DPARSF Cov_regression produces weired images

Dear Dr. Yan,

I am using DPASFA v4.3 to run resting-state image, within the step : ARSWDCF. The detrended images are good, but something goes wrong after the regression.  Bellowing are the settings I used, the Detrended image, and Detrended + Covregessed .  If I put the cov_regressed images into ICA, it will produce even more weird results.   

Any Advice is highly appreciated!