Episodic memory performance was measured in the fMRI task and in various neuropsychological tests, using either incidental or intentional memory formation. Our analyses where based on a large sample of 106 young and 153 older subjects ( Soch et al., 2021a). We compared SVM-based prediction of chronological age versus prediction of memory performance from behavioral data, task-based fMRI, resting-state fMRI, and structural MRI markers associated with increasing age. Thus, data from different modalities may differentially predict chronological age and memory performance, respectively. Several studies evaluated potential structural, functional, physiological and behavioral predictors of age-related cognitive decline (Gross et al., 2011 Hou et al., 2020 Chen et al., 2021), but only few studies systematically compared different predictors and their joint predictive value (e.g., Woodard et al., 2010).Ĭomparing the predictive value of MRI biomarkers for chronological age versus individual memory performance appears to be a promising endeavor, because “successful aging” may reflect dissociable neural mechanisms: differences in the manifestation of age-related physiological changes (“brain maintenance”) and/or differences in cognitive processing (“cognitive reserve” Nyberg et al., 2012). The wide range of cognitive functioning even within narrowly defined age groups suggests that chronological age and cognitive performance might be predicted by different modalities. One such study found that a combination of ApoE genotype and functional MRI was the most effective predictor for future cognitive decline ( Woodard et al., 2010). In contrast to the abundant literature on age prediction from structural MRI ( Cole et al., 2017, 2018 Luders et al., 2016 Steffener et al., 2016 Soch, 2020), few studies have been devoted to predicting cognitive function, particularly memory performance, from neuroimaging data. They have been employed to predict chronological age from structural magnetic resonance imaging (MRI Cole et al., 2017, 2018), to estimate brain age ( Bashyam et al., 2020 Habes et al., 2021) or to distinguish health from disease ( Dyrba et al., 2021 Eitel et al., 2021). Machine learning-based techniques such as support vector machine (SVM) classification and regression provide promising approaches to differentiate normal from pathological neurocognitive aging. An early assessment of changes in cognitive performance can help to determine people at risk of pathological aging, such as various forms of dementia, and allows for early medical and behavioral interventions ( Naismith et al., 2009 Cabeza et al., 2018 Whitty et al., 2020). Notably, a subpopulation of older adults show “successful aging”, with memory performance comparable to that of younger adults ( Nyberg et al., 2012 Nyberg and Pudas, 2019).

Our results suggest that superior memory performance in healthy older adults is better explained by efficient recruitment of memory networks rather than by preserved brain structure.Įpisodic memory performance peaks in young adulthood and declines with increasing age. In contrast, memory performance was best predicted from task-based fMRI contrasts and particularly single-value fMRI scores, whereas gray matter volume has no predictive power with respect to memory performance in healthy adults. While age group could be reliably decoded from all variables, chronological age within young and older subjects was best predicted from gray matter volume. The candidate variables were (i) behavioral response frequencies in an episodic memory test, (ii) recently described fMRI scores reflecting preservation of functional memory networks, (iii) whole-brain fMRI contrasts for novelty processing and subsequent memory, (iv) resting-state fMRI maps quantifying voxel-wise signal fluctuation and (v) gray matter volume estimated from structural MR images.

Prediction was performed using support vector machine (SVM) classification and regression across and within two samples of young (N = 106) and older (N = 153) adults. Here, we approached this question by testing multiple candidate variables from structural and functional neuroimaging for their ability to predict chronological age and memory performance, respectively. Alternatively, the brain might recruit existing resources more efficiently or employ compensatory cognitive strategies. One explanation for this could be higher brain structural integrity in these individuals. However, some older adults show “successful aging”, that is, relatively preserved cognitive ability in old age. Human cognitive abilities decline with increasing chronological age, with decreased explicit memory performance being most strongly affected.

Functional age vs chronological age