If you've been told graded motor imagery might help your CRPS, or you've tried it before and it didn't feel right, this is for you. I'll explain what it is, what it's for, and why I think of it as principles rather than a fixed set of steps.
Graded motor imagery is one of the more established approaches for CRPS, and it's one I know from the inside. Alongside Lorimer Moseley and David Butler, I co-authored the original Graded Motor Imagery Handbook, and I've taught it to clinicians across the world for many years since.
My own thinking on it has moved on since then, and I want to explain how.
(If you haven't read the piece before this one, CRPS: When a Limb Stops Feeling Like Part of Your World sets up why this matters.)
From cortical retraining to person-world reconnection
When I first wrote about this work, the language I reached for was largely about the brain: body schema, cortical representations, prediction errors triggering a "saliency network." That framing isn't wrong exactly, but it puts the action entirely inside someone's skull, as though CRPS were a software fault to be patched by retraining a map.
In plain terms, the explanation was almost all about what the brain was doing, and very little about you, living in a world you're trying to act in.
The fuller picture was there all along. Ronald Melzack's neuromatrix model, described in his 1999 paper "From the gate to the neuromatrix", proposed that pain is the output of a widely distributed network he called the body-self neuromatrix, producing the sensory, affective and evaluative dimensions of pain together. When that was shortened into a teachable rationale for GMI, it became graded cortical activation, and in my experience the body-self part often got left behind. GMI 2.0 is, in part, an attempt to put it back.
What I now think GMI is actually doing is something closer to this:
Safely reopening a person's relationship with a limb that's stopped feeling like a trustworthy part of their world.
That's a person-world problem, not just a brain problem, and it changes how I use the tools.
The three tools, briefly
GMI is built from three tools, usually introduced in this order, though not rigidly:
- Left/right judgement. You're shown images of a limb and have to quickly decide if it's a left or a right. It sounds simple, but it requires you to mentally position your own limb to match the image you're looking at, without actually moving. For many people with CRPS, this becomes noticeably slower or less accurate on the affected side.
- Explicit imagined movement. Imagining moving the affected limb through specific tasks, without physically doing so yet.
- Mirror therapy. Watching the reflection of the unaffected limb in place of the affected one, so the brain and body receive visual evidence of normal, pain-free movement.
Each tool asks a person to re-engage with the affected limb a little more directly than the last, at a pace that stays inside what currently feels safe.
Context matters more than it might seem. Someone struggling to picture their hand moving against a blank, clinical background might find it far easier to imagine reaching for something that actually means something to them, a mug of tea, a dog's lead, the strings of a guitar. The task is the same.
What changes is whether it feels like theirs.
Principles, not a protocol
The original protocol was built as a six-week programme, two weeks on each stage: left/right judgement, then imagery, then mirror work. It already included a form of grading within each stage too, tasks were sequenced from whatever the person expected to hurt least through to whatever they expected to hurt most. Both of those ideas, a defined timeline and grading tasks in order, are still genuinely useful, and I still use them.
What I've never really taught is the idea that either of those has to be rigid. What I felt was missing, even early on, was room: room to grow into the work at a person's own pace, time for changes to actually bed in, and the wriggle room to go sideways or even backwards when that's what someone needs, rather than pressing on with the next stage because the plan says so.
There's also more than one thing worth grading tasks by. Expected pain is one useful axis, but it's rarely the only one that matters. Tasks can just as usefully be sequenced by how safe they feel, how comfortable, or how believable, alongside whatever else turns out to be significant for that particular person in that particular context.
What actually seems to matter is a set of underlying principles that the three tools exist to serve:
- Safety. Nothing here should feel threatening. If it does, it's being pushed too hard.
- Believability. Not just believing in the process, but believing you are genuinely the one doing the acting, and that the limb attempting it is genuinely yours. Without that felt believability, the buy-in that actually matters isn't there.
- Agency. More than being an active participant. It's a felt, knowing will to act, a sense that you are the one directing what happens next, reinforced each time visual or sensory feedback confirms that the action actually succeeded.
- Curiosity. Approaching the limb with interest rather than dread changes the experience of the exercise itself.
- Meaning. Tasks that connect to something the person actually cares about tend to work better than generic ones.
- Embodiment. The goal is a felt, lived reconnection with the limb, not just a cognitive exercise.
- Prediction updating. Each attempt is a small test: what did you expect to happen, and what actually happened. The gap between the two is where change tends to come from.
- Graded exploration. Small, safe steps outward, not big leaps.
- Ownership. Rebuilding the sense that the limb genuinely belongs to the person, not just that it's attached to them.
- Context. Where, when, and how a task is done can matter as much as the task itself.
Different people need different amounts of each tool, in a different order, at a different pace. The principles stay constant. The protocol doesn't have to.
Sometimes what shifts things isn't the exercise itself but something small around it. A compression sleeve, a different fabric resting against the skin, holding the limb in a slightly different position, these can sometimes make left/right judgement noticeably easier, simply by changing how present the limb feels in that moment.
Small, unglamorous adjustments like these are often where the real movement happens.
What this looks like in practice
One of my patients, whose story I've written up in more detail elsewhere, used graded motor imagery as part of the "reconnect" phase of her recovery from CRPS following a knee injury, alongside sensory discrimination work and whole-body awareness training. The aim wasn't to desensitise the leg, but to genuinely reconnect with it. You can read the full case study here: CRPS Several Months On: Work, Sleep and Returning to Sport.
Where this fits, and where it doesn't
Graded motor imagery won't be the whole answer for everyone, and it isn't meant to replace individualised clinical reasoning. It's one tool among several for reopening what has closed down around a limb, what I’ve called a field of affordances: the range of things that feel possible and doable in life. Used well, and used with the person rather than at them, it can be a genuinely powerful part of that process.
If you're trying any of this on your own and it stirs things up, take that as a sign to slow down rather than push on, and it's worth talking it through with someone who knows the approach.
Next, I want to look at what reopening that wider field actually looks like once someone's ready to start.
Working with me
If you're living with CRPS and this sounds like it might fit, I'd be glad to talk it through. I see a small number of people in person in Tunbridge Wells, and work online with people further afield. The first step is a free 15-minute discovery call, where we can look at where you are and whether I can help. You can book a call, or email enquiries@timbeames.com.
If you're a clinician with someone who isn't moving forward, I'm happy to help there too. That might be talking a case through together, or working alongside you and your patient. Get in touch at enquiries@timbeames.com.