2024 Volume 11 Issue 6 Pages 289-296
An electrocardiogram (ECG) is useful for diagnosing heart diseases, particularly arrhythmias; however, it requires time to detect rare abnormalities that occur infrequently in normal conditions. This study aimed to detect abnormalities in the ECG by evaluating the performance of the matrix profile (MP) algorithm, which is used to detect outliers in data with repetitive patterns. Female hairless rats (HWY/Slc) were used, and all procedures were performed under isoflurane anesthesia. The tricyclic antidepressant amitriptyline hydrochloride (50 mg/kg) was administered intraperitoneally, and ECG measurements were obtained using carbon nanotube yarn as the surface electrode. The measurements were collected by an analog-to-digital converter at a sample rate of 2 kHz. Jupyter Lab 4.0.9 was used for the Python 3.9.1 script environment in the MP analysis, with the following related libraries: Numpy 1.19.5 and Pandas 1.2.1 for data processing, Matplotlib 3.3.4 for data visualization, and Matrixprofile 1.1.10 as the implementation library for the MP algorithm. A data size of 2.5 or 25 sec was applied without prior baseline adjustment, labeling, or normalization. The MP analysis did not detect abnormalities in cases with slow changes in the waveform, even when the waveforms were abnormal. However, it detected heart rate and amplitude fluctuations that rarely occurred in normal conditions. In particular, the MP analysis detected discord during a sudden change, such as a cardiac arrest. The MP algorithm showed excellent performance in terms of the time required for analysis and demonstrated the potential to detect ECG abnormalities in real-time during toxicity testing.
