Author(s): CHATTERJEE, P., CARROLL, T., TWEEDY, S., BECKMAN, E., Institution: THE UNIVERSITY OF QUEENSLAND, Country: AUSTRALIA, Abstract-ID: 1746

INTRODUCTION:
Para sport classification aims to minimize the impact of impairment on competition outcomes (1). Athletes with impaired strength must produce maximal efforts during classification. Intentional misrepresentation (IM) of maximal effort may lead an athlete to be assigned to a class with more impaired competitors, gaining potential competitive advantage. No objective methods for detecting IM currently exist (2). This study developed a force-matching protocol requiring athletes to match target forces from submaximal to supramaximal levels under variable gain visual feedback. It was hypothesized that force and myoelectric features from the protocol would provide markers to discriminate IM from genuine effort, as consistent misrepresentation may become difficult when the relationship between force exerted and visual feedback is unpredictable across trials.
METHODS:
Twenty-six non-impaired participants performed wrist extension and flexion tasks with electrical activity recorded from the Extensor Carpi Radialis Brevis and Flexor Carpi Radialis muscles. Participants first performed alternate maximal voluntary wrist extension and flexion (MVC) efforts. Twitch interpolation quantified voluntary activation (VA) of the wrist extensor. Participants then performed force matching efforts at 20%,40%,60%,80%,100%, and 120% of MVC in random order under three different visual feedback gains, alternating extension and flexion. Participants completed three blocks with genuine effort and three blocks with deceptive submaximal effort on different days. Forces produced were fitted against the target forces with piecewise linear regression comprising a ramp-up and a plateau phase. Extracted features captured force matching consistency and accuracy at near-maximal and maximal targets, the scaling of produced force with increasing target force, and the time of force development at supramaximal targets. A logistic regression model trained on a separate cohort in a previous study to discriminate genuine efforts from IM was applied to the current data and validated using leave-one-out cross-validation, then refitted with beta values derived from the current dataset.
RESULTS:
VA during genuine extension was 94.7±5.4% versus 74±14% during IM. The precomputed model produced no false positives in either movement but detected only 15% of IM efforts. After refitting, sensitivity rose from 15% to 73% and 77% for extension and flexion, respectively, representing approximately a fivefold improvement. Similarly, overall accuracy increased from 58% to 75% and 83%.
CONCLUSION:
This force matching protocol with variable gain feedback provides early evidence that objective detection of IM during isometric strength testing is feasible. Since classifiers have no objective methods of identifying deceptive effort, even modest detection capability is a practical step forward. Future studies should validate this protocol in athletes with strength impairments.
References:
1. Tweedy & Vanlandewijck (2011)
2. Deuble et al. (2015)