Detection of apnoea from respiratory time series data using clinically recognizable features and kNN classification

Area of Science:

• Biomedical Engineering
• Sleep Medicine
• Signal Processing

Background:

• Sleep apnoea is a common sleep-related breathing disorder characterized by temporary upper airway closure during sleep.
• Accurate detection of apnoea is crucial for timely diagnosis and treatment.

Purpose of the Study:

• To develop and evaluate a novel system for detecting sleep apnoea using time series analysis of respiratory impedance signals.
• To assess the performance of a kNN classification algorithm for identifying apnoeic episodes.

Main Methods:

• Time series analysis of one-minute epochs of respiratory impedance signals.
• Development of a MATLAB algorithm for feature extraction from respiratory signals.
• kNN classification of extracted features for apnoea detection.

Main Results:

• The system successfully detected apnoeic episodes in eight eight-hour patient records from PhysioNet.
• Achieved a specificity of 88.1% and a sensitivity of 95.7%.
• ROC analysis yielded an area under the curve of 0.9604, indicating high classification performance.

Conclusions:

• The developed system demonstrates a highly sensitive and specific method for detecting sleep apnoea using respiratory impedance signals.
• The kNN classification approach shows promise for clinical application in sleep apnoea diagnosis.
• Future work will focus on larger datasets and improved result presentation for physicians.