Fig. 3. Voxel-wise statistical analysis of H²¹⁵O-PET data during administration of saline (  A ) and low-dose remifentanil (0.05 μg · kg⁻¹· min⁻¹;  B ). Pain induced increases in brain activation are shown (categorical comparison: painful heat > nonpainful heat). The statistical parametric maps (thresholded at  P < 0.05, false discovery rate corrected) are overlaid on skull-stripped normalized structural magnetic resonance images in sagittal planes (average of 27 T1-weighted scans of the same individual from the Montreal Neurological Institute) as provided by SPM99. The results of the moderate-dose remifentanil condition (0.15 μg · kg⁻¹· min⁻¹) are not shown because no significant activation clusters could be identified. 1 = Thalamus; 2 = cingulofrontal cortex; 3 = secondary somatosensory cortex; 4 = insula; 5 = caudate nucleus.