Association between Self-Reported Prior Night's Sleep and Single-Task Gait in Healthy, Young Adults: A Study Using...

Area of Science:

• Biomechanics
• Sleep Science
• Machine Learning

Background:

• Insufficient sleep (7–9 hours) is linked to injuries in various age groups.
• Existing research on sleep and gait primarily focuses on older adults and clinical populations.
• The impact of sleep duration on gait in healthy young adults remains underexplored.

Purpose of the Study:

• To investigate the relationship between prior night's sleep duration (deprivation or extension) and gait characteristics in healthy young adults.
• To identify individuals with varying sleep durations using single-task gait analysis.
• To assess the efficacy of machine learning in predicting sleep status from gait parameters.

Main Methods:

• 123 healthy young adults (24.3 ± 4.0 years) participated.
• Self-reported sleep duration (categorized as <7h, 7–9h, or >9h) was collected.
• Gait data was acquired using inertial sensors during a 2-minute walk test.
• Machine learning (Random Forest Classifier) and ANCOVA were employed to analyze group differences and correlations.

Main Results:

• A significant correlation (r = 0.79) was found between gait parameters and prior night's sleep duration.
• The Random Forest Classifier achieved a mean accuracy of 65.03% using the top 9 gait features.
• Good sleepers (7–9 hours) demonstrated more asymmetrical gait patterns and superior gait speed maintenance compared to poor sleepers (<7h or >9h).

Conclusions:

• Single-task gait parameters can be indicative of prior night's sleep duration in young adults.
• Individuals sleeping 7–9 hours exhibit distinct gait characteristics compared to those with shorter or longer sleep durations.
• Further research with larger cohorts is warranted to enhance the accuracy of machine learning models for sleep status prediction from gait data.