Alterations in Generalization Ability and Neuroimaging Markers with Respect to Aging and Alzheimer's Disease: A Comprehensive Meta-Analysis Dataset

Abstract: 

This dataset supports a comprehensive metadata analysis aimed at exploring emerging evidence for generalization deficits as a sensitive early measure of cognitive and neural decline in Alzheimer's disease (AD). It includes data on neuroimaging biomarkers, cognitive performance metrics related to generalization, and study-specific metadata extracted from multiple peer-reviewed publications. The data was curated and standardized to facilitate comparisons across studies, focusing on multimodal approaches to understanding aging and AD. This dataset serves as a resource for researchers investigating the interrelationships between generalization and neuroimaging, contributing to early AD diagnosis and interventions. It can be used for further meta-analyses, training predictive models, and hypothesis generation in aging and neuroimaging research. 

Dataset Description: 

This dataset encompasses information extracted from a systematic review on associative learning (generalization) in the context of aging and AD. Data was systematically collected from studies published between 2004 and 2024, covering neuroimaging measures linked to cognitive function and aging. Key elements of the dataset include publication details (authors, year, journal), sample sizes, neuroimaging modalities used, DOI links, and reported outcomes. The dataset integrates neuroimaging findings from structural MRI, resting-state functional MRI (rs-fMRI), task-based fMRI, and Diffusion Tensor Imaging (DTI) with cognitive performance across generalization tasks. Studies predominantly utilized MRI scanners at field strengths of 1.5 Tesla, 3 Tesla, 4 Tesla, and 7 Tesla, with a focus on regions such as the medial temporal lobe, hippocampus, frontal and temporal lobes, parietal lobe, precuneus, subcortical and thalamic regions, cingulate and caudate, as well as the cerebellum.  

Methodology 

The protocol for the metadata analysis was pre-registered on AsPredicted (#187862) and followed a rigorous systematic review process. Searches were conducted on PubMed and EMBASE to identify peer-reviewed studies published in English between 2004 and 2024 that investigated associative learning in relation to aging and AD. Inclusion criteria required that studies involved participants aged 45 years or older and included clinical diagnoses of cognitive impairment such as mild cognitive impairment (MCI), AD, or dementia, along with neuroimaging techniques. Exclusion criteria were applied to filter out animal studies, review papers, case studies, studies unrelated to aging and AD, those without MRI data, and those involving only participants younger than 45 years. 

The search strategy utilized a broad set of keywords related to generalization, cognitive decline, neuroimaging, AD, and memory processes, ensuring the identification of all relevant studies. Each stage of the systematic review—database search, duplicate removal, title and abstract screening, and full-text evaluation—was documented and summarized in a PRISMA flow diagram. Extracted data included authors, year of publication, sample size, neuroimaging modalities, reported outcomes, and journal titles. The dataset was meticulously organized and unrelated studies were excluded based on the predefined criteria, leading to a comprehensive compilation of studies examining the impact of aging and neurodegeneration on generalization and associated neuroimaging markers.