Analysis of neuroreceptor PET-data based on cytoarchitectonic maximum probability maps: a feasibility study.

Three-dimensional maximum probability maps (MPMs) of cytoarchitectonically defined cortical regions based on postmortem histological studies have recently been made available in the stereotaxic reference space of the Montreal Neurological Institute (MNI) single subject template. This permits the use of cytoarchitectonic maps for the analysis of functional in vivo datasets, including neuroreceptor positron emission tomography (PET) studies. In this feasibility study, we used 5-hydroxytryptamine 2A (5-HT2A) receptor PET to test the applicability of maximum cytoarchitectonic probability maps for quantitative analysis. As the outcome parameter, we extracted local distribution volume ratios (DVRs) from 19 cytoarchitectonically defined volumes of interest (VOIs) per hemisphere from five healthy subjects. The experimental design included a forward ('PET to atlas' normalization) and a backward ('atlas to PET' normalization) procedure to double-check the stability of transformation and overlay. Resulting DVRs were compared with receptor densities (RDs) obtained from postmortem [3H]ketanserin autoradiography of multiple areas. Correlations between the bi-directional normalization procedures (r = 0.89; 38 VOIs) as well as between in vivo and vitro data (nine VOIs; r = 0.64 and r = 0.47 for forward and backward procedure, respectively) suggest that the implementation of cytoarchitectonic maximum probability maps is a promising method for an accurate and observer-independent analysis of neuroreceptor PET data.