Part I Experimental Models of Brain Disease: MRI Studies Louise van der Weerd 1,2 , David Thomas 3 , John Thornton 4 , ManKin Choy 1 , and Mark F. Lythgoe 1 1 RCS Unit of Biophysics, Institute of Child Health, University College London, London WC1N 1EH, UK; 2 Medical Molecular Biology Unit, Institute of Child Health, University College London, UK; 3 Wellcome Trust High Field MR Research Laboratory, Department of Medical Physics and Bioengineering, University College London, London WC1N 3AR, UK; and 4 Lysholm Department of Neuroradiology, National Hospital for Neurology and Neurosurgery, UCLH NHS Trust, London WC1N 3BG, UK Introduction MRI of intact animals was first employed in the 1970s. With the advent of dedicated animal scanners, the use of MRI for in vivo animal studies has mushroomed, and currently thousands of papers are published every year. Though a wide variety of laboratory animals is used, including monkeys, songbirds, and piglets among many others, the majority of MRI studies concerns rodent models of disease. In particular the use of mice has increased substantially over the past few years due to the development of a large range of transgenic mouse models. This chapter is dedicated to the use of MRI in small animal models of brain disease, and will de- scribe some of the practical issues surrounding the use of MRI, and review the role of MRI in investigating the pathophysiology of the most common neurological disor- ders. Practical Issues Signal-to-noise The size difference between humans and rodents is about a factor 10–20. To achieve a spatial resolution equiv- alent to that obtained in clinical imaging, this means that the voxel volume has to be reduced at least a fac- tor of 10 3 , implying in turn that the sensitivity has to be increased correspondingly to obtain a similar signal-to- noise-ratio (SNR). Much of the sensitivity can be gained by using RF receiver coils that are optimized for the size of the animals. Further sensitivity can be obtained by increasing the magnetic field strength; currently, hor- izontal small animal MRI systems with field strengths up to 11.7 T are commercially available. A third pos- sibility to improve SNR is to increase the acquisition time up to several hours, which currently is used mostly for anatomical imaging of either in vivo or fixated ani- mals. Probes To acquire data within the MR scanner, the animal should be immobilized in a non-magnetic probe with ear and bite bars, allowing the head to be positioned reproducibly rela- tive to the gradient coils to attain the correct imaging slice or region of interest. Several MR-compatible stereotaxic frames have been designed to permit reproducible posi- tioning in the scanner for repeat measurements. In MR magnets with a small bore, the use of ear and bite bars may not be possible, in which case the skull may be glued directly to the probe to reduce movement artifacts [1]. Temperature control is commonly attained through the use of either warm air blown over the animal, an electri- cally heated mat, or a warm water jacket or mat. Animals are readily anaesthetized via administration of anesthet- ics through a nose cone for spontaneously breathing ro- dents, intraperitoneal injection, or mechanical ventilation (see below). Physiological monitoring of cardiac and res- piration rate may be performed simultaneously using im- planted chest electrodes [2] or conventional EE electrodes in combination with an air pressure cushion to monitor respiration. Blood pressure is conventionally monitored using an intra-arterial line or, more recently, a piezo- electric pulse transducer [3]. Intravenous lines may be inserted for blood gas measurements or administration of drugs. With the recent development of numerous transgenic mouse models, the demand for high-throughput pheno- typic and physiological screening of large numbers of an- imals has soared. Therefore, probes have been designed that can hold several mice at the same time, with indi- vidual decoupled and shielded transmitter and receiver coils, anesthetics, and physiological monitoring [4]. Up to 16 mice can thus be measured in a single session. To record EEG signals during an MRI experiment, sil- ver or tungsten electrodes can be implanted. Alternatively, non-magnetic graphite electrodes, which reduce suscep- tibility effects, may be used EEG [5]. Gradient switching causes large artifacts on all EEG measurements, and data C 801 Graham A. Webb (ed.), Modern Magnetic Resonance, 801–821.  2008 Springer.