Non-human models of neurodegenerative diseases have potential for the identification of key pathways in pathogenesis and for the more rapid assessment of therapeutic candidates. While there are legitimate concerns about the physiological differences between the rodent and human motor systems, mice expressing the 'G93A' superoxide dismutase-1 gene mutation are a predictable and robustly-characterized model for amyotrophic lateral sclerosis (ALS). This model has provided evidence for an important role of inflammatory processes during the pre-clinical phase, a stage currently inaccessible for human study in what is largely a sporadic disease. While magnetic resonance imaging is now an established and leading modality for the identification of ALS biomarkers in humans, it can also be increasingly applied to rodent models to probe structural, functional and biochemical changes throughout the course of the disease, with additional potential to generate surrogate markers for the efficacy of therapeutic interventions. Targeted MRI contrast agents, through tagging of various cell types and even individual molecules, will deliver an era of in vivo molecular neuroimaging, with greater specificity for the most relevant pathological processes. These are potentially important steps towards the ultimate goal of human therapeutic translation.