Structural brain changes after rotarod training in mice

Target Audience This study is targeted at Neuroscientists interested in experience-dependent structural brain plasticity. Purpose This study investigates experience-related structural plasticity in response to a standard motor training task in adult mice. To cover a wider array of structural changes we test for volume changes based on T2-weighted images and microstructural changes based on maps of fractional anisotropy (FA) derived from diffusion-weighted MRI. Firstly, this study aims to establish a standard motor skill task that reliably elicits structural changes in adult mice. Here we chose the well-tested rotarod task [1]. Secondly, this study aims to identify a mouse strain that might serve as a background for genetic modification necessary to get at the cause of the cellular changes. Here we use C57BL/B6 mice, the most widely used background strain for genetically modified mice. Methods Behavioural Protocol: A total of 48 mice (C57BL/B6, 27 female), age 16 weeks participated. 25 mice (11 female) were trained on the accelerating rotarod (4 to 40 rpm over 5 min) ten times a day for eight days (80 trials). Mice were placed on a rotating drum and the time was measured until they fell off or the maximum trial length was reached. Controls were handled 2 min a day. Image acquisition: Brains were fixated and scanned within skulls on a 7 T, 40 cm diameter bore magnet (Varian Inc. Palo Alto, CA). A T2-weighted sequence was used to obtain images for deformation based morphometry (3DFSE, TR=2000 ms, echo train length=6, TEeff=42 ms, FOV=25×28×14 mm, matrix size=450×504×250, time=12 h) and a diffusion-weighted FSE sequence was used to obtain FA maps (30 directions, b=1917 s/mm2, 5 b0-images, TR = 325 ms, echo train length=6, first TE=30 ms, TE=6 ms for remaining 5 echoes, 2 averages, FOV=25×14×14 mm, matrix size=192×108×108, time=14 h). Brains were distortion corrected and aligned using nonlinear image registration with iterative template refinement [2,3]. All brains were rigidly registered towards a pre-existing mouse brain average. To test specifically for bilateral changes the left-right flipped version of each rigidly aligned brain was included in the following (nonlinear) registration steps. FA maps obtained with FSL's dtifit were registered separately. Statistics: Group differences between trained and control mice and correlations with average latency within trained mice were estimated using linear mixed effects modelling at each voxel with gender as a co-variate. Hemisphere was included as a random effect to test for bilateral changes/correlations. Clusters were localize using a number of atlases [4,5,6].