Dual-Task Training Improves Gait in Older Adults

Adding a real-time mental task to physical training can improve gait speed and functional mobility in healthy older adults—but only when both tasks happen at the same time. A 2023 meta-analysis of 22 randomized trials found that simultaneous exercise-cognitive combined training increased gait speed by 0.11 m/s (95% CI, 0.07 to 0.15), a change the authors considered clinically meaningful. Programs that put the cognitive task before or after exercise did not significantly improve gait speed.

That distinction changes the practical advice. Solving arithmetic problems while balancing on an unstable surface trains a different demand than doing balance exercises, resting, and then completing a puzzle. The first forces the brain to allocate attention while the body must stay upright and move accurately. The second may be worthwhile exercise, but it does not reproduce the training condition linked to better gait outcomes.

Why timing changes the training stimulus

In Li et al.'s systematic review and meta-analysis, the population was cognitively healthy older adults, and the evidence came from 22 randomized controlled trials involving 1,091 participants. Simultaneous exercise-cognitive training improved gait speed by 0.11 m/s, balance by a standardized mean difference of 0.40, and functional mobility by 0.85 seconds. For functional mobility, lower time is better.

The sequential protocols did not produce a statistically significant improvement in gait speed. This does not prove that every sequential activity is useless or that every simultaneous program will work. It does show that the trials did not find equivalent gait benefits when the cognitive and physical tasks were separated in time.

Researchers call this dual-task training because the participant must manage a motor task and a cognitive task together. During normal life, that might mean walking while planning dinner, watching for a curb while talking, or avoiding an obstacle while carrying groceries. Training the two demands separately may miss the bottleneck: attention has to be allocated while movement is already underway.

A tested 24-week version of the idea

A small randomized trial gives the meta-analysis a more concrete shape. In Oliveira et al.'s 2021 trial, 50 community-dwelling healthy older adults trained three times per week for 24 weeks. One group performed unstable-surface training with simultaneous cognitive tasks; the comparison group completed unstable-surface training alone. The physical work used challenging surfaces such as balance boards and foam pads.

The combined-training group improved dual-task Timed Up and Go performance by 4.80 seconds versus its comparator outcome (95% CI, 0.95 to 8.65 seconds faster). The exercise-only group did not achieve equivalent dual-task gains. Timed Up and Go measures the time needed to rise from a chair, walk, turn, return, and sit; adding a cognitive task makes the test closer to the divided-attention conditions that often expose mobility problems.

The result is encouraging, not a final protocol prescription. The trial had only 50 participants, and its confidence interval was wide. The lower bound barely excludes no effect. It also does not tell us whether verbal fluency, arithmetic, Stroop-style inhibition, or visuospatial tasks work best, nor does it identify the shortest effective program. Still, it establishes a replicable example: challenging balance work plus an active cognitive task, performed concurrently and sustained for months.

Passive screen time during treadmill walking does not necessarily create the same cognitive competition. The cognitive task has to require attention while the movement task is happening.

The prefrontal explanation is plausible, but incomplete

The dorsolateral prefrontal cortex helps support working memory, inhibition, and cognitive flexibility—functions that become especially relevant when walking stops being automatic. In a 2023 controlled experiment by Matsuura et al., anodal transcranial direct-current stimulation of this region improved performance on a dual-task obstacle-avoidance gait task. Because the study altered prefrontal activity rather than merely measuring it, it supports a causal contribution of the dorsolateral prefrontal cortex to cognitive-motor integration.

It does not prove that this brain region alone explains training benefits. Other cortical and motor-control systems also contribute to gait. A 2025 review by Wollesen et al. places combined cognitive-motor exercise within the broader executive-function literature, covering inhibition, working memory, and cognitive flexibility. The mechanism fits the training result, but no trial in this evidence base directly compared simultaneous and sequential training while measuring prefrontal adaptation over time.

More practice may help, but the dose is still unknown

The strongest practical dose evidence is the 24-week, three-sessions-per-week trial. At the other end, low-volume incidental balance practice has failed. In Granacher et al.'s 2021 trial, approximately two minutes per day of balance exercises during tooth brushing did not improve balance or muscle strength in healthy older adults. Casual exposure is not a substitute for progressive training.

Two secondary analyses by Tait and colleagues extend the timeline without defining a dose-response curve. A 6-month cluster-randomized trial analysis examined cognitive outcomes along with neurological and inflammatory markers after dual-task, power-based exercise. An 18-month cluster-randomized trial analysis assessed whether dual-task high-velocity exercise improved cognition over a longer period. Together, they are consistent with sustained or accumulating benefits, but they used different designs and populations and were not dose-finding trials. They cannot tell readers when benefits plateau, whether more training always helps, or what minimum dose applies outside the 24-week trial.

Who should be cautious

Older adults generally experience greater dual-task interference during functional mobility than younger and middle-aged adults, according to Tolotti et al.'s cross-sectional study. That makes dual-task capacity relevant to aging, but higher baseline interference does not guarantee a larger response to training.

Individual response can go in the other direction. In a 2022 randomized study by Mack et al., some older adults retained high multitasking costs despite combined training, particularly participants with low baseline executive function. No validated screening test can yet identify likely nonresponders before training starts.

A stroke-rehabilitation trial offers a related warning, although its participants were not healthy older adults. In Collett et al.'s secondary analysis and imaging substudy, dual-task walking training was not superior to single-task training after stroke. Participants with the greatest baseline prefrontal engagement during walking benefited least from the dual-task approach. The population prevents a direct extrapolation, but the finding supports a sensible concern: when walking already consumes most available attention, adding a mental task can create interference instead of useful challenge.

People with vestibular dysfunction, early cognitive symptoms, neurological disease, or a high fall risk should not turn a balance drill into a self-administered obstacle course. Start with a stable physical task and a low cognitive load, progress only if performance remains safe, and seek clinician or qualified exercise-professional guidance when falls or neurological symptoms are part of the picture. Controlled safety data for high-attentional-load groups remain limited.

What the evidence supports

For healthy older adults, the best-supported approach is concurrent training: perform a physically demanding balance or mobility exercise while completing an active cognitive task. The evidence supports better gait speed, balance, and functional mobility at the group level; it does not establish the best cognitive task, an individualized dose, or guaranteed benefit for every person.

The cleanest starting template is the one tested by Oliveira and colleagues: three supervised or carefully structured sessions per week for 24 weeks, using progressive unstable-surface or balance work with simultaneous mental arithmetic, category-word generation, or another task that requires sustained attention. Keep the activity safe enough to perform well, but difficult enough that the brain cannot put the movement on autopilot. Wearables may help organize practice and track adherence, yet no randomized trial here shows that wearable-delivered training matches supervised programs.

Dual-task training is not a shortcut around ordinary exercise. It is a specific way to train the attention-and-movement problem that ordinary life keeps presenting.