The presence of cerebral activation may bias motion correction estimates when registering FMRI time series. This problem may be solved through the use of specific registration methods, which incorporate or down-weight cerebral activation confounding signals during registration. In this paper, we evaluate the performance of different registration methods specifically designed to deal with the problem of activation presence. The methods studied here yielded better results than the traditional approaches based on least square metrics, almost totally eliminating the activation-related confounds. © Springer-Verlag 2004.
CITATION STYLE
Freire, L., Orchard, J., Jenkinson, M., & Mangin, J. F. (2004). Reducing Activation-Related Bias in FMRI Registration. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 3150, 278–285. https://doi.org/10.1007/978-3-540-28626-4_34
Mendeley helps you to discover research relevant for your work.