Additional supporting information can be found in the online version of this article.

ANA_22387_sm_suppinfo.doc33KSupporting Information
ANA_22387_sm_suppinfofig1.pdf1357KSupporting Information Supplementary figure 1. Individual data plots for the white matter (WM) and ventromedial forebrain (vmFB) voxel ratios for Lac/Cr (A-D and S-V, respectively), Cho/Cr (F-I and Y-AB), NAA/Cr (K-N and AE-AH) and for vmFB Lac/NAA (AK-AN) for each animal (#47-85) randomly assigned to untreated, Xe, hypothermia and combined Xe+H groups. The regression slopes are shown in the 5th column (E, K, Q, W, AC, AI, AO). Comparisons of the Areas Under the Curve/AUC (Areas Over the Curve/AOC for NAA/Cre) for each animal with the Voxel matched Lac/NAA are shown in the last column (F, L, R, X, AD, AJ); in the case of vmFB Lac/NAA (AP) the comparison is with the WM Lac/NAA AUC. Each animal is represented by a single data point. In both WM and vmFB voxels, injury elicited prominent effect in untreated controls for Lac/NAA and Lac/Cr, decreasing successively in Xe, HT and Xe+HT groups. There was no difference between groups for Cho/Cr. Overall, the vmFB traces showed very similar results to those observed with WM. In the 5th column, the parabolic fits of metabolite ratios vs Lac/NAA indicate comparative sensitivity to injury: Cho/Cr (L, AD) showed no correlation, for the WM as well as for the vmFB voxel. In both voxels, Lac/Cr (F, X) was also more sensitive to lower degrees of injury than Lac/NAA with parabolic fit bulging to the left. The reverse was true for NAA/Cr vs Lac/NAA (R, AJ) where the fit bulged to the right. Finally, the vmFB Lac/NAA showed similar results to the WM Lac/NAA as evidenced by the straight line in the parabolic fit (AP).
ANA_22387_sm_suppinfofig2.pdf764KSupporting Information Supplementary figure 2. Individual data plots of whole brain 31P MRS metabolite peak area ratios vs WM Lac/NAA: Pi/EPP (A-D), PCr/EPP (F-I), NTP/EPP (K-N), pH (P-S) for each animal (#47-85) randomly assigned to untreated, Xe, hypothermia and combined Xe+H groups. The regression slopes are shown in the 5th column (E, K, Q, W). The parabolic fits of all metabolite ratios vs WM Lac/NAA indicate relative sensitivity to injury (F,L,R,X). The most sensitive 31P MRS biomarker was Pi/EPP (F). All parabolic fits vs Lac/NAA showed a bulge to the right indicating a lower sensitivity to injury versus Lac/NAA.
ANA_22387_sm_suppinfofig3.pdf773KSupplementary figure 3 A-L: Correlation of MRS biomarker AUCs (area under the curve, X-axis) with the average density of TUNEL+ cells across the forebrain (Y-axis), each data point represents a single animal. All correlations were positive, the highest R2 values were seen with WM Lac/NAA (I, R2=0.57) and Pi/EPP (A, R2= 0.52), decreasing gradually, for Lac/Cr, NAA/Cr, NTP/EPP and PCr/EPP, till comparatively low levels for pH (D, R2 = 0.14). Cho/Cr AUCs showed no correlation. M-P: Effects of averaging, correction for biomarkers and treatment groups on the negative correlation of TUNEL (Y-axis) with the microglial ramification index. M: Averaging TUNEL and Ramification across 7 different forebrain regions resulted in considerably higher covariance (R2=0.27) compared with the primary data in figure 3F (R2=0.11). N,O: Correction of TUNEL and Ramification for the WM Lac/NAA (O) or vmFB Lac/NAA (P) biomarkers using mixed linear model resulted in an almost complete removal of the correlation between the two histological outcome measures, reducing residual R2 to 0.05 and 0,04, respectively (p<2%, F-test). P: Correction for just the 4 treatment groups (control, Xe, hypothermia, combined) had, at most, a much milder effect (R2=0.22, p=16% for the reduction in variance).

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.