The route mirror feedback takes through the brain.

What question was the study asking?

Everyone agrees mirror therapy produces a compelling illusion. The scientific dispute has been about plumbing: when the illusion helps a stroke-affected limb, which pathway carries the effect to the damaged hemisphere? Two candidate stories existed. In the first, watching the mirrored movement drives the healthy motor cortex, which passes activation directly across to its damaged counterpart, motor cortex to motor cortex. In the second, the mirrored hand is processed as an observed action, engaging parietal regions of the action-observation network, and it is the parietal cortex that recruits the damaged motor cortex.

The distinction is not academic. If the parietal route is the working one, then what matters in therapy is the perceptual believability of the observed action, exactly the thing embodiment-focused VR design tries to maximize, rather than sheer motor drive from the healthy side.

How did the researchers test it?

Fifteen chronic stroke patients were scanned with fMRI while moving their unaffected hand; 14 were analysed after one exclusion for head motion and one for a lesion in sensorimotor cortex. During movement, a virtual reality display showed either veridical feedback, the hand that was truly moving, or mirrored feedback, the movement displayed as though the affected hand were performing it.

The analytic tool, dynamic causal modeling, goes beyond asking which regions light up. It compares competing models of directed influence between regions, here a four-node network of the two motor cortices and the two intraparietal sulci, and asks which wiring diagram best explains the measured data. The team compared 30 candidate models.

The result was unusually decisive for this kind of analysis: the winning model for the mirrored condition carried a posterior probability of 1.0. In it, the significant mirror-driven influence on the damaged hemisphere's motor cortex came from the opposite side's parietal cortex (the rostral intraparietal sulcus), the only connection to reach significance (t(12) = 2.3, p = 0.041). No model won for veridical feedback, and direct motor-cortex-to-motor-cortex influence was not significant in either condition.

What does the finding mean, and what doesn't it mean?

What it means: mirrored visual feedback is not generic visual stimulation. It selectively engaged circuitry for understanding observed action, and through that circuitry, reached the motor cortex that stroke had cut off from normal practice. The fact that truthful feedback produced no comparable organized pattern is the study's sharpest point, the therapeutic ingredient appears to be the illusion itself.

The limits, stated plainly

• Small sample. Fourteen patients, with the authors themselves acknowledging the risk of inflated effect sizes at this n.
• No healthy controls. Without them, one cannot say whether this parietal route is the brain's normal response to mirrored feedback or a post-stroke compensation.
• A four-node model. Dynamic causal modeling only adjudicates among the regions and connections you feed it; a richer network might tell a richer story.
• No clinical outcomes. Nobody's arm function was measured over time. A single-session brain response is a mechanism, not a recovery.

Why does a stroke imaging study matter for chronic pain?

Because the intervention it examined, visual feedback that changes what the brain believes the body is doing, is the shared engine of mirror therapy for phantom limb pain and CRPS, of graded motor imagery, and of embodiment-based VR for chronic pain. Evidence that the mirrored image is processed through action-understanding circuits, and that this processing reaches motor areas the person cannot engage directly, is evidence about the mechanism all of these treatments lean on.

It also quietly supports a design principle: if the parietal action-observation route is the carrier, then the quality of the illusion matters. A first-person virtual body that moves believably, the kind of feedback tested here and built into Karuna's VR training, is aiming at the pathway this study mapped.

Karuna's own small mechanistic study in chronic shoulder pain, described on our mirror therapy page, found the perceptual effect runs both ways, what people believed they were moving changed their pain-free range of motion. Saleh's work suggests where in the brain that kind of effect may be brokered.