The Use Of Micro-doppler Radar To Differentiate Sex By Functional Movement

Authors: Will Zofkie1, Cunhao Zeng2, Ram Narayanan2, Cayce Onks1 Institutions: Penn State College of Medicine1, Penn State College of Engineering2 The Use of Micro-Doppler Radar to Differentiate Sex by Functional Movement PURPOSE: Micro-Doppler radar (MDR) coupled with machine learning is a promising alternative to traditional motion analysis. There is interest to understand if MDR can be used to identify individuals at risk for musculoskeletal injury (MSKI). Here we attempt to use MDR to identify female sex, which is a known risk factor associated with Anterior Cruciate Ligament Injury (ACLI). If we can use MDR to identify common risk factors such as sex, then it may be possible to identify those at risk for specific MSKI such as ACLI. METHODS: This is an IRB approved case control study comparing MDR signatures for participants aged 18-40 y/o who were asked to perform a sit to stand maneuver in the presence of MDR. The participants were consented, data was collected, and analysis would seek to determine male vs female. The data sets were analyzed using established machine learning techniques with 80/20 training/test split over multiple trials to establish the accuracy of the model RESULTS: A total of 102 subjects participated in this project. Of these, 82 subjects were used in an 80/20 train/test split to develop and evaluate the model, while a separate 20 subjects (10 males, 10 females) were used as an independent validation set to assess final accuracy. The model achieved an overall accuracy of 74.7% on the validation set. Additionally, precision for males was 80.5%, indicating that when the model predicted "male," it was correct 80.5% of the time. For females, precision was 70.7%, reflecting the model’s accuracy when it predicted "female." CONCLUSION: This is the first attempt to use micro-Doppler technology to determine an individual’s sex based on movement patterns alone. We were successful in determining male and female with good accuracy. This confirms that MDR coupled with machine learning can identify common risk factors for ACLI such as female sex. Future studies should seek to understand if MDR can be used not only to identify risk factors, but individuals at risk for specific injuries such as ACLI. Supported by: This work was supported by the Department of Defense through the FY21, Peer Reviewed Orthopaedic Research Program under Award No. W81XWH2210684. Opinions, interpretations, conclusions and recommendations are those of the author and are not necessarily endorsed by the Department of Defense.