Graduation Project
Start Date : 01 Oct 2023-01 Jul 2024
Cardiorespiratory dysregulation in newborn infants is a main concern that influences the infant’s hospital discharge. Therefore, it is common practice to delay discharge by five consecutive days free of any cardiorespiratory events in order to avoid sudden unexplained life-threatening events after discharge, as such occurrence may result in re-hospitalization. Bradycardia is one of these cardiorespiratory events that is common in preterm infants and may occur spontaneously for several reasons. Bradycardia is defined as a heart condition, where the heart beats at a slower rate than normal. Automatic detection of such a serious health condition is expected to help achieve improved diagnostic and focused management strategies for preterm-born neonates. Hence, this project aims to present a signal processing-based approach for detecting bradycardia in preterm infants. For this, a dataset comprising electrocardiogram (ECG) and respiratory signals recorded from a number of preterm infants admitted to the hospital’s neonatal intensive care unit (NICU) will be used. MATLAB software will be utilized as a simulation tool to incorporate different signal processing techniques with the purpose of analyzing the recorded ECG and respiratory signals. The principal role of employing these signal processing techniques is to extract appropriate features from the collected physiological signals, which will be subsequently used to train a classification model. The capability of this classification model in distinguishing infants with bradycardia from the control group will be tested in terms of different classification performance parameters. Moreover, we look forward to coming up with a hardware prototype of the model that will be developed during the project according to the time plan and the progress of the participating students.