Template-based approach for detecting motor task activation-related hyperperfusion in pulsed ASL data

Arterial spin labeling (ASL) permits the non-invasive measurement of quantitative values of cerebral blood flow (CBF) and is thus well adapted to study inter- and intra-subject perfusion variations whether at rest or during an fMRI task. In this study, a template approach to detect brain activation as a CBF difference between resting and activated groups was compared with a standard generalized linear model (GLM) analysis. A basal perfusion template of PICORE-Q2TIPS ASL images acquired at 3T from a group of 25 healthy subjects (mean Age 31.6±8.3 years) was created. The second group of 12 healthy subjects (mean age 28.6±2.7 years) performed a block-design motor task. The template was compared with the mean activated image of the second group both at the individual and at the group level to extract activation maps. The results obtained using a GLM analysis of the whole sequence were used as ground truth for comparison. The influences of spatial normalization using DARTEL registration and of correction of partial volume effects (PVE) in the construction of the template were assessed.
Results showed that a basal perfusion template can detect activation-related hyperperfusion in motor areas. The true positive ratio was increased by 2.5% using PVE-correction and by 3.2% using PVE-correction with DARTEL registration. On average, the group comparison presented a 2.2% higher true positive ratio than the one-to-many comparison.