Investigation of novel control strategies for promoting motor learning in the upper limb with a haptic computer exercise system in able-bodied adults and those with motor impairments

Alexoulis-Chrysovergis, Aris Christos (2017) Investigation of novel control strategies for promoting motor learning in the upper limb with a haptic computer exercise system in able-bodied adults and those with motor impairments. Doctoral thesis (PhD), Manchester Metropolitan University.

Preview

Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (19MB) | Preview

Abstract

Motor impairments caused by stroke and cerebral palsy (CP) are common and often affect the function of the upper limb, which to be restored requires rehabilitation. As positive outcome is correlated to how early and intensive therapy is and since the resources of the healthcare providers are limited, robotic devices have been introduced to provide adjunctive therapy. The algorithms that control the manner those devices apply forces to the impaired limb are called haptic control algorithms (HCA) and to this date there has not been conclusive evidence as to what the behaviour of these algorithms should be. One type of HCAs is error augmentation (EA) which is a rather understudied but promising approach. This work presents to the literature two novel control strategies of the EA type that incorporate adaptive features namely Error Augmenting Adaptive(EA) and Error Augmenting Proportional (EA). Those two algorithms were implemented for and deployed to a single point of attachment robotic rehabilitation system. The effectiveness in inducing motor learning of the developed algorithms was evaluated in a trial with able-bodied participants and compared against a third more established assistive HCA namely Assistance As Needed (AAN) and a control condition (no forces). Four groups (one per condition) practised reaching movements with a speed and accuracy requirement using their non-dominant arm to interact with the robot under a visual rotation of a 100o. To assess learning kinematic measures were collected to measure their performance on reaching and circle-drawing movements. Also, bilateral transfer to the arm that did not receive practice was assessed. Changes in the participants’ valence, arousal and dominance were assessed with a Self-Assessment Manikin questionnaire. All groups learned to move their non-dominant arm under a visual perturbation showing comparable improvements in all key measures (p<0.05). Passive movements and EAP led to greater improvement in movement smoothness (p<0.05) and resulted in more retention of the improvements after a washout block (p<0.05) was introduced. Conversely, EAA showed a better effect on improving mean velocity (p<0.05). All groups performed similarly in terms of improving movement error and duration but EAA and AAN achieved peak performance faster (p<0.05). Similar improvements were measured on the arm that did not receive any training which were fully retained post-washout indicating that bilateral transfer occurred and led to better retention (p<0.05). The findings of this work indicate that different attributes can be exploited from the developed HCAs to induce motor learning and improve different aspects of the movement suggesting that multimodal training protocols tailored to the needs of the patient are the way forward. Also, this work showed that bilateral transfer training has great potential in upper limb rehabilitation and the positive effects of the different HCAs on the arm that received practice transfer to the one that did not receive training. It is recommended that the findings of this work to be further investigated in experimental therapy protocols for those who suffer from neurological impairments such stroke and CP.