A platform is provided for generating 3D models of a tumor segmented from a series of 2D medical images and for identifying from these 3D models, radiomic features that may be used for diagnostic, prognostic, and treatment response assessment of the tumor. The radiomic features may be shape-based features, intensity-based features, textural features, and filter-based features. The radiomic features are compared to remove sufficiently redundant features, thereby producing a reduced set of radiomic features, which is then compared to separate genomic data and/or outcome data to identify clinically and biologically significant radiomic features for diagnostic, prognostic, and treatment response assessment, other applications.

There is described a method of associating a pulsed signal to a corresponding reference pulse shape. The method generally has accessing reference data having a plurality of reference pulse shapes, each reference pulse shape having a sparse array of average coefficients; receiving a pulsed signal having an array of amplitude values, including generating a sparse array of instantaneous coefficients based on said pulsed signal for example using a discrete wavelet transform; calculating a plurality of first distances between said instantaneous coefficients of said sparse array and the average coefficients of each one of said reference pulse shapes, said first distances having a first minimal distance identifying a closer one of the reference pulse shapes; and upon determining that said first minimal distance is below a first distance threshold, associating said sparse array of instantaneous coefficients to the closer one of the reference pulse shapes.

Catheterization of the heart is carried out by inserting a probe having electrodes into a heart of a living subject, recording a bipolar electrogram and a unipolar electrogram from one of the electrodes at a location in the heart, and defining a window of interest wherein a rate of change in a potential of the bipolar electrogram exceeds a predetermined value. An annotation is established in the unipolar electrogram, wherein the annotation denotes a maximum rate of change in a potential of the unipolar electrogram within the window of interest. A quality value is assigned to the annotation, and a 3-dimensional map is generated of a portion of the heart that includes the annotation and the quality value thereof.

Cardiac activity (e.g., a cardiac electrogram) is analyzed for local abnormal ventricular activity (LAVA), such as by using a LAVA detection and analysis module incorporated into an electroanatomical mapping system. The module transforms the electrogram signal into the wavelet domain to compute as scalogram; computes a one-dimensional LAVA function of the scalogram; detects one or more peaks in the LAVA function; and computes a peak-to-peak amplitude of the electrogram signal. If the peak-to-peak amplitude does not exceed a preset amplitude threshold, then the module can compute one or more of a LAVA lateness parameter for the electrogram signal using one of the one or more peaks detected in the LAVA function and a LAVA probability parameter for the electrogram signal.

Systems and methods are provided for evaluating physiological signal quality. A physiological signal, based on a series measurements on a subject, may be received. A quality of the physiological signal received may be evaluated, and an analysis of the physiological signal may be based at least in part on the quality evaluation.

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating embeddings of EEG measurements. One of the methods includes obtaining a two-dimensional time-frequency electroencephalogram (EEG) representation corresponding to one or more EEG signal measurements of a user; processing the time-frequency EEG representation using a first neural network having a plurality of first network parameters to generate an embedding of the time-frequency EEG representation, wherein the first neural network has been trained using transfer learning; and providing the embedding of the time-frequency EEG representation to a downstream neural network to generate a mental health prediction for the user.

The present application relates to a device, system, and method for determining patient body temperature based on a multi-modal wearable biosensor applied to the surface of the body that measures multiple physiological, physical, and skin-surface/microclimatic thermal parameters, derives additional instantaneous parameters, learns contextual parameters based on temporal dynamics, and employs ensemble model fusion method to estimate core body temperature.

The present disclosure relates generally to systems, methods, and devices for closed loop deep brain stimulation. In particular, a neural signal is measured and provided to software. The software includes a feature generator and a brain network model that takes the neural signal and estimates other neural signals that are not directly measured, and operates as a model of the brain. The software determines a stimulation signal to be sent to stimulating electrodes. Estimated signals by the brain network model are continuously compared to actual signals from the brain. The closed loop feedback system advantageously allows for electrical stimulation levels and patterns to be continuously updated while delivered to a patient.

A Cheyne-Stokes (CS) diagnosis system classifies periods of CS-like breathing by examining a signal indicative of a respiratory parameter. For example, nasal flow data is processed to classify it as unambiguously CS breathing or nearly so and to display the classification Processing may detect and display: apnoeas, hypopnoeas, flow-limitation and snore. The signal may be split into equal length epochs and event features are extracted. Statistics are applied to these primary feature(s) to produce secondary feature(s) representing the entire epoch. Each secondary feature is grouped with other feature(s) extracted from the entire epoch rather than from the epoch events. This final group of features is the epoch pattern. The epoch pattern is classified to produce a probability for possible event classes (e.g., Cheyne-Stokes breathing, OSA, etc.). The epoch is assigned to the class with the highest probability, which may both be reported as an indication of disease state.

A method of identifying respiratory anomalies includes obtaining respiratory data over a first time period and a second time period that is different than the first time period, identifying at least one type of sound associated with respiration in the respiratory data over the first time period, identifying the at least one type of sound associated with respiration in the respiratory data over the second time period, and identifying abnormal respiration based on a comparison of the at least one type of sound associated with respiration in the respiratory data over the first time period to the at least one type of sound associated with respiration in the respiratory data over the second time period. The at least one type of sound associated with respiration in the respiratory data over the first time period is identified using a first set of features generated by a first processing method.