Predicting the initiation of minimum-jerk submovements in three-dimensional target-oriented human arm trajectories

Characterizing and predicting submovements during human three-dimensional arm reaches

We have demonstrated that 3D target-oriented human arm reaches can be represented as linear combinations of discrete submovements, where the submovements are a set of minimum-jerk basis functions for the reaches. We have also demonstrated the ability of deterministic feed-forward Artificial Neural Networks (ANNs) to predict the parameters of the submovements. ANNs were trained using kinematic data obtained experimentally from five human participants making target-directed movements that were decomposed offline into minimum-jerk submovements using an optimization algorithm. Under cross-validation, the ANNs were able to accurately predict the parameters (initiation-time, amplitude, and duration) of the individual submovements. We also demonstrated that the ANNs can together form a closed-loop model of human reaching capable of predicting 3D trajectories with VAF >95.9% and RMSE ≤4.32 cm relative to the actual recorded trajectories. This closed-loop model is a step towards a practical arm trajectory generator based on submovements, and should be useful for the development of future arm prosthetic devices that are controlled by brain computer interfaces or other user interfaces.

Training modalities in robot-mediated upper limb rehabilitation in stroke: a framework for classification based on a systematic review

Robot-mediated post-stroke therapy for the upper-extremity dates back to the 1990s. Since then, a number of robotic devices have become commercially available. There is clear evidence that robotic interventions improve upper limb motor scores and strength, but these improvements are often not transferred to performance of activities of daily living. We wish to better understand why. Our systematic review of 74 papers focuses on the targeted stage of recovery, the part of the limb trained, the different modalities used, and the effectiveness of each. The review shows that most of the studies so far focus on training of the proximal arm for chronic stroke patients. About the training modalities, studies typically refer to active, active-assisted and passive interaction. Robot-therapy in active assisted mode was associated with consistent improvements in arm function. More specifically, the use of HRI features stressing active contribution by the patient, such as EMG-modulated forces or a pushing force in combination with spring-damper guidance, may be beneficial.Our work also highlights that current literature frequently lacks information regarding the mechanism about the physical human-robot interaction (HRI). It is often unclear how the different modalities are implemented by different research groups (using different robots and platforms). In order to have a better and more reliable evidence of usefulness for these technologies, it is recommended that the HRI is better described and documented so that work of various teams can be considered in the same group and categories, allowing to infer for more suitable approaches. We propose a framework for categorisation of HRI modalities and features that will allow comparing their therapeutic benefits.