A fused-image-based approach to detect obstructive sleep apnea using a single-lead ECG and a 2D convolutional neural network

• DOI: 10.1371/journal.pone.0250618
• Publication Date: 2021-04-26T18:12:50Z
• Authors: S. M. Isuru Niros...; Xin Zhu; Keijiro Nakamura; Wenxi Chen
• Keywords: Lead (geology)

Obstructive sleep apnea (OSA) is a common chronic disorder that disrupts breathing during and associated with many other medical conditions, including hypertension, coronary heart disease, depression. Clinically, the standard for diagnosing OSA involves nocturnal polysomnography (PSG). However, this requires expert human intervention considerable time, which limits availability of diagnosis in public health sectors. Therefore, electrocardiogram (ECG)-based methods detection have been proposed to automate procedure reduce its discomfort. So far, most approaches rely on feature engineering, calls advanced knowledge experience. This paper proposes novel fused-image-based technique detects using only single-lead ECG signal. In approach, convolutional neural network extracts features automatically from images created one-minute segments. The comprises 37 layers, four residual blocks, dense layer, dropout soft-max layer. study, three time–frequency representations, namely scalogram, spectrogram, Wigner–Ville distribution, were used investigate effectiveness approach. We found blending scalogram spectrogram further improved system’s discriminative characteristics. Seventy recordings PhysioNet Apnea-ECG database train evaluate model 10-fold cross validation. results study demonstrated classifier can perform an average accuracy, recall, specificity 92.4%, 92.3%, 92.6%, respectively, fused spectral images.