Breath-Hold Induced Cerebrovascular Reactivity Measurements Using Optimized Pseudocontinuous Arterial Spin Labeling

Figures

• 

  FIGURE 1. Regressors used in the CVR calculation, shown here for the first 100 images of a representative volunteer: two boxcar regressors (in red), one adjusted in amplitude to a mean PETCO2 change through the whole scan (top), and one adjusted to recorded PETCO2 change in each BH task (bottom) and two ramp regressors (in blue), one varying in amplitude from 0 to the mean ΔPETCO2 recorded in the whole scan (top), and the other from 0 to the recorded ΔPETCO2 for each BH task (bottom). The mean GM CBF (in ml/100 g/min) is also shown for the same subject (discontinuous line).

• 

  FIGURE 2. CBF signal increase due to breath-holding in a single participant. (A) Consecutive CBF (ml/100 g/min) maps (images 61 to 68) overlayed in anatomical T1-weighted images and shown in neurological orientation. (B) Mean CBF (ml/100 g/min) signal in gray matter (red), superimposed on the respiratory trace (blue), where the response delay can be observed. The oscillation in the respiratory trace during the breath-hold period is due to pumping motion of the heart against the chest wall. Gray discontinuous lines represent the period of image acquisition. Maps displayed in (A) are indicated by image numbers on top of (B). (C) Boxplots showing the measured delays for the GM and WM in the group of volunteers.

• 

  FIGURE 3. (A) Group mean baseline GM CBF (in red and primary axis) and mean baseline PETCO2 (in blue and secondary axis) are presented for the rest periods between apneas, with period 1 being the initial period of normal breathing at the start of the experiment. (B) Group mean breathing rate (in green and primary axis) and amplitude (in orange, secondary axis and in arbitrary units) for each rest period. Error bars represent the standard deviation.

• 

  FIGURE 4. Group mean CVR maps, calculated by averaging normalized individual CVR maps from all ten participants, obtained with the four different tested models. The maps have been overlayed in anatomical T1-weighted template images. The maps display the expected contrast between gray and white matter. Higher values of CVR were obtained using the ramp regressors, with both mean and adjusted ramp amplitudes. Images are shown in neurological orientation.

• 

  FIGURE 5. Boxplots of mean GM (A) and WM (B) CVR computed in the four different models used. Significant differences between the models are indicated. P-values have been corrected for multiple comparisons using the Bonferroni method.

• 

  FIGURE 6. Voxel wise SPM{T} maps obtained in the comparison between the CVR maps estimated with the ramp mean ΔPETCO2 and the boxcar adjusted ΔPETCO2 models (p < 0.05 FEW cluster corrected) overlaid on anatomical T1-weighted images. Clusters shown in color scale represent areas of significant CVR increase for the ramp mean ΔPETCO2 model compared to the boxcar adjusted ΔPETCO2 model. Images are shown in neurological orientation.