Effects of chest movements while sitting on Navon task performance and stress levels

Area of Science:

• Biomedical Engineering
• Human-Computer Interaction
• Cognitive Science

Background:

• Remote work often involves prolonged sitting and limited physical activity.
• Understanding subtle body movements during sedentary tasks is crucial for health and performance monitoring.
• Developing accurate methods to classify physical activity in remote work settings is needed.

Purpose of the Study:

• To develop a machine learning model for classifying chest movements during remote work.
• To investigate the impact of different chest movements on cognitive performance and stress levels.
• To validate a body motion classification model using real-world remote work scenarios.

Main Methods:

• Experiment 1: Developed a Long Short-Term Memory (LSTM) neural network model using data from five participants to classify six types of chest movements.
• Experiment 2: Recruited 34 participants to validate the LSTM model and assess the effects of chest movements on cognitive performance (Navon task) and stress.
• Utilized sensor data to capture and analyze subtle chest movements.

Main Results:

• The LSTM model achieved an 83.8% accuracy in classifying six distinct chest movement categories.
• Swinging chest movement was significantly correlated with improved cognitive performance, indicated by reduced reaction time and increased accuracy on the Navon task.
• Stress measurement results were unreliable, preventing definitive conclusions on the relationship between body movements and stress.

Conclusions:

• LSTM classification effectively distinguishes subtle chest movements, with 'swinging' showing a positive impact on cognitive performance.
• Chest movements, particularly swinging, can enhance cognitive functions like reaction time during sedentary tasks.
• Further research is needed to reliably measure the impact of physical activity on stress during remote work.