Good morning, my name is Ahmed Jorge. I come from the University of Pittsburgh and the title of my talk is Evidence for the Hyperdirect Pathway in the Human Speech Network. I have no conflicts of interest. So to talk about this, we want to talk about previous studies that have shown the hyperdirect pathway. So these studies have shown direct axonal projections from the cortex to the SDN and obviously have been implicated in numerous human behaviors and also some of them could be of therapeutic interest to deep brain stimulation. So three of these studies conducted in rats and humans have shown this characteristic peak waves that we see around 2 milliseconds, 3 milliseconds, and 5 milliseconds after the first stimulus artifact that you see at time zero. Specifically, the last one to show this has been a paper bat by Milosevic et al. in 2018 where they've shown this distinctive three peaks, the figure I have on the right. So for our experimental setup, we did something similar. Now we have 21 patients with Parkinson's and we are stimulating the SDN at 1 hertz in between 1 and 3 milliamps. We have then recorded on an ECOG array in the premotor M1, S1, and SDG and in the frontal operculum. Of novelty, we are recording from the SDG and frontal operculum, which hasn't been shown before. So on top of that, we are trial averaging 30 trials for 30 seconds on that 1 hertz stimulation on the ECOG array for all of our 126 ECOG pairs that we have. In addition to that, we are calculating the statistics that will let us know if a peak is rising above the noise, the significant noise that we're seeing on the baseline on that ECOG array. Here on the figures, A and B, we have the setup of one ECOG during the surgery. B is the reconstruction of all the ECOGs we have. C and D are the baseline and the signal after stimulation at time zero. And E and F are the same signal but now bounded by our t-statistics. Again, that let us know if the signal goes above noise levels. So you can see on the baseline here, none of those peaks in figure E are rising above the red line. None of them are rising above significant levels. But on the right, on figure F, you see on those two asterisks, you see two significant peaks that are clearly rising above that red t-statistics that we have calculated. And we delve a little deeper in the abstract. So we see different responses across the cortex. We see, for example, here in the motor strip, we're seeing two peaks around 2 and 6 milliseconds. Similarly, in the frontal operculum and the STG, we see some peaks rising above the significant levels. And of course, in the ECOG array, the ECOG contacts that are on the sulci, on any sulci region, aren't really recording anything significant, as you can see on the lower left plot. It's pretty much all noise there and nothing is rising above significant levels, which is not even scale in that Y scale. So apart from doing that, we then grouped all those signals into a simple histogram to see what type of latencies and outputs we're dealing with. And not surprisingly, because we have seen this in previous publications, but we see a distribution that centers around 2 milliseconds of latency, around 3 milliseconds, and also around 5 to 6 milliseconds. And those correspond to those red asterisks that we saw in the previous figure, and now we see here again on the left side. This is also consistent with previously published literature on where the peak latencies are seen. So apart from that, we also explore what happens when we stimulate different regions in the STN. For example, here we are stimulating in the central tract of the STN, and we see that the latency is much shorter in M1 and S1 compared to the STG region we see, and perhaps towards the front in the frontal opercular region. When we stimulate in the posterior tract, however, we see that the latencies become much slower, and this could be of clinical significance. In addition to that, we stimulated the STN and the GPI to see if there was any hyperdirect pathway in the GPI. We did not see any. We did see, however, the orthodromic stimulation when we stimulate in the STN and in the GPI. Again, if you're familiar with the literature, you will see those orthodromic stimulation amplitudes in the 40 to 60 range for the STN and the 20 to 40 range for the GPI, and that's clearly what we see here as demarcated by those black arrows in figures C and F. So in short, we have seen motor and sensory cortex being stimulated by the STN for those evoked potentials that are related to the hyperdirect pathway because of timing. In addition to that, we have seen evoked potentials in the superior temporal gyrus and also in the frontal opercular region, which is of novelty. And in addition to that, obviously, this hyperdirect pathway and their timing of the signal could have implications for different implant sites and also for tailoring multiple symptoms in patients with DBS, in patients with Parkinson's.

hyperdirect pathway

human speech network

direct axonal projections

deep brain stimulation

Parkinson's symptoms