An intercampus collaboration has resulted in the creation of new software called “Brainstorm,” designed to locate and track brain activity in both space and time with unprecedented precision.
Physicist Manbir Singh worked with Richard Leahy, a professor in the School of Engineering’s Signal and Image Processing Institute (SIPI), on integrating brain imaging data from Functional Magnetic Resonance (fMRI), and Magnetoencephalogram (MEG) imaging techniques, which measure the extremely weak magnetic field produced by brain electrical activity.
Singh, who holds professorships in both the School of Engineering’s Department of Biomedical Engineering and the Medical School department of radiology, pioneered the use of both MEG and fMRI imaging at HSC and is co-principal investigator on Leahy’s NIH-funded project.
Through sophisticated manipulation of the imaging protocols for anatomical MRI, Singh and other pioneer researchers have learned in recent years to produce images that reflect local changes in blood flow associated with neural activity.
These images track changes in magnetic properties between oxygenated and deoxygenated blood, and can be used to zero in on the sites of neural activity. Combining these spatial fMRI images with MEG’s high temporal resolution makes possible accurate localization of neural sources in time as well as space.
MEG technology has developed rapidly over the past decade and systems are available today that contain arrays of several hundred ultrasensitive detectors that can cover the entire head.
But, Leahy says, development of computing tools for processing MEG data has lagged behind the instrumentation. Leahy’s group worked on improving these tools, developing a multi-step process in which information from different imaging modalities (including the better-known electroencephalogram, or “EEG”) are successively integrated into volumetric and cortical surface representations of the brain taken from anatomical MRI scans.
Singh worked closely with Leahy’s group to integrate fMRI data into the mix to check and improve the spatial resolution of the MEG array calculations
Leahy and his collaborators, including Sylvain Baillet of USC SIPI and John Mosher of Los Alamos National Laboratory, then consolidated and automated their improved process to create Brainstorm, a Matlab-format software toolbox that is able to produce maps and movies showing changes in cortical neural activity on a millisecond time scale.
Leahy and Singh believe the new software is potentially accurate enough to be useful in medical applications, notably cases of epilepsy that anti-seizure medications are unable to control.
The brain area causing the seizures now must be mapped for surgery by using implanted electrodes, an invasive and potentially dangerous procedure.
“We hope that Brainstorm may be equally accurate, but less invasive,” says Leahy.
Brainstorm is being presented at HBM 99, the Fifth International Conference on the Functional Mapping of the Human Brain, in Duesseldorf, Germany, June 22-26. It is currently in beta testing and will be available soon for widespread use.