Reconstruction of 12-Lead Electrocardiogram from a Three-Lead Patch-Type Device Using a LSTM Network

Abstract

Reconstructing a standard 12-lead electrocardiogram (ECG) from signals received from electrodes packed into a patch-type device is a challenging task in the field of medical instrumentation. All attempts to obtain a clinically valid 12-lead ECG using a patch-type device were not satisfactory. In this study, we designed the hardware for a three-lead patch-type ECG device and employed a long short-term memory (LSTM) network that can overcome the limitations of the linear regression algorithm used for ECG reconstruction. The LSTM network can overcome the issue of reduced horizontal components of the vector in the electric signal obtained from the patch-type device attached to the anterior chest. The reconstructed 12-lead ECG that uses the LSTM network was tested against a standard 12-lead ECG in 30 healthy subjects and ECGs of 30 patients with pathologic findings. The average correlation coefficient of the LSTM network was found to be 0.95. The ability of the reconstructed ECG to detect pathologic abnormalities was identical to that of the standard ECG. In conclusion, the reconstruction of a standard 12-lead ECG using a three-lead patch-type device is feasible, and such an ECG is an equivalent alternative to a standard 12-lead ECG.

Keywords: LSTM network; cardiovascular monitoring; hypertension; patch-type device; reconstructed electrocardiogram; synthesized ECG; ubiquitous healthcare.

Figure 1

(a.1) Picture of the device attached on a patient. (a.2) Customized circuit. (a.3) Components of the front side. (a.4) Components of the back side. (b) Block diagram showing the signal processing and data acquisition of the device.

Figure 2

(a) Screenshot of three-lead ECG (electrocardiogram) data acquisition program. (b) Screenshot of ECG reconstruction and analyzing program using ECGPUWAVE.

Figure 3

System for the clinical validation of a reconstructed ECG from a patch-type device. ECG: electrocardiogram.

Figure 4

Flow of the proposed algorithm (x1: lead1 ECG data, x2: lead2 ECG data, x3: lead3 ECG data, MSE: mean square error). ECG: electrocardiogram.

Figure 5

Structure of employed LSTM (X1: lead1 ECG data, X2: lead2 ECG data, X3: lead3 ECG data). LSTM: long short-term memory.

Figure 6

The loss values of reconstruction models based on the numbers of hidden nodes.

Figure 7

Correlation coefficients between standard 12-lead ECGs and synthesized ECGs. (a) Linear regression. (b) LSTM network. LSTM: long short-term memory

Figure 8

False negative result for LVH.

Figure 9

(a) False positive for ST elevation and (b) false negative result for ST depression.

Figure 10

(a) False positive wide QRS in the reconstructed ECG using LR. (b) Large differences in voltage between limb and precordial leads were usually seen in patients who showed the false positive wide QRS in the reconstructed ECG using LR. LR: linear regression, ECG: electrocardiogram

Figure 11

Reproduction of the ventricular ectopic beats that are not present in the 5 s training data.