Skip to main content
. 2017 Mar 21;12(3):e0173289. doi: 10.1371/journal.pone.0173289

Fig 4. Power spectral density plotted for every data cleaning strategy, per dataset.

Fig 4

ICA-AROMAsoft, MWC and FIXsoft showed the highest power spectral density amplitudes. MWC and FIXsoft also showed residual power in the non-BOLD frequencies. FIXagg and ICA-AROMAagg exhibited a power spectral density reduction at high frequencies and a drop of the very low frequencies. ME-ICA cleanup efficiently reduced the power spectral density variability, suppressed the high frequencies and preserved the ultra-slow frequencies. ME-Uncleaned and ME-AROMAagg power spectral densities were very flattened at low and high frequencies.