Next up is Dr. Tarapour, who will be discussing MEG imaging of recurrent gliomas, reveals functional plasticity of language networks. Hello, good afternoon. Thank you for being here and for hearing me talk. So we're going to shift gears a little bit and talk about glioma. My interest is in studying how the brain reorganizes itself over the course of time in patients with low-grade gliomas. A little bit of background. Language networks, as we know now, thanks to the work of Dr. Berger, my mentor, Dr. Defoe, and several other people, we know that these reorganize over time. And insults tend to encourage plasticity and reorganization. So recurrent low-grade glioma is actually an ideal condition in which to study the reorganization of language cortex. So the objectives of this study were to characterize how the language system itself reorganizes in patients with glioma and to see what those patterns look like, to see how does hemispheric language laterality change with the presence of a tumor in one hemisphere or the other. A little bit of background on how we do this. Language laterality, as we know, has been studied for decades, most originally with the WADA test. More recently, we have started using a technique that involves MEG, magnetoencephalography, which is a non-invasive test that looks at the function of the brain. I'm going to go into that a little bit more in the next couple of slides. But using MEG, we're able to calculate basically a language laterality index, which gives us a very good idea of which hemisphere is language dominant. We use this now exclusively in place of WADA testing at our institution. We do not do WADA testing any longer. So MEG, or electromagnetic brain imaging, is non-invasive. You get high temporal resolution down to the millisecond, also known as magnetic source imaging. This is what an MEG scanner looks like. It looks somewhat similar to an MRI without the huge magnet. The patient sits in that chair, and the sensors are contained within that large cylinder that sits atop the patient's head. The sensors are comprised of these little devices called SQUIDs, which are superconducting quantum interference devices. And they're able to detect infinitesimal magnetic fields that are generated from the electrical activity of neurons firing. So they can detect all the way down to the femto-tesla, which is extremely small magnetic signal. So MEG-based brain mapping has been around for a while. We have been working on it at our institution for about 15 years. These are some of the original papers. Using very complex mathematics, one can actually identify sources within the head of various patterns of magnetic fields. And we're now using it to identify both language laterality and to look at functional connectivity and at motor systems within the cortex. This is how the language laterality is actually calculated. We have a laterality index that's LI, which is calculated basically as the absolute difference of the left and the right activity. A language laterality index greater than zero is left-sided and less than zero is right-sided. And when you're in that middle minus 0.1 to 0.1 range in the middle there, we consider the patients to be bilateral. This is a language task that we do. It's a very simple verb generation task where there's this auditory noun given as a stimulus. And then the patient responds with a verb that is related to the noun that has been used as a prompt. So this is our overall composition in this paper. We have 73 patients who've been followed prospectively. Overall, the gender balance was pretty even. Tumor side, the majority of patients were left-sided. Tumors formed about 34% of the patients. Average age was about 42. Most of the patients were right-handed with a couple of ambidextrous patients in there. And similar to the general population, the language dominance pattern was about 85% left, about 5% right, and about 10% bilateral. The tumor locations were as described here in this table. And the majority of patients had either grade 2 or grade 3 gliomas, with about 20% having glioblastomas. This is a scatterplot of the actual change in language laterality index versus the initial. So I'm going to go through this chart a little bit in detail. On the x-axis, we have the initial laterality index. So if you see, most of the dots here fall on the right side of the chart, which is the positive side of the chart. That's between 0 and 1, indicating that most folks have a left-sided predominance in their laterality. The y-axis here indicates the change. So a positive shift indicates a left-sided movement, and a negative shift indicates a right-sided movement. So the folks that basically fall in this upper right quadrant are people who had a initial left-sided laterality and who became, if you will, more left-sided over the course of the time that we were following them. You'll notice that the majority of patients actually fall on this right lower quadrant, indicating that these folks had initial left-sided laterality but became more right-sided as we followed them. Similarly, amongst the few right-lateralized patients that we followed, the majority of them shifted leftward over the course of the study. We also broke this thing down by the sidedness of the tumor. So we were asking the question, does the location of the tumor influence the direction of the laterality shift? And you'll see this data aggregated a little bit later on, but you can appreciate that there is more blue circles here on the right lower quadrant, and there are more red diamonds up here in the right upper quadrant, indicating that folks who had left-sided language dominance and left-sided tumors tend to move to the right, and folks that had right-sided tumors actually moved in the opposite direction. This is a case-by-case scatterplot, again, demonstrating the change in laterality but with regard to the laterality of the tumor. So is the tumor ipsi or contralateral to language dominance? Here are a couple representative cases. So this is a patient who had a left-sided language laterality index initially. That's the top line here, Visit 1, and you'll notice that the left-sided blue areas of interest represent the activation upon completion of this language test that they were doing. And in the bottom row, you'll see that the second visit, there's less blue on the left side and more blue on the right side. And so what we're seeing here is an actual visual representation of the movement of language away from the tumor towards the contralateral hemisphere. We saw a similar effect in patients who were right-lateralized. So if you focus here on this top box, for example, you see a pretty large activation here in the frontal area. On the second visit, that area has gone away, and you see more extensive activation in the similar area on the contralateral side. So that's movement from right to left. And then we see a few people who actually went from left-sided to a more bilateral representation of language. So again, we see that left-sided blob here becoming smaller. You see the left-sided area here becoming smaller, and you see a greater representation on the right side. So when we look at this aggregated, we get this result here, which basically shows that when your language is ipsilateral to your tumor, which is this first bar, the laterality shift is significantly greater than when language and tumor are on contralateral hemispheres. So this, in some ways, is a intuitive prediction. One would expect that the brain would reorganize function away from tumors and toward a more healthy brain, and we're actually seeing quantitative evidence of this. So in conclusion, in language function at least, there is significant plasticity that can take place. These lesions grow slowly, and the brain actually has the ability and the time to move away from the affected hemisphere. This effect is seen both in left- and right-sided hemispheric dominance and left- and right-sided lesions. So it seems to hold up across lateralities. Thank you very much. Thank you.

language reorganization

patients

recurrent low-grade gliomas

hemispheric language laterality

MEG imaging