Deep Learning Unravels Differences Between Kinematic and Kinetic Gait Cycle Time Series from Two Control Samples of Healthy Children Assessed in Two Different Gait Laboratories

• Published in: Sensors (Basel, Switzerland) Vol. 25; no. 1; p. 110
• Main Authors: de Gorostegui, Alfonso, Kiernan, Damien, Martín-Gonzalo, Juan-Andrés, López-López, Javier, Pulido-Valdeolivas, Irene, Rausell, Estrella, Zanin, Massimiliano, Gómez-Andrés, David
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 01.01.2025; MDPI
• Subjects: Algorithms; Biomechanical Phenomena - physiology; Child; children; Classification; Datasets; Deep Learning; Ethics; external validity; Female; Gait; Gait - physiology; Gait Analysis - methods; Humans; Kinematics; Kinetics; Laboratories; Machine learning; Male; Medical history; Neural networks; Orthopedics; Pediatrics; Software; Time series; Ireland; Spain; deep learning; children; gait; external validity
• ISSN: 1424-8220; 1424-8220
• DOI: 10.3390/s25010110

We investigate the application of deep learning in comparing gait cycle time series from two groups of healthy children, each assessed in different gait laboratories. Both laboratories used similar gait analysis protocols with minimal differences in data collection. Utilizing a ResNet-based deep learning model, we successfully identified the source laboratory of each dataset, achieving a high classification accuracy across multiple gait parameters. To address the inter-laboratory differences, we explored various pre-processing methods and time series properties that may have been detected by the algorithm. We found that the standardization of the time series values was a successful approach to decrease the ability of the model to distinguish between the two centers. Our findings also reveal that differences in the power spectra and autocorrelation structures of the datasets play a significant role in the model performance. Our study emphasizes the importance of standardized protocols and robust data pre-processing to enhance the transferability of machine learning models across clinical settings, particularly for deep learning approaches.