Neuroplasticity at Any Age: What Juggling Teaches Us About the Learning Brain

For most of the twentieth century, the dominant view of the adult brain was that it was largely fixed. You finished developing in your early twenties, and after that you mostly held your ground or lost it. Recovery from injury was the exception that proved the rule.

The juggling literature is one of the cleanest contradictions of that view. It is also one of the more useful, because the experimental design is simple enough to draw clear conclusions from.

First measurable change

after starting practice

50-67

Age range

where effects still appear

~equal

Effect size by age

older adults match younger

Hours

What predicts change

not skill achieved

The “use it or lose it” finding, restated

The Draganski 2004 paper showed that learning the cascade produces grey matter changes in motion-processing regions over three months. The follow-up by Driemeyer and colleagues in 2008 sharpened the timeline: the changes begin within seven days of starting practice. They are not a slow accumulation that finally crosses a threshold. The brain begins restructuring almost immediately.

The reverse is also true. When practice stops, the structural changes regress. The Boyke et al. 2008 study with adults aged 50-67 ran the same protocol Draganski used and showed both sides: grey matter increased over the practice period and partially regressed when practice ended.

This is what “use it or lose it” actually means in the juggling literature. It is not a moral statement about effort. It is a description of a feedback system. The brain maintains structure proportional to current demand. Stop the demand and the structure begins to retreat.

The age finding in full

The Boyke study is worth reading carefully because it is often summarised as “older adults can still juggle” when the more interesting finding is that they show comparable structural changes.

In their cohort, adults aged 50-67 produced grey matter increases in V5/MT and the hippocampus that were not statistically distinguishable from the increases seen in younger adult cohorts on the same protocol. The skill acquisition was slower - older participants took longer to reach the 60-second cascade criterion - but the structural response to the practice was preserved.

A 2022 systematic review by Vetter and colleagues looked across 11 juggling-training neuroimaging studies and found the same pattern repeatedly: structural changes consistently across age groups, including participants over 60.

Process drives change, not expertise

One of the more counterintuitive findings is that the structural changes are not driven by skill achieved. They are driven by time spent practising, regardless of whether the cascade ever stabilises.

Scholz and colleagues (2009) made this explicit. In their six-week white-matter study, the diffusion tensor imaging changes in the right intraparietal sulcus correlated with hours of practice rather than with cascade duration achieved. Participants who never reached a clean cascade showed adaptation comparable to those who did.

This matters because it inverts a common assumption about learning. The folk model is that the brain rewards success - that the structural benefit of practice accumulates when you finally get it right. The juggling data says the opposite. The brain rewards engagement. The struggle is the practice. The drops are part of what produces the change.

The structural benefit of practice accumulates from the attempt, not from the achievement. The brain restructures itself in response to effortful engagement with a new motor problem - regardless of whether the attempt succeeds.

The parallel to technical learning

The same pattern appears in studies of adults learning unfamiliar technical domains. Engineers learning a new cloud architecture, a new programming language, or a new system topology show the steepest learning curves at the start, when the material is most unfamiliar and the error rate is highest.

The pedagogical implication mirrors the juggling finding: the value of early practice is not in producing correct output but in restructuring the substrate that will eventually produce correct output. A team in week two of learning Kubernetes is not yet useful for production work. They are also doing the most important work of their first three months, because the structural learning happens fastest when the material is hardest.

This is hard to honour in practice. The visible signal of progress is competence on the deliverable, not adaptation of the underlying capability. Organisations that judge new-skill teams by their week-two output systematically discount the learning that is actually happening.

What “incremental” means in this context

A theme that runs through the juggling literature is that motor learning happens in small additive units. Not three balls at once. One ball, two balls, then three. The brain restructures around each manageable unit of new demand.

This is structurally identical to how production systems get migrated to new architectures. Not the entire stack at once. One service, then a second service, then a coupled pair. The cognitive load is bounded at each step, and the substrate adapts to handle the next step.

The juggling cascade exists as a learning target precisely because it is the smallest stable juggling pattern. It is the minimal unit that requires the full coordination loop - throw, catch, anticipate, repeat - and once you can run that loop on three balls, the loop scales. Five balls is the same loop, tighter timing. The whole skill ladder is a story of incremental loop adaptation.

The same is true of distributed systems, of organisational change, of any sustained adult learning. The mechanism the juggling literature documents - structural adaptation in response to small, sustained, novel demand - is the mechanism that underlies most adult skill acquisition.

It is just unusually visible in juggling because we can put people in MRI machines before and after.