REAL-TIME BIOFEEDBACK INTEGRATED INTO NEUROMUSCULAR TRAINING INCREASES BRAIN FUNCTIONAL CONNECTIVITY AND REDUCES HIGH-RISK KNEE BIOMECHANICS

Background: Anterior cruciate ligament (ACL) injury is associated with alterations in the central nervous system and resultant sensorimotor control (Courtney et al., 2005; Grooms et al., 2017). Our prospective data indicates that altered knee-motor functional brain connectivity is associated with increased risk for ACL injury (Diekfuss et al., revisions invited), revealing novel neural targets for neuromuscular training interventions. Specifically, interventions that integrate concomitant sensorimotor feedback with injury prevention techniques have the potential to enhance brain functional connectivity to optimize ACL injury risk reduction strategies. To deliver concomitant sensorimotor feedback, we have developed an augmented neuromuscular training (aNMT) system that utilizes interactive, real-time biofeedback to simultaneously target multiple biomechanical variables associated with ACL injury risk (Bonnette et al., in press; Figure 1A). aNMT calculates and maps key biomechanical parameters to an interactive graphical shape that responds in real time as a function of participants’ movements. Participants are instructed to perform exercises to achieve a goal shape, which equates to producing biomechanical parameters associated with ACL injury risk reduction, while deviations toward injury risk factors result in specific shape distortions. We hypothesized that aNMT would significantly improve biomechanics associated with ACL injury risk and also increase knee-motor functional connectivity. We further predicted that the identified connectivity changes would be associated with the hypothesized changes in biomechanics. Methods: Over six weeks of training, participants (n = 25) performed a series of aNMT-based progressive exercises (e.g., squat, overhead squat, squat jump, tuck jump, single-leg Romanian dead lift, pistol squat) and completed a drop vertical jump (DVJ) task while fully instrumented for 3D motion analysis pre- and post-training. Peak knee abduction moment (pKAM; bilateral average) from the DVJ was used as the biomechanical outcome variable. Resting-state functional magnetic resonance imaging (fMRI) scans were also collected pre- and post-training on a subset of the cohort (n = 17). Thirteen additional participants were recruited to serve as untrained controls and completed the DVJ and resting-state fMRI on two testing sessions separated by approximately 6 weeks. Twenty-five knee-motor regions of interest (ROIs) were created based on the areas of brain activation derived from previously published data (Grooms et al., 2015; Kapreli et al., 2007). Paired-samples t tests with a false discovery rate correction for multiple comparisons determined differences in functional connectivity among these 25 ROIs (post > pre). Fisher-transformed Pearson correlation coefficients between the average residual blood oxygen level dependent (BOLD) signal time series extracted from ROIs that demonstrated significant group level changes were associated with pKAM in DVJ task at pre- and post-training. The pre- and post-training Pearson correlation coefficients were subsequently compared using the cocor package (Diedenhofen & Musch, 2015) to determine if the two relationships were significantly different. Results: Results showed that pKAM in the aNMT group was significantly lower following aNMT (p < .05), while no significant changes were found between the two time points for controls (p > .05). Results also revealed significantly greater functional connectivity between the right supplementary motor area (SMA) and the left thalamus at post-training relative to pre-training for the aNMT group, t(16) = 3.37, p = .04 (Figure 1B). No significant differences between the two time points were observed for the controls (all p > .05). The association between pKAM and the right SMA and left thalamus at pre-training (r = -.22; Figure 1C) was significantly different compared to that at post-training (r = .26; Figure 1D), p < .05, with a positive relationship between pKAM and SMA and thalamus activation following aNMT biofeedback. No similar changes in pKAM and right SMA and left thalamus connectivity were observed for the untrained controls, p > .05. Conclusions/Significance: The right SMA is involved in the planning and coordination of movement, and the left thalamus is associated with neuromotor control. The increased functional connectivity between these regions, combined with the reduction in pKAM, which is associated with reduced risk of ACL injury, indicate a possible neural mechanism for improved motor adaption associated with aNMT biofeedback. These findings have distinct implications for ACL injury prevention strategies. Biofeedback tools such as aNMT can be designed to target specifically the neural drivers of aberrant movement biomechanics underlying increased ACL injury risk. [Figure: see text]