Lateral ventricle volume trajectories predict response inhibition in older age - a longitudinal brain imaging and machine learning approach

BioRxiv : the Preprint Server for Biology
Astri J LundervoldArvid Lundervold


Objective: In a three-wave 6 yrs longitudinal study we investigated if expansion of lateral ventricle (LV) volumes (regarded as a proxy for brain parenchyma loss) predicts performance on a test of response inhibition. Participants and Methods: Anatomical trajectories of left (LH) and right (RH) lateral ventricle volumes across the three study-waves were quantified using the Longitudinal Stream in Freesurfer 5.3 and modelled using a linear mixed-effects (LME) algorithm. All participants (N = 74, mean age 60.7 yrs at inclusion, 48 females) performed the Color-Word Interference Test (CWIT). Response time on the third condition was used as a measure of response inhibition (RI) and divided into three classes (fast, medium and slow). The Extreme Gradient Boosting (XGBoost) algorithm was used for calculating the relative importance of selected LV volume features from the LME model in predicting RI class. Finally, the two most important extracted features were fed into a 10-fold cross-validation framework, estimating the accuracy, precision and recall of the RI class prediction. Results: Four LME based features were selected to characterize LV volume trajectories: steepness of LV volume change and the LV volume at the time of inclusion...Continue Reading

Related Concepts

Biological Markers
Cell Growth
Luteinizing Hormone
Longitudinal Studies
Biologic Preservation
Lateral Ventricle Structure
Imaging of Brain

Related Feeds

BioRxiv & MedRxiv Preprints

BioRxiv and MedRxiv are the preprint servers for biology and health sciences respectively, operated by Cold Spring Harbor Laboratory. Here are the latest preprint articles (which are not peer-reviewed) from BioRxiv and MedRxiv.