Skip to main content
. Author manuscript; available in PMC: 2015 Apr 1.
Published in final edited form as: Neuroimage. 2013 Dec 16;89:314–330. doi: 10.1016/j.neuroimage.2013.12.013

Table 6.

Mean explained variance by physiological noise at different p values for unsmoothed data

p X1 (%) HF cardiac X2 (%) HF resp X3 (%) LF cardiac X4 (%) LF resp [X1 X2 X3 X4] (%) Affected voxels (%)
0.05 1.5±0.2 3.2±0.9 1.4±0.3 1.3±0.2 9.8±0.9 20.3±8.7
(0.05) (1.5±0.2) (3.3±0.8) (1.2±0.3) (1.2±0.2) (9.6±0.8) (19.7±8.7)
0.01 1.7±0.3 4.3±1.3 1.7±0.6 1.6±0.4 12.5±1.0 7.3±4.6
(0.01) (1.8±0.3) (4.6±1.3) (1.4±0.5) (1.5±0.4) (12.3±0.9) (7.1 ±4.6)
0.001 2.3±0.6 6.0±2.1 2.2±0.9 1.9±0.7 16.2±1.4 2.2±1.9
(0.001) (2.3±0.5) (6.5±2.1) (1.7±0.8) (1.7±0.7) (15.9±1.0) (2.2±1.9)

Note: The columns refer to the type of physiological noise averaged over all subjects (X1. high-frequency cardiac waveform, X2: high-frequency respiratory waveform, X3: low-frequency cardiac waveform, X4: low-frequency respiratory waveform). The rows refer to different p values (uncorrected). The numbers without parenthesis refer to the optimized model using the low-frequency waveforms in this research, whereas the numbers in parenthesis refer to the low-frequency model of Chang et. al. (2009) and Bim et. al. (2008). The last column refers to the percentage of the voxels that are affected by all 4 regressors.