Analyzing the Sleep and Anxiety States of an Individual with Anxiety Disorder by Utilizing Deep Neural Networks and Heart Rate Variability Based on Fingertip Pulse

Anxiety disorder is a prevalent psychological condition that can impact individuals' blood pressure and sleep patterns. This study employs a photoplethysmographic sensor to gather fingertip pulse data from a patient with an anxiety disorder over the course of one year. Both heart rate variability and deep learning methodologies are utilized to investigate the inherent correlations between the patient's anxiety levels, sleep quality, and blood pressure. A total of 824 five-minute samples of fingertip pulse data are utilized, and the data is categorized based on sleep quality and anxiety levels. After preprocessing steps such as noise reduction, both time-domain and frequency-domain heart rate variability indices are calculated. The analysis is carried out across various categories, illustrating that heart rate variability can effectively discern physiological states based on fingertip pulse data. Finally, the EfficientNetB0 neural network is employed to predict blood pressure values, and trends in blood pressure during different anxiety and sleep states are analyzed. The results indicate that blood pressure varies across different sleep and anxiety states. This research provides a non-invasive method for understanding individuals with anxiety disorders' sleep issues and anxiety states, offering valuable insights for clinical treatment and intervention.