MRI data

qAIM-MRI and image analysis In the presence of Mn2+ in the extracellular solution, highly active neurons should accumulate larger amounts of Mn2+ than those of weakly active neurons. Mn2+ also has the characteristic of shortening the longitudinal relaxation time (T1) of H+. Differential accumulation of Mn2+ in active and silent brain regions is generally assessed using T1-weighted images. The Mn2+ concentration can be quantified by measuring the absolute T1 (R1 = T1-1) value. Mice were grouped into stimulation and non-stimulation groups and further into two groups acquired at 1 and 4 weeks and at 2 weeks post-injury. Mice were injected with MnCl2 solution (0.2 mmol/kg, intraperitoneally) 48 h before MRI acquisition. Mice were electrically stimulated using the same methods as the electrophysiological assessment using H-reflexes 2 h after MnCl2 injection. Furthermore, spinal stretch reflexes related to spasticity were evoked. Acquisition methods for MRI data were similar to those previously described. MRI data were acquired using an AV400WB 9.4-T, 89-mm spectrometer equipped with a 150 G/cm gradient insert (Bruker) and an 18-mm 1H volume coil (Takashima Seisakusho), and signals were recorded using ParaVision 5.1 software (Bruker). Mice were anaesthetised with isoflurane (induction: 3−4%, maintenance: 1−2%) in a mixture of N2O and O2 (N2O:O2 = 7:3). Body temperature was maintained using circulating heated water. The temperature and the respiration were continuously monitored during data acquisition and remained within normal ranges. For T1 measurements, rapid acquisition with relaxation enhancement (RARE) with a variable repetition time (RARE-VTR) pulse sequence with seven repetition times (TR: 625, 850, 1,200, 2,000, 3,000, 5,000, and 8,000 ms) was used with an effective echo time (TEeff) = 6.7 ms, RARE factor = 2, the field of view = 16 × 14 mm2, matrix size = 160 × 140, slice thickness = 0.5 mm, number of slices = 24, and the number of averages (NEX) = 2. Multislice, fast spin-echo T2-weighted images (RARE, TR = 3000 ms, TEeff = 30 ms, RARE factor = 4, NEX = 3) were acquired and used to co-register images to the mouse brain template29.The MRI data analysis was performed using ANTx2 30 (https://github.com/ChariteExpMri/antx2) and custom-made scripts running in MATLAB (Math Works). The RARE-VTR images for T1 measurements were re-aligned to the T2-weighted images using ANTx2. Through the ANTx2 pipeline, T2-weighted images were processed using SPM12 (https://www.fil.ion.ucl.ac.uk/spm/) and non-linear warping of tissue probability maps in ELASTIX (Klein et al., 2010), and these were registered in the Allen Mouse Atlas 2017 (CCFv3) in combination with Hikishima Templates26,29,31,32. Thereafter, the RARE-VTR images were co-registered to the atlas using the ANTx2 pipeline. After the registration, parametric T1 maps were calculated pixel-by-pixel by fitting with the following equation using the MATLAB scripts (http://freddy.odille.free.fr/index.php/2020/01/14/vectorized-levenberg-marquardt-optimization-for-t1-mapping-mri-matlab-and-gpu-code/):S(TR) = A (1 – exp(-TR/T1)),where S(TR) is the signal intensity of each pixel at a given TR and A is a scaling factor.Pixels with T1 values >6,000 ms or <500 ms were excluded from the analysis, and the T1 map was filtered using a 3 × 3 median filter. As R1, the T1 inverse, is proportional to Mn2+, the R1 maps were calculated pixel-by-pixel by inverting the T1 values of the T1 maps. For region of interest (ROI) analyses, we used the mouse brain atlas to identify anatomical locations and determine the ROIs on the T1 maps. We averaged the T1 values of each pixel within the ROIs, inverted the results, and used them as R1 values for the subsequent analysis.When comparing the stroke and sham groups, all group R1 values were adjusted using the R1 values in the non-stimulus sham group to exclude reduced neuronal activity in the entire brain due to ischaemic damage. The substantia innominate (SI) has no direct neuronal connection to the motor cortex. Thus, the means R1 value (0.517 ± 0.030, Table 1)of the sham group without stimulus was used as the calibration standard, and the R1 values of SI in all other mice were standardised.