Can Brain Training Improve Gait in Older Adults?

Computerized brain training may improve how healthy older adults walk when attention is divided. That is the surprising result of an 8-week randomized trial: participants trained cognition only, yet gait was a primary outcome. The finding does not make a puzzle app a replacement for exercise. It does show that walking through the real world—a sidewalk, conversation, traffic, and grocery list included—depends partly on cognitive systems that training can reach.

The more practical question is how to use that overlap without creating a fall hazard. The evidence supports structured cognitive, physical, or combined training. It does not support doing calf raises while brushing your teeth and calling it a protocol.

Walking is partly an executive-function task

Easy walking can feel automatic. Add mental arithmetic, a conversation, or an unexpected obstacle, and the brain must allocate attention, inhibit distractions, and regulate movement. A 2023 fNIRS study by Salzman et al. found that older adults recruit the prefrontal cortex during cognitive-motor dual-task walking, with hemodynamic patterns that differed by cognitive status. Because the study was cross-sectional, it cannot show that prefrontal activation causes better or worse gait. It does make the brain-to-body connection biologically plausible.

That shared circuitry helps explain a single-blind randomized trial by Marusic et al. in 63 cognitively healthy, community-dwelling Slovenian adults. Participants averaged 67.5 years old and had a mean Montreal Cognitive Assessment score of 27.2, consistent with intact cognition. The intervention was computerized cognitive training with CogniFit: 40 minutes per session, three times weekly, for eight weeks. Researchers measured single- and dual-task gait speed as primary outcomes. They also studied EEG event-related potentials in a 45-person subgroup during a lower-limb motor-reaction task.

The trial found that cognitive training transferred to mobility, particularly under attentionally demanding dual-task conditions. The publication information available here does not provide an effect estimate or confidence interval, so the size and precision of the gait change cannot be judged from a single number. Still, the design matters: there was no physical-training component to credit for the mobility result. The EEG work also found training-related changes in motor-relevant neural activity, including N2 and P300 components. Those signals are supportive mechanistic evidence, not proof that the neural changes caused the gait improvement.

Combined programs make more sense as impairment rises

In mild cognitive impairment (MCI), the problem is often broader than divided attention during a walk. Strength, balance, executive function, and everyday function can all decline together. A 2024 parallel-arm randomized trial by Castellote-Caballero et al. enrolled 98 adults older than 65 with MCI, defined in the study as MMSE below 25. Twice each week, the combined-training group completed 45- to 50-minute sessions that paired resistance, balance, and flexibility work with tasks targeting memory, language, executive function, and sustained attention. The comparator received cognitive stimulation alone.

The combined program improved physical and cognitive outcomes, including balance, strength, and executive function, relative to cognitive stimulation alone. That comparison supports adding physical training when the alternative is cognitive work by itself. It does not settle whether simultaneous dual-task practice is superior to completing equally matched cognitive and exercise sessions separately.

Other MCI trials point in the same general direction but come with an important complication. In a 2024 randomized trial by Lau et al., interactive computerized cognitive training delivered with transcranial direct current stimulation (tDCS) improved cognition and gait performance. A 2021 randomized trial by Liao et al. found that tDCS combined with Tai Chi reduced dual-task gait cost and fall risk in older adults with MCI. Neither design can separate the contribution of tDCS from that of the cognitive or movement program. These studies support the possibility of improving both domains, but they do not show that brain stimulation is the ingredient doing the work.

Falls are the endpoint that counts

Gait speed and dual-task cost are useful measures, but falls change lives. A 2026 six-month randomized trial by Rice et al. found that home-based exercise reduced fall risk in older adults with MCI who had sustained a hip fracture. This was a severely vulnerable group, and the outcome was clinically meaningful rather than a laboratory measure. The available abstract summary does not provide the absolute fall rate or effect estimate, so it cannot tell readers how large the reduction was.

Cognitive frailty raises the stakes further. In a 16-week randomized trial, Lai et al. studied 72 Chinese community-dwelling older adults with cognitive frailty. Participants trained three times weekly for 50 to 60 minutes, and the trial assessed frailty status, cognition, physical performance, and dual-task cognitive load. The research is early and the sample is small, but this is one of the few trials designed specifically for cognitive frailty rather than treating it as ordinary healthy aging with a longer intake form.

Two boundaries keep the advice honest

First, training needs a real dose. Across the central trials, structured programs ran for 8 to 16 weeks, two or three times a week, with sessions lasting roughly 40 to 60 minutes. That pattern does not prove a universal minimum dose, but it is the range with direct supportive evidence here. By contrast, a 2021 randomized trial by Granacher et al. found no significant improvement in balance or muscle strength when healthy older adults performed balance exercises while brushing their teeth. Daily-life embedding can be convenient; convenience did not turn it into enough training.

Second, dual-task training requires a motor foundation. A secondary analysis and imaging substudy by Collett et al. of a post-stroke dual-task gait trial found that baseline single-task walking performance was associated with dual-task walking response and with different interference profiles. The study does not establish a gait-speed cutoff, and stroke survivors are not interchangeable with people who have MCI or cognitive frailty. Still, the practical implication is sensible and cautious: when walking alone already demands substantial attention, add strength, balance, and single-task gait work before asking someone to walk while solving problems.

This caution has direct safety relevance. Lai et al. cite a randomized trial in adults older than 70 in which dual-task training produced more falls than multicomponent exercise alone. The studies may differ in participant ability, program design, and comparator, but the signal argues against treating divided-attention drills as harmless enrichment for every older adult.

What the evidence supports now

For cognitively healthy older adults, structured computerized cognitive training has credible randomized evidence for improving mobility under attentional demand after eight weeks. Add it if you enjoy it and can sustain the schedule; keep aerobic, resistance, and balance exercise in the program because cognitive training did not test their broader health benefits.

For MCI, combined physical-cognitive sessions have stronger practical support than cognitive stimulation alone. For people with poor balance, slow walking, prior falls, cognitive frailty, or recent fracture, professional assessment and a staged program are wiser than immediately layering a memory task onto every step. The evidence supports integration, but it also sets limits: dose matters, baseline motor competence matters, and no current trial proves that simultaneous dual-task training is better than well-designed sequential training.