The neural mechanisms of drug cue‐reactivity regarding the temporal fluctuations of functional connectivity, namely the dynamic connectivity, are sparsely studied. Quantifying the task‐modulated variability in dynamic functional connectivity at cue exposure can aid the understanding. We analyzed changes in dynamic connectivity in 54 adult cannabis users and 90 controls during a cannabis cue exposure task. The variability was measured as standard deviation in the (a) connectivity weights of the default mode, the central executive, and the salience networks and two reward loci (amygdalae and nuclei accumbens); and (b) topological indexes of the whole brain (global efficiency, modularity and network resilience). These were compared for the main effects of task conditions and the group (users vs. controls), and correlated with pre‐ and during‐scan subjective craving. The variability of connectivity weights between the central executive network and nuclei accumbens was increased in users throughout the cue exposure task, and, was positively correlated with during‐scan craving for cannabis. The variability of modularity was not different by groups, but positively correlated with prescan craving. The variability of dynamic connectivity during cannabis cue exposure task between the central executive network and the nuclei accumbens, and, the level of modularity, seem to relate to the neural underpinning of cannabis use and the subjective craving.
11 January 2025 in The Journal of Infectious Diseases
Compatible deployment of AI-aided methods for PET quantification is an important prerequisite to maximizing their benefits. We focus on partial volume correction (PVC), which can substantially improve the precision of radiotracer uptake quantification in brain PET for intricate and atrophic regions. Conventional post-reconstruction PVC requires anatomical MR images that are often unavailable or of inadequate quality. We address this limitation by systematically evaluating two AI-aided strategies: (1) indirect PVC, which uses synthesized MR images for anatomical guidance, and (2) direct PVC, which predicts corrected PET images without anatomical processing. Multiple AI architectures were assessed under both strategies, using paired 18F-FDG PET + CT + MR datasets from multiple scanner sites. Indirect PVC consistently outperformed direct approaches across all tested architectures with the Denoising Diffusion Probabilistic Model yielding the best overall performance while preserving compatibility with standard PET processing pipelines. Both AI-aided approaches increased the utility of standalone 18F-FDG PET in clinical and research applications without requiring anatomical MR images. Indirect PVC showed advantages in transparency and performance for quantification in smaller anatomical regions, whereas direct PVC may be more suitable for rapid assessment in larger brain regions.