrocardiogram waveform and / or at least one dynamic offset. The one or more segmented electrocardiogram waveforms may be detrended to filter linear trends and / or drift. One or more peaks in a P-wave, a Q-wave, an R-wave, an S-wave, and / or a T-wave may be identified in the one or more segmented ECG waveforms. Using the fitted Gaussian curve, one or more of the following may be determined: a cycle index, a cycle trend, an R-squared, a root-mean squared error (RMSE), a global center, a global left index, a global right index, cycle-relative indices and times, gaussian fit features for the P, Q, R, S, and / or T-waves, voltage and / or morphology metrics for the P, Q, R, S, and / or T-waves, one or more intervals, and / or one or more inter-wave voltage differences. Using the fitted Gaussian curve, at least one rise time, at least one decay time, at least one rise decay symmetry, and / or at least one sharpness may be determined. A treatment may determined or applied based on the one or more parameters and / or the waveform shape parameters. PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0016] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0017] Filing Date: August 15, 2025 Customer No.: 39564

[0018] [7] The details of one or more variations of the subject matter described herein are set forth in the accompanying drawings and the description below. Other features and advantages of the subject matter described herein will be apparent from the description and drawings, and from the claims.

[0019] BRIEF DESCRIPTION OF THE DRAWINGS

[0020] [8] The accompanying drawings, which are incorporated in and constitute a part of this specification, show certain aspects of the subject matter disclosed herein and, together with the description, help explain some of the principles associated with the disclosed implementations. In the drawings,

[0021] [9] FIG. 1 depicts an example of an ECG processor, in accordance with some embodiments;

[0022]

[0010] FIGs. 2A and 2B depicts examples of processes for ECG parameterization, in accordance with some embodiments;

[0023]

[0011] FIG. 2C depicts another example of a process for ECG parameterization, in accordance with some embodiments;

[0024]

[0012] FIG. 3A depicts a plot of a raw ECG waveform and a high pass filtered ECG waveform, in accordance with some embodiments;

[0025]

[0013] FIG. 3B depicts a plot of a single epoch of an ECG waveform including P, Q, R, S, and / or T waves, in accordance with some embodiments;

[0026]

[0014] FIG. 3C depicts an example of the parameters obtained from a Gaussian curve fitted to an ECG waveform segmented to a single epoch, in accordance with some embodiments;

[0027]

[0015] FIG. 3D depicts detection of R-peaks within a segment of an ECG time series, in accordance with some embodiments;

[0028]

[0016] FIG. 3E depicts the alignment of ECG cycles to their corresponding R-peaks to create a set of segmented ECG waveforms, in accordance with some embodiments; and

[0029]

[0017] FIG. 4 depicts an example of a system, in accordance with some embodiments.

[0030] DETAILED DESCRIPTION

[0031]

[0018] In some embodiments, there is provided electrocardiogram (ECG) waveform analysis and parameterization that, among other things, analyzes ECG waveforms to discern for example certain differences in individual ECG components. For example, the ECG components PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0032] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0033] Filing Date: August 15, 2025 Customer No.: 39564 may include the P-, Q-, R-, S-, and T-waves. The waveform shapes of the P-, Q-, R-, S-, and T- waves may be correlated to certain cardiovascular health issues, health risks, mortality, and / or other related health issues. The precise analysis of ECG components (including the P-, Q-, R-, S-, and T-waves) may extend beyond certain cardiac measures, such as heart rate variability and / or the like, to other linked maladies or conditions, such as diabetes, general health in aging populations, and mortality from various cardiac and / or non-cardiac causes.

[0034]

[0019] In some embodiments, there is provided an ECG processor that provides advanced parametrization of ECG components. For example, there may be provided advanced parametrization of ECG components to enable quantifying biophysically interpretable features of the ECG waveforms. This parameterization may not only capture the stereotypical alterations associated with cardiac aging and disorders but also harnesses certain variability of ECG signals to unveil subtle physiological differences among even seemingly healthy individuals. An objective of the parameterization by the ECG processor is to establish connections between specific ECG features indicated by the parameterization and the application of treatment and other health outcomes, such as neural activity, cognitive functions, and / or the like — thereby offering a more nuanced understanding of cardiovascular health on other health related issues including for example brain health.

[0035]

[0020] FIG. 1 depicts an example of a patient 102, such as a human (although non-human patients, such as animals may be a patient as well). In the example of FIG. 1, the patient 102 may have one or more ECG leads 104 (e.g., 12 lead, 3 lead, 2 leads, and / or the like) attached to the torso of the patient to collect an ECG waveform. The ECG leads may collect the ECG waveform (which includes the noted P-, Q-, R-, S-, and T-waves), and the ECG waveforms are provided to at least an ECG processor 106 for the advanced ECG waveform parametrization, in accordance with some embodiments.

[0036]

[0021] In some embodiments, the ECG processor 106 may comprise for example at least one processor and at least one memory including instructions, which when executed cause one or more of the ECG processing operations disclosed herein. The ECG processor may also include (or be coupled to) one or more analog filters, one or more digital filters, one or more amplifiers, one or more digital-to-analog converts, one or more digital signal processing circuitry, one or more user interfaces, and / or other components / circuitry. The ECG waveform received by the ECG processor may be an analog waveform, which may be digitized by an analog to digital PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0037] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0038] Filing Date: August 15, 2025 Customer No.: 39564 converter. Alternatively, or additionally, the ECG waveform received by the ECG processor may be a digital waveform representative of the electrical activity of the patient 102.

[0039]

[0022] In some implementations of the ECG processor 106, it may be used to provide fast, precise quantification of ECG features to enable a detailed examination of biological rhythms and their variation within and across individuals using, for example, real-time ECG data for real-time health status monitoring. The ECG features determined from the ECG parameters may facilitate, as noted, deeper investigation into the physiological underpinnings of health, cognition, and disease, uncovering novel biomarkers for cardiovascular and neural health assessment.

[0040]

[0023] In some implementations of the ECG processor 106, it may include time-domain ECG shape parameterization that provide a detailed characterization of each waveform for a heartbeat, wherein the heartbeat’s waveform may be minimally filtered into a set of clinically interpretable ECG parameters. For example, the ECG processor 106 may identify beat-to-beat waveform shape ECG features to offer a level of detail and clinical relevance not previously available. Examples of the ECG parameters (also referred to as “features” herein) and wave shape features that may be determined by the ECG processor are included in Table. 1. These parameters may provide a more nuanced understanding of heart health and may bridge a significant gap in cardiac care by ensuring diagnostic outcomes are not only accurate but also clinically meaningful.

[0041]

[0024] Table 1 PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0042] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0043] Filing Date: August 15, 2025 Customer No.: 39564 PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0044] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0045] Filing Date: August 15, 2025 Customer No.: 39564 PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0046] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0047] Filing Date: August 15, 2025 Customer No.: 39564

[0048]

[0025] FIG. 2A depicts an example process 200 for ECG parameterization, in accordance with some embodiments. The description of process 200 may also refer to FIGs. 1, 2 A, 3 A, 3B, and 3C.

[0049]

[0026] At 202, the process 200 may include receiving at least one ECG waveform (which includes the noted P-, Q-, R-, S-, and T-waves), in accordance with some embodiments. For example, a patient, such as the patient 102, may be fitted with the ECG leads 104 to enable collection of an ECG waveform from the patient. This ECG waveform may then be received by the ECG processor 106. And, the ECG waveform may be received, as noted, as an analog signal and / or a digital signal.

[0050]

[0027] At 203, the process 200 may include filtering the ECG waveform signal through a filter, such as a high-pass filter (e.g., at 0.05 Hz which passes frequencies above 0.05 Hz) to remove certain artifacts, such as slow-moving artifact noise. Alternatively, or additionally, the filtering may include a notch filter to eliminate interference caused by electrical power (e.g., in countries with 50 Hz power, a 50 Hz notch is applied, while in countries with 60 Hz power, a 60 Hz notch filter is applied). For example, the ECG processor 106 may process a channel containing the ECG waveform by using a high-pass filter and / or a notch filter. Alternatively, or additionally, a power spectrum may be generated to determine whether additional filtering is needed of the ECG waveform. In some embodiments, the preprocessing and, in particular, ECG waveform cleaning provided by the filtering is kept to a minimum to preserve the integrity of the original, raw ECG signal.

[0051]

[0028] FIG. 3A depicts an original, raw ECG waveform 333A from a given patient, such as patient 102. As shown at FIG. 3A, the ECG waveform includes multiple heartbeats. FIG. 3A also depicts at 333B a high pass filtered version of the ECG waveform 333A.

[0052]

[0029] Referring again to FIG. 2A, the process 200 may include, at 204, the ECG processor 106 detecting R-peaks from the ECG waveform. After the filtering at 202, the ECG PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0053] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0054] Filing Date: August 15, 2025 Customer No.: 39564 processor 106 may detect the R-peaks by identifying one or more locations of maximal upward deflection within the QRS complex, which corresponds to ventricular depolarization. The R- peak serves as a robust fiducial marker for cycle segmentation because it is typically the largest and most distinct feature of the ECG waveform, even in noisy recordings. The ECG processor 106 may identify one or more, if not all, of the R-peaks in the ECG waveform. The identification of R-waves enables precise segmentation of the ECG cycles.

[0055]

[0030] At 206, process 200 may include the ECG processor 106 segmenting, based on the detected R-peaks and / or dynamic offsets, the ECG waveform into one or more epochs. For example, the ECG processor 106 may apply a dynamic offset calculation (e.g., via wavelet-based methods) to define the start and end points for each cycle relative to the R-peak position. In this way, each epoch (e.g., one cardiac cycle) has the ability to capture the complete P, Q, R, S and T wave components. The phrase “dynamic offset calculation” refers to determining a duration for each ECG segment. This duration determination may be adaptive in the sense that the cardiac cycles can differ in duration, so parameterization should account for minor duration differences from epoch (cycle) to epoch (cycle). For example, a wavelet decomposition (or other decomposition or transform) may be used to isolate the QRS-complex in an ECG epoch and then identify bounds (e.g., start and end of and ECG epoch) based on signal energy. Referring to FIG. 3B, it depicts the ECG waveform for a single epoch that has been segmented from the ECG waveform of FIG. 3A.

[0056]

[0031] At 208, the process 200 may include the ECG processor 106 detrending each waveform for an epoch to remove, for example, a linear trend from the epoch’s ECG waveform. For example, the ECG processor 106 may detrend each of the epoch’s ECG waveforms by applying, for example, a high pass filter (or using another detrend technology) to remove from the ECG waveform signals drifts (which may be caused by perspiration, lead contact, patient movement, and / or the like). Here, the detrending may ensures some if not all peaks and troughs are above and below a zero-crossing, respectively, for optimal peak / trough detection.

[0057]

[0032] At this point, the patient’s ECG waveform (which includes the P-, Q-, R-, S-, and T-waves) is preprocessed (e.g., filtered, detrended, etc.) and segmented into a plurality of epochs (each epoch representing a single heartbeat).

[0058]

[0033] At 210, the process 200 may include the ECG processor 106 identifying within each of the epochs peaks in the P-, Q-, R-, S-, and / or T-waves. To determine an R-peak, it can PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0059] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0060] Filing Date: August 15, 2025 Customer No.: 39564 be identified by the ECG processor by locating the maximum voltage sample within the cycle. The ECG processor may identify the Q peak and S peak by searching for local minima before and after the R-peak, respectively, within bounds (e.g., dynamically offset bounds). The ECG processor may determine the P peak and T peak by searching for local maxima within intervals preceding the Q-peak and following the S-peak, respectively.

[0061]

[0034] At 212, the process may include the ECG processor 106 processing of each of the epochs to obtain one or more parameters for each of the epochs.

[0062]

[0035] At 212 for example, the ECG processor 106 may process each epoch to determine, at 218 of FIG. 2B, one or more initial estimates for one or more Gaussian parameters for each of the five main ECG components, in accordance with some embodiments. In the case of the P-wave, for example, an initial estimate of the mean, height (e.g., magnitude), and width (e.g., standard deviation) may be determined. The processing may further include generating initial Gaussian parameter estimates for curve fitting. A detrended electrocardiogram cycle is processed to detect candidate waveform components. Detection is performed using polarityspecific peak searches within time windows derived from wavelet-based offsets, including: (a) identifying an R-peak by locating the maximum voltage sample within the cycle; (b) determining Q and S peak candidates by searching for local minima before and after the R-peak, respectively, within dynamically offset bounds; and (c) determining P and T peak candidates by searching for local maxima within intervals preceding the Q-peak and following the S-peak, respectively. The detected peaks are validated against positional and amplitude criteria to exclude physiologically implausible or low-prominence peaks. For each validated component, a center parameter is assigned corresponding to the detected peak index, a height parameter is assigned corresponding to the signal amplitude at the center parameter, and a width parameter is determined by locating left and right half-height points relative to the center parameter, computing the full width at half maximum, and converting the result to a Gaussian standard deviation. The width parameter is optionally constrained to a minimum value for numerical stability. The center, height, and width parameters for the validated components are assembled in component order to form the initial parameter vector.

[0063]

[0036] Next, upper and lower bounds for the Gaussian initial estimates may be determined by the ECG processor 106 at 220 of FIG. 2B, in accordance with some embodiments. For the Gaussian initial estimates (which were determined at 218), the ECG processor 106 may PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0064] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0065] Filing Date: August 15, 2025 Customer No.: 39564 determine an upper bound and a lower bound. In the case of the height of the P-wave for example, the upper and lower bounds may be determined as 0 and 10, respectively. The upper and lower bounds may be determined for each of the estimates generated at 218. For each Gaussian component, the ECG processor 106 may use for example a bound_f actor parameter (which may have a default value of 0.2, although other values may be used as well). The parameter bounds may be defined as: (1) Center: + / - (bound_f actor x standard deviation) from the initial center; (2) Height: + / - (bound_factor x initial height), with sign inversion if the peak is negative; (3) Width: + / - (bound_factor x initial width), constrained to be greater than or equal to le'2. In some embodiments, the bound factor may be selected empirically based on the variability of ECG morphology in the target population, the sampling rate of the input signal, and the desired trade-off between fitting flexibility and parameter stability. The computed upper and lower bounds may then be provided to a curve fitting process to constrain the optimization within physiologically plausible ranges, thereby improving robustness and preventing overfitting to noise.

[0066]

[0037] Once the upper and lower bounds are determined, the ECG processor 106 may curve fit, at 222 of FIG. 2B, to fit five curves, such as 5 Gaussians to the ECG waveform data, in accordance with some embodiments. This curve fitting process may model each of the P-, Q-, R-, S-, and T waves. To illustrate further, Gaussian curve fitting may be used to fit 5 Gaussian curves to the ECG waveform data for a given ECG epoch. At the end of the curve fitting, 5 Gaussian curves are generated, one Gaussian curve for the P-wave, one Gaussian curve for the Q-wave, one Gaussian curve for the R-wave, one Gaussian curve for the S-wave, and one Gaussian curve for the T-wave. The estimates determined at 218 for the mean, height, and width for the P-wave, Q-wave, R-wave, S-wave, and T wave and the upper and lower bounds determined at 220 are used during curve fitting to converge (or find) a fit for the 5 Gaussian curves for each of the P-wave, Q-wave, R-wave, S-wave, and T wave of a given ECG epoch. In some embodiments, an additional skew parameter may be included for one or more components to quantify deviation from perfect symmetry around the center parameter. This allows the fitted waveform to capture asymmetries commonly present in physiological ECG morphology. Although Gaussian curves are described, alternative fitting methods may be employed, such as non-linear least squares optimization (e.g., scipy.optimize.least_squares, Imfit), global optimization algorithms (e.g., differential evolution, genetic algorithms), probabilistic parameter PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0067] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0068] Filing Date: August 15, 2025 Customer No.: 39564 estimation methods (e.g., Markov Chain Monte Carlo, Bayesian inference), or alternative waveform basis functions (e.g., Gabor functions, Hermite functions, log-normal curves, or spline-based models). The choice of method may depend on computational constraints, robustness to noise, or the ability to model specific waveform morphologies.

[0069]

[0038] After the curve fitting to the P-, Q-, R-, S-, and T waves, the ECG processor 106 may determine, at 224, waveform shape parameters for each waveform components, in accordance with some embodiments. The fitted Gaussian corresponding to each component may be used as a model (e.g., mathematical representation) of that waveform, from which quantitative parameters are computed directly from the model.

[0070]

[0039] FIG. 3C depicts an example of some of the wave shape parameters obtained from a curve fitted model, such as a Gaussian curve. For each waveform component, parameters may include: the Gaussian mean (center location in samples or milliseconds), height (peak amplitude), standard deviation (o), peak center index, peak width in samples and milliseconds, full width at half maximum (FWHM), rise time (center to right-half point), decay time (center to left-half point), rise-decay symmetry ratio, peak sharpness during rise and decay phases, and voltage integral (area under the curve). Additional derived features may include left- and rightsided slopes, inter-deflection voltage differences, and local curvature metrics. Certain of these parameters are well-established as clinical markers of health and disease (e.g., prolonged ST segments as indicators of acute myocardial infarction, QRS widening as a marker of conduction abnormalities). Others, such as rise-decay symmetry, sharpness, and localized curvature of the P-wave or T-wave, have not been systematically quantified in prior ECG analysis methods. The disclosed parameterization enables these features to be measured with high reproducibility across beats and patients.

[0071]

[0040] In some embodiments, these parameters may be used to support clinical decisionmaking. For example, detection of asymmetry or abnormal sharpness in the T-wave may indicate early repolarization abnormalities and prompt evaluation for arrhythmic risk. Similarly, prolonged rise time in the P-wave could be used to identify atrial conduction delay, supporting early intervention in patients at risk for atrial fibrillation. Detection of reduced R-wave sharpness may aid in monitoring myocardial ischemia progression. These uses are illustrative and may be applied to other diagnostic or monitoring scenarios as appropriate. PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0072] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0073] Filing Date: August 15, 2025 Customer No.: 39564

[0074]

[0041] Referring again to FIG. 2B, alternatively, or additionally, the ECG processor 106 may determine, at 226, an indication of the quality of the fit between each of the curves (which are generated at 222) and the ECG data. For example, a coefficient of determination (e.g., r2values) may be generated for each of the five Gaussian curves and the underlying ECG data to provide an indication of the quality of the fit. In the case of r2values for example, an r2value closer to “1” indicates a better fit, when compared to a lower r2value. Other statistical or signaldomain measures of fit quality may be used in addition to, or in place of, R2, including rootmean-square error (RMSE), mean absolute error (MAE), normalized residual sum of squares, or correlation coefficients. In some embodiments, fit quality metrics such as R2or RMSE are computed for the combined multi-Gaussian reconstruction of the entire ECG cycle.

[0075]

[0042] The process 200 may include the ECG processor 106 using the five Gaussian curves for the P-, Q-, R-, S-, and T waves, additional features, in accordance with some embodiments. For example, the ECG processor 106 may determine additional wave shape features across a plurality of ECG epochs (e.g., using the corresponding Gaussian fit curves for each epoch). In some embodiments, the additional features include component intervals, which are temporal measurements between defined fiducial points on the ECG waveform. Examples include the PR interval (from the onset of the P-wave to the onset of the QRS complex), the QRS duration (from the onset of the Q-wave to the end of the S-wave), the QT interval (from the onset of the Q-wave to the end of the T-wave), and the ST segment (from the end of the S-wave to the onset of the T-wave). These intervals are well-documented in predicting specific clinical states, such as conduction abnormalities, ischemia, electrolyte disturbances, and risk for arrhythmias, as well as in assessing general cardiac function.

[0076]

[0043] In some embodiments, for ECG signals meeting a minimum criterion of at least sixty valid RR intervals (where an RR interval refers to a time between two consecutive R- waves), which may correspond to for example approximately one minute of data at normal heart rates, additional temporal features may be derived from the RR intervals. These include average heart rate and multiple heart rate variability (HRV) metrics, such as SDNN (standard deviation of normal-to-normal RR intervals), RMSSD (root mean square of successive differences between normal-to-normal RR intervals), and NN50 (count of pairs of successive normal-to-normal RR intervals differing by more than 50 milliseconds). Other HRV-derived measures, such as pNN50 (the percentage of NN50 intervals relative to the total number of intervals), may also be PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0077] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0078] Filing Date: August 15, 2025 Customer No.: 39564 computed. In further embodiments, cross -component features may be derived, including interdeflection voltage differences (e.g., R-S amplitude difference), morphological ratios (e.g., T- wave height to R-wave height), and repolarization symmetry indices. Such features may provide additional diagnostic or prognostic information, for example, detecting subtle conduction delays, early repolarization abnormalities, or autonomic dysfunction.

[0079]

[0044] Alternatively, or additionally, the ECG processor 106 may determine an autocorrelation of each feature to extract autocorrelation function (ACF) features of interest. The ACF quantifies how similar a feature’s value is to its past self at different time lags. While the epoch-level shape features inform instantaneous morphology of the heart, the ACF of each feature determines how long that morphology stays similar across beats and can reveal any potential recurring patterns over time. For example, the P-wave amplitude may periodically peak ever)' 10-12 beats, which could reflect respiratory modulation of atrial depolarization. Similarly, oscillations in T-wave morphology at characteristic low-frequency (about 0.1 Hz) and high- frequency (about 0.25 Hz) bands may correspond to autonomic nervous system activity, enabling quantification of sympathetic and parasympathetic influences on ventricular repolarization. ACF analysis of RR intervals can reveal periodic heart rate modulations consistent with sleep- disordered breathing or other cyclical patterns in autonomic tone. In some embodiments, these ACF-derived features may be used to detect physiological rhythms, identify abnormal periodicities (e.g., premature atrial or ventricular contraction patterns recurring at regular intervals), or monitor changes in rhythm modulation over time. Such information could aid in detecting conditions such as respiratory sinus arrhythmia, autonomic dysfunction, or early signs of conduction system disease, and may guide further diagnostic evaluation or therapeutic adjustment.

[0080]

[0045] At 232, the process 200 may include applying or recommending treatment based on the determined features. In some embodiments, diagnostic mappings are built from feature patterns that reflect underlying physiological states, informed by prior clinical evidence. For example, a flattened T-wave may be associated with myocardial ischemia or electrolyte imbalance, prolonged QT interval may indicate elevated risk for ventricular arrhythmia, and abnormal P-wave duration or morphology may suggest atrial enlargement or conduction delay. From the diagnostic phase, the process may triage to recommended next steps, such as ordering confirmatory diagnostic tests (e.g., serum electrolyte panel, cardiac enzyme testing, PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0081] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0082] Filing Date: August 15, 2025 Customer No.: 39564 echocardiography, etc.), initiating acute interventions (e.g., supplemental oxygen, anti-ischemic therapy, etc.), adjusting medications (e.g., modifying QT-prolonging drug regimens, initiating rate-control therapy, etc.), or providing lifestyle and monitoring recommendations (e.g., home ECG monitoring, exercise modification, dietary adjustments, etc.). In further embodiments, the system may integrate with electronic health records or remote monitoring platforms to automate alerts when parameters exceed predefined clinical thresholds, enabling earlier clinician intervention. For example, detection of a recurring abnormal T-wave morphology pattern via autocorrelation analysis could prompt evaluation for evolving ischemia, while a gradual increase in P-wave duration over weeks could trigger screening for atrial fibrillation risk.

[0083]

[0046] FIG. 2C depicts another example of a process 299 for ECG parameterization, in accordance with some embodiments. The process 299 may be similar in some response to the processes of FIG. 2A and 2B.

[0084]

[0047] At 201, the process 299 may include receiving a plurality of segmented ECG waveforms, in accordance with some embodiments. For example, the ECG processor 106 may receive a plurality of segmented ECG waveforms. Each of the plurality of segmented ECG waveforms may correspond to an epoch (e.g., heartbeat cycle) of ECG waveform data collected from a patient, such as patient 102. The received segmented ECG data may be pre-processed before 201 with some filtering, such as high pass filtering for detrending and / or a notch (to remove power related noise).

[0085]

[0048] At 218-232, the process 299 may proceed as noted above at FIGs. 2A and 2B.

[0086] For example, at 218 at FIG. 2C, it may be the same or similar as 218 at FIG. 2B; and so forth for each of 220, 222, 224, 226, and 232.

[0087]

[0049] FIG. 3B depicts a plot of a single epoch segmented by the ECG processor 106, in accordance with some embodiments. As noted at 210 (FIG. 2A), the ECG processor may detect or identify the P wave peak 305B, the Q wave peak 305C, the R wave peak 305 A, the S wave peak 305D, and / or the T wave peak 305E.

[0088]

[0050] FIG. 3D shows an example of the ECG processor 106 detection of R-peaks within a segment of an ECG time series. The vertical dashed lines mark the sample indices corresponding to detected R-peaks, and the overlaid markers indicate their precise locations in time and amplitude. This step identifies the fiducial points used for cycle segmentation and PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0089] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0090] Filing Date: August 15, 2025 Customer No.: 39564 subsequent feature extraction. Accurate R-peak identification ensures consistent temporal anchoring of all waveform components across beats.

[0091]

[0051] FIG. 3E shows alignment of ECG cycles to their corresponding R-peaks to create a set of epoched waveforms. Each cycle is temporally aligned such that the R-peak occurs at the same reference point (vertical dashed line), enabling direct comparison of morphology across beats. The bold dashed trace represents the average cycle waveform, and the shaded region represents variability across aligned cycles. This alignment facilitates computation of average waveform shapes, beat-to-beat variability, and statistical aggregation of Gaussian-derived parameters.

[0092]

[0052] By successfully fitting five Gaussian functions to the ECG data, the ECG processor 106 may be used to analyze ECG waveform shape features and, in some embodiments, to generate realistic ECG simulations by inverting the fitting process. In this approach, the fitted Gaussian parameters for each waveform component (P, Q, R, S, T) are used to reconstruct an individual’s ECG waveform with high fidelity. Certain parameters, such as amplitude, width, rise time, or decay time, may be modified to simulate specific physiological or pathological changes. This capability enables the creation of personalized “digital twin” ECG models in which subtle, controlled deviations can be introduced to explore the diagnostic impact of morphological alterations. Such simulations allow clinicians and researchers to isolate the effect of a single waveform feature on diagnostic metrics, without the confounding factors present in real- world patient recordings. This is not feasible with conventional ECG acquisition, where morphology changes occur in an uncontrolled and multifactorial manner. In some embodiments, the system may be used to model the progression of disease states, test the sensitivity of diagnostic algorithms to early morphological changes, or design patient-specific thresholds for clinical alerts. For example, simulated prolongation of the QT interval in a patient’s digital twin could be used to predict arrhythmia risk under certain drug regimens, or gradual widening of the QRS complex could be modeled to determine the earliest point at which conduction abnormalities become detectable. By enabling controlled, patient- specific waveform manipulation, this approach supports personalized diagnostics, proactive monitoring, and hypothesis testing in both clinical and research settings.

[0093]

[0053] Moreover, simulations may be used to explore how waveform alterations impact diagnostic features for each individual, including changes in segment and interval durations or PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0094] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0095] Filing Date: August 15, 2025 Customer No.: 39564 shifts in repolarization patterns. Furthermore, this may be used to offer insights into how changes in waveform shape could influence the contributions to the power spectra of the ECG. Mapping waveform shape changes to the frequency domain allows clinicians and researchers to test hypotheses about the spectral impact of specific morphological changes before they occur in a patient, and identify spectral biomarkers linked to disease states.

[0096]

[0054] In addition, different rhythmic and random processes, such as Poisson distributions, may be used to insert simulated ECG cycles into a signal. This method may enable an evaluation of the effects of different, or even random, interbeat intervals and heart rate variability on the power spectra, enhancing the understanding of these dynamics. For example, by directly testing the influence of beat-to-beat rate and variability in the frequency domain, ratedependent spectral changes may be separated from morphological / waveform shape based spectral changes. Many draw conclusions about heart rate variability from the ECG power spectrum alone, often using ratios between high- and low-frequency bands. However, these spectral changes may be driven by other waveform shape modifications. Without direct simulation and systematic tweaking of these parameters, it may not be possible to reliably separate the two effects.

[0097]

[0055] In some implementations, some of the aspects disclosed herein may be implemented in a computing system, such as a computing system 500 depicted at FIG. 4. For example, the ECG processor 106 may include aspects of the computing system 500, such as a processor 510, a memory 520, a storage device 530, and / or an input / output device 540. The processor 510, the memory 520, the storage device 530, and the input / output device 540 can be interconnected via a system bus 550.

[0098]

[0056] The processor 510 is capable of processing instructions (such as the instruction to implement opioid missus detection) for execution within the computing system 500. In some implementations of the current subject matter, the processor 510 can be a single-threaded processor. Alternately, the processor 510 can be a multi-threaded processor. The processor 510 is capable of processing instructions stored in the memory 520 and / or on the storage device 530 to display graphical information for a user interface provided via the input / output device 540.

[0099]

[0057] The memory 520 is a computer readable medium such as volatile or non-volatile that stores information within the computing system 500. The storage device 530 is capable of providing persistent storage for the computing system 500. The storage device 530 can be a PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0100] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0101] Filing Date: August 15, 2025 Customer No.: 39564 floppy disk device, a hard disk device, an optical disk device, or a tape device, or other suitable persistent storage means.

[0102]

[0058] The input / output device 540 provides input / output operations for the computing system 500. In some implementations of the current subject matter, the input / output device 540 includes a keyboard and / or pointing device. In various implementations, the input / output device 540 includes a display unit for displaying graphical user interfaces. According to some implementations of the current subject matter, the input / output device 540 can provide input / output operations for a network device. For example, the input / output device 540 can include Ethernet ports or other networking ports to communicate with one or more wired and / or wireless networks (e.g., a local area network (LAN), a wide area network (WAN), the Internet).

[0103]

[0059] In some implementations of the current subject matter, the computing system 500 can be used to execute various interactive computer software applications that can be used for organization, analysis and / or storage of data in various formats. Alternatively, the computing system 500 can be used to execute any type of software applications. Upon activation within the applications, the functionalities can be used to generate the user interface provided via the input / output device 540. The user interface can be generated and presented to a user by the computing system 500 (e.g., on a computer screen monitor, etc.).

[0104]

[0060] One or more aspects or features of the subject matter described herein can be realized in digital electronic circuitry, integrated circuitry, specially designed ASICs, field programmable gate arrays (FPGAs) computer hardware, firmware, software, graphics processing units (GPUs), artificial intelligence (Al) circuitry, neural network circuitry, and / or combinations thereof. These various aspects or features can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0105]

[0061] These computer programs, which can also be referred to as programs, software, software applications, applications, components, or code, include machine instructions for a programmable processor, and can be implemented in a high-level procedural and / or object- oriented programming language, and / or in assembly / machine language. As used herein, the term “machine-readable medium” refers to any computer program product, apparatus and / or device, PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0106] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0107] Filing Date: August 15, 2025 Customer No.: 39564 such as for example magnetic discs, optical disks, memory, and Programmable Logic Devices (PLDs), used to provide machine instructions and / or data to a programmable processor, including a computer or machine-readable medium that receives machine instructions as a machine-readable signal. The term “machine-readable signal” refers to any signal used to provide machine instructions and / or data to a programmable processor. The machine-readable medium can store such machine instructions non-transitorily, such as would a non-transient solid-state memory or a magnetic hard drive or any equivalent storage medium. The machine- readable medium can alternatively or additionally store such machine instructions in a transient manner, such as for example, as would a processor cache or other random access memory associated with one or more physical processor cores.

[0108]

[0062] To provide for interaction with a user, one or more aspects or features of the subject matter described herein can be implemented on a computer having a display device, such as for example a cathode ray tube (CRT) or a liquid crystal display (LCD) or a light emitting diode (LED) monitor for displaying information to the user and a keyboard and a pointing device, such as for example a mouse or a trackball, by which the user may provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well. For example, feedback provided to the user can be any form of sensory feedback, such as visual feedback, auditory feedback, or tactile feedback; and input from the user may be received in any form, including acoustic, speech, or tactile input. Other possible input devices include touch screens or other touch-sensitive devices such as single or multi-point resistive or capacitive track pads, voice recognition hardware and software, optical scanners, optical pointers, digital image capture devices and associated interpretation software, and the like.

[0109]

[0063] In the descriptions above and in the claims, phrases such as “at least one of’ or “one or more of’ may occur followed by a conjunctive list of elements or features. The term “and / or” may also occur in a list of two or more elements or features. Unless otherwise implicitly or explicitly contradicted by the context in which it used, such a phrase is intended to mean any of the listed elements or features individually or any of the recited elements or features in combination with any of the other recited elements or features. For example, the phrases “at least one of A and B;” “one or more of A and B;” and “A and / or B” are each intended to mean “A alone, B alone, or A and B together.” A similar interpretation is also intended for lists including three or more items. For example, the phrases “at least one of A, B, and C;” “one or PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0110] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0111] Filing Date: August 15, 2025 Customer No.: 39564 more of A, B, and C;” and “A, B, and / or C” are each intended to mean “A alone, B alone, C alone, A and B together, A and C together, B and C together, or A and B and C together.” Use of the term “based on,” above and in the claims is intended to mean, “based at least in part on,” such that an unrecited feature or element is also permissible.

[0112]

[0064] In view of the disclosure above, various examples are set forth below. It should be noted that one or more features of an example, taken in isolation or combination, should be considered within the disclosure of this application.

[0113]

[0065] Example 1. A system for processing electrocardiogram waveforms comprising: at least one processor; and at least one memory including instructions, which when executed by the at least one processor causes operations comprising: receiving a segmented electrocardiogram waveform, wherein the segmented electrocardiogram waveform is for a heartbeat epoch; determining, for the segmented electrocardiogram waveform, initial estimates for mean, height, and standard deviation for a Gaussian curve; determining, for the initial estimates, upper bounds and lower bounds; fitting, using the initial estimates, the upper bounds, and the lower bounds, the Gaussian curve to the segmented electrocardiogram waveform; and determining, using the fitted Gaussian curve, one or more parameters and / or waveform shape parameters to enable application of a treatment.

[0114]

[0066] Example 2. The system of Example 1, wherein the fitting comprises fitting 5 Gaussian curves to the P-wave, Q-wave, R-wave, S-wave, and T wave of the segmented electrocardiogram waveform.

[0115]

[0067] Example 3. The system of any of Examples 1-2, wherein the receiving comprises receiving a plurality of segmented electrocardiogram waveforms, wherein each of the plurality of segmented electrocardiogram waveforms is for a corresponding heartbeat epoch.

[0116]

[0068] Example 4. The system any of Examples 1-3, wherein the operations further comprise receiving an electrocardiogram waveform and filtering the electrocardiogram waveform using at least one of a high-pass filter and / or a notch filter. PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0117] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0118] Filing Date: August 15, 2025 Customer No.: 39564

[0119]

[0069] Example 5. The system of any of Examples 1-4, wherein the operations further comprise detecting one or more R-peaks from the electrocardiogram waveform based on one or more upward deflections in the electrocardiogram waveform.

[0120]

[0070] Example 6. The system of any of Examples 1-5, wherein the operations further comprise segmenting the electrocardiogram waveform to form one or more segmented electrocardiogram waveforms including the received segmented electrocardiogram waveform, wherein the segmenting is based on the detected one or more R-peaks in the electrocardiogram waveform and / or at least one dynamic offset.

[0121]

[0071] Example 7. The system of any of Examples 1-6, wherein the operations further comprise detrending the one or more segmented electrocardiogram waveforms to filter linear trends and / or drift.

[0122]

[0072] Example 8. The system of any of Examples 1-7, wherein the operations further comprise identifying, in the one or more segmented ECG waveforms, one or more peaks in a P- wave, a Q-wave, an R-wave, an S-wave, and / or a T-wave.

[0123]

[0073] Example 9. The system of any of Examples 1-8, wherein the determining, using the fitted Gaussian curve, further comprises determining one or more of the following: a cycle index, a cycle trend, an R-squared, a root-mean squared error (RMSE), a global center, a global left index, a global right index, cycle-relative indices and times, gaussian fit features for the P, Q, R, S, and / or T-waves, voltage and / or morphology metrics for the P, Q, R, S, and / or T-waves, one or more intervals, and / or one or more inter-wave voltage differences.

[0124]

[0074] Example 10. The system of any of Examples 1-9, wherein the determining, using the fitted Gaussian curve, further comprises determining at least one rise time, at least one decay time, at least one rise decay symmetry, and / or at least one sharpness.

[0125]

[0075] Example 11. The system of any of Examples 1-10, further comprising determining a treatment based on the one or more parameters and / or the waveform shape parameters.

[0126]

[0076] Example 12. A method for processing electrocardiogram waveforms comprising: receiving a segmented electrocardiogram waveform, wherein the segmented electrocardiogram waveform is for a heartbeat epoch; determining, for the segmented electrocardiogram waveform, initial estimates for mean, height, and standard deviation for a Gaussian curve; PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0127] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0128] Filing Date: August 15, 2025 Customer No.: 39564 determining, for the initial estimates, upper bounds and lower bounds; fitting, using the initial estimates, the upper bounds, and the lower bounds, the Gaussian curve to the segmented electrocardiogram waveform; and determining, using the fitted Gaussian curve, one or more parameters and / or waveform shape parameters to enable application of a treatment.

[0129]

[0077] Example 13. The method of Example 12, wherein the fitting comprises fitting 5 Gaussian curves to the P-wave, Q-wave, R-wave, S-wave, and T wave of the segmented electrocardiogram waveform.

[0130]

[0078] Example 14. The method of any of Examples 12-13, wherein the receiving comprises receiving a plurality of segmented electrocardiogram waveforms, wherein each of the plurality of segmented electrocardiogram waveforms is for a corresponding heartbeat epoch.

[0131]

[0079] Example 15. The method of any of Examples 12-14, further comprising receiving an electrocardiogram waveform and filtering the electrocardiogram waveform using at least one of a high-pass filter and / or a notch filter.

[0132]

[0080] Example 16. The method of any of Examples 12-15, further comprising detecting one or more R-peaks from the electrocardiogram waveform based on one or more upward deflections in the electrocardiogram waveform.

[0133]

[0081] Example 17. The method of any of Examples 12-16, further comprising segmenting the electrocardiogram waveform to form one or more segmented electrocardiogram waveforms including the received segmented electrocardiogram waveform, wherein the segmenting is based on the detected one or more R-peaks in the electrocardiogram waveform and / or at least one dynamic offset.

[0134]

[0082] Example 18. The method of any of Examples 12-17, further comprising detrending the one or more segmented electrocardiogram waveforms to filter linear trends and / or drift.

[0135]

[0083] Example 19. The method of any of Examples 12-18, further comprising identifying, in the one or more segmented ECG waveforms, one or more peaks in a P-wave, a Q- wave, an R-wave, an S-wave, and / or a T-wave.

[0136]

[0084] Example 20. The method of any of Examples 12-19, wherein the determining, using the fitted Gaussian curve, further comprises determining one or more of the following: a cycle index, a cycle trend, an R-squared, a root-mean squared error (RMSE), a global center, a PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0137] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0138] Filing Date: August 15, 2025 Customer No.: 39564 global left index, a global right index, cycle-relative indices and times, gaussian fit features for the P, Q, R, S, and / or T-waves, voltage and / or morphology metrics for the P, Q, R, S, and / or T- waves, one or more intervals, and / or one or more inter-wave voltage differences.

[0139]

[0085] Example 21. The method of any of Examples 12-20, wherein the determining, using the fitted Gaussian curve, further comprises determining at least one rise time, at least one decay time, at least one rise decay symmetry, and / or at least one sharpness.

[0140]

[0086] Example 22. A non-transitory computer-readable storage medium including program code, which when executed by at least one processor, causes operations comprising: receiving a segmented electrocardiogram waveform, wherein the segmented electrocardiogram waveform is for a heartbeat epoch; determining, for the segmented electrocardiogram waveform, initial estimates for mean, height, and standard deviation for a Gaussian curve; determining, for the initial estimates, upper bounds and lower bounds; fitting, using the initial estimates, the upper bounds, and the lower bounds, the Gaussian curve to the segmented electrocardiogram waveform; and determining, using the fitted Gaussian curve, one or more parameters and / or waveform shape parameters to enable application of a treatment.

[0141]

[0087] The subject matter described herein can be embodied in systems, apparatus, methods, and / or articles depending on the desired configuration. The implementations set forth in the foregoing description do not represent all implementations consistent with the subject matter described herein. Instead, they are merely some examples consistent with aspects related to the described subject matter. Although a few variations have been described in detail above, other modifications or additions are possible. In particular, further features and / or variations can be provided in addition to those set forth herein. For example, the implementations described above can be directed to various combinations and subcombinations of the disclosed features and / or combinations and subcombinations of several further features disclosed above. In addition, the logic flows depicted in the accompanying figures and / or described herein do not necessarily require the particular order shown, or sequential order, to achieve desirable results. For example, the logic flows may include different and / or additional operations than shown without departing from the scope of the present disclosure. One or more operations of the logic PCT / US25 / 42168 15 August 2025 (15.08.2025)

[0142] Via Patent Center Docket No.: 24636-766WO 1 / 2024-342-2

[0143] Filing Date: August 15, 2025 Customer No.: 39564 flows may be repeated and / or omitted without departing from the scope of the present disclosure.

[0144] Other implementations may be within the scope of the following claims.

Claims

Via Patent Center Docket No.: 24636-766WO1 / 2024-342-2Filing Date: August 15, 2025 Customer No.: 39564CLAIMS1. A system for processing electrocardiogram waveforms comprising: at least one processor; and at least one memory including instructions, which when executed by the at least one processor causes operations comprising: receiving a segmented electrocardiogram waveform, wherein the segmented electrocardiogram waveform is for a heartbeat epoch; determining, for the segmented electrocardiogram waveform, initial estimates for mean, height, and standard deviation for a Gaussian curve; determining, for the initial estimates, upper bounds and lower bounds; fitting, using the initial estimates, the upper bounds, and the lower bounds, the Gaussian curve to the segmented electrocardiogram waveform; and determining, using the fitted Gaussian curve, one or more parameters and / or waveform shape parameters to enable application of a treatment.

2. The system of claim 1, wherein the fitting comprises fitting 5 Gaussian curves to the P-wave, Q-wave, R-wave, S-wave, and T wave of the segmented electrocardiogram waveform.

3. The system of claim 1, wherein the receiving comprises receiving a plurality of segmented electrocardiogram waveforms, wherein each of the plurality of segmented electrocardiogram waveforms is for a corresponding heartbeat epoch.

4. The system of claim 1, wherein the operations further comprise receiving an electrocardiogram waveform and filtering the electrocardiogram waveform using at least one of a high-pass filter and / or a notch filter.

5. The system of claim 4, wherein the operations further comprise detecting one or more R-peaks from the electrocardiogram waveform based on one or more upward deflections in the electrocardiogram waveform.Via Patent Center Docket No.: 24636-766WO1 / 2024-342-2Filing Date: August 15, 2025 Customer No.: 395646. The system of claim 5, wherein the operations further comprise segmenting the electrocardiogram waveform to form one or more segmented electrocardiogram waveforms including the received segmented electrocardiogram waveform, wherein the segmenting is based on the detected one or more R-peaks in the electrocardiogram waveform and / or at least one dynamic offset.

7. The system of claim 6, wherein the operations further comprise detrending the one or more segmented electrocardiogram waveforms to filter linear trends and / or drift.

8. The system of claim 6, wherein the operations further comprise identifying, in the one or more segmented ECG waveforms, one or more peaks in a P-wave, a Q-wave, an R-wave, an S-wave, and / or a T-wave.

9. The system of claim 1, wherein the determining, using the fitted Gaussian curve, further comprises determining one or more of the following: a cycle index, a cycle trend, an R- squared, a root- mean squared error (RMSE), a global center, a global left index, a global right index, cycle-relative indices and times, gaussian fit features for the P, Q, R, S, and / or T-waves, voltage and / or morphology metrics for the P, Q, R, S, and / or T-wavcs, one or more intervals, and / or one or more inter- wave voltage differences.

10. The system of claim 1, wherein the determining, using the fitted Gaussian curve, further comprises determining at least one rise time, at least one decay time, at least one rise decay symmetry, and / or at least one sharpness.

11. The system of claim 1, further comprising determining a treatment based on the one or more parameters and / or the waveform shape parameters.

12. A method for processing electrocardiogram waveforms comprising:Via Patent Center Docket No.: 24636-766WO1 / 2024-342-2Filing Date: August 15, 2025 Customer No.: 39564 receiving a segmented electrocardiogram waveform, wherein the segmented electrocardiogram waveform is for a heartbeat epoch; determining, for the segmented electrocardiogram waveform, initial estimates for mean, height, and standard deviation for a Gaussian curve; determining, for the initial estimates, upper bounds and lower bounds; fitting, using the initial estimates, the upper bounds, and the lower bounds, the Gaussian curve to the segmented electrocardiogram waveform; and determining, using the fitted Gaussian curve, one or more parameters and / or waveform shape parameters to enable application of a treatment.

13. The method of claim 12, wherein the fitting comprises fitting 5 Gaussian curves to the P-wave, Q-wave, R-wave, S-wave, and T wave of the segmented electrocardiogram waveform.

14. The method of claim 12, wherein the receiving comprises receiving a plurality of segmented electrocardiogram waveforms, wherein each of the plurality of segmented electrocardiogram waveforms is for a corresponding heartbeat epoch.

15. The method of claim 12, further comprising receiving an electrocardiogram waveform and filtering the electrocardiogram waveform using at least one of a high-pass filter and / or a notch filter.

16. The method of claim 15, further comprising detecting one or more R-peaks from the electrocardiogram waveform based on one or more upward deflections in the electrocardiogram waveform.

17. The method of claim 16, further comprising segmenting the electrocardiogram waveform to form one or more segmented electrocardiogram waveforms including the received segmented electrocardiogram waveform, wherein the segmenting is based on the detected one or more R-peaks in the electrocardiogram waveform and / or at least one dynamic offset.Via Patent Center Docket No.: 24636-766WO1 / 2024-342-2Filing Date: August 15, 2025 Customer No.: 3956418. The method of claim 17, further comprising detrending the one or more segmented electrocardiogram waveforms to filter linear trends and / or drift.

19. The method of claim 17, further comprising identifying, in the one or more segmented ECG waveforms, one or more peaks in a P-wave, a Q-wave, an R-wave, an S-wave, and / or a T-wave.

20. The method of claim 12, wherein the determining, using the fitted Gaussian curve, further comprises determining one or more of the following: a cycle index, a cycle trend, an R-squared, a root-mean squared error (RMSE), a global center, a global left index, a global right index, cycle-relative indices and times, gaussian fit features for the P, Q, R, S, and / or T- waves, voltage and / or morphology metrics for the P, Q, R, S, and / or T-waves, one or more intervals, and / or one or more inter-wave voltage differences.

21. The method of claim 12, wherein the determining, using the fitted Gaussian curve, further comprises determining at least one rise time, at least one decay time, at least one rise decay symmetry, and / or at least one sharpness.

22. A non-transitory computer-readable storage medium including program code, which when executed by at least one processor, causes operations comprising: receiving a segmented electrocardiogram waveform, wherein the segmented electrocardiogram waveform is for a heartbeat epoch; determining, for the segmented electrocardiogram waveform, initial estimates for mean, height, and standard deviation for a Gaussian curve; determining, for the initial estimates, upper bounds and lower bounds; fitting, using the initial estimates, the upper bounds, and the lower bounds, the Gaussian curve to the segmented electrocardiogram waveform; and determining, using the fitted Gaussian curve, one or more parameters and / or waveform shape parameters to enable application of a treatment.

Images