Epilepsy and math

Anatoly Buchin
Jul 4, 2017 · 4 min read
Image source: Parvizi lab


Recently I have started doing epilepsy research again due to my job, yet I did not know that studying the disease could tell us a lot about our brains. Recently in our institute there was a talk by Josef Parvizi from Stanford. He was talking about mapping of brain functions in epilepsy patients. It was so curious talk, so I would like to share my thoughts here.

Usually before doing surgery to remove the epileptic region patients there is an opportunity to record the activity of their brains without anaesthesia. It might sound a bit scary, but it is not such a big thing. In fact, our brains by themselves do not have pain receptors on their surface, which allows to put the electrodes without directly on it. From medical perspective these recordings are used to localise the seizure focus to make surgery more efficient. Before cutting out the piece of the brain responsible for epilepsy you should know precisely where it is.

One of the methods to record the activity directly from the brain surface is called the electrocorticography or EcoG. In this case the doctors install the grid of electrodes on the cortex surface involving one or two hemispheres. Then the patient is asked to move their arm or think about something. It allows to determine the activity associated with a particular action. In fact, neurons in the brain are always active, therefore it is necessary to make the brain do particular action. Then you can see the activity beyond the baseline and correlate it with this action. This process is called brain mapping and there are multiple brain maps that could be generated depending on the method being used, example 1 and example 2. Often these maps have a lot of intersections because there is no perfect method that could show everything at once.

For scientist it is an exceptional opportunity to record the activity of human cortical neurons on the detailed level. If not epilepsy, it would not be possible to do any of these recordings. Each EcoG electrode records the activity of a few millions of neurons on the surface of about 5 mm2 at very high temporal resolution. Other methods such as EEG or MRI do not have such a good temporal and/or spatial resolution. This does not allow to record the activity of human cortex so precisely.

In the experiment of Josef Parvizi they installed the electrode grid in various regions, including the area of intraparietal sulcus and other adjacent areas. This area is special, it is being activated when human thinks, speaks or sees the numerals. But in the previous experiments it was not possible to record from this area so precisely during natural conditions. It is actually a good question what natural conditions for humans are. Spending a day in a hospital with the electrodes installed into your brain is hardly natural. But if you want to record from human cortex at that precision, there are no other alternatives.

After recording of the activity the electrical signal has been processed and decomposed into different frequencies. In particular the frequencies between 80 and 200 Hz were found to be associated with numerals. See the video, each time when the patient is talking about numbers during her telephone talk, it leads to activation of the intraparietal sulcus, which is shown in red.

Moreover the further experiments have shown that the same area is active when the human sees the numbers and makes the computations. For example, when there is first 3 then + and 5 on the screen and then the response 8. In this case all these numbers could be seen in the activity of the cortex. Specifically what Josef Parvizi have shown is that the same area is being activated during the natural speech.

It is important to mention that brain mapping is very different from the pseudo-scientific belief like phrenology. In this belief people used to think that all parts of human soul are precisely located in various parts of the brain. Something like friendship in the occipital lobe, memory is in the frontal lobe etc. In reality everything is much more complicated that our naïve believes. Every part of the cortex is interconnected with many others. The information processing in the brain look more like an internet web rather than a library where everything is organised in shells. The brain imaging methods could at their best localise the most active brain area involved into the particular task, involved into numerals processing for example. In other words, every active region is just a part of the whole network, but if it is active it allows us to measure it and correlate with the function.

Another examples of the important network nodes involved into speech are the zones of Broca and Wernicke’s. But if you look into the brain activity map during listening to speech, you will see that the whole cortex is being active, not just Wenicke zone. But this would be another story.

In fact connecting brain functions with structure is a holy grail for neuroscience research. People are trying to solve it for several thousands of years and apparently it is not simple and straightforward. One part of this complexity is the distribution of functions over the cortex. But we hope that new brain imaging technics and data analysis tools would allow us to understand what our inner world is about in terms of brain activity patterns.

Written by

computational scientist at Allen Institute, eternal learner

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade