Huntington disease (HD) is an incurable neurodegenerative disorder caused by an expansion in CAG trinucleotide repeats in the huntingtin gene [1] and is characterized by neurodegeneration particularly in striatum and cortex [2]. Recently, memantine (MMT) was found to be effective in reducing aberrant extrasynaptic NMDA receptor activity delaying the progression of disease phenotypes in a mouse model of HD [3]. But how does the drug alter the brain function and connectivity is not clear. Here we applied BOLD fMRI to evaluate connectivity in HD and to understand the effects of MMT treatment and its behavioral correlates.The animal study was approved by the Institutional Animal Care and Use Committee (A*STAR, Singapore). YAC128 HD mice and their littermate controls on the FVB/N strain were used (10 WT, 9 HD and 6 MMT-treated HD). The mice were treated with MMT or vehicle for 8 months and scanned at 10 months of age. Behavioural assessments of treated mice included the novel object recognition (NOR) and the climbing test. The mice were anesthetized with medetomidine (0.3 mg/kg bolus, 0.15 mg/kg infusion) and isoflurane (0.25-0.4%). MRI was performed on a 9.4T MRI (Bruker BioSpec, Germany). 16 axial slices were acquired using gradient-echo EPI with TR=1s, TE=15 ms, thickness=0.5mm, matrix=64 × 64, and FOV=20×20mm2. Resting-state fMRI was acquired for 10 min and then evoked fMRI was conducted with electrical forepaw stimulation [4].The data was pre-processed by motion correction (SPM). Nuisance signals from the ventricles and muscles, and 6 motion parameters were decomposed by SVD to determine the singular vectors that contributed fluctuation of whole brain and then removed from the data. The data was filtered between 0.01 and 0.1Hz, co-registered to a labeled brain template, and spatially smoothed by a Gaussian kernel (FWHM=0.6 mm). The functional connectivity (FC) maps were calculated by correlation analysis based on seed points and 66 regions defined by the atlas [5].

BOLD activation by stimulation showed reduced activity in both MMT and vehicle treated HD mice though the difference was not significant (Fig.1).

The resting-state data shows that FC in caudate putamen (CPu) and S1 forelimb (SLFL) was reduced in HD mice while increased in medial parietal association cortex (MPtA), and retrosplenial cortex (RSP). MMT treatment increased all of them (Fig.2). Whole-brain correlation matrix showed that connectivity within the default mode (DMN)-like network was the same but reduced elsewhere (Fig.3). The MMT-treated HD mice had generally higher FC throughout the brain especially outside of DMN-like network. Strong correlation between FC and behavior was observed (Fig.4). The latency to climb was correlated with FC between lateral parietal association cortex (LPtA) and lateral globus pallidus (GPe), which is an area involved in the regulation of voluntary movement. NOR correlated with FC between CPu and ventral anterior cingulated area (ACAv), which is known to associate with novelty detection.

The results of forepaw stimulation reduced evoked responses though significance was hampered by large individual variation. Interestingly, functional connectivity outside of DMN, but not within DMN, was decreased by HD. With MMT treatment, the connectivity increased, suggesting MMT may improve excitatory transmission [6]. Less change in the DMN-like network may be due to the lesser involvement of DMN with the striatum. The FC relevant to the behavioral test also showed behavioral correlates. This study shows that resting-state fMRI can be a potential biomarker for detecting disease-related change and treatment response in transgenic mouse model.