Structural differences between the left and right cerebral hemispheres—collectively, brain asymmetry—have established clinical and functional relevance. New work summarized here investigated whether variation in global hemispheric asymmetry associates with normal ageing and with age-related neurological conditions. The authors of the primary study aimed to move beyond regional asymmetry measures by modelling broad left–right cortical differences using machine-learning approaches and then testing those models in older and disease-specific cohorts.
Investigators trained models to capture patterns of left–right cortical variation using imaging data from young adults. The trained models were then applied to imaging data from ageing cohorts to detect deviations from the patterns learned in youth. The summary in Nature Reviews Neurology describes this strategy but does not provide technical specifics in the piece: details such as the exact imaging modalities, preprocessing pipelines, neural-network architecture, training procedures, cross-validation methods, or performance metrics were not reported in the summary and must be obtained from the original Sci. Adv. article for full appraisal.
The trained model was evaluated in multiple ageing cohorts that included individuals undergoing healthy ageing as well as participants with age-related neurological diseases. The ageing cohorts included people with clinical diagnoses such as Alzheimer disease (AD) and Parkinson disease. The Nature Reviews Neurology write-up highlights that the model was used to identify changes in hemispheric asymmetry that emerge with ageing and to compare patterns across diagnostic and genetic subgroups.
When applied at the group level to participants with AD, the machine-learning model indicated disease-dependent changes in global brain asymmetry. The summary emphasises that these changes were detectable at the group level, implying consistent directional differences between AD groups and reference or healthy-ageing groups. Specific quantitative results (for example, the magnitude of asymmetry differences, statistical significance, or predictive accuracy for diagnosis) are not reported in the Nature Reviews Neurology summary and require reference to the primary article for detail.
Among older individuals without diagnosed disease, the Nature Reviews Neurology summary reports that hemispheric asymmetry patterns differed by APOE genotype. This observation suggests a potential link between genetic risk for AD and variation in global brain asymmetry even in apparently healthy ageing individuals. The summary does not report the number of participants stratified by APOE genotype, nor does it report effect sizes or whether the genotype–asymmetry relationship remained after adjustment for confounders.
The Nature Reviews Neurology piece notes inclusion of Parkinson disease among the age-related conditions assessed, with the models applied to such cohorts to probe asymmetry differences. The summary does not provide a detailed account of Parkinson disease–specific findings or comparative results across different diseases beyond noting that asymmetry changes differed between groups. For disease-specific outcomes, the reader should consult the original Sci. Adv. article.
The summary stresses that the associations suggested by the machine-learning models require further validation. Important methodological and interpretive limitations remain unresolved in the summary: cohort sizes, demographic composition, imaging acquisition parameters, handling of confounders, replication in independent samples, and mechanistic explanations were not described in the Nature Reviews Neurology account. As the authors and reviewers indicate, additional work is needed to determine whether the detected asymmetry variations represent robust biomarkers, to exclude technical or sampling artefacts, and to identify biological mechanisms linking asymmetry changes to ageing and disease.
If validated, group-level changes in global brain asymmetry could inform understanding of how ageing and genetic risk factors such as APOE relate to brain structure and to vulnerability for neurodegenerative disease. However, the current summary frames the findings as preliminary: translation to clinical practice would require replication, demonstration of individual-level diagnostic or prognostic utility, and mechanistic insights. Future research priorities identified implicitly by the summary include independent replication, detailed methodological transparency, longitudinal analyses to assess change over time, and studies designed to probe biological underpinnings of asymmetry variation.
The Nature Reviews Neurology piece summarises the primary publication: Hiu H. et al., “Variations of global brain asymmetry are associated with aging and related diseases,” Sci. Adv. (2026). For full methodological details, numerical results, and disease-specific analyses, readers should consult that original article.