The Treatment of Akinesia using Virtual Images

by Jerrold D. Prothero

Future Research

This study has demonstrated the effectiveness of both a laser pointer and the VV Sport display as a method for providing portable, non-tangible visual cues for akinetic patients. A comparison of their relative effectiveness and possible modifications can be found in Subsection 6.4.1. The results are sufficiently encouraging to justify a more extensive study.

The work done so far has demonstrated that non-tangible cues are capable of producing kinesia paradoxa, and has clarified some of the associated issues. A program of study should address at minimum the following questions:

Confounding factors

Having run only two subjects to date, we know little about the variability of the effectiveness of different sorts of cues. Future quantitative studies will have to involve controlling more carefully for visual cues than was possible for the current study. For studies performed in the University of Washington's Biomechanics Laboratory, it should be possible to cover the force plate with a material which is both plain (to provide no visual cues) and conductive (so that foot strike information can be recorded). Either metalized mylar or wire screening over a formica sheet would probably work.

Inter-subject variability

Again because of the small sample size, we know very little about the affect of variables such as progressive state of the disease, age, etc.

Different hardware configurations

As discussed in Chapter 6, a different design for the VV Sport display would probably have been more effective, in particular one with a longer vertical field-of-view. A laser pointer which projected multiple spots might also be advantageous.

Monocular display preference

The question of whether the display should be presented below the dominant or non-dominant eye, and whether the decision is affected by which side is more affected by Parkinson's disease, needs to be investigated.

Visual cue stabilization

The parameters of the visual cues presented in the head-up display need to be examined systematically. It seems clear that the most crucial issue is spatial stabilization, i.e., providing an image which appears to be fixed on the ground. However, we do not know how accurate this stabilization has to be, particularly across subjects. We produced approximate stabilization by choosing the rate at which images moved down the VV Sport display to approximate the walking rate of the subject. It is quite possible that more precise stabilization, driven by a gait tracker, would produce better results.

Visual cue realism

We have not explored to what extent a more realistic appearance (images which look like real bricks, etc.) would improve the effectiveness of a space stabilized cue, nor whether images with an apparent depth would be significantly more effective than flat images. In the T.R. trials we tried horizontal and vertical polygons as well shadowed boxes, without noticing a difference, but did not look at these factors systematically.

Shaping

The extent to which T.R.'s experience of learning new ambulatory skills can be replicated needs to be resolved with future trials.

A practical assistive device must deal with two issues: speed and direction determination; and graphics generation. We envision a product which could be strapped to one's waist and in which speed and direction could be controlled by dials or a joystick. Images could either be generated in real-time by a simple graphics unit, or else played back from a waist-mounted CD-ROM player. The images would be fed into the VV Sport display (or something similar).

Perhaps the most striking aspect of this research was that, quite by accident, we taught T.R. significant (if unfortunately short-lived) new ambulatory skills. The fundamental process may have been that we provided him with visual cues which were steadily less realistic, and as a result he gradually became less dependent on them.

It would seem that underlying this success is the plasticity of the human nervous system: while one cognitive technique for doing some task (in this case walking) may be most natural or obvious, other techniques are also possible. By gradually moving away from the current technique (in this case, relying on tangible visual cues) we are able to develop new techniques.

In general, one can imagine using virtual environments to gradually change a task which a subject can do into one which the subject cannot currently do. In the process, the subject may be able to learn a completely new way of tackling the problem. The potential for facilitating this type of shaping may turn out to be an important contribution of virtual environments to medicine.