Image-space based estimation of undersampled respiration noise in multislice fMRI data

As demonstrated in previous work by Ziyan et al. (2002), it is possible to estimate and remove respiration-induced signal fluctuations from multi-slice image-domain (magnitude and phase) fMRI data. However, this technique was limited to the analysis of over-sampled respiratory noise within a single imaged slice. This work presents a procedure to estimate the undersampled respiration waveform in multi-slice fMRI data when the slice acquisition rate is below the Nyquist rate. This technique is based upon the single-slice image-space (magnitude and phase) estimation and removal procedure developed by Ziyan et al., and is presented as an alternative to the mathematically complex technique of Frank et al. (2001) that uses k-space analysis to achieve similar ends.