Scale-space analysis of event-related fMRI data using higher harmonics

Scale-space analysis is able to detect activation at various scales from a functional MR brain slice. Though scale space analysis has been used in functional MRI, its use has been limited to periodic block-based experiments. This paper extends scale space analysis to event-related experiments. Furthermore, the method uses higher harmonics of stimulus frequency, which contain considerable energy especially in periodic event-related data. The use of these harmonics results in a chi-square random field of higher degree and improves accuracy in activation detection.