Scientists and clinicians interested in new MRI techniques for epilepsy that could improve seizure localization and improve our understanding of the underlying metabolic changes.Epilepsy is a chronic medical condition characterized by unpredictable, recurrent seizures. The balance between normal and abnormal neuronal function in specific brain regions afflicted with epilepsy is determined by the maintenance and manipulation of neuron membrane potential through sodium-potassium exchange and voltage-gated sodium membrane channels1. Voltage-gated sodium channels in fact are the most commonly recognized cause of genetic epilepsy and a common target for seizure prevention drug therapies2. The detection and localization of sodium tissue abnormalities in patients with epilepsy may have potential to improve seizure localization, identify effective pharmacotherapy and/or provide prognostic information for individual patients. Here, we report initial results evaluating an efficient approach for performing 23Na MRI at 3-T in in clinically acceptable imaging times. Two adult and one pediatric patient with chronic localization-related epilepsy (ages 13-49 years old, 1 male) consented to participate in this study under institutional review board approval. Eight adult healthy control subjects also were imaged. Sodium MRI scans were performed on a clinical 3-T scanner (PRISMA, Siemens Healthcare, Erlangen, Germany), with a custom-built 8-channel dual-tuned (1H-23Na) Tx/Rx head coil3. The twisted projection imaging (TPI) sequence4 was used for data acquisition with FOV=220mm, matrix size=64, 3D isotropic, RF duration=0.5ms, TE/TR=0.3/100ms, flip angle=90°, rings=28, p=0.4, averages=4, and TA=10.3min. Total sodium MR images were visually inspected in conjunction with conventional proton T1-weighted structural imaging by a board-certified neuroradiologist. Relative signal intensity was inspected along a line placed on axial images perpendicular to the falx at the level of the centrum semiovale, two slices cranial to the lateral ventricles to avoid ventricular CSF volume averaging. In particular, signal intensity in the region suspected to represent the anatomic epileptogenic focus by semiology, EEG and/or conventional MRI was examined relative to CSF signal in the interhemispheric fissure and compared to homologous brain regions. A custom-built 8-channel dual-tuned (1H-23Na) Tx/Rx head coil offers improved signal-to-noise for obtaining sodium MRI at 3-T in epilepsy patients. We successfully imaged 3 subjects in this preliminary study. A pediatric patient demonstrated a large malformation of cortical development in the left posterior superior frontal gyrus, paracentral lobule and medial precentral gyrus. Semiology and EEG results were concordant with this abnormality. We observed a qualitative decreased sodium concentration in the malformation compared the homologous region in the right cerebral hemisphere, which is possibly explained by changes in cell packing within the malformation. Two additional patients with suspected temporal lobe epilepsy did not have detectable lesions on conventional MRI – there were no striking abnormalities on sodium MRI.This pilot study demonstrates the feasibility of using sodium MRI for imaging epilepsy patients in clinically acceptable imaging times. The SNR and image quality are high (Fig. 1B) – decreased sodium associated with an MRI-positive cortical malformation is readily obvious both on the image and when signal intensity is mapped along a line through the two cerebral hemispheres. In contrast, control subjects demonstrate highly symmetric sodium concentrations in the two cerebral hemispheres. In two MRI-negative patients, qualitative interpretation by a neuroradiologist did not detect obvious sodium abnormalities. For quantitative sodium imaging, spatially-varying modulations of the B1+/- fields associated with Tx/Rx array coils need to be corrected and were not performed in the current analysis of the data– we present a potential solution elsewhere at this meeting. Quantitative assessment of sodium MRI may prove to be a helpful biomarker for seizure localization, identifying effective pharmacotherapise or provide prognostic information for individual patients. In this setting, combination with a concurrently acquired multiple-quantum-filtered sodium scan5 could prove useful to improve localization of seizure foci. Future efforts will continue to study these techniques in epilepsy patients with clearly established seizure localization and lateralization and improving the analysis via quantitative evaluation of the pixel data against an atlas derived from age-matched normal human volunteers.