% IMPORTANT: The following is UTF-8 encoded.  This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.

@INPROCEEDINGS{Zhang:1048931,
      author       = {Zhang, Mingxian and Bittner, Nora and Mendl-Heinisch,
                      Camilla and Miller, Tatiana and Moebus, Susanne and Dragano,
                      Nico and Caspers, Svenja},
      title        = {{L}everaging lifestyle clusters and multimodal brain
                      features to enhance cognitive prediction models in healthy
                      older adults},
      reportid     = {FZJ-2025-05028},
      year         = {2025},
      abstract     = {Developing neuroimaging markers for normal cognitive aging
                      is challenging due to variability in the brain and behavior
                      among older adults, complicating the identification of
                      predictors. However, modifiable lifestyle factors may help
                      link underlying group differences in brain structure and
                      function. Examining brain differences across distinct
                      lifestyle groups may better identify informative features
                      for predicting cognitive performance rather than relying
                      solely on data-driven methods. This study explored whether
                      lifestyle-related brain features could predict cognitive
                      function in healthy older adults at baseline and after ~4
                      years, using multimodal MRI data from 563 participants of
                      the 1000BRAINS cohort. We performed KModes clustering
                      analysis on eight lifestyle factors to identify four
                      distinct lifestyle groups and conducted univariate analyses
                      to find significant between-group brain differences. These
                      differences were used in a lifestyle-related model for
                      machine learning, compared to data-driven models for
                      predicting 13 cognitive tests. The lifestyle-related brain
                      model better predicted visual and episodic memory than
                      data-driven models but showed limited generalization.
                      Correlations between predicted and actual cognitive scores
                      were significant at both baseline and follow-up. This study
                      highlights the potential for integrating lifestyle
                      information as a form of feature selection to help improve
                      predictive models of cognitive performance during aging,
                      pending further external validation.},
      month         = {May},
      date          = {2025-05-07},
      organization  = {Aging and Cognition Conference, Pavia
                       (Italy), 7 May 2025 - 10 May 2025},
      cin          = {INM-1},
      cid          = {I:(DE-Juel1)INM-1-20090406},
      pnm          = {5251 - Multilevel Brain Organization and Variability
                      (POF4-525) / HBP SGA3 - Human Brain Project Specific Grant
                      Agreement 3 (945539)},
      pid          = {G:(DE-HGF)POF4-5251 / G:(EU-Grant)945539},
      typ          = {PUB:(DE-HGF)1},
      url          = {https://juser.fz-juelich.de/record/1048931},
}