Stroke is one of the most common forms of acquired brain injury (ABI), with around 60,000 new and recurrent strokes occurring every year in Australia alone . The clinical outcome of stroke is variable but often includes persistent upper-limb motor deficits, including weakness, discoordination, and reduced speed and mobility , and cognitive impairments in information processing and executive function [3, 4]. Not surprisingly, stroke is a leading cause of disability worldwide, and the burden of stroke across all levels of the International Classification of Functioning (ICF) – body structures/function, activity, and participation – underlines the importance of interventions that can impact multiple domains of functioning [5, 6].
Recovery of functional performance following stroke remains a significant challenge for rehabilitation specialists [7, 8], but may be enhanced by innovation in the use of new technologies like virtual reality [9, 10, 11, 12]. A critical goal is to find compelling ways of engaging individuals in their therapy by creating meaningful, stimulating and intensive forms of training . The term, virtual rehabilitation (VR), is used to describe a form of training wherein patients interact with virtual or augmented environments, presented with the aid of technology [14, 15]. The technologies can be either commercial systems (e.g. Nintendo Wii, Xbox Kinect) or those customised specifically for rehabilitation. VR offers a number of advantages over traditional therapies, including the ability to engage individuals in the simulated practice of functional tasks at higher doses [16, 17], automated assessment of performance over time, flexibility in the scaling of task constraints, and a variety of reward structures to help maintain compliance .
While evaluation research is still in its infancy, recent systematic reviews and meta-analyses show that VR can enhance upper-limb motor outcomes in stroke [10, 11, 19], yielding treatment effects of medium-to-large magnitude [10, 11], and complementing conventional approaches to rehabilitation. VR has been shown to engender high levels of engagement in stroke patients undergoing physical therapy [20, 21] and training of even moderate intensity can afford functional benefits at the activity/skill level [9, 19]. In the specific case of upper-limb VR, however, there is little available evidence that these benefits transfer to participation . Furthermore, most available data is on patients in chronic stages of recovery, with less on acute stroke . Notwithstanding this, use of VR has begun to emerge in clinical practice, recommended in Australian and international stroke guidelines as a viable adjunct in therapy to improve motor and functional outcomes [22, 23, 24].
Until recently, most VR systems have been designed to improve motor functions, with cognitive outcomes often a secondary consideration in evaluation studies [9, 10, 11]. Notwithstanding this, treatments that target both motor and cognitive functions are indicated for stroke, given evidence that cognitive and motor systems overlap at a structural and functional level [25, 26], and work synergistically in a “perception-action cycle”  in stroke patients undergoing rehabilitation . Recent studies provide preliminary evidence of improved attention and memory in stroke patients following motor-oriented VR [29, 30, 31, 32], amounting to a small-to-medium effect on cognition . When designed to address aspects of cognitive control and planning, VR has the potential to enhance dual-task control, resulting in better generalization of trained skills to daily functioning .
While evaluation research is still in its infancy, several recent customized systems (like Elements, the system evaluated here) have been deliberately designed to exploit factors known to enhance training intensity and motor learning. Informed by neuroscience and learning theory [for a recent review see 12], the Elements VR system was designed to enhance neuro-plastic recovery processes via: (1) an enriched therapeutic environment affording a natural form of user interaction via tangible computing and surface displays , which engage both the cognitive attention of participants and their motivation to explore training tasks; (2) concurrent augmented feedback (AF) on performance  offering participants additional information on the outcome of their actions to assist in re-building a sense of body position in space (aka body schema) and ability to predict/plan future actions; and (3) scaling of task challenges to the current level of motor and cognitive function , ensuring dynamic scaffolding of participants’ information processing and response capabilities. The Elements system, described in detail below and in earlier publications [37, 38], consists of a large (42 in.) tabletop surface display, tangible user interfaces, and software for presenting both goal-directed and exploratory virtual environments. Previous evaluations of the system in patients with traumatic brain injury showed improvements in both motor and cognitive performance, with transfer to activities of daily living [37, 39]. However, the impact of Elements in other forms of ABI, such as stroke, has not been evaluated.
The broad aim of current study was to evaluate the efficacy of the Elements VR interactive tabletop system for rehabilitation of motor and cognitive functions in sub-acute stroke, compared with treatment as usual (TAU). We were particularly interested in motor and cognitive outcomes, their relationship, and the transfer and maintenance of treatment effects. Training-related changes at the activity/skill level on standardized measures of motor and cognitive performance were investigated, together with functional changes. By offering an engaging, principled and customized form of interaction, we predicted that the Elements system would effect (i) greater changes on both motor and cognitive outcomes than with TAU alone; (ii) sustained benefits, as assessed over a short follow-up period, and (iii) transfer to everyday functional performance (i.e. participation).[…]