Mild neurocognitive disorder, or mNCD, is the clinical stage in which cognitive impairment is noticeable but daily independence remains largely intact. It is also the stage at which functional brain changes, detectable with functional magnetic resonance imaging, precede structural atrophy and are considered system-level biomarkers of disease modification. Alterations in the hippocampus, precuneus, and cingulate cortex are hallmark features, and therapies aimed at secondary prevention are expected to normalize, preserve, or compensate for these changes rather than merely slow symptom progression.

The trial, conducted in Switzerland between May 2022 and February 2024, randomized participants with clinically diagnosed mNCD to receive Brain-IT in addition to usual care or usual care alone. The functional imaging sub-study analyzed data from 32 participants with an average age of about 72 years, most of whom had biomarker-supported Alzheimer’s disease etiology, confirmed primarily through cerebrospinal fluid amyloid-beta ratios. The Brain-IT program prescribed at least 120 minutes of training per week, delivered in at least five home sessions, using a pressure-sensitive platform running nine serious games that combined reactive stepping tasks with motor-cognitive challenges and biofeedback-guided resonance breathing designed to modulate the vagus nerve.

What distinguishes Brain-IT from generic exercise programs is its algorithmic personalization. Training focus and game content were tailored to four neurocognitive domains: learning and memory, executive function, complex attention, and visuospatial skills. Exercise intensity was anchored to 40 to 60 percent of each participant’s individual heart rate reserve, and progression through game levels depended on performance precision rather than speed, an incentive structure the researchers say encouraged participants to favor accuracy over haste. A cloud-based system allowed remote telemonitoring of adherence, and between nineteen and twenty-four sessions were supervised in person by an exercise instructor. Over the twelve-week period, the intervention group completed an average of about 71 sessions, totaling roughly 1,689 minutes of training, with no intervention-related serious adverse events.

The imaging results centered on two complementary analyses. During resting-state scans, participants in the Brain-IT group showed a significant reduction in functional connectivity between the hippocampus and the left middle cingulate gyrus, a change that emerged only in the intervention arm and carried a large effect size. During the task scans, which used a validated face-occupation matching paradigm to probe episodic memory encoding and retrieval, the Brain-IT group exhibited a significant decrease in hippocampal activation during memory encoding, with the Group by Time interaction reaching family-wise error corrected significance. The researchers interpret this reduction as a potential normalization of hippocampal hyperactivation, a compensatory state seen in early mNCD that is linked to amyloid accumulation and may herald clinical worsening if left unchecked.

Alongside the reduced hippocampal overdrive, the intervention group showed compensatory activation increases in the right precuneus and middle cingulate gyrus. Critically, these increases correlated with changes in memory retrieval accuracy: participants whose retrieval improved the most displayed the largest activation gains in these regions. The precuneus plays a central role in episodic memory retrieval and attention during movement, and precuneus atrophy and hypoactivation are characteristic of mNCD. The authors argue that both criteria for genuine neural compensation are met here, because the enhanced activation was tied to better cognitive performance and emerged in a region whose resources are diminished in the disorder.

The cingulate findings are more nuanced. The resting-state decrease in hippocampal connectivity with the left middle cingulate gyrus could reflect a reduced reliance on compensatory rerouting of hippocampal signals through executive centers, consistent with theories of aging-related neural scaffolding. Meanwhile, the task-related activation increase in the right middle cingulate may represent a hemispheric shift in compensation, in line with the HAROLD model of aging, possibly reinforced by the training design itself, which rewards precise task execution and appeared to shift participants toward more deliberate response strategies during neuropsychological assessment.

These functional results build on two previous analyses from the same trial. A hypothesis-driven study published in Alzheimer’s & Dementia in 2024 reported clinically meaningful, large-effect improvements in global cognition and verbal recall in the Brain-IT group, and an exploratory structural imaging analysis published in Alzheimer’s Research & Therapy in 2025 documented gray matter improvements that correlated with those cognitive gains. Taken together, the three analyses suggest that two hours per week of tailored exergame training for twelve weeks may exert a brain-protective effect that is visible at both structural and functional levels and causally linked to behavioral improvement, though the authors caution that analyses from a single trial do not constitute independent replication.

The researchers are careful to frame these findings as exploratory. The imaging sub-study was not powered for neuroimaging endpoints, baseline imbalances across imaging markers may have inflated or obscured effect sizes, and the modest sample of 32 participants limits the stability of the estimates. Because usual care was delivered by the recruiting memory clinics, the study also cannot isolate exergame-specific effects from other components of the intervention or demonstrate superiority over conventional exercise. Nonetheless, the design choices, including assessor blinding, prospective registration, biomarker-supported etiological characterization, and harmonized imaging protocols, strengthen the credibility of the signal.

The broader implications are striking. Rigorous secondary prevention could potentially avert or delay up to half of dementia cases, yet exergame therapies remain overlooked in motor-cognitive neurorehabilitation, with no published studies evaluating their long-term implementation in healthcare for mNCD. The authors call for hybrid implementation-effectiveness trials with long-term follow-up measuring biomarker and clinical progression, and suggest that future versions of Brain-IT could incorporate artificial intelligence-assisted, human-in-the-loop algorithms to automate tailoring decisions while preserving clinical oversight. If confirmed in adequately powered studies, the vision of a physician prescribing a personalized video game as medicine for the aging brain moves from intriguing possibility toward clinical reality.

Subject of Research: Functional brain adaptations linked to cognitive changes following personalized exergame training in mild neurocognitive disorder

Article Title: Exergame as medicine? Functional brain adaptations linked to cognitive changes in neurocognitive disorders: exploratory randomized controlled trial

Article References: Manser, P., Rosio, M., Schmidt, A., Michels, L., & de Bruin, E. D. (2026). Exergame as medicine? Functional brain adaptations linked to cognitive changes in neurocognitive disorders: exploratory randomized controlled trial. GeroScience. https://doi.org/10.1007/s11357-026-02540-0

Image Credits: AI Generated

DOI: 10.1007/s11357-026-02540-0

Keywords: exergame, mild neurocognitive disorder, Alzheimer's disease, hippocampus, precuneus, fMRI, functional connectivity, neuroplasticity, dementia prevention, digital health, motor-cognitive training, GeroScience