Poster No:
868
Submission Type:
Abstract Submission
Authors:
Natalia Vilor-Tejedor1, Patricia Genius1, Blanca Rodríguez-Fernández1, Tavia Evans2, Carolina Minguillon1, Manel Esteller3, Arcadi Navarro1, Hieab Adams4, Juan Domingo Gispert1
Institutions:
1BarcelonaBeta Brain Research Center, Barcelona, Catalonia, 2Erasmus University Medical Center, Rotterdam, Zuid Holland, 3Josep Carreras Leukaemia Research Institute, Barcelona, Catalonia, 4Department of Genetics, Radboud University, Nimejgen, Holland
First Author:
Co-Author(s):
Tavia Evans
Erasmus University Medical Center
Rotterdam, Zuid Holland
Manel Esteller
Josep Carreras Leukaemia Research Institute
Barcelona, Catalonia
Arcadi Navarro
BarcelonaBeta Brain Research Center
Barcelona, Catalonia
Hieab Adams
Department of Genetics, Radboud University
Nimejgen, Holland
Introduction:
Brain imaging genetic studies investigate genetic influences on brain structure and function by integrating neuroimaging-based features and genetic data. While many studies focus on individual genetic correlations with single brain measurements, the field emphasizes the necessity of multivariate methods. We conducted a detailed analysis of how genetic predisposition to Alzheimer's disease (AD) influences multivariate hippocampal subfields modeling. We compared the results with those from univariate analyses underscoring the significance of multivariate methodologies in unraveling genetic influences of hippocampal structures.
Methods:
A total of 1,411 cognitively unimpaired participants from the ALFA study (55.8yo in average; 62.4% women), with available information on both genetics and magnetic resonance imaging were included. Volumes for 8 hippocampal subfields were quantified from ultra-high resolution dual-echo Inversion-Recovery MRI scans covering the hippocampal formation and acquired in the direction of the planum temporale with the following parameters: Voxel Size = 0.4 x 0.4 x 2.0 mm; field of view: 230 x 184 x 78 mm; TR: 3000/8000 ms; TE = 26 ms; Refocusing Angle = 120º using Automatic Segmentation of Hippocampal Subfields (ASHS). Briefly, ASHS involves deformable registration of the T1- and T2-weighted images, multi-atlas joint label fusion, and voxelwise learning-based error correction. This process is employed to transfer anatomical labels from a collection of manually labeled training images to an unlabeled imageGenetic predisposition to AD was estimated by calculating polygenic risk scores using PRSice version 2. Canonical Correlation Analysis (CCA) and Compositional Data Analysis (CODA) were applied for identifying joint patterns of variation in hippocampal subfields volumetric changes influenced by genetic predisposition to AD. CCA enables the identification of hippocampal subfields simultaneously associated with a higher genetic predisposition to AD. CODA identified an optimal multivariate hippocampal structural signature involving the joint modulation of hippocampal subfields associated with higher genetic predisposition to AD.
Results:
CCA revealed that higher volumes of CA1, GC-ML-DG, and CA3 were simultaneously associated with higher genetic predisposition to AD. Applying CODA, the optimal hippocampal signature associated with higher genetic predisposition to AD was defined by the joint modulation of CA3, CA1 as well as CA4, hippocampal fissure, and hippocampal tail. Finally, when evaluating univariate effects no significant results were found after multiple comparison correction.
Conclusions:
The study underscores the importance of employing multivariate methodologies in investigating genetic influences on hippocampal structures, particularly in the context of AD. Overall, our study contributes valuable insights to the field of brain imaging genetics, emphasizing the better performance of multivariate analyses in unraveling the complex interplay between genetics and brain hippocampal volumes in the context of AD.
Genetics:
Genetic Modeling and Analysis Methods 1
Modeling and Analysis Methods:
Multivariate Approaches 2
Keywords:
Data analysis
Modeling
MRI
Statistical Methods
STRUCTURAL MRI
1|2Indicates the priority used for review
Provide references using author date format
NA