The Effect of Band-Pass Filter Range and Global Signal Regression on the Amplitude of Low-Frequency Fluctuations: A Pilot Study

摘要

Aim: The amplitude of low-frequency fluctuations has been widely used in neuroscience. However, the stability should be checked before clinic implementation. We explored the effect of band-pass filtering range and global signal regression application on amplitudes of low-frequency fluctuations in the resting-state fMRI study. Materials and methods: The dataset were from longitudinal test-retest neuroimaging data from healthy young adults in southwest China (SLIM). All data collected from resting-state fMRI scan. Twenty-eight young healthy subjects were randomly selected from this dataset. One-way analysis of variance was used to compare amplitudes of low-frequency value among three different band-pass filter ranges: 0.01-0.03 Hz, 0.03-0.07 Hz, and 0.01-0.1 Hz. Two-sample paired t-tests were used to compare amplitudes of low-frequency value between data with and without global signal regression fixing the filter ranges. Results: Amplitudes of low-frequency fluctuation value were changed in the rectus gyrus, caudate, thalamus, putamen, Heschl's gyrus, temporal lobe, and supramarginal gyrus among different filtering ranges. Most alterations appeared in the 0.01-0.03 Hz and 0.03-0.07 Hz range. Global signal regression only influenced the amplitudes of low-frequency fluctuations in the different filter ranges and had no influence when the band-pass filter range was fixed. Conclusion: Filter range selection affected part of brain amplitudes of low-frequency value, such as rectus gyrus, caudate, thalamus, putamen, Heschl's gyrus, temporal lobe, and supramarginal gyrus. Besides, global signal regression influenced amplitudes of low-frequency fluctuations among different filter ranges as well, but no effect when filter range fixed. It should be carefully designed when ALFF applied in clinical work.