US 12,350,019 B2
Methods and systems for determining abnormal cardiac activity
Shamim Nemati, Atlanta, GA (US); Gari Clifford, Atlanta, GA (US); Supreeth Prajwal Shashikumar, Atlanta, GA (US); Amit Jasvant Shah, Atlanta, GA (US); and Qiao Li, Atlanta, GA (US)
Assigned to Emory University, Atlanta, GA (US)
Filed by Emory University, Atlanta, GA (US); and Georgia Tech Research Corporation, Atlanta, GA (US)
Filed on Jan. 7, 2022, as Appl. No. 17/571,298.
Application 17/571,298 is a continuation of application No. 16/472,818, abandoned, previously published as PCT/US2017/068029, filed on Dec. 21, 2017.
Claims priority of provisional application 62/437,457, filed on Dec. 21, 2016.
Prior Publication US 2022/0125322 A1, Apr. 28, 2022
Int. Cl. A61B 5/0205 (2006.01); A61B 5/00 (2006.01); G16H 10/60 (2018.01); G16H 50/20 (2018.01)
CPC A61B 5/0205 (2013.01) [A61B 5/7221 (2013.01); A61B 5/7253 (2013.01); A61B 5/7267 (2013.01); G16H 10/60 (2018.01); G16H 50/20 (2018.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method for using machine learning to determine abnormal cardiac activity of a subject, the method comprising:
receiving one or more periods of time of cardiac data for a subject from one or more cardiac sensors of a wearable device and motion data for the subject from one or more motion sensors of the wearable device, each period of time including more than one window of the cardiac data and the motion data;
determining one or more signal quality indices for the cardiac data and for the motion data for each window of the cardiac data and the motion data of the one or more periods of time;
extracting one or more cardiovascular features for each period of time using at least the cardiac data, the motion data, and the one or more signal quality indices for the cardiac data and the motion data;
applying a tensor transform to the cardiac data and/or the motion data to generate a tensor for each window of the one or more periods of time;
applying a trained deep learning architecture to each tensor of the one or more periods of time to classify each window and/or each period into one or more classes using at least the one or more signal quality indices for the cardiac data and the motion data and cardiovascular features;
the deep learning architecture including a convolutional neural network, a bidirectional recurrent neural network, and an attention network;
the one or more classes including abnormal cardiac activity and normal cardiac activity; and
generating a report including a classification of cardiac activity of the subject for the one or more periods based on the one or more classes.
 
8. A non-transitory computer-readable storage medium storing instructions for using machine learning to determine abnormal cardiac activity of a subject, the instructions comprising
receiving one or more periods of time of cardiac data for a subject from one or more cardiac sensors of a wearable device and motion data for the subject from one or more motion sensors of the wearable device, each period of time including more than one window of the cardiac data and the motion data;
determining one or more signal quality indices for the cardiac data and for the motion data for each window of the cardiac data and the motion data of the one or more periods of time;
extracting one or more cardiovascular features for each period of time using at least the cardiac data, the motion data, and the one or more signal quality indices for the cardiac data and the motion data;
applying a tensor transform to the cardiac data and/or the motion data to generate a tensor for each window of the one or more periods of time;
applying a trained deep learning architecture to each tensor of the one or more periods of time to classify each window and/or each period into one or more classes using at least the one or more signal quality indices for the cardiac data and the motion data and cardiovascular features;
the deep learning architecture including a convolutional neural network, a bidirectional recurrent neural network, and an attention network;
the one or more classes including abnormal cardiac activity and normal cardiac activity; and
generating a report including a classification of cardiac activity of the subject for the one or more periods based on the one or more classes.
 
14. A system for using machine learning to determine abnormal cardiac activity of a subject, comprising:
a wearable device, the wearable device including:
one or more cardiac sensors configured to collect cardiac data; and
one or more motion sensors configured to collect motion data;
a cardiac activity processing device, the cardiac activity processing device including:
a memory; and
one or more processors, wherein the one or more processors is configured to cause:
receiving one or more periods of time of the cardiac data for a subject from one or more cardiac sensors of a wearable device and the motion data for the subject of from one or more motion sensors of the wearable device, each period of time including more than one window of the cardiac data and the motion data;
determining one or more signal quality indices for the cardiac data and for the motion data for each window of the cardiac data and the motion data of the one or more periods of time;
extracting one or more cardiovascular features for each period of time using at least the cardiac data, the motion data, and the one or more signal quality indices for the cardiac data and the motion data;
applying a tensor transform to the cardiac data and/or the motion data to generate a tensor for each window of the one or more periods of time;
applying a trained deep learning architecture to each tensor of the one or more periods of time to classify each window and/or each period into one or more classes using at least the one or more signal quality indices for the cardiac data and the motion data and cardiovascular features;
the deep learning architecture including a convolutional neural network, a bidirectional recurrent neural network, and an attention network;
the one or more classes including abnormal cardiac activity and normal cardiac activity; and
generating a report including a classification of cardiac activity of the subject for the one or more periods based on the one or more classes.