Fig. 4: Hyperpolarized ¹²⁹Xe MR thermometry in mice.

a Example of ¹H spectrum acquired from one of the mice showing the lipid peaks along with the CH₂ resonance frequency line used as internal reference to measure absolute temperature in fat. b ¹²⁹Xe spectrum acquired from the same mouse showing the temperature-sensitive lipid-dissolved xenon resonance frequency line (LDX). c Coronal image of one of the obese mice showing the relative location of the shimmed voxel (green box) and the voxel from which the CH₂ spectrum was acquired (red box). d Axial image of one of the obese mice showing the relative location of the shimmed voxel (green box) and the voxel from which the CH₂ spectrum was acquired (red box) along with the location of the ¹²⁹Xe surface coil (white line) and of the optical temperature probe (yellow arrow), running through the middle of the voxel. e Scatter plot showing correlation between tissue temperature measured by using the CH₂-referenced LDX frequency \({\omega }_{{rLDX}}\) and tissue temperature measured by the optical temperature probe (n = 54 independent measurements acquired on the four animals and reported in Supplementary Data 2). Data collected above 40.5 °C originated from a single animal that probably entered a febrile state before being euthanized. Error bars represent measurement experimental uncertainties derived from error propagation accounting for the experimental uncertainty in the CH₂ and LDX peak positions, as well as on the uncertainty associated with the temperature coefficients of Eq. (2). The median difference between the two measurements was 0.12 °C and ranged from −0.32 °C to 0.14 °C, while the standard deviation was 0.16 °C.