Resting-state fMRI after experimental hemispherectomy in rats : changes in functional connectivity and network synchronization

Introduction Hemispherectomy is a last resort treatment for children with catastrophic hemispheric epilepsy. It abolishes seizures and improves quality of life. The remarkable motor recovery after hemispherectomy, observed in both clinical and experimental settings, reflects the plastic capacities of the brain, especially at young age. Direct cortical stimulation of the healthy hemisphere in adult rats two weeks after anatomical hemispherectomy has demonstrated that the latent capacity to control motricity bilaterally was brought into function after surgery. Other studies have reported on the functional role of the basal ganglia in motor function improvement after hemispherectomy. Recently it has been shown that the healthy brain features functional complex network properties, such as dense local clustering, highly connected hubs and optimal synchronizability, that are concordant with a small-world topology. Graph theoretical analysis of brain network properties improves our understanding of higher cerebral functioning, and the functional impact of focal lesions on global network configuration. In this study we assessed the functional consequences of right hemispherectomy on network characteristics of the contralesional hemisphere in rats, at an acute and chronic time-point after surgery, using resting-state functional MRI (rs-fMRI), and interregional connectivity and graph theoretical analyses. Methods A hemicraniotomy, followed by opening of the dura mater and microsurgical hemispherectomy of the right hemisphere was performed in adult male Sprague-Dawley rats (n=7). The cortex, white matter, hippocampus and substantial part of the basal ganglia were resected with preservation of the diencephalon. In hemispherectomized rats, sensorimotor function was measured longitudinally by scoring neurological deficiency (NDS). Structural MRI and rs-fMRI were conducted in hemispherectomized rats at 7 and 49 days after surgery, and in age-matched controls (n=10). Rats were mechanically ventilated with 2.5% isoflurane in air/O2 (2:1). MRI measurements were conducted on a 4.7 T horizontal bore Varian MR system. For at least 10 minutes, end-tidal isoflurane was reduced to 1%. Subsequently, rs-fMRI was performed during 10 minutes, using a T2*-weighted gradient echo EPI sequence (35° flip angle; TR/TE=500/19 ms; 64×64 matrix; 7 coronal slices; 0.5×0.5×1.5 mm voxels; 1200 images). In addition, gradient echo 3D MRI (GE3D; TR/TE=10/2.6 ms; 20° flip angle; 128×128×256 matrix; 0.3×0.3×0.3 mm voxels) and multi-echo multi-slice T2weighted MRI (TR/TE=3000/17.5 ms; echo train length=8; 128×128 matrix; 19 coronal slices; 0.25×0.25×1.0 mm voxels) was included for registration purposes. Images from the rs-fMRI time-series were non-rigidly aligned to the high-resolution GE3D image. The GE3D anatomical images were registered non-linearly to a 3D model of a stereotaxic rat brain atlas. The inverse b-spline transformations were estimated, which enabled mapping of atlas regions-of-interest (ROIs) to the original functional time-series space. Rs-fMRI preprocessing included spatial smoothing (1mm FWHM kernel), rigid-body motion correction, band-pass filtering (0.01More