A Machine Learning-Based Initial Difficulty Level Adjustment Method for Balance Exercise on a Trunk Rehabilitation Robot.

Trunk rehabilitation exercises such as those for remediating core stability can help improve the seated balance of patients with weakness or loss of proprioception caused by diseases such as stroke, and aid the recovery of other functions such as gait. However, there has not yet been any reported method for automatically determining the parameters that define exercise difficulty on a trunk rehabilitation robot (TRR) based on data such as the patient's demographic information, balancing ability, and training sequence, etc. We have proposed a machine learning (ML)-based difficulty adjustment method to determine an appropriate virtual damping gain (Dvirtual) of the controller for the TRR's unstable training mode. Training data for the proposed system is obtained from 37 healthy young adults, and the trained ML model thus obtained is tested through experiments with a separate population of 25 healthy young adults. The leave-one-out cross validation results (37 subjects) from the training group for validation of the designed ML model showed 80.90% average accuracy (R2 score) for using the given information to predict the desired difficulty levels, which are represented by the level of balance performance quantified as Mean Velocity Displacement (MVD) of the center of pressure. Statistical analysis (Repeated measures analysis of variance) of subject performance also showed that ground truth difficulty levels from the training data and predicted difficulty levels did not differ significantly under any of the three exercise modes used in this study (Hard, Medium, and Easy), and the standard deviations were reduced by 16.39, 41.39, and 25.68%, respectively. Moreover, the Planar Deviation (PD) of the center of pressure, which was not the target parameter here, also showed results similar to the MVD, which indicates that the predicted Dvirtual affected the difficulty level of balance performance. Therefore, the proposed ML model-based difficulty adjustment method has potential for use with people who have varied balancing abilities.

Study on the Effects of Different Seat and Leg Support Conditions of a Trunk Rehabilitation Robot.

Performance of trunk rehabilitation exercises while sitting on movable surfaces with feet on the ground can increase trunk and leg muscle activations, and constraining the feet to move with the seat isolates control of the trunk. However, there are no detailed studies on the effects of these different leg supports on the trunk and leg muscle activations under unstable and forcefully perturbed seating conditions. We have recently devised a trunk rehabilitation robot that can generate unstable and forcefully perturbed sitting surfaces, and can be used with ground-mounted or seat-connected footrests. In this study, we have evaluated the differences in balance performance, trunk movement and muscle activation (trunk and legs) of fourteen healthy adults caused by the use of these different footrest configurations under the different seating scenarios. The center of pressure and trunk movement results show that the seat-connected footrest may be a more suitable choice for use in a balance recovery focused rehabilitation protocol, while the ground-mounted footrest may be a more suitable choice for a trunk movement focused rehabilitation protocol. Although it is difficult to make a clear selection between footrests due to the mixed trends observed in the muscle activation results, it appears that the seat-connected footrest may be preferable for use with the unstable seat as it causes greater muscle activations. Furthermore, the results provide limited evidence that targeting of a particular muscle group may be possible through careful selection of the seat and footrest conditions. Therefore, it may be possible to utilize the trunk rehabilitation robot to maximize the training outcomes for a wide range of patients through careful selection of training protocols.

Comparative Study on Overground Gait of Stroke Survivors With a Conventional Cane and a Haptic Cane.

The conventional cane (single cane) is widely used to promote gait ability of stroke survivors as it provides postural stability by extending the base of support. However, its use can reduce muscle activity in the user's paretic side and cause upper limb neuropathies due to the intermittent and excessive loading of the upper limb. The provision of low magnitude support and speed regulation may result in collective improvement of gait parameters such as symmetry, balance and muscle activation. In this paper, we developed a robotic Haptic Cane (HC) that is composed of a tilted structure with motorized wheels and sensors to allow continuous haptic contact with the ground while moving at a regulated speed, and carried out gait experiments to compare the HC with an Instrumented conventional Cane (IC). The results show that use of the HC involved more continuous ground support force of a comparatively lesser magnitude than the IC, and resulted in greater improvements in the swing symmetry ratio and significant improvements in the step length symmetry ratio. Percentage of Non-Paretic Activity (%NPA) of paretic muscles (vastus medialis obliquus (VMO), semitendinosus (SMT), tibialis anterior (TBA) and gastrocnemius medialis (GCM)) in swing phase was significantly improved by the use of either device at fast speed. However, the use of HC improved %NPA of paretic VMO and SMT more than the use of IC at both preferred and fast speeds. It also significantly improved %NPA of paretic GCM in stance phase. Furthermore, comfortable speed with the HC was higher than with the IC and exhibited better RMS of anteroposterior (AP) tilt. Thus, the developed device with a simple and intuitive mechanism can provide efficient assistance for overground gait of stroke patients with a high possibility of widespread use.

Effects of Vibrotactile Biofeedback Coding Schemes on Gait Symmetry Training of Individuals With Stroke.

Variations in biofeedback coding schemes for postural control, in recent research, have shown significant differences in performance outcomes due to variations in coding schemes. However, the application of vibrotactile biofeedback coding schemes to gait symmetry training is not well explored. In this paper, we devised various vibrotactile biofeedback modes and identified their efficacy during gait symmetry training of individuals suffering from hemiparesis due to stroke. These modes are composed of variations in vibration type (on-time or intensity), and relation type (proportional or inversely-proportional) with the error in symmetry ratio. Eight individuals with stroke participated in walking trials. From dependent t-tests on the collected data, we found improved achievement of temporal gait symmetry while utilizing all the provided biofeedback modes compared to no biofeedback (P < 0.001). Furthermore, two-way repeated measures ANOVA revealed statistically significant difference in symmetry ratio for main effect of vibration type (P-value = 0.016, partial eta squared = 0.585). The participants performed better with modes of biofeedback with varying vibration on-times. Furthermore, participants showed better performance when the biofeedback varied proportionally with the error. These findings suggest that biofeedback coding schemes may have a significant effect on the performance of gait training.

Evaluating the Effects of Kinesthetic Biofeedback Delivered Using Reaction Wheels on Standing Balance.

Aging, injury, or ailments can contribute to impaired balance control and increase the risk of falling. Provision of light touch augments the sense of balance and can thus reduce the amount of body sway. In this study, a wearable reaction wheel-based system is used to deliver light touch-based balance biofeedback on the subject's back. The system can sense torso tilt and, using reaction wheels, generates light touch. A group of 7 healthy young individuals performed balance tasks under 12 trial combinations based on two conditions each of standing stance and surface types and three of biofeedback device status. Torso tilt data, collected from a waist-mounted smartphone during all the trials, were analyzed to determine the efficacy of the system. Provision of biofeedback by the device significantly reduced RMS of mediolateral (ML) trunk tilt (p < 0.05) and ML trunk acceleration (p < 0.05). Repeated measures ANOVA revealed significant interaction between stance and surface on reduction in RMS of ML trunk tilt, AP trunk tilt, ML trunk acceleration, and AP trunk acceleration. The device shows promise for further applications such as virtual reality interaction and gait rehabilitation.