Using virtual environment (VE) technology in rehabilitation may facilitate the rehabilitation process by (i) allowing the systematic presentation of a varied set of rehabilitation exercises, tailored to the individual patient and his or her specific deficit, (ii) providing the therapist with additional quantitative assessment tools for diagnosis and progress monitoring, and (iii) providing the patient with purposeful and motivational environments which make it easier and more fun to adhere to a certain treatment protocol. For the rehabilitation of motor skills, a combination of VEs with haptic feedback and robotic technology is frequently used, in order to support and guide movements [1–4]. During the course of implementing VE rehabilitation program, clinicians are faced with many technology options and choices, that are not necessarily well understood in terms of their effects on a clinical population. One such option is whether or not to utilise stereo visualization as a way to present the rendered environment. Stereoscopic visualization has been associated with improved judgments of three-dimensional spatial relationships e.g., , more precise localization of objects , faster performance in visually guided reaching tasks e.g., , and a more natural mapping of people’s movement in the VE . However, negative effects associated with stereoscopic displays include people reporting a feeling of eyestrain, blurred vision, focusing problems, all of which can impact the usefulness of stereoscopic displays [for an overview see: . While using stereoscopic presentation of the VE could have beneficial effects for patients with neuropsychological deficits, it is also a potentially more vulnerable group which may already suffer from visual and spatial cognition deficits. It is unclear to what extent visualization and graphic rendering of the VE is beneficial or disruptive to patients’ navigation and experience of rehabilitation training. In the current study we investigated responses of patients with a clinical diagnosis of MS or CVA to manipulations of depth cues used within a rehabilitation VE. Specifically, we manipulated the dimensionality of the graphic environment (2.5D vs. 3D), and the use of stereoscopic presentation of the VE (monoscopic vs. stereoscopic).
Our findings illustrate that stereoscopic visualization of a VE can be beneficial over monoscopic visualization across patient groups. These benefits are, however, specific to the task that is being performed (reach, lift, and transport). Stereoscopic visualization was found to have positive effects on average speed, time spent on the tasks, and accuracy when the task itself included movement in depth (i.e., a reaching task). With a decrease in time, an increase in speed and an increase in accuracy it can be concluded that stereoscopy aided patients to reach the target more efficiently. When the task itself did not include movement in depth (lifting and transporting), these benefits were less marked. Although lacking functional use of the depth cues during lifting, there was evidence of increased speed and reduced time resulting from stereoscopy, but no effects were found regarding accuracy. This makes sense as the additional information provided by the stereoscopic visualization was to a lesser extent needed for the target localization during the vertical lifting task.
While these findings may lead to a conclusion that stereoscopic visualization is to be recommended in motor skills training for the patient groups under study, stereoscopic presentation did, however, result in more eyestrain in both MS and stroke patients. In addition, patients having weak stereovision reported higher fatigue. In this way, stereoscopy can also impede VR training. These are side effects that should be weighed carefully, as they potentially inhibit people's motivation to engage in training, leading to reduced training duration and reduced potential benefits for the recovery of arm-hand function. Since the potential motivational benefits of using VEs for rehabilitation training can be regarded as one of the key benefits from the patient’s point of view, negative effects on visual comfort may nullify such benefits. As such, current research regarding 3D visualizations and visual comfort are welcomed, while examinations with more recent systems are needed as stereoscopic systems continue to improve in image quality and visual comfort.
Besides influences of stereoscopy on time, speed and accuracy we also found evidence indicating that the graphic rendering environment (2.5D vs. 3D) influences these parameters. For the lifting task it was found that participants spent less time on the task and the average speed was higher in the 2.5D than in the comparative 3D environment. However, the stereoscopic ability varied considerably in each patient group, and clearly influenced the results. There was evidence that these benefits of the 2.5D environment over the 3D environment were most marked for those with weak stereoscopic vision. In contrast, for participants with good stereoscopic vision (SFT 5 or higher) the results suggest that the effects are opposite. For them, it appeared that the 3D environment reduced time on the task, increased average speed, and improved accuracy (reduced the average overshoot) as compared to the 2.5D condition. Although this finding has clear implications for the importance of tailoring the VE to the stereoscopic ability of individual patients, we do not have a readily available explanation for this finding. One potential explanation for this finding is that although visualization of a VE in 3D (as compared to 2.5D) improves the virtual environment in realism, by doing so part of the simplicity of the flat planes in the 2.5D environment is lost. Possibly, for those weak in stereoscopic ability, translating a richer VE (including a continuous 3D visualization for instance) might introduce an extra delay in interpreting object and target locations from the VE.
In developing virtual tools for rehabilitation training, we need to consider both patient characteristics (e.g., level of stereoscopic ability), and define the functional parameters of the tasks (does the task require depth localization?). When depth cues like stereoscopic visualizations and 3D VE do not add value in terms of functionality (e.g., depth cues are not needed per se when moving in the x-y plane), they may best be removed as we also found evidence that the use of stereoscopy increased visual fatigue for those participants with weak stereovision. If virtual learning environments are applied in patients with severe arm dysfunction, with the aim to aid patients in continued and intensive repetitive training, balancing person traits and task dependent functionality become important factors to consider.
A limitation to our current study is that participants used the stereoscopic mode of presentations for a limited duration only. The effect of stereoscopy on task performance and stamina in training when this system is used for prolonged periods of time is an important question, given that long training programs are needed to obtain functional training effects. It is a question that needs to be addressed in future research. Further, the effects of stereoscopy on task performance could increase when using larger disparities. In the current study we used a low disparity (0.1 degrees), which has been found to have relatively limited effects on performance measures . A higher disparity would likely result in more pronounced effects on our objective performance measures, although it may also lead to a higher level of visual discomfort. Care should be taken that disparities should not exeed the 1 degree comfort limit , although it should be noted these limits have been established in relation to normal, unimpaired viewing populations, and not in relation to neuropsychological patient populations. Based both on our review of the literature and the findings of our study, we have reason to believe that patient populations suffering from functional deficits following brain damage may require a more conservative disparity limit so as to ensure visual comfort.