Assessment of Brain Function After 240 Days Confinement Using Functional Near Infrared Spectroscopy.

Area of Science:

• Neuroscience
• Space Medicine
• Machine Learning

Background:

• Astronauts face significant stressors on long-duration space missions.
• Early detection of mental stress is critical for mission success and crew health.
• Current methods for monitoring astronaut mental stress are limited.

Purpose of the Study:

• To identify and quantify mental stress levels in astronauts during a 240-day confinement scenario.
• To evaluate the effectiveness of functional near infrared spectroscopy (fNIRS) combined with machine learning for stress assessment.
• To explore various physiological and performance indicators of stress.

Main Methods:

• Utilized functional near infrared spectroscopy (fNIRS) to measure brain activity.
• Collected stress indicators: salivary alpha amylase (sAA), reaction time (RT), target detection accuracy, power spectral density (PSD), and functional connectivity networks (FCN).
• Employed Fast Fourier Transform (FFT) for PSD estimation and partial directed coherence for FCN analysis.
• Applied five machine learning classifiers (KNN, LDA, NB, DT, SVM) to differentiate stress levels.

Main Results:

• Salivary alpha amylase (sAA) levels increased throughout the 240-day mission.
• Reaction time (RT) and target detection accuracy showed significant fluctuations.
• Power spectral density (PSD) increased, while functional connectivity networks (FCN) decreased in specific brain regions.
• Machine learning models achieved high accuracy rates, with Support Vector Machine (SVM) and k-Nearest Neighbor (KNN) exceeding 96%.

Conclusions:

• fNIRS combined with machine learning provides an effective method for assessing astronaut mental stress.
• The study demonstrates a promising approach for early stress detection in prolonged space missions.
• This technology can contribute to astronaut health monitoring and mission safety.