Magnetic resonance imaging

Each examined animal was placed onto a dedicated holder, and inhalation anaesthesia (2% isoflurane in air) was applied while the respiration rate was maintained within the 50–70 BPM range. All MRI experiments were performed at 37 °C using a 4.7 T direct-drive horizontal scanner (Agilent, Yarton, UK) equipped with a 400-mT/m gradient and 120-mm bore diameter insert. A Rapid Biomed (Rimpar, Germany) quadrature volume coil transmitter with an i.d. of 72 mm and a STARK Contrast (Erlangen, Germany) surface coil receiver were used for the signal detection. A diffusion-weighted sequence (FSEMSDW) was used with the following parameters: TR/TE_eff = 3000/36 ms, b-value = 1000 s/mm², echo train length = 8, echo spacing = 10 ms, 10 averages, FOV = 25 × 25 mm², and matrix size = 128 × 128. Eighteen contiguous 1-mm thick coronal slices were scanned to cover the brain regions of interest. The overall scan time was 8 min.

Both the data processing and volumetry were performed using ImageJ software, version 1.47³³. A median filter with a 2-pixel radius was applied to reduce the noise and delineate the boundaries of the image. The overall lesion size was calculated by summing the hyperintensive area of all slices and multiplying by the slice thickness. Both the ipsilateral and contralateral lesion volumes were divided by the appropriate hemisphere volume. The measurement reliability was determined at the level of the slice for the oedema (n = 88) and brain (N = 108) measurements by evaluating the intra-class correlation coefficient using MedCalc for Windows, version 16.4 (MedCalc Software, Ostend, Belgium).