The number of neuroimaging studies has grown exponentially in recent years and their results are not always consistent. Meta-analyses are helpful to summarize this vast literature and also offer insights that are not apparent from the individual studies. While a number of suitable voxel-based meta-analytic methods for neuroimaging data had been developed at the time this thesis was conceived, they also suffered from a series of important drawbacks such as the separate analyses of positive (e.g. grey matter volume increases) and negative (e.g. grey matter volume reductions) findings, not accounting for the effect size, not taking sample size, intra-study variance or between-study heterogeneity into account, and not allowing combination of reported peak coordinates and statistical parametric maps. The aim of this thesis was the development of a series of voxel-based meta-analytic methods and software tools for neuroimaging studies, which overcame some of the limitations of previous methods. Specifically, this thesis includes: a) the development of a new voxel-based meta-analytic method, named signed differential mapping (SDM), which adopted and combined the various positive features of previous methods and also introduced a series of improvements and novel features; b) the subsequent development and adaptation of the method to allow addressing additional research questions, such as meta-analyses comparing several disorders, meta-analyses of white matter volume or fractional anisotropy studies, and combination of various imaging modalities; and c) examples of applications of these methods. The methods and software derived from this thesis have been well received by the field. As of September 2012, more than thirty meta-analyses using SDM have been published, and the first study introducing the methods has been cited more than a hundred times. Suggestions for future research and further methodological development are discussed.
|Date of Award||Jan 2013|
|Supervisor||David Mataix-Cols (Supervisor)|