Combined Exercise May Support Brain Health in Type 2 Diabetes

Combined aerobic and resistance exercise may do more for adults with type 2 diabetes than improve blood sugar and fitness. A 2024 meta-analysis found a modest but statistically significant improvement in global cognition: standardized mean difference (SMD) 0.34, with a 95% confidence interval of 0.13 to 0.55. That is a real signal, but it does not tell us whether memory, executive function, or processing speed improved—or whether the benefit lasts once training stops.

The metabolic case is more settled. The same exercise approach improved HbA1c, aerobic capacity, physical function, and health-related quality of life in randomized trials. Think of the cognitive finding as a probable co-benefit of an exercise prescription that already has strong reasons to exist, not as proof that a particular workout prevents dementia.

A modest cognitive gain in a larger metabolic package

Zhang et al.'s 2024 systematic review and meta-analysis pooled 16 randomized controlled trials involving 2,426 middle-aged and older adults with type 2 diabetes. Combined aerobic-plus-resistance training improved global cognition versus the comparison conditions, with an SMD of 0.34 (95% CI, 0.13–0.55). The confidence interval excludes zero, so the pooled result was statistically significant.

It also found lower HbA1c (SMD −0.35), higher VO2max (SMD 0.58), and significant improvements in physical function and health-related quality of life. Those outcomes matter on their own. Better aerobic capacity can make daily activity easier; lower HbA1c is a standard clinical target in diabetes care. The cognitive result adds interest, but it is not the most mature part of the evidence.

An SMD of 0.34 generally falls in the small-to-moderate range. It does not translate cleanly into points on a memory test because the trials used different cognitive instruments. More importantly, a statistically significant pooled effect does not establish a clinically meaningful improvement for every participant.

Why the biology makes sense—and where it stops short

Type 2 diabetes often travels with insulin resistance, chronic inflammation, and vascular risk. Each could plausibly contribute to poorer brain health over time. Exercise could affect several of those pathways at once, but plausibility is not proof of a brain-specific mechanism.

A separate 2024 meta-analysis by Zaki et al. examined six randomized trials of concurrent exercise training in people with type 2 diabetes. It reported reductions in tumor necrosis factor alpha (TNF-α; p<0.01) and HOMA-IR, a measure of insulin resistance (p=0.02). The review also found improvements in glucose, HbA1c, body composition, triglycerides, waist circumference, and peak VO2. These are useful metabolic and inflammatory findings, but TNF-α and HOMA-IR are biomarkers and surrogate measures. They do not show that either change caused the cognitive result.

Other studies make the hypothesis more credible without closing that gap. A 2026 systematic review of exercise and arterial stiffness found that aerobic and resistance training can reduce arterial stiffness in older adults, a vascular adaptation consistent with better cerebrovascular health. A 2025 network meta-analysis in people with depression found that aerobic exercise can raise brain-derived neurotrophic factor, or BDNF. Neither analysis studied type 2 diabetes-specific brain outcomes. They support a reasonable hypothesis, not a demonstrated chain from workout to biomarker to preserved cognition in diabetes.

The dose finding is a useful clue, not a prescription carved in stone

Zhang et al. identified stronger cognitive outcomes in a responder subgroup at at least 135 minutes of exercise per week, including at least 60 minutes of resistance training. That sounds usefully specific, but the analysis did not use formal dose-response meta-regression. It was a descriptive subgroup finding, so it cannot establish that 135 minutes is a biological threshold or the best target for every person.

For contrast, the 2025 Lancet Public Health step-count meta-analysis characterized dose-response associations for mortality and cardiovascular outcomes across large adult populations. It did not address cognitive outcomes in type 2 diabetes. The comparison illustrates the difference between a well-developed dose-response literature and an early signal that needs replication.

That does not make the 135-minute figure useless. It is a sensible starting point to discuss with a clinician or qualified exercise professional, especially because resistance work has benefits that walking alone may not provide. It simply should not be sold as a precision brain-health dose.

"Combined exercise" is a broad label

The pooled cognitive estimate hides an important problem: the 16 trials used heterogeneous training protocols and cognitive tests. The meta-analysis reported global cognition, not separate effects on memory, executive function, or processing speed. The lone type 2 diabetes-specific randomized trial with domain-level cognitive data, Ghodrati et al. (2023), directionally supported benefit from combined training in women with type 2 diabetes, but its quality score was low. It cannot reliably settle which domains improve.

This concern is not academic nitpicking. A 2025 Fit4Alz study in older adults with existing cognitive decline found that strength, aerobic, and cognitive-training combinations produced variable results across cognitive domains. Its population differed from the diabetes trials, but it shows why a pooled global score may average together protocols that work differently.

Likewise, a systematic review in adults over 60 with mild cognitive impairment found that exercise effects varied by cognitive domain and intervention type. And a 2025 review in healthy adults aged 60 and older found that combined training did not consistently outperform single-modality exercise for VO2max. No head-to-head type 2 diabetes trial has established that combined training is uniquely superior to aerobic-only or resistance-only exercise for cognition.

Combined training has strong support for metabolic management in type 2 diabetes. Its cognitive benefit is promising, statistically significant in one meta-analysis, and still too nonspecific to promise a better memory or a lower dementia risk.

What to do with this evidence

For a person managing type 2 diabetes, the practical conclusion is straightforward. Combined aerobic and resistance exercise already has evidence for improved glycemic control, insulin sensitivity, inflammation-related biomarkers, aerobic fitness, physical function, and quality of life. The cognitive finding gives one more reason to make training consistent.

Keep expectations proportionate. The available trials do not show which cognitive skills improve, whether benefits persist after training ends, or whether one exact mix of aerobic and resistance work beats another. They also do not prove that reduced inflammation or insulin resistance protects the brain.

Exercise is still a remarkably efficient investment here: the metabolic benefits justify it, while the possible cognitive gain is a credible bonus worth testing in better, longer trials.