Translator Disclaimer
14 April 2005 Compensation of intra-frame head motion in PET data with motion corrected independent component analysis (MCICA)
Author Affiliations +
Independent Component Analysis (ICA) has proved a powerful exploratory analysis method for fMRI. In the ICA model, the fMRI data at a given time point are modeled as the linear superposition of spatially independent (and spatially stationary) component maps. The ICA model has been recently applied to positron emission tomography (PET) data with some success (Human Brain Mapping 18:284-295(2003), IEEE Trans. BME, Naganawa et al, in press). However, in PET imaging each frame is, in fact, activity integrated over a relatively long period of time, making the assumption that the underlying component maps are spatially stationary (and hence no head movement has taken place during the frame collection) very tenuous. Here we extend the application of the ICA model to 11C-methylphenidate PET data by assuming that each frame is actually composed of the superposition of rigidly transformed underlying spatial components. We first determine the “noisy” initial spatially independent components of a data set under the erroneous assumption of no intra or inter-frame motion. Aspects of the initial components that reliably track spatial perturbations of the data are then determined to produce the motion-compensated components. Initial components included ring-like spatial distributions, indicating that movement corrupts the statistical properties of the data. The final intra-frame motion-compensated components included more plausible symmetric and robust activity in the striatum as would be expected compared to the raw data and the initial components. We conclude that 1) intra-frame motion is a serious confound in PET imaging which affects the statistical properties of the data and 2) our proposed procedure ameliorates such motion effects.
© (2005) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Martin J. McKeown, Marwa Gadala, and Rafeef Abu-Gharbieh "Compensation of intra-frame head motion in PET data with motion corrected independent component analysis (MCICA)", Proc. SPIE 5746, Medical Imaging 2005: Physiology, Function, and Structure from Medical Images, (14 April 2005);

Back to Top