Understanding the source and network of signals as the brain functions is a central goal of brain research. Now, Carnegie Mellon engineers have created a system to obtain high-density EEG images of the origin and pathway of normal and abnormal brain signals.
Bin He, head of the Department of Biomedical Engineering at Carnegie Mellon University, and his colleagues are working on a core National Institutes of Health initiative called Brain Research by Advancing Innovative Neurotechnologies (BRAIN).
The group’s work with high-density EEG, published in the journal Nature Communications,1 takes an important step toward realizing one of the central goals of the BRAIN initiative: “to produce a revolutionary new dynamic image of the brain that, for the first time, shows how individual cells and complex neural circuits interact in both time and space.” In other words, develop a way of observing how the brain does what it does.
“Electroencephalography (EEG) has been used for decades to track brain signals,” explained Shumin Wang, Ph.D., director of the NIBIB program in Bioelectromagnetic Technologies. “Dr. He and his team developed a more powerful, high-density version of EEG that can track brain signals in much larger areas of the brain than was previously possible. Artificial intelligence is then used to identify where these stronger signals originate and travel through the brain, with great precision.”
The research team tested the high-density EEG system in patients with epilepsy. Individuals experience seizures caused by errant pulses of brain activity, known as the epileptogenic zone. Many patients’ seizures can be controlled with medications. For those who are resistant to medications, surgical removal of the epileptogenic zone is a clinical option. The ability of high-density EEG to accurately identify the epileptic source would allow for greatly improved brain surgeries to efficiently remove only the problem area and preserve surrounding brain tissue.
The team tested their method, called FAST-IRES for spatiotemporal iteratively reweighted edge sparsity, on 36 epilepsy patients undergoing preoperative testing. Non-invasive FAST-IRES brain recordings were obtained over several days using a special high-density EEG array. The fact that FAST-IRES is non-invasive is extremely important, given that standard preoperative testing requires an invasive EEG, which is necessary to identify the epileptogenic region but carries increased risks of infection, complications, and costs.
In the 36 patients, the FAST-IRES method analyzed more than 1,000 EEG spikes and 86 seizures occurred. The high-density recordings were then analyzed using the artificial intelligence piece of the FAST-IRES method and compared to standard invasive preoperative EEG testing and surgical findings.
The FAST-IRES method was extremely accurate in identifying the position and extent of the epileptogenic zone in patients, which was confirmed with the surgical data. “Our results clearly demonstrated that FAST-IRES could identify the epileptogenic zone with high precision, using high-density non-invasive scalp EEG recordings,” He said. Study collaborators at the Mayo Clinic are considering implementing the system in the future.
The team believes that the FAST-IRES technique could be used for the diagnosis and treatment of Alzheimer’s, Parkinson’s, stroke and even depression. “It is extremely satisfying to see our work having a significant impact on meeting one of the core goals of the BRAIN initiative,” He said. “In addition to treating patients, we hope our work will help researchers make significant advances in our understanding of human neuroscience.”
The research was supported by grant EB021027 from the National Institute of Biomedical Imaging and Bioengineering (NIBIB) and grants from the National Institute of Neurological Disorders and Stroke, the National Institute of Mental Health, and the National Center for Advancing Translational Sciences.
