Impact of Global Mean Normalization on Regional Glucose Metabolism in the Human Brain

Figure 1

Validation of quantified CMR_glc (qCMR_glc). (a) Comparison of absolute CMR_glc from Hyder et al. [36] (aCMR_glc-HYD) with qCMR_glc from five sites for the HAEC condition (Table 1). Bars represent the mean across all subjects in each group for gray matter regions (Table S1), where error bars are one standard deviation. All qCMR_glc and aCMR_glc-HYD values were very similar both relatively between regions and in terms of mean value, suggesting across-site comparisons are possible with our procedure for quantified CMR_glc. The Pearson correlation and the Euclidean distance (Table 2) suggest high similarity and low difference of CMR_glc for the HAEC group across all sites. (b) Scatter plots, from left to right, for aCMR_glc between Des1%, Sev0.25%, or Sev0.5% groups and equivalent HAEC groups (Table 1). Each point is one gray matter region (Table S1). Slope and from a linear fit are shown, and units are μmol/g/min, where a slope of less than 1 corresponds to a lower CMR_glc in the anesthetized group compared to the control. (c) Same as in (b), except using qCMR_glc from each group. The slopes are almost identical in (b) and (c), indicating that the calculation of qCMR_glc does not alter the relationship between groups for aCMR_glc.

(a)

(b)

(c)