Next, we have Dr. Hudgens speaking on mesial temporal lobe and cingulate event-related potentials signal memory recall errors. Good afternoon. Thank you for that introduction. Thank you all for being here. Appreciate it. So I'd like to talk with you guys today about a topic near and dear to my heart, looking at mesial temporal lobe and cingulate event-related potentials signaling memory recall error. I have no disclosures. It's remarkable to me that the human eye has roughly two degrees of visual angle field of view. So I want to put this into perspective a little bit. So do me a favor and hold your arms out in front of you at about arm's length and look at your thumbnail. I encourage you to do this. It really puts it in perspective. Look at your thumb. That's two degrees. Two degrees of visual angle. That's remarkable that we're able to actually perceive what's happening in the world. And of course, our eyes are constantly scanning the environment. But it's likely that our brains have developed strategies to alert us to the salient events in our environment to direct our attention to those events to find out what they really are. And it also kind of underscores the point that we probably, you know, it really matters what we're putting in that two degrees of visual space. And just looking at a few people walking down the street, it really puts that message to heart where you see people running into each other in places that don't have texting lanes. So it is a bit of a struggle. Our brains are now trying to comprehend all of this extra sensory information, make meaning of it, and then also allow us to walk down the street. And in many modern cities, there are a lot of hazards that can really upset our lives if we're focused solely on our phones. And so it really is no simple task to manage all of this constant barrage of sensory information. And up to this point, I've only talked about visual information. But there's also a barrage of auditory information that's equally, if not more so important than visual information, at least in terms of guiding our attention towards those important events in our environments. And if I were to take EEG leads and place them on your scalp and show you something completely unexpected on the next slide, I'd be able to detect a voltage change on just scalp EEG, which has been studied for over 50 years, this signal, this event-related potential. And if I showed you something more surprising on that next slide, that signal would be even greater of a response. So the more surprising the signal, doesn't matter what it is, if it's sound or vision, I'd be able to detect a change in voltage, likely a coordinated activity of many neural networks. And so it's possible that some of this network was developed as a means of helping us navigate our environments, evading predators, seeking out food, and that sort of thing. But an interesting subset of surprising events are errors. We don't often want to make a mistake. But when we do, we probably consider that an unexpected and unfortunate surprise. And a lot of people who have studied this over the last 50 years have shown that it requires perception of the error, that you're aware the error happened, and the faster you're aware that that error happened, the larger this potential, the larger this error potential. And so looking over the data from the last 50 years or so, investigating novelty, surprising events and errors, there's remarkable similarity to the signals that are recorded from surface EEG. And there is mounting evidence that suggests they share the same neural generators. And now, since we have intracranial EEG, we have a new platform and a new strategy to study this information with a little bit more spatial resolution than scalp EEG, which is what we've attempted to do with our study. And so we've taken a series of subjects who have intracranial EEG for epilepsy monitoring, and we've had them perform a task. It's a pretty simple task. Just requires them to listen through headphones and hear a few cues. They have a couple of buttons in either hand, and they can press either the left or the right button, depending on what they're asked to do. They're given a series of immediate response requests up front, and then they have to register that into some form of memory, and then repeat that sequence when they're asked to recall it. They'll get immediate feedback after each button press. And if they get the button press correctly, they get a positive encouraging tone. If they get it incorrectly, they get an error tone. And if you look at some of the data of scalp EEG that have used dipole mapping to source the neural generators or potential neural generators for these signals, most of the data suggests that they may arise from areas in cingulate cortex. But again, the limited spatial resolution of scalp EEG limits what we can conclude. And so, in our study, we're hopeful to at least narrow down where the neural generator for these signals may arise. So we have seven subjects, four men, three women, mean age was 48 years. And on an interesting subgroup analysis, two subjects had memory impairment on preoperative neuropsychiatric evaluations. And their event-related potentials are different, which we thought was pretty unique and remarkable. And in our study, we saw what others have seen, that there is an event-related potential to the feedback tone. In this graph and all subsequent plots, the vertical black line at time point zero is both the time point the subjects press the button in the memory recall phase, as well as the time point of the onset of a feedback tone. And so it is a response, it's a significant response, it's not a very robust one. But it does happen in the early trials. But what we found was interesting and surprising to us is that over time, this signal changed. And so if we looked at the later trials of these subjects, this voltage change seemed to occur before the subjects even made their decision on what button to press. And it's a much greater potential change in these subjects, happening before time point zero where they make their decision. And it only happens in the later trials, doesn't happen in the earlier trials in these subjects, which we thought was pretty unique and interesting. And so we looked at other areas that are related, including mesial temporal lobe, where we have a vast amount of electrode coverage, and we saw a very similar response. But the interesting thing to note was that in mesial temporal lobe structures, this response occurred even earlier, around 1400 milliseconds before the subjects made their choice. And so we did some signal detection theory analysis and quantified that, and showed that indeed, in a robust manner, the mesial temporal lobe structures detect error before cingulate structures in our data set. And it's interesting to note that if you look at the correct trials in these subjects, there's really not much voltage change at all, even before the subject makes an action, or after they've made their action and they've received their correct feedback tone. So it's pretty interesting that mesial temporal lobe and cingulate electrodes in our study seemed to prefer error than accuracy. And probably has something to do with the fact that learning by error might be a fundamental strategy by which we learn how to interact with our environment, and the sensory cues our environments give us. And to the neurosurgeons and proceduralists in the room, we know this quite well in our own personal experience, and this is a subject that Henry Marsh has written about very well, as well as other surgeons, and it's true that when you make an error that leads to a bad outcome, you remember that. It stays with you. And so our next question was, is the error event-related potential widely distributed, or is it just in mesial temporal lobe and cingulate? And we looked at all of our electrodes, and the greatest coverage we had was in frontal cortex. And so plotted here in green is frontal cortex in the early trials, with red and blue showing mesial temporal lobe and cingulate. And we see that the frontal cortex in all electrodes, there really do not seem to be any discernible event-related potentials in the frontal cortex leads. And similarly, in those late trials where we saw that robust response in the memory retrieval stage for these subjects, we again didn't see much activity in the frontal cortex. So it does seem to be a unique part of mesial temporal lobe and cingulate structures. Looking at the individual statistics of each subject, six subjects had feedback potentials. That's the potential that happens after they've chosen either the left or the right one, and they've received either correct or incorrect feedback, depending on how well they performed. And then six subjects had recall. So-called recall potentials would happen before the onset of the feedback tone and before the subject's action. But the two subjects that didn't have either one of these had very unique potentials, and this is just one of them. This is a subject who performed not as well in their preoperative neuropsychiatric evaluation. And it was pretty apparent that there was still event-related potential, but it happened only to feedback, and only in the later trials, and not in the earlier trials, which was unique in our data. And so the conclusions that we've drawn at this point are that this event-related potentials, while we call it activity, there are still many steps we have to go before we can clearly state that these event-related potentials truly are active neurons, since these are macroelectrodes. But we do think that it does reflect some kind of coordinated network, given that we see them in mesial, temporal lobe, and cingulate, but not as in frontal cortex and other electrodes. And that these areas are likely more concerned with error and not accuracy. And one interesting side note is that direct stimulation of these structures might actually engage error circuitry. As some recent evidence suggests, direct stimulation of human interrenal region and hippocampus actually impairs memory. So just an interesting side note. I do have a lot of people to thank on this project. Dr. Grady, our chair, has been an ardent supporter. I've got to thank all of the patients who donated their time. Tim Lucas for his mentorship and working in his lab. I've got to thank my wife, of course, Justine Hudgens, a great supporter for me. And then I also have to thank Casey Halpern, who's here in the room. Appreciate all your help and mentorship over the years. And thank you all for being here and I'd love to take any questions if you have.

mesial temporal lobe

cingulate event-related potentials

memory recall errors

narrow field of view

visual focus