Pleasure to be here. A wonderful course, highly successful, so congratulations to all of you. I've missed a good portion of the earlier parts of this course, so I apologize if some of this is redundant to you. I have no pertinent disclosures relative to this course. These are some of the abbreviations I'll be using during my presentation, and I know all of these have been introduced thus far in this course. But I want to spend a few minutes on Magnetoencephalography, or MEG. We have the world's largest experience with MEG, particularly when applied to children who have undergone surgery. Just a reminder, in the presence of a magnetic field, an electric current is generated that's perpendicular according to the right-hand rule in physics, and it detects this electrical disturbance that can sometimes find a seizure focus. It's also very important for us to brain map. It has the best spatial and temporal resolution of all the imaging modalities. Here you see MEG. Here you see the other modalities. So we've become quite familiar with this. So we've linked this to navigation over time, where you define what's called a cluster on MEG, as you can see here, a small cluster, scatters, where there's really no well-defined MEG focus. And we published this not too long ago on our usage of this in the pediatric population. One of the criticisms about MEG is the fact that it's mostly interictal data that you're getting. However, occasionally you can get an ictal MEG, and I'm showing you an example here of an ictal MEG in a patient who had resmus encephalitis, and there you can see the very tight spike focus that's detected on MEG and hemispherectomy that was done accordingly in this patient. We've published also on the use of MEG to attain these spike cluster excisions. This child had a post-traumatic injury. You can see the spike cluster that's located just in front of the area of encephalomalacia that's depicted there. We went ahead, did surgery, resected that. This patient's seizure-free off medications. We didn't have to do grid excision. We didn't have to do neuromonitoring at that time. So it's useful in several instances where you can do guided resections of an MEG spike cluster to help us with these particular cases. We're also using MEG now for motor mapping. We can do it for visual field mapping. We can do the language mapping, auditory cortex mapping on MEG. But this example of motor cortex mapping by MEG, and soon the entire homunculus can be mapped by using MEG and its capabilities. So just in conclusion for MEG, this aspect of the talk, it's useful for many surgery candidates, especially those with epilepsies and tumors. We use it to provide functional maps linked to neural navigation. We can excise these very tight spike clusters to get excellent outcomes. And it's helped us, even in cases that are quite severe, like status epilepticus. And I didn't show you the data, but we've operated on several children who've had status epilepticus based on the MEG data. These are the modalities that we use at SickKids Hospital. We have a very good team of monitoring neurophysiologists. This is Sam Stranzas, but basically, we are very fortunate to have this team available. And these are all the techniques that we use. I know you've already talked about phase reversals. I won't emphasize that point. We also use direct cortical mapping with stimulators. We choose to use the monopolar train-of-five technique as opposed to the Penfield technique for reasons I'll talk about. You'll recall the Penfield technique is relatively low frequency compared to the train technique, which is a higher frequency form of stimulation. And the reason why is because we looked at incidents of seizures in our series at SickKids, and we found that the seizure incidents when using stimulation using the Penfield technique was higher than it was using the train technique, although other series I acknowledge are different. We found it safer to use the train technique, so we've not been using the Penfield technique for some time. This is our technique for continuous motor monitoring and using the strip electrode. And as we're resecting lesions, we have continuous information that's coming back to us from our neuromonitoring neurophysiologists. It was mentioned in the previous talk by Dr. Wu, I think, that what happens if you get a decrement in activity, and this is what we would do. So we're operating near the motor tract. We get an indication from our neurophysiologist that there's absence or decrement in the recordings. This was, in this case, due to excessive brain retraction, and so we released the retractor, and you can see the signals coming back a short time afterwards. But I can tell you, and you all know, those of you who've been in this situation, that it's a long wait at times when you're very close to the cortical spinal tract. We're also using subcortical stimulation. We use a threshold of four milliamps as a warning for being near the motor tracts and some recent articles that relate to the use of subcortical stimulation. Here's a case that I did recently, a child with quite extensive polymicrogyra. You can see how this insular cortex here is quite profoundly disturbed. And as we were resecting all of this contained within this pink or violet circle here, we were approaching or getting close to the cortical spinal tracts, and so having subcortical stimulation available was extremely helpful to us in doing this extensive resection of polymicrogyra. What do you do if you have motor changes? So we've already talked about pausing, waiting, allowing for the amplitudes to return, readjusting retractors, application of pharmacotherapeutic agents, irrigating the field, and then possibly terminating what you're in the middle of doing for fear of inducing a long-term neurological deficit. Because I think we would all agree that a disabling new post-operative deficit, such as paresis, may cancel out or outweigh the benefit that's achieved from a maximal surgical resection of a tumor or a seizure focus. So over the years, we've published quite extensively on cortical mapping in children, and just to show that these techniques are available and are used in the pediatric population here, some of our publications in this field. I just want to spend just a brief moment on awake craniotomy in children. We've done that in children as young as age four. For that, you need a very good neuropsychologist, a good neuroanesthesiologist, good neuromonitoring, and all the team associated with that to do awake craniotomy in children. Putting it all together, we use invasive subdural grid monitoring on occasion for epilepsy cases, sometimes for tumor cases, and this is all the information you can map onto the surface of the brain. And I'm showing this because I think we were among the first to report the use of digital camera capturing. You've seen all these wonderful images during this course, but back in the 90s, we reported on this, and we've used this then to add on to how we do the resections in the pediatric population. I don't know if you've talked about high-frequency oscillations in this course thus far, but I thought I'd spend just a moment talking about this, because we've been using this quite frequently now at SickKids Hospital. Just to remind you, the normal brain rhythm is on the order of 30 to 50 hertz, and so I'm talking about fast frequency, so over 200 hertz types of oscillations detected by invasive monitoring, needing special mathematical formula to translate that information. But then that is thought to be an indicator of where a seizure focus starts, and the number of electrodes that are expressing these high-frequency oscillations correlates with seizure onset zone and frequency. So I'll just show you some data here that hopefully this video will play. So it'll show in real-time cartoon animation of how the high-frequency oscillations occur in this area of the brain, making it suspicious for the seizure onset zone. And so with this cartoon animation, then we take this information into the operating room. And I'll show you a case that depicts this. So just some of our publications on high-frequency oscillations that you can see and how this is useful, particularly in determining seizure foci in patients with epilepsy. Some of the limitations, we don't understand the mechanisms by which HFOs are produced. The normal cortex on its own can have HFOs produced. And how HFOs actually cause seizures is unknown, and long-term data follow-up are not yet available. But this is the current model on epileptogenesis and high-frequency oscillations. So I'm just going to conclude with a couple of examples and show you how we use this information clinically. And again, you see on the screen here, neuromonitoring, we have high-frequency oscillations, diffusion tensor imaging, subdural invasive grid monitoring, subdural depth electrode implantation, and the like. But this is a child who had this lesion, you can see, in what was the dominant hemisphere, very close to the cortical spinal tract. There it is outlined in pink. So this is a four-year-old, so right focal motor seizures, sensory seizures, and some degree of intractability, had this lesion as shown on the left side. We did functional MRI, but at age four, it was equivocal, we didn't get as much information from that as possible. So we went ahead to do invasive monitoring. This is before we placed the grid in place, and what you can see here is the area of basically motor sensory here. This is temporal frontal, and this is the area where the lesion was situated. Here we have the subdural grid, and nowadays, supplementing that with depth electrode implantation that you can see here, right into the lesion. All of the information that we can capture is listed here on the right-hand side of this image. This is the video of this child who is pushing the button because she's feeling sensory in her right hand, looking at the camera, and having a bit of a right facial twitch. This was the seizure that she had related to this lesion. This is the high-frequency oscillation that lights up the area that's involved in epileptogenesis, and here's the real-time video of the high-frequency oscillation coming into view here, as you can see right where the depth electrodes are inserted. This was not a tough case to figure out in terms of the origin of where the seizures were coming from and the lesion that was there. It was the management of the lesion afterwards because of the importance of dominant hemisphere, Rolandic area, and unwillingness to impart a motor tract injury, corticospinal tracts are nearby. So that's the area of the resection that was proposed, that was done, and in the presence of continuous train-of-five neural monitoring, you can see, so as we're resecting, we're continuously getting feedback from our neurophysiologists, and as we've done that, we've done a complete resection of this lesion. And because the motor tracts were intact throughout continuous neural monitoring, despite the fact the patient woke up with a post-operative hemiparesis and was dysphasic, we knew that that should recover by rights given the information that we had, and it did within a very short time frame after surgery, and the good news, this was a case of focal cortical dysplasia. And then the last case I'll show you is just this one, which is a very small lesion located just anterior to the corticospinal tract. You can see it here. These are, I find, challenging cases to localize and then to operate upon because they are not tumors. This was a case of, again, focal cortical dysplasia. Ten-year-old, failed multiple medications. We went ahead in this case because of uncertainty, exactly the zone of resection that would be required for that small lesion that was suspected as being the primary onset zone. So we subdural grid, depth electrodes placed here, all the information that we can garner from this. Here we are at the time of surgery going back. This is the navigation telling us that we are exactly where this small lesion is located, but we used the implantation of the subdural depth electrodes to help us localize these lesions because those of you who have been involved in these cases know it's not a straightforward, not a trivial matter to find these small lesions because there's nothing remarkable about them when you get there. And we resected it, and this patient has now been seizure-free nine months after surgery, and I thought a very technically challenging case, and proving it with all of this technology was very worthwhile. So in conclusion, we've used advances in neural imaging. These sequences are linked to neural navigation. I've shown you some data to suggest that the combination of surface and depth electrodes rather than one modality by itself may be interesting and perhaps superior to just using one. I know a lot of groups are moving towards stereo EEG. I'm sure you've had a lot of discussion about that during this course. We are characterizing the safe regions even better now for resection, including the Rolandic area. I talked about HFOs, which may be new to this group to hear about with fast ripples and resections of those areas, and a weight craniotomy, even in a pediatric population, can be performed with a good expectation of outcome after surgery. This is the team at SickKids Hospital. It's a very multidisciplinary team, and I'm listing many of the members there, and thank you again for the opportunity to present to you today. Thanks.

neurosurgery

magnetoencephalography

MEG

seizure foci detection

surgical outcomes