A Novel Tool for Standing Balance using Functional Electrical Stimulation with Visual Feedback Training

Grabke, Emerson P.1; Apostoli, Christopher2; Chow, Kelvin4; Lee, Jae W.1; Yoo, Jaeeun1; Andrysek, Jan1, 5; Musselman, Kristin E.3; Masani, Kei1, 3

1. Institute of Biomaterials and Biomedical Engineering, University of Toronto; 2. Department of Engineering Science, University of Toronto; 3. Toronto Rehabilitation Institute, University Health Network; 4. Department of Mechanical and Industrial Engineering; 5. Bloorview Research Institute, Holland Bloorview Kids Rehabilitation Hospital

Background: Incomplete spinal cord injury (iSCI) impairs muscle control, mobility and quality of life. Standing without upper body support can increase personal autonomy of individuals with iSCI, increasing their quality of life. Here we developed a visual feedback training (VFT) rehabilitation system using closed-loop controlled functional electrical stimulation (FES) of ankle plantarflexors and dorsiflexors, and subsequently validated its functionality on able-bodied individuals. VFT involves displaying participant center of pressure (COP) onscreen, and tasking the participant with moving their COP to satisfy onscreen tasks. FES induces noninvasive muscle contractions and has been successfully used for upper limb rehabilitation of individuals with iSCI. A PD controller with gravity compensation and asymmetric leg biasing was designed to apply FES to ankle plantarflexors and dorsiflexors in a manner mimicking the human physiological postural controller. VFT and lower-limb FES have been previously shown in our lab to independently improve postural control rehabilitation of individuals with iSCI. Fusing these two methods is hypothesized to improve posture control rehabilitation in individuals with iSCI.

Methods: 1) To optimize the FES controller, ten able-bodied young adults (9M1F, A=20.8±1.9, H = 176.7±9.7 cm, M = 74.1±16.4 kg) completed each of the four VFT games three times apiece, in a randomized order among participants, during which electromyography (EMG) data of their tibialis anterior, medial gastrocnemius and soleus was collected. FES controller parameters were tuned by fitting controller outputs to the EMG data. 2) To validate the system, ten able-bodied young adults (7M3F, A=21.9±2.6, H= 173.7±10.7cm, M=67.8±14.4kg) completed the experiment 1) protocol twice: (a) without FES, and (b) with FES as designed in experiment 1). 30 minutes of rest time was taken between segments to minimize the effect of fatigue.

Results: In experiment 1), we obtained the optimized generalized controller parameters, including a proportional gain of 529 ± 310, derivative gain of 113 ± 44 and asymmetric bias percentage of 20 ± 9. The average Pearson correlation between the simulated generalized controller output and EMG data was r = 0.53 ± 0.11. In experiment 2), we found that no performance score nor COP travel distance was statistically different between trials with and without FES in any games (paired t-tests with p < 0.05, p = 0.28~0.94 and 0.55~0.78 respectively).

Conclusion: The FES controller exhibited good correlation with healthy adult EMG data, suggesting that the FES controller successfully activates the target muscles in the physiological manner. Additionally, the use of the FES controller does not hinder the performance of able-bodied individuals completing the VFT games. We assume that the FES controller can improve the game performance score in case of individuals with iSCI who have unstable standing balance, which will be tested in the next step.