7 Tesla Clinical Functional Magnetic Resonance Imaging

Brain surgery is the best therapeutic option for many patients with epilepsy and brain tumors. Functional Magnetic Resonance Imaging (fMRI) is used to define brain regions which must be left intact when patients undergo surgery to remove tumors or regions of the brain which cause epileptic seizures. fMRI works by measuring changes in a series of images which come about when groups of neurons fire. The images are acquired in rapid succession while patients perform tasks such as moving the fingers or talking. Ultra-high field (7 Tesla) MR scanners are more sensitive to the changes in image signal caused by neuronal activation than the 3 Tesla scanners which are in routine operation. This higher sensitivity means that 7 Tesla fMRI has the potential to generate more reliable and accurate maps of brain function. These will improve the quality of resection, increasing both life expectancy and quality of life by ensuring that patients retain the ability to move, speak and remember after surgery. Clinical fMRI at 7 T is in its infancy, and some problems need to be solved before its full potential is realized. We propose to use a new fast fMRI sequence called IDEA EPI to eliminate a common artifact called Nyquist ghosting, which particularly affects images acquired at 7 Tesla. This will improve the signal-to-noise ratio of images and the ability to detect activation. As a second aspect of the project, we will continuously assess the magnetic throughout the brain during the fMRI measurement, allowing us to remove image distortions. This will ensure accurate mapping of brain function for neurosurgeons. Finally, an exploratory data analysis method, Independent Component Analysis, will be applied to isolate activation in language tasks which have proven difficult with conventional analysis methods. Characteristics of activation will be identified which will enable the clinician to rapidly isolate activation related to the task. Finally, these methods will be translated to clinical 3 T protocols which are already routinely used presurgically. The developments outlined in this proposal will increase the sensitivity, specificity and localization accuracy of clinical fMRI, so that full advantage can be taken of the potential of the technique to preserve vital brain function in tumor and epilepsy patients who need to undergo an operation. This will reduce post-operative complications significantly, benefitting both patients and society as a whole.

This project has contributed to improvements in the sensitivity and accuracy of ultra-high field (7 Tesla) functional Magnetic Resonance Imaging (fMRI) in presurgical planning, in which fMRI is used to map vital brain function in patients with brain tumors and epilepsy patients who need to undergo an operation. Brain surgery is the best therapeutic option for many patients with epilepsy and brain tumors. fMRI is used to define brain regions which must be left intact when tumors or regions of the brain which cause epileptic seizures are removed. Blood flow changes when neurons are active, and fMRI can detect these changes in a series of rapidly-acquired images, allowing us to see which areas of the brain are active when patients perform tasks such as moving the fingers or talking. Ultra-high field MR scanners are more sensitive to the changes in image signal caused by neuronal activation than the 3 Tesla scanners which are in routine use. This higher sensitivity means that 7 Tesla fMRI has the potential to generate more reliable and accurate maps of brain function. These can provide more sensitive and more accurate information to the brain surgeon, increasing both life expectancy and quality of life by ensuring that patients retain the ability to move, speak and remember after the operation. Clinical fMRI at 7 T is in its infancy, however, and some problems have needed to be solved to realize its full potential. In this project, we have developed methods for correcting distortions in the activation maps so these are more accurate. These corrections work by calculating the magnetic field from the images themselves so that the distortions which these create can be corrected. We have also developed methods for identifying neuronal activation in images even when because of pathology the signal response is very untypical. Rather than looking for the same time course of signal changes which we observe in healthy subjects, we search only for patterns of signal changes which are the same in each execution of a task, even though these might be unique to each patient. This project has led to 14 publications in top peer-reviewed journals and the patenting and development of tools which are being distributed to other neurologists, neurosurgeons and medical physics researchers worldwide. Most importantly, the improvements we have made in ultra-high field clinical fMRI will help reduce post-operative complications, benefitting both patients and society as a whole.