Thank you so much, Dr. Mainali, for the very kind introduction. Good afternoon, everybody. It's an honor to be here. Some of the work that we'll be discussing was funded by a variety of federal foundation and departmental sources, and I have no financial disclosures or conflicts of interest to report. There are four objectives for today's talk. First, we'll discuss the clinical motivation for the development of targeted therapies aimed at promoting recovery of consciousness in the ICU. Second, the state of the science in this field. Third, the targets that these therapies are trying to engage. And then finally, we'll talk about some future directions, where we go from here. The clinical motivation is that we aim to accelerate the recovery of consciousness in the ICU. Doing so has the potential to decrease ventilator days for our patients, to decrease the risk of complications that we're all familiar with that are associated with immobility, complications like deep vein thrombosis, ventilator-associated pneumonias, UTIs, bed sores, and others. And perhaps most importantly, by accelerating recovery of consciousness, we can potentially prevent premature withdrawal of life-sustaining therapies. In other words, our therapies can give clinicians and families a more accurate prognostic picture about a given patient's chances of long-term recovery. So let's highlight this point about why early recovery of consciousness is so important in the ICU. These are data from Giacino and Kalmar in 1997, looking at 70 patients with severe traumatic brain injury. On the x-axis, we're looking at the number of months post-injury, the y-axis, the mean disability rating scale score, where higher scores indicate worse disability. And the blue bars represent patients who are in a vegetative state when they got to inpatient rehabilitation. The pink bars indicate those patients who were in a minimally conscious state when they got to inpatient rehab. Not surprisingly, both groups have high levels of disability at one month post-injury. But as we get out to one year, these bars have dissociated such that the patients who were initially in a minimally conscious state when they left the ICU and got to rehab have far less disability. In fact, 10% of this group, these are patients who are just starting to localize to noxious stimulation, are just starting to track with their gaze. 10% of those patients have zero measurable disability at one year post-injury. Fundamentally changing the way that we communicate prognosis to families if we've seen signs of minimal consciousness in the ICU. So where are we now? There has been dramatic progress in our field, progress that Dr. Klassen highlighted in his talk, with advanced technologies that can detect signs of covert consciousness, like task-based functional MRI or task-based EEG. These are techniques that are now endorsed by multiple academic societies and government entities in the US and Europe, as well as New Zealand. But what about therapies? Where are we in that domain? How many treatments are there that are FDA approved in the ICU to promote early recovery of consciousness? The answer is zero. How many treatments do we have that have high level randomized controlled trial evidence for their efficacy? Unfortunately, the answer is also zero. When we take a step back and think about the therapies that have strong evidence to support their use, it's really in the subacute phase. And this is the one therapy, imantadine, that is endorsed by the 2018 US guideline for treating this population. This is by Giacino White and colleagues. And there's a key observation here. These are patients who received imantadine or placebo within four to 16 weeks of their injury, so the subacute phase. And if we look at the four weeks of therapy, we see that those who received imantadine, and again, the disability rating scale score is on the y-axis here, those who received imantadine had an acceleration with respect to their recovery of consciousness and recovery of function. But during a two-week washout period where those who had received imantadine now got placebo, what we see is these lines converging. So while it may be a worthwhile goal to try to fundamentally alter long-term outcomes, it may be more realistic for us to focus on acceleration of recovery, because that's what the imantadine study teaches us. Which treatments are available to us now and are being tested in investigational studies? Well, there are five different domains that we can consider, first of which is pharmacologic, the second being electromagnetic therapies such as deep brain stimulation, transcranial magnetic stimulation. There are mechanical therapies like low-intensity focused ultrasound pulsation or LIFUP, a treatment that's believed to perturb the membrane of the neurons leading to depolarization and activation of those cells. There are sensory therapies like music therapy to stimulate the brain, and then there are regenerative therapies like stem cells. This is a schematic that was created in collaboration with Dr. Gretchen Brophy and Megan Barr who are here, where we tried to show the variety of mechanisms of action that pharmacologic therapies involve. Most of these therapies are either direct agonists that are stimulating, oops, sorry about that, that are stimulating the postsynaptic neuron, or that are reuptake inhibitors increasing the concentration of neurotransmitters in the synapse. What are the networks that these therapies are trying to engage? So there are multiple different networks that contribute to consciousness in the brain, and several of them are shown here. The network that's gotten the most attention, one that's been discussed in the prior talks, is the default mode network, whose nodes are shown here in red. This is a network that is apparently associated with introspection and self-awareness, and is thus believed to play a critical role in consciousness. We can see the frontal and parietal nodes of the default mode network here, shown on the schematic, and in this circuit diagram, we can get a sense of the subcortical regions that these nodes are connected to. Regions like the striatum, the globus pallidus, the brainstem, and the central thalamus. This is some of the circuitry that we are trying to engage with our therapies. Now there is a very important principle that is informing and guiding the development of new therapies in this field, and that is the cortical-subcortical funnel effect. The idea is, if we stimulate the cerebral cortex, let's imagine transcranial magnetic stimulation of this specific cortical gyrus. We activate that local region, and that region of the cortex is certainly connected to other regions, like through U-fibers to nearby gyri, or even across the corpus callosum to the contralateral hemisphere. So certainly, when we activate a local gyrus, we do also activate downstream areas. But now let's compare that to a subcortical region, like the thalamus, where here we are looking at a color-coded functional connectivity map derived from fMRI, and here are the cortical correlates of this map. The idea is that if we stimulate the region shown in red in the anterior thalamus and the central thalamus, we can activate a much broader expanse of the cerebral cortex, as shown here in red on the lateral and medial surfaces of the brain. This funnel effect, the basic idea is that electrical information is condensed and dereferenced as you get from the cortex to the subcortex, such that stimulation of a single subcortical node can activate a broad expanse of the brain. Which types of networks are our treatments trying to engage? So the default mode network, again, is the one that's most commonly being targeted by all of the different therapeutic domains, except the sensory one. But it's also important to point out that we don't yet understand which of these neurotransmitter systems are targeting specific networks. They appear to be nonspecific in many of their effects, but we don't have models to show us the differential stimulation parameters of dopamine as compared to serotonin, norepinephrine, or acetylcholine. What about safety? There may be plausible mechanisms by which these therapies may help our patients, but are they safe to give in the ICU? Some of the theoretical risks include seizures. Many of you, like me, may have been taught in medical school or training that any drug that increases the firing properties of a neuron may also lead to a higher risk of seizures. Interestingly, if you look at the imantadine study, there was not a higher risk of incidence of seizures in the group that received imantadine compared to placebo, but this has not been rigorously studied in the ICU environment yet. What about excitotoxicity? Releasing glutamate and stimulating these postsynaptic neurons may lead to excitotoxic damage to those neurons. If we are releasing a lot of new transmitters like dopamine, could we deplete those neurotransmitters by stimulating our patients? And then finally, many of these pharmacologic stimulants have effects on the heart, the cerebral vasculature, the peripheral vasculature, and thus that is another category of side effects that we have to consider. The data here are few. One pilot randomized controlled trial of imantadine looking at 15 ICU patients with severe TBI found no serious adverse events, very small numbers, preliminary data. The jury is still out. These are data from a study that was led by Megan Barra where we looked retrospectively at two level one trauma centers, Mass General and Brigham in Boston, and we looked at the utilization and the safety profile of pharmacologic stimulants in patients with traumatic brain injury. 608 total patients of whom 8% received a stimulant, the most common of which was imantadine. The median day of drug initiation was 11 with a broad range. The median Glasgow Coma Scale score was nine also with a broad range. And when we looked at adverse events, which were defined again retrospectively based on drug discontinuation or dose reduction, we found 10 events in seven patients, a total of 15% of the overall population. The most commonly observed side effect was agitation. We also saw anxiety, insomnia, rash, and urinary retention. And you can see that some of these side effects were observed in the ICU, whereas others were observed in the floor. And again, want to recognize Dr. Barra's leadership in this study. As far as our therapeutic targets, there are two main ones that I'd like to highlight today. And the first is the central thalamus. We have to pay homage to the landmark study by Schiff, Giacino, and colleagues, 2007. This is deep brain stimulation of a patient in the chronic minimally conscious state years after a traumatic brain injury. The deep brain stimulation electrodes as seen here, when they were on, they led to increased responsiveness, more interactions between the patient and the examiners, proof of principle that it is possible to engage these networks even years after a brain injury. A group at UCLA led by Martin Monte, Caroline Schnakers, Paul Vespa, and colleagues are now trying to stimulate the central thalamus, but non-invasively with the life of therapy that we talked about earlier. The second major target is in the brainstem, the ventral tegmental area in the midbrain as shown here. And we can see in this tractography figure that the VTA has diffused projections as shown in pink to the frontal region and the parietal region, those nodes of the default mode network that I mentioned earlier, which are connected by the cingulum bundle shown here in blue. And there are particular connections to the posterior cingulate, which is the core hub node of the default mode network. Some very elegant fMRI work by Spindler and colleagues in Cambridge in the UK looked at VTA functional connectivity with the posterior cingulate. First in healthy volunteers receiving propofol and found that the functional connectivity between these two nodes decreased as subjects got more propofol. And in patients with severe brain injury where lower levels of consciousness were associated with lower levels of connectivity between these nodes. In an independent cohort of patients with TBI, they then gave methylphenidate and observed that the functional connectivity increased. They could up-regulate or modulate this particular pathway. So where do we go from here? What are some of the clinical trials that are being performed in this field now? We recently launched at Mass General Hospital an intravenous methylphenidate study where we're testing the hypothesis that patients who have partially preserved connections between the dopaminergic ventral tegmental area and rostral sites in the diencephalon, forebrain and cortex. These are the patients whom we predict will respond to methylphenidate because you need to have some intact axons releasing dopamine at the synapse for this reuptake inhibitor to be efficacious. And I wanna acknowledge here Dr. Tom Bleck who's playing a key leadership role in the study as the independent medical monitor. We are also partnering with the UCLA group to test the hypothesis that life up stimulation of the central thalamus is likely to be efficacious in patients who have partially preserved thalamocortical connections. We call this the connectome-based clinical trial platform. And here is an example of a patient whom we predict would respond to methylphenidate. We're looking from a lateral perspective at diffusion tractography data from a clinical scanner where we see that in a patient with severe TBI as compared to a control, there are far fewer connections between the brainstem and cortex. But we also see that there are some preserved connections with the cortex suggesting that this is the type of patient who might respond to a dopamine reuptake inhibitor. Different patient, 53-year-old woman falls down a flight of stairs, dies in a coma three days after her arrival to our ICU. These are post-mortem imaging data that we acquired in her formalin-fixed brain. We're looking from a posterior perspective at connections between the brainstem and the thalamus. In a control subject, we see extensive connectivity between the brainstem and thalamus. This patient has a near complete disconnection of the brainstem from the thalamus as shown here in the schematic connectogram. Yet when we look at the same patient's connectivity map between her thalamus and her cortex as compared to a control, we see multiple areas of preserved connectivity between the thalamus and the temporal lobe, the orbitofrontal region, the supplementary motor area, and the parieto-occipital region. And if we think of that funnel mechanism we talked about earlier, you could imagine how this is the type of patient who might be a responder to thalamic stimulation. And this is work I wanna recognize Rebecca Folkerth and Hannah Kinney, two mentors and collaborators on this project. This is our moonshot MRI scan, the 100-micron MRI, post-mortem ex vivo acquisition over 100 hours of scanning. So why do I show this here? This was part of a big collaboration with the team at the MGH Martino Center. Well, it's because we're now using this ultra-high resolution imaging dataset to inform our selection of subcortical targets. These are fMRI data that were analyzed by Andrew Lee, a postdoctoral fellow in our group, and we superimpose these connectivity data onto the 100-micron MRI scan. And we ask the question, what are the subcortical regions that are most strongly connected to the cortical nodes of the DMN? We see these connections in the dorsal raphe, we see them in the ventral tegmental area, in the lateral hypothalamus. We see them as we come up rostrally in the central thalamus, giving us a precise map of functional connectivity superimposed on the 100-micron MRI scan. Andrew Horn at Brigham and Women's Hospital then co-registered these 100-micron MRI data into standard stereotactic space and released them as part of the LeadDBS software program. This is all freely available, anybody can use it. And we can now identify the precise anatomic targets of stimulation therapies like deep brain stimulation, LifeUp, and others. The Curing Coma Campaign that Dr. Olson mentioned earlier convened a coma science group to look at therapies and to perform a gap analysis for where this field needs to go in the future. And here's some of the contributors to that group. And when we looked at the five different domains, we identified a number of directions that we need to go in the future. First, with respect to the pharmacologic therapies, we need tools to measure imbalances in these neurotransmitters. With respect to electromagnetic and mechanical therapies, we don't know yet the optimal stimulation sites or the optimal parameters to stimulate those sites. With respect to sensation, we need to understand how activation of association cortices lead to improvements in level of consciousness. And with regenerative therapies, how do these stem cells integrate into these neural networks that control consciousness? More broadly, we need a unifying conceptual framework for how to evaluate these mechanisms of action. We certainly need large-scale randomized controlled trials, as well as predictive biomarkers to identify likely responders and pharmacodynamic biomarkers to measure therapeutic effects. To conclude, there is currently a broad spectrum of available therapies that are being tested in clinical trials, but we fundamentally do not understand yet the safety profile of these therapies in critically ill patients, nor do we have definitive efficacy data. Those data are only limited to small-scale proof-of-principle studies. And finally, as far as our clinical motivation and the ethical imperative to move forward as a field, we have the potential with these therapies to accelerate recovery and thereby prevent premature withdrawal of life-sustaining therapy. With that, I'd like to thank you for your attention and happy to take questions during the Q&A session.

recovery of consciousness

ICU

therapeutic domains

safety considerations

randomized controlled trials