Application Of Meta-analysis In Medical Image Data

Meta-analysis is a kind of quantitative combined analysis aiming at a series of independent research results. Since1976, this method has come forward by Glass on psychological research. It has been applied in many subject area broadly, especially in medical domain. This paper aims at quantitative description of subjects’activated brain regions in finger-tapping task and introduction of ALE method applicated in meta-analysis for functional magnetic resonance imaging data.ALE meta-analysis is a kind of activation likelihood estimate method based on random effects model, with quantitatively analysis about previous fMRI research results. Its basic thought is that locating those stable activated brain areas, in order to estimate clusters of concordance of them from subjects under the condition of tasks, by calculating the ALE maps of every region voxels. Concrete steps are as follows:Firstly, the paper selects33literatures (18fMRI and15PET), reviewed in Medline, Embase, CNKI and so on by Sleuth, as the data in this paper. Secondly, it analyzes data using GingerALE. Finally, it shows activated brain areas by Mango. The paper applies selected data for three analysis. The first pools the results from all of literatures. For the second, studies are divided into three groups based on the type or lack of pacing stimulus employed:auditory stimulus (20papers;381foci), visual stimulus (6papers;95foci), and no stimulus (10papers;157foci). The final analysis divides the studies into three groups determined by the complexity of the tapping task used:right hand index finger (19papers;261foci), RH multi-finger sequence (14papers;240foci), and bimanual (4papers;88foci).The results find that clusters of concordance are identified within the primary sensorimotor cortices, supplementary motor areas, premotor cortex, bilateral inferior parietal cortices, bilateral basal ganglia, and bilateral anterior cerebellum in all finger tapping tasks. Distribution of activated brain areas based on pacing stimulus and the complexity of the tapping task are wider than the others. In addition, the results can make a conclusion that the different complexity of the tapping task can lead to have a significant impact on areal distribution of the activated brain.