Tracking in VR rehabilitation: from the pool to the clinic

Opening

A woman stands in a quiet clinic room, wearing a headset. To anyone watching, she's simply walking slowly on the spot, turning her head now and then. But in her eyes, she's in a supermarket. Shelves stretch away on either side. Somewhere ahead is the item on her list. A trolley rattles past. A voice comes over the tannoy.

She pauses at the end of an aisle, looks left, then right, and chooses. Every step, every turn of the head, every hesitation is being recorded. Not to catch her out, but to measure how she's recovering.

It's a long way from a rat in a pool of cloudy water. And yet the idea underneath is surprisingly similar.

Where tracking went next

This is our final episode in this first series, and it's about where tracking behaviour has gone next. From animal mazes, to virtual mazes for people, to virtual reality used in rehabilitation after stroke and other conditions. What can virtual reality really do for recovery? What does the evidence say? And what does tracking add?

From the pool to the clinic

Let's start with the thread that runs through this whole series. In 1984, Richard Morris described an early version of video tracking for his water maze, following a hooded rat on camera. From then on, a computer could record where an animal was, moment by moment, and turn that into measures of learning and memory.

In episode three we heard how that idea crossed over to people, with the virtual water maze. Instead of swimming, a person moves through a computer-generated pool and searches for a hidden platform. The same logic applies. Where did they go, how directly, and how close did they search?

Rehabilitation takes the next step. If you can track a person's path through a virtual space, you can also track how they reach, balance, turn, and find their way. And if the space is virtual, you can control it completely. Make it calm or busy, quiet or noisy, simple or full of distractions. Then repeat it exactly, session after session, to see whether things are improving.

Virtual reality in rehabilitation

Virtual reality in rehabilitation is a broad field. It ranges from commercial gaming consoles used for arm exercises, to custom-built systems in which patients walk through realistic environments. The common idea is to make practice more engaging, more varied and more measurable, while keeping people safe.

Stroke is where the evidence is strongest, thanks to a series of Cochrane reviews led by Kate Laver at Flinders University in Adelaide. Cochrane reviews are systematic summaries of randomised trials, and they're among the most trusted sources of evidence in medicine.

The review published in 2017 included seventy-two trials with two thousand four hundred and seventy participants. Its findings were cautious. Compared with conventional therapy, virtual reality was not significantly better for arm function. But when virtual reality was added on top of usual care, giving patients more therapy time overall, there was a significant improvement. There was also a benefit for activities of daily living. The evidence was mostly low quality, because many trials were small and poorly reported. Adverse events were few and relatively mild. The reviewers also spotted a trend. Higher doses, more than fifteen hours of total intervention, seemed preferable, and so did virtual reality programmes built specifically for rehabilitation, rather than off-the-shelf games.

The 2025 update

In 2025, the same team published a major update. It included one hundred and ninety trials, with seven thousand one hundred and eighty-eight participants. One hundred and nineteen of those trials were new. That tells you how quickly the field is growing.

The picture shifted a little. Compared with alternative therapy, virtual reality may slightly improve arm function and activity. It may also slightly improve balance, and probably reduces limitations in daily activities, with moderate-certainty evidence for that last point. There may be little or no effect on walking speed, participation, or quality of life.

When virtual reality was added to usual care, so that people spent more time in therapy, it probably improved arm function and activity, again with moderate certainty. And it may help balance.

The update also noted that most studies were still small. Only about one in five involved more than fifty participants, and the largest recruited one hundred and fifty-two. Fifty-nine studies monitored for side effects. A small number of people reported pain, headaches, or feeling faint or dizzy, and no serious unwanted effects were reported. In plain terms, the reviewers suggested that virtual reality may be most beneficial when people spend many hours using it.

So the honest summary is this. Virtual reality isn't a miracle, but it's a reasonable tool, especially as a way to increase the amount of practice people get.

Screens, headsets and immersion

There's also a question about what kind of virtual reality we mean. Much early research used a screen and a joystick or keyboard. You sit still, and the world moves past. That's useful, but it leaves out something important. When we move through real space, our bodies tell us about it. The balance organs in the inner ear sense turning and acceleration, and our muscles and joints sense each step.

Immersive systems try to put some of that back. With a headset, turning your head turns your view, just as it would in real life. Add a treadmill or a pressure-sensitive plate underfoot, and walking can drive movement through the virtual world. For navigation, that matters, because people use those body-based signals to keep track of where they are. It's one reason researchers are interested in moving from desktop tasks to more immersive ones, both for studying memory and for rehabilitation.

Where tracking fits

Where does tracking fit in? In many rehabilitation systems, measurement is what turns an exercise into an assessment. If the system records exactly how someone moves, it can show progress that's hard to judge by eye. How straight was that walk? How far did they lean before losing balance? How long did they hesitate at a junction? Did they follow a sequence of instructions in the right order?

There's a natural parallel with the water maze here. Way-finding, finding your way to a goal through a space, is exactly what the maze measures in animals. In people recovering from stroke or brain injury, getting lost in familiar places can be a real problem. A virtual environment can test way-finding safely, and measure it in the same kinds of ways, path length, directness, and errors.

One example of this approach is HVS 6D VR. According to its makers, it's fully immersive, and lets clinicians and researchers set tasks such as way-finding, walking in a straight line with or without a cognitive task, balance exercises, and a virtual version of mirror therapy. In mirror therapy, a patient moves one hand and sees both hands moving, which is used in stroke rehabilitation. The same system can also run virtual water maze tasks for research. Like every system in this field, its value in any particular clinical use is something that trials need to show.

The limits

And what are the limits? The first is the evidence itself. Even after nearly two hundred trials, most are small, and many are poorly reported. The systems used vary enormously, from games to fully immersive headsets, which makes it hard to say which features matter.

Standardisation is a problem in research too. A 2021 review by Thornberry, Cimadevilla and Commins looked at more than two hundred papers using virtual navigation tasks in humans. The virtual water maze was the most popular task. But set-ups varied widely from study to study, and the authors called for more standard methods. Without them, results from different labs are hard to compare.

There are practical issues too. Some people feel dizzy or unwell in virtual reality. Headsets need to fit, and some patients may find the technology confusing at first. Clinicians need training. And a good result in a virtual supermarket doesn't automatically mean a good result in a real one. Showing that skills transfer to everyday life is one of the field's biggest challenges.

Finally, measurement is only useful if the measures mean something. Just as in the water maze, a fast time can hide a poor strategy. A patient might complete a task quickly by a route that wouldn't be safe in real life. Good tracking needs good interpretation. Numbers on a screen are a starting point for a clinician's judgement, not a replacement for it.

Back to the clinic

Back to that clinic room, and the woman choosing an aisle in a supermarket that isn't there. Every hesitation is recorded, just as every turn of a rat in a pool was recorded four decades ago. The technology has changed beyond recognition, but the question hasn't. Where are you going, how do you get there, and what does that tell us about the brain? That's been the story of this whole series, and we hope you've enjoyed it.

You'll find the papers we mentioned, and a transcript, at water-maze.com.