Gast ::
| Hauptseite > Publikationsdatenbank > Multivariate Statistical Approaches to investigate Sex Differences in Brain and Cognition |
| Hauptseite > Publikationsdatenbank > Multivariate Statistical Approaches to investigate Sex Differences in Brain and Cognition |
| Dissertation / PhD Thesis | FZJ-2025-01510 |
2024
This record in other databases:
Please use a persistent id in citations: doi:10.34734/FZJ-2025-01510
Abstract: Decoding individual variability in cognition and brain organization is essential to enhanceour understanding of heterogeneity in the brain and behavior. Individual variability is oftenrelated to specific demographic phenotypes, with sex being a prominent phenotypecontributing to individual variability. Examining how differences between males andfemales are reflected in cognitive and neuroimaging data advances the understanding of sexdifferences in cognitive processing, brain organization, and the heterogeneity ofneuropsychological and mental diseases. To characterize common sources of variability suchas sex, the present work aims to present multivariate statistical methods as powerful tools toidentify patterns in complex datasets such as neuroimaging or cognitive data (commentary).By using multivariate statistical approaches, the present work examines sex differences inneuropsychological (study 1) and brain imaging data (study 2 & study 3). Specifically, study1 examined sex-specific cognitive profiles derived from a battery of neuropsychological testsusing structural equation modeling. Studies 2 and 3 supplement this investigation byexamining sex-related variability in the functional (study 2) and structural (study 3) brainorganization using machine learning (ML) approaches. Additionally, methodologicalconsiderations in ML were taken into account such as the influence of training samples onthe generalization performance of ML models (study 2) and the influence of confoundingvariables (study 3).The commentary highlighted the importance of new methodological approaches such asmultivariate statistical learning to enhance our understanding of the complex nature of sexdifferences in rich data. Study 1 identified sex-specific cognitive profiles pertaining to sexdifferences in component solutions in cognitive processing strategies. Results of study 2revealed sex differences in the functional brain organization for some, but not all brainregions, with the highest generalization performance when sex classification models weretrained on a large and heterogeneous sample comprising the data of multiple datasets. Study3 demonstrated sex differences in the structural brain organization by accurately classifyingsex with ML models that were debiased for the confounding influence of brain size bymatching males and females for brain size. In sum, the present studies demonstrated thatmultivariate statistical approaches can effectively decode sex-related variability in cognitiveas well as structural and functional brain imaging data while incorporating importantmethodological considerations.
|
The record appears in these collections: |
PDF