Analysis of dynamic radioligand displacement or "activation" studies

We present a simple way of assessing dynamic or time-dependent changes in displacement during single-subject radioligand positron emission tomography (PET) activation studies. The approach is designed to facilitate dynamic activation studies using selective radioligands. These studies are, in principle, capable of characterising functional neurochemistry by analogy with the study of functional neuroanatomy using rCBF activation studies. The proposed approach combines time-dependent compartmental models of tracer kinetics and the general linear model used in statistical parametric mapping. This provides for a comprehensive, voxel-based and data-led assessment of regionally specific effects. The statistical model proposed in this paper is predicated on a single-compartment model extended to allow for time-dependent changes in kinetics. We have addressed the sensitivity and specificity of the analysis, as it would be used operationally, by applying the analysis to 11C-Flumazenil dynamic displacement studies. The activation used in this demonstration study was a pharmacological (i.v. midazolam) challenge, 30 min after administration of the tracer. We were able to demonstrate, and make statistical inferences about, regional increases in k2 (or decreases in the volume of distribution) in prefrontal and other cortical areas.