von Ellenrieder, N.,Dan, J., Frauscher, B., Gotman, J. “Sparse Asynchronous Cortical Generators Can Produce Measurable Scalp EEG Signals.” NeuroImage, vol. 138, Elsevier Inc., 2016, pp. 123–33, doi:10.1016/j.neuroimage.2016.05.067.
Dr. Jean Gotman, engineer, professor, and neuroscientist at the Montreal Neurological Institute, is the director of his laboratory and namesake, The Gotman Laboratory. I recently had the pleasure of sitting down with Dr. Gotman, a pioneer in the analysis of the electroencephalogram (EEG) for epilepsy, and Dr. von Ellenrieder, an expert in signal processing, to discuss their paper entitled “Sparse asynchronous cortical generators can produce measurable scalp EEG signals.” Their lab focuses on the analysis of the EEG and sleep, mechanisms of generation of epileptic discharges in EEG, seizure generation and the treatment and study of epilepsy. As Dr. Gotman explains, “the lab has been involved in studies using EEG and studying where the EEG signal comes from for many years, particularly in the context of the evaluation of patients with epilepsy. We use the EEG as a primary tool because it reflects neuronal activity in the brain and it shows very specific abnormalities in patients with epilepsy.”
Both coming from electrical engineering backgrounds, Dr. Gotman and Dr. von Ellenrieder are now driven to understand “what goes on inside the brain and how it relates to what you see on scalp EEG.” Currently, their work aims to improve our understanding of the brain’s neuronal activity, and how to interpret various representations of this activity for the study of epilepsy, sleep, and other phenomena.
This particular paper questions a fundamental basis for scalp EEG interpretation which dates back many decades and was formulated in 1985. Since then, it has been commonly accepted that scalp EEG activity, which is a result of a weighted average of cortical generators, is a result of the synchronous activation of a smooth patch of cortex. In this study, they perform simulations to compare scalp EEG activity generated by both synchronous and asynchronous, as well as in-phase and in random phase generator activation. Their results show that sparse asynchronous cortical generators can produce measurable scalp EEG activity and therefore, standard interpretations of scalp EEG might be based on an oversimplification of the true cortical activity.
How it came about
Having spent years and years studying all forms of EEG, it was only a matter of time before the Gotman Lab found peculiarities and inconsistencies in standard EEG analysis and, as a result, started to revisit the accepted rules, changing how we normally think about EEG. In a previous study in which intracranial EEG (iEEG) was recorded and analyzed in conjunction with scalp EEG, the team states that they “observed something, which to [them] was surprising. [They] observed a lack of synchrony between the intracerebral channels in different regions of the brain during sleep where this particular pattern of sleep spindle is very prominent.” So what does this mean? While the idea that a large region of synchronized cortical activity is necessary to observe oscillations in scalp EEG is a widely accepted assumption, Dr. von Ellenrieder states that “even though there was no synchrony intracerebrally, you could still see the same kind of oscillations on the scalp.” This, along with similar observations reported in the literature, is what motivated the team to challenge this common assumption and ultimately demonstrate, through simulations, that scalp EEG oscillations could also come from asynchronous cortical generators.
“Question … This is what we do”
When asked about the importance of going back and questioning long-standing assumptions or theories before building on them, Dr. Gotman responded with: “You should always question. This is the basis of what we do.” He goes on to say that, in his opinion, we probably don’t question quite enough the things that we believe and the reason being that “it’s difficult, it’s not comfortable and often it’s not necessarily glorious. You know, we live with writing grant applications that people have to find exciting and very often, premises are not questioned, not because they have been well established, but because that’s the way it seems to be.” Of course, if we questioned every scientific theory before building upon it, science would advance at a glacial rate. However, when there is a reason to question, and one does not, nothing which rests on these inconsistent assumptions would be true or worth doing. This was exactly the case for Dr. von Ellenrieder after he began gathering evidence, before simulating his demonstration. He “tracked the origin [of the theory] to an article from Gloor (1985) which says ‘if, then’, but it doesn’t say the reverse! With time, and people citing and citing, they kind of built up the idea ‘Ok, this is how it has to be.’”
The impact of the paper
Dr. Gotman and Dr. von Ellenrieder emphasize the fact that their study is, by no means, a proof that scalp EEG oscillations are a result of non-synchronous cortical activation, but rather a demonstration that it is theoretically possible. They also acknowledge that it is possible that this situation is very unlikely to occur in a real brain and that perhaps the initial assumption does stand. As Dr. von Ellenrieder explains, “this study is not discrediting everything that was right before. It’s just trying to build awareness of the many possibilities that we should be aware of.” In other words, scientists should always keep in mind that there could be an alternate explanation for their results, whatever it may be. When asked what they hope readers take from this paper, both authors agree that “the most important outcome is that we have a different way of understanding the scalp EEG. Now, when we look at the scalp EEG and we try to understand what’s happening inside the brain at that time, we think differently about what’s happening. We have made this demonstration and hopefully, this will alter the way people think when they do EEG.”
What now?
So what is the Gotman Lab working on since publication of the article? One of the biggest challenges associated with studying the functionality of the human brain is our very limited window into what is going on inside. One of the reasons for this being that intracranial EEG is used exclusively in patients with pharmaco-resistant epilepsy. iEEG is a measure which allows us to localize the source of their seizures and plan for surgical resection of the problematic region, also known as the epileptogenic region. This, however, is the only justification for implanting intracranial electrodes in a person. It is an invasive procedure with some risk and is not done unless it is the only promising solution for treatment of the disease. So, the MNI, in collaboration with other centers, got the idea “that you can use these implanted electrodes to study normal brain function when you have no other way to study with this level of precision. So, what [they] have been working on for several years now, and recently completed, is what we call an atlas of the normal intracerebral EEG.” Obviously, even implanted patients present limitations, the biggest one being the covered region of the brain. Implanted electrodes are generally concentrated in the epileptic or abnormal region of the brain but, in many patients, a small portion of the contacts are implanted in what is called a normal brain region. What they did is collect continuous data from over 100 patients in 3 centers (2 in Montreal and 1 in France). Dr. Gotman explained that they collected a few contacts from each patient which were in normal regions of the brain and “combined one minute of wakefulness from each contact distributed in different positions into one atlas of a human brain.” The team just recently “published the atlas on the web so that it will be available to the world to look at.”
As impressive as this project is, it is not the first time that Dr. Gotman went above and beyond to bring the tools for EEG analysis to the hands of researchers around the world. In 1986, he formed Stellate, a company which developed and sold equipment and software for EEG recording and analysis around the world.
The future of epilepsy
Being a new researcher to the field myself, I could not help but wonder what the experts in epilepsy research are passing down to the next generation of brain explorers. The answers I received were promising:
“I think that people need to better understand what you can and cannot do with source analysis. All the measures that we do today are not the measures of what we are interested in, they’re surrogates. You look at an MRI image and see a lesion. This is not the epilepsy, this is a physical lesion. Then we look at fMRI, which is the BOLD signal and, again, not the electrical signal, which is the epileptic discharge. What is most important is that you need to start with knowing what you don’t know. And then, hopefully, progress from there.”
Asked about the potential of using nanoparticles to directly image current, he replied:
“If you could do the imaging of the currents in MRI, then you don’t have to make the assumptions that you have to make with scalp or source EEG. If you could make a direct measurement of current everywhere in the brain, that would be a real breakthrough.”