Realignment strategies in ASL timeseries
Jespersen A, Lund T, Sidaros K
Copenhagen University Hospital Frederiksberg
Realignment parameter curves for ASL timeseries sometimes exhibit high-frequency modulations that could arise from the small perfusion-related difference between tag and control images. We compare three realignment strategies and test the detection of respiration and pulsation induced motion as well systematic tag-control differences. We find that the high frequency modulations in the realignment parameters are in fact due to respiration-induced motion and that systematic tag-control differences in the realignment parameters only arise in one of the compared methods. Furthermore, we find that not modelling the respiration effects in ASL timeseries leads to non-white noise in the residuals.