Tracing diurnal differences in brain anatomy with voxel-based morphometry - associations with sleep characteristics and data quality

Abstract

AbstractIn recent decades magnetic resonance imaging (MRI) has established itself as the golden standard for studying human brain structure. As a result, it has become increasingly more important to identify technical and physiological factors that may influence study outcomes and contribute to misleading conclusions. This work set out to investigate and compare the variability in voxel-based morphometry (VBM) associated with data quality and time-of-day (TOD) when images are collected using a large longitudinal dataset (N = 72). To test the replicability of the results connected to each factor, the analyses were conducted in two comparable cohorts (each N = 36). The replicated TOD differences were further tested for their correlation with indices of sleep. We report that TOD is a significant and partially replicable predictor of the variability in VBM with the magnitude of the observed differences regionally connected to sleep length. Although data quality also has a prevalent impact on VBM, the affected areas were inconsistent across the two datasets. Interestingly, certain overlapping regions were identified in both analyses, showing that longitudinal results could be misattributed to image quality unless this factor is accounted for. All in all, both TOD and data quality are crucial variables to be considered and controlled for at the stage of experimental planning, as well as data acquisition and analysis. Inclusion of these parameters in the study designs can greatly improve the reliability and replicability of the reported MRI results.