A Multi-Task Group Bi-LSTM Networks Application on Electrocardiogram Classification

Abstract

Background: Cardiovascular diseases (CVD) are the leading cause of death globally. Electrocardiogram (ECG) analysis can provide thoroughly assessment for different CVDs efficiently. We propose a multi-task group bidirectional long short-term memory (MTGBi-LSTM) framework to intelligent recognize multiple CVDs based on multi-lead ECG signals.

Methods: This model employs a Group Bi-LSTM (GBi-LSTM) and Residual Group Convolutional Neural Network (Res-GCNN) to learn the dual feature representation of ECG space and time series. GBi-LSTM is divided into Global Bi-LSTM and Intra-Group Bi-LSTM, which can learn the features of each ECG lead and the relationship between leads. Then, through attention mechanism, the different lead information of ECG is integrated to make the model to possess the powerful feature discriminability. Through multi-task learning, the model can fully mine the association information between diseases and obtain more accurate diagnostic results. In addition, we propose a dynamic weighted loss function to better quantify the loss to overcome the imbalance between classes.

Results: Based on more than 170,000 clinical 12-lead ECG analysis, the MTGBi-LSTM method achieved accuracy, precision, recall and F1 of 88.86%, 90.67%, 94.19% and 92.39%, respectively. The experimental results show that the proposed MTGBi-LSTM method can reliably realize ECG analysis and provide an effective tool for computer-aided diagnosis of CVD.

Keywords: ECG; attention mechanism; bidirectional long short-term memory network; multi-task learning.

FIGURE 1.

Multi-Task Aided Diagnosis Model Based on Group Bi-LSTM. The Global Bi-LSTM (colored in blue) transmits the hidden state of each time step to Intra-group Bi-LSTM (colored in yellow) through the global access structure (colored in red). and obtained after the output feature of Bi-LSTM passes attention structure, the final fully connected layer (FC_1, FC_2) respectively generate the category distribution of two tasks.

FIGURE 2.

The architecture of the Res-GCNN module. The “X3” in the figure represents three consecutive convolutional layers with kernel sizes of , , and . “X2” represents two consecutive similar residual modules.

FIGURE 3.

Cell Structure of Global Bi-LSTM. We employed a peephole connections, i.e., the extra blue connections.

FIGURE 4.

Cell Structure of Intra-group Bi-LSTM. is the inter-lead feature information introduced from the global Bi-LSTM through the global access structure.

FIGURE 5.

Bi-LSTM Model with Attention Mechanism. represents the sum of the final hidden layer state values for each independent direction in Bi-LSTM, which is called the final state of Bi-LSTM. denotes the attention probability distribution of the hidden layer unit state to the final state at all times. denotes the final ECG feature vector weighted by attention in Bi-LSTM.

FIGURE 6.

Validation performance graph of comparison networks.

FIGURE 7.

Normalized confusion matrix of model results as percentages for the 15-classproblems. The percentage of all possible records in each category is displayed on a color gradient scale.