Denoising and DA release: effect of denoising on the ability to identify voxel-level neurophysiological response

Parametric kinetic models such as lp-ntPET are used to estimate spatiotemporal patterns of endogenous neurotransmitter release, such as smoking-induced dopamine (DA) release. Applicability of such models is often limited by noise in the voxel-level time activity curves (TAC). Previous work has demonstrated the usefulness of HYPR-based spatial post-processing to increase neurotransmitter release detection sensitivity. Here we extend on this work by developing more realistic noise modelling and including additional denoising approaches. Using simulations and data acquired in non-human primates (NHP), we demonstrate that HYPR-based denoised reconstructions coupled with HYPR post-processing enables a ∼ 7 fold reduction in the detectable size of the region with neurotransmitter release, less biased parameters of interest, and favours bolus vs. bolus + constant infusion protocols.