Next, we'll have Dr. Wang speaking to us on identification, I'm sorry, induction and quantification of plasticity in human cortical networks. The microphones were off. That's why they don't work. Thank you so much. My name's Danny and I'm a third year medical student at Stanford University. I had the opportunity to work with Dr. Ash Mehta over the last two years. And today I'll be talking about the induction and quantification of excitability change in human cortical networks. No conflict of interest to disclose. And my talk will consist of a background, method, results and conclusions. So to start, brain simulation therapy have been playing an increasing role in the treatment of various neuropsychiatric diseases. Invasive measures such as DVS, RNS and VNS are controlling symptoms in patients who previously medications and or surgeries have failed and or are contraindicated. And in addition, non-invasive simulation measures such as TMS and TCDS, transcranial direct current simulation are already effective therapies for patients with depression. So given these advances in neuromodulation techniques, it's absolutely critical to understand the precise effects of simulation on the human cortex. Using TMS as an example, it is used at high frequency greater than 10 hertz to potentiate cortical excitability. And this is thought to restore hypo connected prefrontal networks in patients with depression. On the other hand, if you use it at low frequency less than one hertz, it is thought to suppress neuronal activity and studies have been done to show it might decrease instance of seizure in patients with epilepsy. So given this, however, there's actually not a definitive answer on how TMS works. And most of the studies on previously were done in the motor cortex. And traditionally studies that were done to look at cortical excitability following simulation have been done through neuroimaging methods such as FRMI, PET-CT and EEG, which have provided important insight into this field. However, these methods have low spatial temporal resolution and limit the conclusions of causal effects. In patients with medically intractable epilepsy who are undergoing seizure localization, this is a great opportunity to study the brain. And specifically, we can use the implanted electrodes to simulate the human cortex in a way that's consistent with non-invasive therapies such as TMS. This allows us to record neuronal activity with unrivaled spatial temporal resolution and offer mechanistic insight into how these therapies work to treat depression and or epilepsy. So the purpose of my study was to look at the effects of repetitive simulation with high spatial temporal resolution using cortical-cortical evoked potential or CCEP mapping. So how CCEP mapping works is it involves the injection of a current directly onto the brain parenchyma and subsequently measuring the simulation invoked the response at different recording electrodes. The CCEP itself consists of usually two large deflections, the amplitude of which we can quantify and some marker of cortical excitability. We can also look at latency, which then is a indicator of the synaptic length between the brain regions. So in our study, we recruited eight patients with medically intractable epilepsy at North Shore University Hospital. And specifically, we recorded the dorsolateral prefrontal cortex as this region is important in the treatment of various neuropsychiatric diseases. However, if this site is not available, then we undertook a stepwise approach in terms of site selection. Our simulation paradigm consists of two phases. We recorded CCEP before and after simulation by delivering biphasic test pulses at one second interval. Our simulation pulses consists of 10 hertz trains at 100% of motor threshold. And in a cohort, in our study, four patients we implanted with electrodes targeting the prefrontal cortex. And the pre-op MRI and post-op CT here shows the location of the electrodes, as well as with the brain masks showing the electrodes and their associated pre-simulation CCEP voltages. And on visual inspection, we can clearly see that pre-simulation CCEPs are robustly evoked local to the simulation sites as indicated by red. So our first question then is, does a single session of repetitive electrical simulation induce cortical excitability change in the regions that we record? So in one patient with prefrontal stimulation, we found that immediately after 10 hertz simulation, there was suppression of cortical excitability with a gradual return to baseline over the next 10 minutes. And this is visually striking even on waveform visualization here. And on statistical testing, it is also significant. Now looking at the whole brain effects of a 10 hertz simulation, what we found was that suppression of cortical excitability mostly occurred local to the simulation site. And some of the regions that were modulated distal to the simulation site presumably had long-range synaptic connections through the simulated region. In the other three patients with prefrontal cortex stimulation, we found potentiation, suppression, and potentiation. And these results suggest that prefrontal cortex stimulation effects are not uniform as in contrast to the non-invasive studies as previously found with TMS. We also had three patients where we implanted electrodes recorded in the motor strip. And we found that after simulation, we mostly saw suppression of the CCEP amplitude, again, local to the motor cortex. And your last patient where we had coverage of the temporal cortex, stimulation of the temporal cortex led to widespread suppression of cortical excitability. So since most of our results have been based on location where we saw decreased or increased cortical excitability local to the simulation site, we wondered if we could predict the location of excitability change based on some pre-simulation parameter. And to do this, we looked at pre-simulation CCEP, specifically quantifying the amplitude, the latency, as well as the recording electrode, its distance to the simulation site. To start with CCEP amplitude, when we looked at electrodes that were affected by repetitive simulation in red compared to electrodes that were non-modulated by repetitive simulation, we can clearly see that those channels that were modulated by simulation had a greater CCEP amplitude. This indicates that effective connectivity may be determined in regulating whether a brain region would be affected by simulation. Now, looking at latency, electrodes or brain regions that had undergone modulation tended to have shorter latency, indicating that perhaps shorter synaptic connections to the simulated region led to modulation by simulation. And lastly, if we look at the proximity of our recording electrodes to the simulation site, then those brain regions that are modulated tended to exist closer to the simulation site, again, consistent with what we observed visually previously. So now to answer the question if we can predict the zones of excitability change following stimulation, we took those pre-simulation CCEP parameters, the amplitude, the latency, as well as the distance from the recording electrodes to the simulation site, and we used a classification algorithm called a support vector machine. And what we were able to do was to predict the location of excitability with, on average, 85% accuracy. The sensitivity ranged from 54% to 67%, and the specificity ranged from 90% to 95%. So to conclude, we found in our study that repetitive stimulation induces local and distal excitability changes, change in a subset of cortical regions recorded. These excitability changes tend to occur in regions that are anatomically closer and functionally connected to the simulated region. And these changes happen in forms of suppression and potentiation, with particular to prefrontal cortex stimulation. We also found that regions of excitability change can be accurately predicted using an individual's pre-simulation connectivity profile, and this lends itself to the idea that potentially we can optimize existing non-invasive therapies by modeling excitability change based on individual's pre-simulation characteristics. Some of our future work involves looking at CCEP within the simulation train, and specifically on preliminary analysis here, we show that within a train, we can track CCEP amplitude in real time as the pulses are being given, and that the change observed within, in real time, can actually track pre- and post-CCEP testing. To conclude on, I would like to acknowledge all the co-authors of this study, in particular Corey Keller and Dr. Ash Mehta. Of course, a big thank you to the nursing and physician staff at North Shore University Hospital for taking excellent care of the patients. Thank you. Questions? I'll start. I have one. This is very interesting work. The latency that I understood from the graph between stimulation and increment in acceptability was in the order of 20 to 40 milliseconds? Yes. Do you think that this is an effect that's really cortical-cortical mediated, or is it cortical-thalamo-cortical? So, in this study, we mostly looked at the early phase of the CCEP, specifically within 100 milliseconds. So, these are likely short-range synaptic connections as opposed to long-range synaptic connections. Samir? Yeah, great work. Turning it on helps, right? Yeah, it does. Yeah, it's remarkable. Yeah, great work. Did you look to see if there's any way of predicting, not even just from the baseline physiology, but from things that you could figure out from the implant, the seizure-onset zone, and or just structural connectivity differences between these areas that lead you to maybe make some predictions about excitability versus not? Absolutely. That's one area that we're trying to explore. We're hoping to use connectivity analysis, perhaps from an imaging standpoint, or perhaps using CCEP connectivity analysis to see if we can, even before simulation, predict areas of modulation. Thank you. Thank you very much. That was a great talk.

excitability change

cortical-cortical evoked potential

medically intractable epilepsy

repetitive electrical stimulation

non-invasive therapies