fMRI and MEG analysis of visceral pain in healthy volunteers

Background Although many studies of painful rectal stimulation have found activation in the insula, cingulate, somatosensory, prefrontal cortices and thalamus, there is considerable variability when comparing functional magnetic resonance imaging (fMRI) results. Multiple factors may be responsible, including the model used in fMRI data analysis. Here, we assess the temporal response of activity to rectal barostat distension using novel fMRI and magnetoencephalography (MEG) analysis. Methods Liminal and painful rectal barostat balloon inflation thresholds were assessed in 14 female healthy volunteers. Subliminal, liminal and painful 40s periods of distension were applied in a pseudo-randomized paradigm during fMRI and MEG neuroimaging. Functional MRI data analysis was performed comparing standard box-car models of the full 40s of stimulus (Block) with models of the inflation (Ramp-On) and deflation (Ramp-Off) of the barostat. Similar models were used in MEG analysis of oscillatory activity. Key Results Modeling the data using a standard Block analysis failed to detect areas of interest found to be active using Ramp-On and Ramp-Off models. Ramp-On generated activity in anterior insula and cingulate regions and other pain-matrix associated areas. Ramp-Off demonstrated activity of a network of posterior insula, SII and posterior cingulate. Active areas were consistent with those identified from MEG data. Conclusions & Inferences In studies of visceral pain, fMRI model design strongly influences the detected activity and must be accounted for to effectively explore the fMRI data in healthy subjects and within patient groups. In particular a strong cortical response is detected to inflation and deflation of the barostat, rather than to its absolute volume.