Julien Cohen-Adad1,2, Serge Rossignol3, Richard D. Hoge4
1INSERM, Univ Paris 6, Pitie Salpetriere Hospital, Paris, France; 2GRSNC, Physiology department, Univ Montreal, Montreal, QC, Canada; 3GRSNC, Physiology department , Univ Montreal, Montreal, QC, Canada; 4UNF, Univ Montreal, Montreal, QC, Canada
To date, spinal cord BOLD-fMRI has generated considerable efforts to obtain robust activation maps. One possible cause of the low sensitivity of this technique in the spinal cord are the relatively narrow activation blobs, making activation maps extremely sensitive to subject motion. This study demonstrates the benefits of 2D versus 3D methods to correct subject motion in axial fMRI time series of the spinal cord. SliceCorr, a spinal cord-dedicated method has been developed and compared with 3D-based methods, using human dataset of the whole cervical spinal cord. Results showed the least residual motion for the SliceCorr-corrected datasets in all subjects.