This secondary analysis of the FIT4BRAIN randomized controlled trial tested whether a remotely delivered, 8-week multicomponent physical exercise intervention influences white matter microstructure in healthy older adults. The study used neurite orientation dispersion and density imaging (NODDI) indices to probe microstructural changes, focusing on the neurite density index (NDI) and the orientation dispersion index (ODI) across major white matter tracts.
The diffusion MRI sample comprised 66 participants (mean age 66.4 ± 3.6 years; 43 females). Participants were randomized to either the multicomponent exercise or an active control intervention. Group sizes were PAG (physical activity group) n = 34 and CON (control) n = 32. The exercise program was delivered remotely and lasted 8 weeks. Further granular details of the intervention content and control activities were not reported in the source abstract.
NODDI provides biophysically informed metrics of white matter microstructure. The analysis focused on two indices derived from NODDI: the neurite density index (NDI), which reflects intracellular volume fraction associated with axons and dendrites, and the orientation dispersion index (ODI), which quantifies angular variation of neurites. The authors evaluated these metrics across major white matter tracts to determine whether short-term exercise modulates microstructural properties.
Intervention effects on NDI and ODI were examined using linear mixed-effects models and Bayesian multilevel models. Analyses were adjusted for age and sex. The models tested Timepoint × Group interactions to identify differential pre-to-post changes between the exercise and active control arms. Results were reported with p-values for hypothesis tests and were further evaluated within the Bayesian framework across 22 white matter tracts.
A significant Timepoint × Group interaction was observed for NDI (p = 0.003), indicating that the exercise group experienced a greater increase in NDI from pre- to post-intervention compared with the active control. Bayesian analyses corroborated this finding across 22 white matter tracts. The reported effect reflects an increase in the NODDI-derived neurite density index following the 8-week multicomponent exercise program in this cohort of older adults.
No significant Timepoint × Group interaction was observed for ODI (p = 0.785). In other words, the orientation dispersion metric did not show differential change between the exercise and control groups over the intervention period. The source did not report tract-level ODI effects beyond the overall null interaction.
The standardized composite VO2 max score increased from pre- to post-intervention within the exercise (PAG) group. However, the Timepoint × Group interaction for VO2 max did not reach statistical significance (p = 0.079). Across all participants combined, changes in mean NDI from pre- to post-intervention correlated positively with changes in VO2 max, indicating that individuals who improved cardiorespiratory fitness also tended to show greater increases in NDI. The magnitude of this association did not differ between the randomized groups according to the reported analysis.
The reported findings indicate that white matter microstructure, as indexed by NODDI-derived NDI, remains responsive to a short-term, multicomponent physical exercise intervention in healthy older adults. The lack of change in ODI suggests that neurite density may be more sensitive than orientation dispersion to short-term exercise-induced microstructural alterations in this sample.
Limitations and specifics not provided in the source abstract include detailed descriptions of the exercise components, adherence metrics, effect sizes for imaging outcomes, and tract-by-tract quantitative results beyond the statement that Bayesian analyses supported NDI increases across 22 tracts. Those detailed data are not reported in the provided source text.
The trial was registered (DRKS00028022) and received ethical approval from the ethics committee of Jena University Hospital (No. 2021-2345-BO). The authors declared no competing interests. The source indicates that data for the main analyses can be accessed via a provided Zenodo link prior to peer-reviewed publication.