Can Mild Cycling Sharpen Executive Function After 68?

Three months of mild stationary cycling improved performance on a Stroop task while prefrontal cortex activation fell in a randomized trial of older adults. That combination is intriguing because better performance usually invites a simple story: the brain worked harder. Here, the authors argue that it worked more economically.

The practical claim needs careful limits. Byun et al.'s 2024 randomized controlled trial supports mild cycling as a plausible way to improve one measure of inhibitory control in a select group of healthy older adults. It does not show that mild exercise outperforms moderate exercise, improves every part of executive function, or reverses brain aging. The imaging result is interesting, but it remains an interpretation rather than proof of a mechanism.

What the cycling trial found

Byun and colleagues randomly assigned 81 cognitively normal, community-dwelling Japanese adults to mild stationary cycling (40 people) or a passive control group asked to continue usual behavior (41 people). Participants had a mean age of 68.2 years (95% confidence interval, 67.0 to 69.4). The exercise group cycled three times a week for three months, with sessions lasting 30 to 50 minutes.

Before and after the intervention, participants completed a color-word matching Stroop task. The exercise group improved its Stroop interference reaction time relative to the passive-control group. Stroop interference tests the ability to suppress an automatic response and select the correct one; it is a useful measure of inhibitory control, but it is not a complete test of executive function. The trial did not report working memory, planning, or cognitive flexibility outcomes as part of this claim.

During the task, the researchers used functional near-infrared spectroscopy, or fNIRS, to measure changes in oxygenated hemoglobin in the prefrontal cortex. After training, the exercise group showed lower prefrontal oxygenated-hemoglobin change alongside better Stroop performance. The authors calculated a neural-efficiency score and interpreted lower activation with better behavior as evidence that participants needed fewer prefrontal resources for the task.

That is a coherent hypothesis. Skilled performance can require less effortful recruitment of control networks than novice performance. But fNIRS measures a hemodynamic signal near the cortical surface, not neurons directly. Lower oxygenated hemoglobin could reflect improved efficiency, but it could also reflect habituation, a less demanding experience of the repeated task, or changes in blood-flow dynamics. Without converging evidence from EEG or fMRI, neural efficiency is the leading explanation, not a settled fact.

The behavioral result fits broader exercise evidence

The narrow cognitive finding is more convincing than the mechanistic one because it agrees with the larger literature. A 2024 systematic review and meta-analysis by Ye et al. found that aerobic exercise improves executive function in healthy middle-aged and older adults, with small-to-moderate effects. Inhibitory-control tasks such as Stroop tests tend to show the most consistent gains.

That context makes the cycling result plausible, but it does not expand the trial's endpoint. A faster Stroop interference reaction time means better performance on that task. It does not establish broad improvements in daily decision-making, multitasking, memory, or independence.

There is also a biologically plausible route by which aerobic exercise could affect the aging brain. A 2025 meta-analysis by Cheng et al. reported that walking, running, and cycling were associated with increases in circulating brain-derived neurotrophic factor, or BDNF, in adults aged 55 and older. BDNF supports neuroplastic processes and could contribute to cognitive adaptation. Still, circulating BDNF is a biomarker, not a readout of prefrontal neural efficiency, and the meta-analysis does not prove that BDNF caused the Stroop improvement in Byun's trial.

Does the result apply only after age 68?

Probably not in the way a headline might imply. Byun et al. reported that statistically significant improvements in Stroop reaction time and the neural-efficiency measure appeared among participants aged 68 and older, but not among those aged 60 to 67. The age cutoff was the sample median, which makes it an exploratory split rather than a pre-specified biological threshold.

The trial's power calculation addressed the main exercise-versus-control comparison, not an age-by-treatment interaction. After dividing 81 participants by age and treatment group, each cell contained roughly 20 people. Such subgroup analyses can produce dramatic-looking differences by chance, especially when the threshold is chosen after examining the sample. The result is worth testing in a larger trial; it is not a reason to change exercise advice at one's 68th birthday.

Mild may help. It has not been tested against harder cycling.

The appealing part of this study is that the intervention was mild. For people who find vigorous exercise impractical or unpleasant, that matters. Yet the trial compared mild cycling with a passive control group, not with moderate or vigorous cycling. It can show that this program performed better than doing no assigned exercise program. It cannot show whether mild cycling is equally effective, more effective, or less effective than a harder aerobic program.

The best available head-to-head evidence in a different population does not settle the question. The 2025 EXERT trial compared moderate-to-high-intensity aerobic exercise with low-intensity stretching, balance, and range-of-motion training in adults with amnestic mild cognitive impairment. It found no significant cognitive advantage for the higher-intensity aerobic program over the active low-intensity program. A related EXERT analysis also found limited separation from usual-care context.

EXERT weakens the tidy idea that more intensity automatically produces more cognitive benefit. But its low-intensity comparison was not mild aerobic cycling, and its participants had amnestic mild cognitive impairment rather than normal cognition. Those differences matter. A stretching-and-balance program may influence cognition differently from low-intensity cycling, and an impaired brain may respond differently from a cognitively normal one.

Other evidence points to dose effects in clinical populations without answering the healthy-aging question. A 2024 meta-analysis by Chen et al. found that exercise parameters, including intensity, modified executive-function outcomes in older adults with mild cognitive impairment or dementia. A 2022 meta-analysis by Li et al. likewise found cognitive benefits from exercise training in adults over 60 with mild cognitive impairment. Neither population directly answers the minimum effective intensity for cognitively normal adults.

Important limits before treating this as brain training

• Attrition was substantial. Of 110 people enrolled, 81 completed the trial: a 26.4% dropout rate. Some participants in both groups could not complete aerobic-fitness testing. Completers may have been fitter or more motivated than the people who left.
• The sample was unusually selective. Participants were right-handed Japanese adults without cardiovascular disease, diabetes, or psychiatric history. Many older adults considering an exercise program have one or more of those conditions.
• The cognitive outcome was narrow. The Stroop task is particularly responsive to exercise interventions. A single favorable test does not establish a general improvement across executive function.
• The mechanism remains indirect. fNIRS cannot distinguish neural efficiency from several other explanations for lower prefrontal hemodynamic activation.

What to do with this result

For a healthy adult in their late 60s or older, regular mild cycling is a reasonable, low-risk option for supporting inhibitory-control performance, provided it fits their health status and medical guidance. The evidence is strongest for the direction of the behavioral result: a three-month mild-cycling program improved Stroop interference reaction time in one randomized trial, consistent with broader aerobic-exercise evidence.

Confidence should drop as the claim becomes more specific. The evidence for a true prefrontal neural-efficiency mechanism is low. The evidence for a special benefit beginning at age 68 is insufficient. And no study described here shows that mild cycling is the optimal intensity. A larger, more diverse trial that directly compares mild and moderate cycling and measures several cognitive domains could answer those questions. This study cannot.

The sensible reading: mild cycling may be enough to improve a Stroop-type measure of inhibitory control. It is not evidence that harder exercise is unnecessary, or that a lower fNIRS signal has decoded the secret economy of the aging brain.