Wearable Sensors and Artificial Intelligence for Sleep Apnea Detection: A Systematic Review

Area of Science:

• Biomedical Engineering
• Computer Science
• Pulmonology

Background:

• Sleep apnea affects millions globally, impacting public health and quality of life.
• Wearable devices and AI offer new avenues for sleep apnea diagnosis and treatment.
• Real-time monitoring and analysis of sleep patterns are crucial for early detection.

Purpose of the Study:

• To systematically review current AI techniques, wearable devices, data types, and preprocessing methods for sleep apnea diagnosis.
• To identify trends and promising methodologies in the field.
• To assess the impact of data quantity and type on diagnostic accuracy.

Main Methods:

• Systematic literature review of 28 studies published between 2020 and 2024.
• Analysis of wearable device types (patches, clocks, rings) and AI techniques, particularly convolutional neural networks (CNNs).
• Evaluation of data types, preprocessing methods, and the role of transfer learning.

Main Results:

• A growing trend integrates wearable devices with CNNs for sleep apnea diagnosis.
• Transfer learning techniques enhance the performance of AI models.
• The quantity and type of training data significantly influence algorithm outcomes.
• Combinations of neural networks with convolutional layers show potential for precise early diagnosis.

Conclusions:

• AI and wearable devices represent a significant advancement in sleep apnea diagnosis.
• CNNs, especially when combined with transfer learning and diverse data, offer promising results.
• Further research into multi-network combinations is recommended for developing highly accurate diagnostic systems.