Head motion is one of the major sources of artifacts in functional Magnetic Resonance Imaging. Head motion is said to cause large spatially varying changes in the signal intensity of the BOLD signal, making it very difficult to separate neuronal activations from the motion artifact. Realignment corrects the changes in brain position, but it does not take into consideration the changes in the image intensity associated with motion. Head Motion, particularly in the direction perpendicular to the slice selection is susceptible to artifacts due to Magnetic Field Inhomogeneity and Spin excitation history effects [1]. Resting state functional connectivity measures the synchronicity of the brain activity in different regions of the brain …show more content…
In Prospective motion correction, the motion is corrected for before the acquisition of the volume, whereas retrospective motion correction correct for motion after the acquisition of the volumes. Realignment, Nuisance signal regression, modelling the effects of the head motion on the BOLD signal using motion parameters and removing the fitted response, filtering, Motion Censoring/Spike regression, Group level correction or some combinations of them are routinely used with varying degrees of success in retrospective motion …show more content…
So we can try to effects of head movement on the BOLD signal by using current motion parameters and the previous motion parameters and remove the fitted response from the BOLD signal. 24 parameters (Rt, Rt2, Rt-1, Rt-12) and 36 parameters (Rt, Rt2, Rt-1, Rt-12, Rt-2, Rt-22) are being used to correct for motion due to the increase in the fit compared to models using a lower number of parameters. However, going further back in time does not justify the increasing fit with the loss of degrees of freedom. Also, a fewer number of parameters must be used for low motion subjects as increasing the number of parameters causes the model to overfit the data