Abstract: Pulmonary function estimation can include detecting one or more cough events from a time series of audio signals generated by an electronic device of a user. Based on the one or more cough events, one or more lung function metrics of the user can be determined.

Abstract: A method includes receiving stress-related measurements collected by one or more stress sensors, where the stress-related measurements represent one or more physiological responses of a user to a stressor. The method also includes receiving context data collected by one or more context sensors, where the context data represents a context associated with the user. The method further includes determining stress profile features associated with the user based on the stress-related measurements. The method also includes providing the stress profile features to a trained stress profile identification machine learning model to select a stress profile from among multiple candidate stress profiles for association with the user. The method further includes providing the selected stress profile and the context data to a trained stress intervention recommendation machine learning model to select a stress intervention activity for the user.

Abstract: In particular embodiments, a method includes accessing, by a computing device, sensor data from a sensor of a client device corresponding to a user. The method further includes determining one or more segments of sensor data and selecting at least one segment of sensor data. The method further concludes determining, based at least on the selected segments of sensor data, one or more features corresponding to the user's pulmonary condition. The method further includes determining, based at least on the determined features, an assessment of the user's pulmonary condition.

Abstract: A method includes estimating, by a consumer electronic device, an atrial fibrillation (AF) burden of a subject based on multiple measurements passively collected by at least one sensor associated with the consumer electronic device while the subject is in a free-living environment. The method also includes outputting, by the consumer electronic device, an AF burden notification associated with the subject based on the estimated AF burden. The method further includes outputting, by the consumer electronic device, an AF burden recommendation based on the estimated AF burden.

Abstract: A method includes receiving, by an electronic device, sensor data during a spirometry test of a user, the sensor data comprising audio data of the user and distance data of a distance from a face of the user to the electronic device. The method also includes obtaining, by the electronic device, an amount of air volume exchange and at least one pulmonary health parameter that are determined based on the audio data and the distance data. The method also includes presenting an indicator on a display of the electronic device for use by the user or a medical provider, the indicator representing the amount of air volume exchange.

Abstract: In one embodiment, a method includes accessing a sequence of images of a portion of a person's skin illuminated by ambient light and determining, from at least one of the images, a plurality of microregions in the portion of the person's skin. The method further includes determining, based on a similarity between particular microregions of the plurality of microregions, one or more regions of interest of the person's skin; determining, for each of the regions of interest, a remote photoplethysmogram (rPPG) signal based on the sequence of images; and determining, based on one or more of the rPPG signals, an estimate of an oxygen saturation in the person's blood.

Abstract: In one embodiment, a method includes accessing a plurality of images of a region of interest of a person's skin and extracting, from the plurality of images, a color signal of the region of interest as a function of time. The method further includes determining, based at least on the color signal, a first quality associated with an estimated vital sign of the person, where the vital sign estimate is determined from the plurality of images.

Abstract: In one embodiment, a method includes accessing, from a first sensor of a wearable device, motion data representing motion of a user and determining, from the motion data, an activity level of the user. The method includes selecting, based on the activity level, an activity-based technique for estimating the breathing rate of the user, which is used to determine a breathing rate of the user. The method further includes determining a quality associated with the breathing rate and comparing the determined quality with a threshold. If the determined quality is not less than the threshold, then the method includes using the determined breathing rate as a final breathing-rate determination for the user. If the determined quality is less than the threshold, then the method includes activating a second sensor of the wearable device and determining, based on data from the second sensor, a breathing rate for the user.

Abstract: A method includes obtaining at least one breathing audio sample of a user captured using earbuds worn by the user. The method also includes converting the at least one breathing audio sample to a breathing spectrogram configured as an image. The method further includes processing the breathing spectrogram using a trained multi-task convolutional neural network (CNN) to identify a breathing rate and a breathing depth of the user. In addition, the method includes outputting the breathing rate and the breathing depth of the user.

Abstract: Detecting an adverse health event and responding to an emergency event can include updating a current context profile of a user based on signals generated by one or more sensors of a device, the signals generated in response to one or more physically measurable phenomena associated with the user. A context-aware baseline profile having a same context as the current context profile can be selected, the context-aware baseline profile selected from a plurality of context-aware baseline profiles having different contexts. Features of the current context profile can be compared with corresponding features of the context-aware baseline profile of the user having the same context as the current context profile. A remedial action can be initiated in response to recognizing, based on the comparing, an occurrence of a possible adverse health event affecting the user.

Abstract: A method includes obtaining, by an electronic device, an audio segment comprising one or more audio events of a target subject. The method also includes extracting, by the electronic device, audio embeddings from the one or more audio events using an embedding model, the embedding model comprising a trained machine learning model. The method further includes comparing, by the electronic device, the extracted audio embeddings with a match profile of the target subject, the match profile generated during an enrollment stage. The method also includes generating, by the electronic device, a label for the audio segment based on whether or not the extracted audio embeddings match the match profile, wherein the label enables correlation of the audio segment with the target subject for monitoring a health condition of the target subject.

Abstract: Adaptive respiratory condition assessment can include capturing signals generated by sensors of a portable device positioned to sense a user's body. The sensors can generate the signals in response to sensor-detected movements of the user's body measured along multiple axes and corresponding to lung activity in the user's body during respiratory cycles. A preferred axis of measurement can be selected using signal processing performed by a signal processor embedded in the portable device. Waveforms generated from signals corresponding to signals generated in response to movements of the user's body measured along the preferred axis can be filtered for extracting biomarkers from the waveforms using the signal processor. The one or more biomarkers extracted correspond to the lung activity. Based on the biomarkers, a respiratory condition of the user can be determined. A notification based on the respiratory condition as determined can be generated and conveyed with the portable device.

Abstract: A method includes receiving multimodal data collected using at least one wearable device during an assessment window. The method also includes extracting biomarker features from the multimodal data, based on changes in the extracted biomarker features. The method also includes detecting that a stress event occurred during the assessment window. The method also includes accessing a plurality of templates of patterns in biomarker features, wherein a first subset of the templates is associated with unhealthy response to stress and a second subset of the templates is associated with healthy response to stress. The method also includes determining whether the stress event corresponds to a healthy response or an unhealthy response based on similarities between a pattern in the extracted biomarker features and the plurality of templates. The method also includes responsive to the stress event corresponding to an unhealthy response, providing a stress management recommendation.

Abstract: Detecting and identifying a predetermined health event can include detecting a potential occurrence of the predetermined health event for a user by processing in real-time motion signals corresponding to motion of the user. A likelihood that the potential occurrence is an actual occurrence of the predetermined health event can be determined based on template matching of the motion signals. In response to determining that the likelihood exceeds a predetermined threshold, audio signals coinciding in time with the motion of the user can be processed using one or more layers of a multilayered audio event classifier.

Abstract: Personalized vertigo rehabilitation and treatment can include generating a sequence of sounds that are modulated to perceptually emanate from one or more distinct locations in three-dimensional space relative to a user. The sequence can be predetermined to guide head movements of the user in a prescribed manner to mitigate vertigo experienced by the user. The user's head movements can be tracked as the user responds to the sequence of sounds. A signal can be conveyed to the user in response to detecting a predetermined head movement by the user.

Abstract: A multimodal, contactless vital sign monitoring system is configured to perform the following operations. Images are received from a video capture device. An image of a subject is identified within the images. The image of the subject is segmented into a plurality of segments. A first analysis is performed on the plurality of segments to identify a color feature. A second analysis is performed of the plurality of segments to identify a motion feature. Using a combination of the color feature and the motion feature a plurality of vital signs for the subject are determined. The first analyzing and the second analyzing are performed in parallel. The plurality of vital signs include one or more of heart rate, respiration rate, oxygen saturation, heart rate variability, and atrial fibrillation.

Abstract: Continuously monitoring heart rhythm can include grouping, using computer hardware, a plurality of inter-beat intervals (IBI) data for a user into a plurality of epochs, wherein each epoch includes a subset of the IBI data corresponding to a predetermined time span. For each epoch, a selected feature set selected from a plurality of feature sets is extracted based on a determination of temporal consistency of the epoch. A plurality of epoch classifications may be generated for the epochs using a selected feature processor, wherein each epoch classification indicates whether arrhythmia is detected for the epoch from which the epoch classification is generated. The selected feature processor is selected from a plurality of different feature processors on a per-epoch basis based on the selected feature set extracted from the epoch. An indication of arrhythmia may be output, via an output device of the computer hardware, based on the epoch classifications.

Abstract: Energy-efficient monitoring and detection of atrial fibrillation using an electronic device can include scheduling, by the electronic device, monitoring periods during which the electronic device intermittently monitors a user of the electronic device for atrial fibrillation. The scheduling can be based on determining an AF risk specific to the user. Time intervals between successive AF monitoring periods can be modulated by the electronic device in response to detecting a change in the AF risk.

Abstract: Device-invariant, frequency-domain signal processing with machine learning includes retrieving with a host device a device-specific alien dataset corresponding to an alien device. The device-specific alien dataset is retrieved from a remote data storage device communicatively coupled with the host device. A plurality of frequency-domain features are extracted from the device-specific alien dataset and a machine learning model is trained using the plurality of frequency-domain features. The host device extracts frequency-domain features from signals generated by sensors operatively coupled with the host device. Real-time frequency bin adaptation of the frequency-domain features extracted by the host device is performed. Based on the frequency-domain features extracted by the host device, as adapted, an inference is performed using the machine learning model.

Abstract: Limb motion tracking using a single sensor can include capturing acceleration data. The acceleration data can be captured in real time by an IMU sensor of a wearable device worn on a limb of a user. Orientation data can be captured in real time by the IMU sensor concurrently with the acceleration data. Estimated positions of joints of the limb can be determined based on the acceleration data and the orientation data, the joints positions estimated using a machine learning model and relative to a coordinate system. Motion of the limb can be tracked based on the estimated positions determined at different times as the user moves the limb.