Hello, my name is David Bonda. I'm a PGY-5 neurosurgery resident from the Zucker School of Medicine at Hofstra Northwell, and I'm going to be describing the research I've been doing for the past two years at Cold Spring Harbor Laboratory. I do not have any disclosures. We'll start by asking the question, how does one overcome an urge? More specifically, how does the brain exert control over reflexive physiologic behavior in the pursuit of long-term or goal-oriented behavior? We are quite familiar with this ability in our everyday life. We also know that overcoming impulsive or reflexive behavior is cognitively effortful. In the past few months, most of us have probably realized that if we're not paying attention, reflexive behavior can get the most of us. Psychiatric disorders are often characterized by dysfunction in this executive control. The nervous compulsions that accompany obsessive compulsive disorder, the motor and verbal tics that are unwanted and describe Tourette syndrome, and the self-injurious behavior that sometimes characterizes autism spectrum disorder can all be thought of as pathologic executive control phenotypes. My goal of this project is to better understand the neurophysiological and computational underpinnings of executive control using a variety of neural tracing, recording, and manipulation techniques in mice. The approach that I've taken first involves the creation of a novel behavioral paradigm that puts into conflict this reflexive behavior with a more goal-oriented behavior. I will briefly describe the paradigm that we've created. In it, the mouse is in an arena, freely moving, with an LED on one end and shelter on the other end. The LED triggers or indicates the beginning of a trial, and the mouse is trained to learn that the LED is accompanied by a water reward if it can maintain its nose in a port for three full seconds. The mice are water-restricted, so they're motivated to learn and participate in this task. You can see the freely moving mouse in the shelter on the right. The LED goes on. It approaches the shelter, puts its nose in, waits three full seconds, and successfully gets a water reward in the situation. Intermittently, we expose the mouse to an aversive stimulus that's been demonstrated in the literature to produce reflexive escape or freezing behavior, what I'll collectively describe as defensive behavior. In these situations, the mouse is exploring the arena, and when it crosses a certain threshold, an aversive stimulus is presented to it. We use auditory and visual, separately, stimuli. They both, as I'll describe shortly, involve the same reflexive midbrain pathway. In this situation, what you're seeing here is when it crosses the threshold that you cannot see, that red dot indicates the auditory stimulus that goes on. Unfortunately, you can't hear it here, but the mouse interrupts what it's doing and reflexively flees to the shelter in that situation. The third part of this behavioral paradigm is the most interesting, and it's where we put into conflict that innate reflexive response with the desire to stay and obtain the water reward. In these situations, the mice go to the LED, put their nose in the port, are awaiting a reward, and then we expose it to this aversive stimulus. You can see here, the LED goes on, the mouse goes to the port, the red dot indicates the auditory stimulus. In this situation, the mouse withdrew and sacrificed the reward to reflexively escape into the shelter. Overall, the paradigm is such that 90% of the time, the mouse goes to the LED, gets the reward if it waits long enough. 5% of the time, it's exposed to this innate aversive stimulus while it's just freely moving and running around, and the other 5% of the time, the mouse is exposed to the stimulus while it's waiting at the port. Behaviorally, we can quantify their responses in these two scenarios using psychometric analyses. You can see here, honing in on the auditory component, if we plot their probability of escape against the salience of the stimulus, so in this case, the loudness of the auditory stimulus, you can see that as the loudness increases in decibels of the stimulus, the mouse has a higher probability of escaping. This is characteristic of many behaviors and one of the very benefits of psychometric analysis. As the mice are exposed through subsequent sessions to the same stimulus, because it is not accompanied by actual mortal danger, the animals have a lower probability at any given threat salience of escaping, and that's something we consider habituation. When we look at the conflicted threat trials, when the animal is awaiting a reward in the port while it is exposed to the stimulus, you see a much different psychometric curve. Here, for any given stimulus salience, there's a much lower probability that the mouse will actually escape. Of course, this changes even lower, more dramatically as habituation proceeds. If you compare conflicted threat with unconflicted threat, the top three panels are the unconflicted situations when the animal is just exploring, the bottom being the conflicted threat. You can see there's a very different psychometric phenomenon. The animals are much less likely for a given decibel level stimulus to escape. The curve is shifted to the right. Behaviorally, we would say that there is a difference between unconflicted threat response and conflicted threat response. I would suggest that the latter is demonstrative of an executive control phenotype. The neural circuitry underlying this behavior begins with the reflexive arm of their response. This group out of UC London, the Bronco group, very nicely demonstrated the midbrain components that orchestrate reflexive escape. Specifically, neurons of the deep medial superior colliculus, once activated, trigger activation of the dorsolateral periaqueductal gray. These neurons decusate and activate medullary cuneiform nucleus neurons, which produce synchronous locomotion and thus escape. Notably, the ventrolateral periaqueductal gray similarly is involved in freezing behavior. Through this work and the work of others, we know that this reflexive component of the behavior is orchestrated by the midbrain, specifically superior colliculus and periaqueductal gray. What's most interesting is that a number of studies in the literature that are tracing-based show that the medial prefrontal cortex projects to both of these regions. The medial prefrontal cortex has been implicated in a number of other studies, including fMRI and EEG studies in humans, to be associated with what we would call executive control or this top-down influence of behavior. The cell type and projection specificity, however, of those neurons is unknown. So, overview of this pathway that I'm investigating here, you can see medial prefrontal cortical neurons projecting to these midbrain regions that govern the reflexive escape or freezing. So, I will briefly show some neuronal tracing experimental data. These are slices from the medial prefrontal cortex of specific genetic mouse lines that have been injected in the superior colliculus and the periaqueductal gray with a retrograde viral vector that contains cre-dependent fluorophores that are expressed only in neurons that project to the sites of the injection. You can see here that what I'm showing is these medial prefrontal cortical regions have a very robust expression of the fluorophores. What this is telling you is that these specific neurons in the medial prefrontal cortex project to the periaqueductal gray and the superior colliculus. And another cell type, using a different genetic line of animal, shows a very similar but not exactly the same distribution of medial prefrontal cortical projections to these midbrain regions. The Bronco Group has collaborated with us and helped us with some voltage clamping acute slice experiments. And their work has suggested that these medial prefrontal cortical neurons that project to superior colliculus and the periaqueductal gray are in fact mostly excitatory. And these neurons are excitatory and they synapse on glutamatergic neurons in those midbrain regions. So a revised cartoon would show that these are excitatory projections from medial prefrontal cortex to the midbrain. Using fiber photometry, which is a technique of recording cell type and projection specific populations of neurons in awake behaving animals, we can get physiologic data from these populations of neurons. You can see these animals are freely moving still. They have optic fibers implanted into the medial prefrontal cortex recording neurons that project to superior colliculus or the periaqueductal gray. And we can record their activity during both conflicted and unconflicted behavioral types. So looking at the neural signatures or recordings from these tasks, I'll start by showing you the fluorescent data from the trials in which the animal simply goes to the port and either gets a reward or doesn't, depending on if it waits long enough. The green on the upper left panel here, the green tracing, is the fluorescent signal from the medial prefrontal cortical neurons when the animal gets the reward. The red when it does not get the reward. You can see there is a minor difference between the two tracings. The GFP control animals demonstrate that this is not simply a motion artifact or other kind of noise. This gets interesting when we look at the unconflicted threat. So we're recording from medial prefrontal cortical neurons that project to superior colliculus or periaqueductal gray. In this case, we're aligning the signals to the onset of the threat. So these animals are running around exploring. A threat is suddenly triggered. The green trace shows the activity of these population of neurons when the animal ignores the threat for whatever reason, whether it doesn't detect it or it's habituated to it or it doesn't care or prefers to proceed with its behavior. The red tracing shows the trials when the animal responded defensively to that threat. So you can see that these neurons seem to ramp up their activity in response to a threat when the animal actually engages defensively. The GFP control animals show this is not simply motion artifact. And most interestingly, in the conflicted threat trials, when the animal is at the port awaiting a reward and is exposed to the threat, you can see a much more dramatic cell activity phenomenon in these NPFC neurons. The green tracing shows when the animal chose essentially to stay and get its reward versus the red tracing, which shows a much more dramatic uptick in neuronal activity that is sustained for the duration of the recording in the trials when the animal interrupts what it's doing and reflexively either freezes or escapes to the shelter. Similarly, this is not an artifact. If we compare the conflicted and the unconflicted threat, the unconflicted threat is on the left panel. The right panel shows the conflicted threat. What you can't see here, unfortunately, is the scale difference is much more significant. And if I overlap the responses of these neurons when the animal engages defensively in response to the stimulus, you can see that the green tracing is the unconflicted threat. These neurons bump up in their activity a little bit in response to the threat, but when the animal is in the port awaiting a reward and it experiences that aversive stimulus, their activity jumps up through the roof essentially and is sustained. And so neurophysiologically, I can conclude here that there is a difference in the population level activity of these NPFC neurons that project superior colliculus and periaqueductal gray when the animal experiences a conflict in motivation. And I would suggest that these neurons really are important for these cases where executive control is necessary. Ongoing experiments that I'm doing, unfortunately I don't have data to show, aim to causally manipulate this population of neurons using channelrhodopsin and halorhodopsin to activate or inhibit these neurons during awake behavior. In summary, I've shown you that this novel assay we've created seems to successfully elicit executive control behaviorally. The neuroanatomic tracing and slice physiology experiments I've shown you demonstrate that the NPFC to superior colliculus and periaqueductal gray neurons are excitatory, glutamatergic from layer 4-5. There's multiple cell types that seem to have distinct populations and that they are both synapsing on glutamatergic neurons primarily, but also on GABAergic in both superior colliculus and periaqueductal gray. The photometry data, most interestingly, demonstrates that this neuronal population seems to care only or mostly when there is some kind of a conflict in behavioral options, and I would suggest that this is reflective of executive control. With that, I will thank you very much. I appreciate your time. I want to thank the members of the Kepich Lab who have helped me tremendously throughout this work, as well as the work of our collaborators in the Austin and Bronco Lab, and of course my mentors, Drs. Scholder and Narayan from Northwell Neurosurgery. Thank you very much.

executive control

reflexive behavior

mice models

neuronal tracing

physiological data