Audio Transcript from April 14, 2026

This transcript is intended to provide an accurate, verbatim representation of the language used by the speakers during the R3 Seminar Impact of Long COVID on neurocognitive function held on April 14, 2026. You can view a timestamped version of the transcript with the video on YouTube.

Watch the recording or read the summary on the R3 Recap page.

Disclaimer: This transcript has been generated using AI technology and lightly edited. Please note that some errors or omissions may have occurred due to the limitations of automated transcription.

Patrick Ahearn:

So, once again, welcome everyone to today’s RECOVER webinar, Impact of Long COVID on Neurocognitive Function. My name is Patrick Ahearn with RTI, and I’ll be helping out with the virtual room today. So, just a few housekeeping notes before we get started. When you log in, your microphones will be automatically muted with web cameras turned off. But if you have any questions for our panelists today, please submit those into the Q&A window at any time. And if you run into any technical issues, please let me know in the Q&A window as well. Closed captions are available during today’s webinar. Just click on the show captions button on your main Zoom toolbar to turn those on. And we will have a brief feedback poll at the end of today’s session, and the RECOVER team would greatly appreciate your responses. At this time, I’ll turn things over to my colleague, Quinn, to kick off today’s session. Thank you.

Quinn Barnette:

All right. Good afternoon, everyone, and welcome to the RECOVER Research Review, or R3 Seminar. My name’s Quinn Barnette. I’m an epidemiologist with the RECOVER Administrative Coordinating Center, and I’ll be your moderator for today’s session. The goal of the R3 Seminar series is to share the RECOVER Initiative’s research findings with researchers and the public. These seminars accelerate scientific discovery by allowing experts to share their latest insights on Long COVID and related conditions. In today’s seminar, our panelists will share RECOVER findings on how Long COVID can affect the brain, but we still need to learn and how researchers are developing better ways to measure and understand these neurocognitive symptoms to help inform possible treatments.

I want to start by thanking those who submitted questions in advance and remind everyone that you can submit any questions during today’s presentation using the Q&A feature in your Zoom menu. After today’s panel, our speakers will answer as many questions as possible. A Q&A document will also be posted with the recording of the seminar on RECOVERcovid.org. The document will include the answers for submitted questions relevant to today’s presentations. Questions about other scientific topics will be addressed in future seminars, and answers to broader questions about RECOVER will be available in the FAQs found at RECOVERcovid.org. As a reminder, we will not be able to answer questions about individual clinical care during this seminar.

I’m pleased to share that our panelists today are Dr. Jacqueline Becker, Dr. Michael VanElzakker, Ms. Liza Fisher, and Mr. Mike Zissis, and our discussion will be Dr. Jennifer Frontera.

Dr. Becker is a clinical neuropsychologist, assistant professor of medicine, and health services researcher in the Division of General Internal Medicine at Icahn School of Medicine at Mount Sinai. Her work focuses on neurocognitive assessment in the context of chronic illness, including Long COVID and ME/CFS [myalgic encephalomyelitis/chronic fatigue syndrome]. She’s MPI [multiple principal investigator] of RECOVER-NEURO, PI [principal investigator] of a clinical trial treating cognitive dysfunction and depression in people with Long COVID, and PI of an award investigating the cognitive effects of Long COVID through a health equity lens. She’s also the neurocognitive lead for RECOVER-TLC clinical trials in Cycle 2 of the RECOVER observational cohort study.

Dr. VanElzakker is an assistant professor of psychiatric neuroscience at Massachusetts General Hospital and Harvard Medical School, an instructor at Tufts University, and co-founder of the PolyBio Research Foundation. He’s a neuroscientist whose research centers on neuroimmunology or the intersection of the nervous and immune systems. He studied infection-associated chronic illness for many years prior to the COVID pandemic and has co-authored an article on core research priorities for studying Long COVID biological mechanisms that became one of the most highly cited in the field and was used by the US Government Accountability Office as the basis for their Long COVID mechanisms technical white paper.

Ms. Liza Fisher is a patient representative involved in the RECOVER Initiative. After 5 weeks of acute COVID infection, she was hospitalized for months, resulting in being disabled by Long COVID and Long COVID–associated conditions. Her eyes were opened to the world of under-acknowledged and -researched complex chronic illnesses affecting millions. She wanted to help with awareness and research, hopefully leading to cures, which led to her involvement with RECOVER. Under RECOVER, Lisa’s co-authored research manuscripts and as part of the interventions task force, the National Community Engagement Group and clinical trial committees.

Mr. Mike Zissis is a dedicated Long COVID advocate who serves on the NIH [National Institutes of Health] RECOVER National Community Engagement Group, the RECOVER Observational Consortium Steering Committee, and the RECOVER-TLC Community Advisory Board. Mike’s initial COVID infection was in January 2021. Since then, many symptoms have subsided, but others have persisted, ranging from hearing loss to chronic pain and cognitive dysfunction. Mike is a Long COVID patient advocate focused not only on the science, but also on the care and needs of patients who can’t receive adequate care.

Finally, Dr. Frontera is a professor of neurology at NYU [New York University] Grossman School of Medicine. Dr. Frontera specializes in neurocritical care, treating conditions such as brain and spine trauma, stroke, and intracerebral and subarachnoid hemorrhage. Her research focuses on post-COVID sequelae, including cognitive outcomes, as well as blood and MRI [magnetic resonance imaging] biomarkers associated with Alzheimer’s disease. She has NIH funding for several projects evaluating neurological sequelae post-COVID and serves as site PI for multiple RECOVER studies. She’s a member of the World Health Organization Brain Health Task Force focusing on neurological sequelae of COVID-19 and serves on the RECOVER Ancillary Studies Committee.

The topic of today’s seminar is the Impact of Long COVID on Neurocognitive Function. Today’s speakers will share research findings on how Long COVID affects the brain, including memory, attention, language, and sense of smell. Panelists will also explore how RECOVER researchers are finding better ways to measure and understand neurocognitive symptoms, which can inform possible treatments. The discussion will include findings on the circulatory system and its possible connections to neurocognitive symptoms. Patient representatives will talk about living with these symptoms and why this research matters for the Long COVID community. Please welcome all of our panelists, and with that, I will turn it over to our first speaker, Dr. Becker.

Dr. Jacqueline Becker: 

Hi, everyone. Thank you all for being here, and thank you to RECOVER for the invitation. I was asked here today to talk about something that I’m deeply invested in, which is the impact of Long COVID on neurocognition. Just going to share my slides here. Okay. Quinn, can you tell me, is everyone able to see my slides correctly?

Quinn Barnette: 

Yes, we’re seeing them.

Dr. Jacqueline Becker: 

Okay, great. Okay. So, when RECOVER first launched back in 2021, the intention was to follow a broad range of symptoms and conditions to begin understanding Long COVID a bit better. But as data accumulated, a lot of different patterns emerged. And so, I’m also going to talk a bit about how Cycle 2 has made a deliberate shift from those initial broad surveillance questions to a more focused inquiry into the symptoms or conditions that most deeply impact patients, placing cognition really, from my perspective, front and center. So, in today’s talk I’m going to start by sort of grounding us in what we mean when we say “neurocognitive symptoms” or “brain fog” and providing a glimpse into why and how we measure cognition. Then, I’ll also walk through some key insights on cognition that we learned from Cycle 1, drawing on what we’ve sort of learned overall and where the field still has some important gaps, and then I’ll talk about how Cycle 2 plans to address those gaps.

So, beginning with what we mean by “neurocognition,” it’s a term we use all the time, but I recognize it’s not always clear to people what it actually includes. So, at its core, neurocognition refers to generally how the brain processes information. It’s what allows us to think and learn and remember and make decisions and function in everyday life. So, this includes several domains, like attention, executive functioning, memory, language, processing speed, visuospatial abilities, and others. And importantly, these systems work together constantly, often seamlessly, in the background, and they aren’t independent of one another.

So, now with that foundation, I just want to talk a bit about the burden—why neurocognition specifically deserves the attention that it’s getting in RECOVER Cycle 2. Well, most of us know that cognitive symptoms or brain fog are actually among the most commonly reported symptoms in Long COVID and are also among the most debilitating, and cognition is thought to be a pretty good predictor of daily functioning and ability to work. So, it really reflects this meaningful functional impact that patients report. But from a neuropsychological perspective, we measure cognition because it’s more than just the symptoms that are described by patients. It’s also a sensitive marker of neuronal integrity and function. When cognition changes, it often signals that something deeper in the system is off. And so, in many ways cognitive symptoms in Long COVID are really the tip of the iceberg. And what lies underneath may reflect some of those downstream effects of multiple interacting processes. For example, neuroinflammation, vascular changes, autonomic dysfunction, et cetera, which makes cognition one of the more sensitive indicators that we have of recovery or persistence in Long COVID.

Now, just going back here for a moment, I’m often asked in the context of Long COVID and other conditions where people describe this concept of brain fog. How does that apply here? So, where does brain fog live? Well, brain fog is not a formal neurocognitive domain. It’s not something we can point to actually on a single test even. Instead, I like to think of brain fog as being one’s experience of disruption across these systems. And so, when someone says, “I have brain fog,” what they’re describing is that their thinking feels slower, less efficient, less reliable. And importantly, it’s not just about whether they can do something; it’s also about how much effort it takes and whether they can sustain it.

And this gets at a really important distinction then in how we measure cognition, because we have 2 fundamentally different types of measurement that capture different things. The first we call objective neurocognitive tests. These are the tests that we typically administer in clinics—and if you’ve been involved in the RECOVER studies—that you have to come in person for. These are measuring cognitive capacity. They assess whether the brain can do something under standardized controlled conditions. So, what we’re really tapping into here is the integrity of brain networks and brain function at a structural and physiologic level. And so, when we see deficits on these tests, it tells us that something is going wrong in the system itself, and then we can look at the pattern of deficits to point to specific mechanisms or pathologic processes.

And then we have the patient-reported outcome measures on the other side, which are often measured through self-report questionnaires. These are capturing something different. They’re capturing the lived experience of people with Long COVID. So, how someone is actually functioning day to day, how much disability they’re experiencing, how their symptoms interact with the demands of real life. And that experience is shaped by many things beyond just raw cognitive capacity, like fatigue, sleep, mood, among many other factors. But importantly, these symptoms are equally valid, real, and very scientifically important. It’s also important to keep in mind here that these 2 types of measurement don’t always agree, and that discordance actually has been well documented for decades and across many conditions, not just Long COVID. So, what that means is that someone can perform sort of within normal limits on an objective test but still be struggling enormously in their daily life. And that’s not because one measure is right, and the other is wrong; it’s because again, they’re measuring different things.

And in the context of Long COVID specifically, this gets even more complex because of the fluctuating nature of the condition and the good days and bad days, the variability that patients describe. So, a clinic visit will capture one snapshot on one day in a controlled distraction-free setting, right? So, it may not always align with someone’s daily experience, especially in the context of distractors and life demands, but I’ll talk about why this discordance even is scientifically meaningful in a little bit.

To ensure that we’re measuring cognition rigorously, our tests need to be as fair as possible. And I just love this cartoon here. It says, “For a fair selection, everybody has to take the same exam: please climb that tree.” So obviously, for obvious reasons, this wouldn’t be a fair test. So, cognitive tests are standardized and calibrated against normative data, and so we consider impairment to be scores that fall 1 to 1.5 standard deviations below the normative mean. But to be really rigorous, we also need to estimate individual baseline abilities to understand if the sort of normal scores are truly normal, or if they reflect a decline for a given individual. And we also need to have performance validity measures to ensure that scores are truly reflecting cognitive capacity and not just reflecting those other factors I mentioned that can influence cognitive performance. And finally, we also have to measure both the objective, cognitive performance and subjective patient-reported outcome so that we can capture, again, different but very complementary information.

Now I want to walk you through how we actually assess cognition in RECOVER Cycle 1, focusing on the adult cohort only here, because I think understanding the measurement approach is really important for interpreting what has been found and for understanding why we made the changes that we did for Cycle 2. So first, not everyone was tested, and that’s largely because Cycle 1 followed a trigger structure for most assessments, including cognition. And what that means is that participants first had to report thinking problems or brain fog in order to then get a validated self-report measure called the Neuro-QoL. And if they were 1 standard deviation below the normative mean on the self-report measure, then they would get an objective measure, the NIH Toolbox, and then only if their toolbox performance was impaired, would they then be eligible for more comprehensive domain-specific testing as part of the full neurocognitive battery in Tier 3.

And importantly, there was also a random sampling of participants who were not triggered, meaning they did not report brain fog, but who also did these tests as well, but this was a smaller sample. And at the time, this approach made sense; it was efficient and scalable, but it also had some important limitations. So, besides the inherent challenges of the trigger structure, the NIH Toolbox has a fairly low ceiling, which means that it’s not so sensitive to detect the more granular and subtle impairment that exists in Long COVID. It also doesn’t quite assess all of those domains I mentioned in a comprehensive way, and in particular, executive functioning, which tends to be the most predominant impairment in Long COVID. And then the full Tier 3 neurocog battery was better, but it still had some major psychometric limitations that made it challenging to detect impairment.

And finally, there was no measure of premorbid ability and no measure of performance validity, which again just makes it difficult to characterize cognitive decline relative to individual baselines and to verify valid performance.

In order to increase the relevance of measuring cognition in RECOVER Cycle 2, we first did away with the trigger structure. So, now all participants will get all measures, and we also added in measures that were missing, added measures with better sensitivity and specificity for Long COVID based on the existing evidence and based on the data from Cycle 1, and we retained some of the measures from Cycle 1 that helped us find impairment where it existed so that we can still capture trajectories from the first cohort over the next several years.

And so, despite some of those challenges, the data from Cycle 1 really gave us a critical foundation from which to expand. One of the clearest findings is that cognitive symptoms are highly prevalent with about 64% of people with Long COVID in our cohort reporting brain fog, and this was based on the Long COVID research index update paper. So, that’s nearly two-thirds of people that met the LCRI [Long COVID Research Index] criteria identifying cognitive symptoms as a real and significant problem. And this estimate is actually really consistent with what we’ve seen across other studies as well. So, our team at Mount Sinai published one of the first studies to use validated in-person neuropsychological testing in 2021 in people who were an average of 8 months post-COVID and also found that there was a high frequency of impairment across several cognitive domains, and particularly in executive functioning and executive aspects of memory.

We recently followed that study up with another one to answer the question, “Does this get better?” And in this recent paper, we tracked cognitive trajectories over 42 months and over 1,500 participants. And while we found that many cognitive domains tended to improve somewhat over time, processing speed and executive functioning still remained at least 1 standard deviation below the normative mean at 42 months. So, that essentially tells us that people continued to struggle with the same impairment that we saw early on 6 years ago.

Back to RECOVER on the objective side, we’re currently finalizing a manuscript with the NIH Toolbox. So, we’ll have a better characterization of what this looks like for the RECOVER cohort probably later this year. And in the meantime, another important finding from Cycle 1 was from the olfactory dysfunction paper, where among infected participants with objectively measured smell loss, 67% also reported cognitive symptoms on the Neuro-QoL, which is the self-reported symptom measure that I mentioned. But because of the trigger structure in Cycle 1, the Neuro-QoL was only given to people who are already reporting brain fog, so we could only see that cognitive association within a preselected group. And so, this is something that we hope to be able to reassess again in Cycle 2 where everyone will get self-reported and objective cognitive testing, so we can really measure whether smell loss predicts objective cognitive impairment, not just self-reported concerns with participants who already told us that they were having brain fog. And in a moment, I’ll explain why this olfactory dysfunction–cognition connection may actually be really important for us to understand.

So, now let me turn to the gaps. What don’t we know? What do we still need to figure out through Cycle 2 from a neurocognitive perspective? I’ve organized these into 4 buckets: biology, trajectory, clinical, and mechanistic. The first is at the biological level. So, is SARS-CoV-2 causing transient functional disruption, or is it potentially injuring the brain in ways that have lasting consequences? On the injury side, the evidence is growing. The most compelling of these is a study from the UK Biobank a few years ago, which had the unique advantage of having brain scans pre- and post-infection in the same individuals. And what they found is that there were differences in brain size and other metrics that were consistent with the cognitive decline from pre to post in exactly the brain areas that we are finding impaired in our studies as well.

Other studies have shown elevated markers of brain injury in people with neurologic Long COVID symptoms. And beyond structural changes in blood-based biomarkers, there’s now direct neuroimaging evidence of neuroinflammation in the brains of Long COVID patients. Dr. VanElzakker, who you’ll hear from in a little bit, has found significantly elevated neuroinflammation across several brain regions in Long COVID, including the ones that are directly relevant to the cognitive domains that we are measuring. So, we’re not just seeing structural changes on MRI or injury markers in the blood; we’re also seeing active ongoing immune activation in the brain itself. And again, Dr. VanElzakker will walk you through those findings in detail.

But so, what we don’t know yet, despite these interesting findings, is the trajectory. So, are these brain changes progressive, meaning is the brain continuing to change months and years out? Do they plateau? Do they reverse? And if systemic inflammation decreases, whether naturally or through treatment, do the cognitive symptoms follow? So, for patients, this is really the difference between “this will get better, and this is something I need to learn to manage.”

And for us as researchers and for clinical trialists, it determines what we should be targeting and when is most critical to intervene. So, in Cycle 2, we’ll have comprehensive cognitive data at 2 time points 18 months apart on all participants, and we’re going to have this alongside detailed biospecimen collection. So, we can really begin to directly address some of these questions. And while neuroimaging isn’t going to be part of the core Cycle 2 protocol, this is where ancillary studies can actually be enormously helpful and of value as well.

So, the second gap builds directly on the first, and it’s quite a big question that many of us have been wondering for quite some time, which is, “Does Long COVID-related cognitive impairment put people at greater risk for conditions like Alzheimer’s, disease, or other dementias?” Let me be clear upfront: we are not claiming that Long COVID causes Alzheimer’s, and we don’t yet have data to make those claims, but there are some preliminary converging signals that warrant our investigation. For example, back to the signal with olfactory dysfunction or smell loss, we don’t yet have enough research on this, but considering the connection that (1) olfactory dysfunction is a common and often persistent symptom in Long COVID, and (2) that it has been recognized as an early marker for neurodegeneration in other populations, often preceding cognitive decline by many years, tracking this over time in an objective way is of interest and is something that we can robustly do in Cycle 2.

The second concern comes from blood-based biomarkers, where recent studies are showing that a protein called phosphorylated tau or pTau, which is sometimes found in some neurodegenerative conditions, is elevated in people with neurologic symptoms of Long COVID compared to both uninfected controls and infected individuals without neurologic symptoms. So, finding it elevated in Long COVID raises the question of whether SARS-CoV-2 infection is accelerating or unmasking neurodegenerative processes in potentially vulnerable individuals. But right now, we have these signals sitting in isolation, right? So, what we don’t have is all of this data linked together longitudinally in the same individuals. We also don’t know which patient characteristics like age, severity of infection, number of reinfections, et cetera, can really modify the risk of cognitive decline over time. And without robust cognitive data, these biomarkers are just numbers, right? The cognitive tests are actually what give these numbers clinical meaning, which is why in Cycle 2, we will hopefully be able to start linking biomarker trajectories to cognitive trajectories.

And a particularly exciting opportunity for ancillary studies is that our Cycle 2 neurocognitive battery is largely harmonized with some of the other well-established datasets that are tracking neurodegenerative conditions, meaning that we can directly compare Long COVID cognitive trajectories to some of those established datasets to see if similar patterns emerge.

And now the third gap is one that I really feel strongly about as a clinician and one that I think resonates, especially with our patients here on the line today. So, how do we characterize brain fog in a way that actually reflects what people are going through? As I discussed at the beginning, brain fog has become sort of a catchall, right? It communicates something very real and meaningful, but scientifically, it’s not very precise. And so, as we collect more data, it’s becoming clear that brain fog isn’t just one thing. It likely includes several different patterns of symptoms and each with different underlying causes. And so, one of our goals moving forward is to sort of break apart this broad idea of brain fog into more specific and meaningful categories. We’re also asking some important questions over time. If someone reports brain fog now, could that be an early warning sign of measurable cognitive problems later?

The hope is that in Cycle 2, we’ll be able to define some cognitive phenotypes a bit better to understand the objective and subjective discordance, and this will hopefully help identify some better intervention targets. And then with ancillary studies, this is also a really good opportunity to validate some innovative cognitive tools and digital cognitive platforms against some of our gold-standard measures to try to really capture symptom variability and fluctuations.

And then finally, the fourth gap that we’re trying to understand is how much cognitive symptoms in Long COVID are connected to problems outside the brain, so specifically, the heart, lungs, and autonomic nervous system. And of course, this will require a lot of multidisciplinary collaboration, which RECOVER is very well positioned for. I think it’s easy to think of cognition as being isolated from the brain, but as we know, that’s often not how people experience cognitive symptoms, right? They experience them sort of in the context of multiple other co-occurring conditions, and the brain really depends heavily on the rest of the body, and even small disruptions can affect how well it works. And so, the key unknowns here are, when someone’s having cognitive problems, what’s driving it? Is it a direct effect on the brain, like neuroinflammation, as I mentioned earlier, or other changes, or is it coming from these broader body systems not supporting the brain properly?

And for many people, it may be a mixture of many different things. And so, this is where things get really important. By collecting data on cognition, heart function, and autonomic function at the same time, we can also start to see what’s most closely linked to cognitive symptoms. For example, do thinking problems track more with heart rate and blood pressure changes, or with markers of neuroinflammation, or both, right? And so, answering these questions will help us understand not just what’s happening, but how to treat it and who will benefit most from which treatment. And so, all of this, again, makes cognition a really powerful downstream functional biomarker, even for when the issue is not in the brain itself but with other systems in the body. And here, ancillary studies will also have the opportunity to study this in new ways, like using different kinds of brain scans that can look at blood flow to the brain and some newer tools.

And so, all of this is a really big reason why Cycle 2 of RECOVER has been organized around essentially these 3 core aims. So, neurocognitive dysfunction, cardiopulmonary dysfunction, and autonomic dysfunction. And what I hope has come through here is that these aren’t 3 separate problems, right? They’re deeply interconnected, and this is what makes cognition such a valuable cross-cutting outcome. Cognition is really sensitive to disruption, regardless of where the disruption originates again. So, whether, again, it’s reduced blood flow or autonomic dysfunction, cognition is really a downstream functional readout of how well the whole system is working. And one of the most important roles of observational research like RECOVER is that all of this can help us identify patterns and uncover mechanisms so that again, we know what to treat and in whom. And so, that’s exactly how RECOVER has informed several clinical trials like RECOVER Neuro and others that will be covered in a separate seminar and how it has informed the upcoming Treating Long COVID trials (or RECOVER-TLC) as well.

All of those gaps I walked you through, the biological mechanisms, the cognitive trajectories, the phenotypes, the autonomic and cardiopulmonary links, they sort of point us toward specific treatment targets, and that’s exactly what RECOVER-TLC is designed to address. RECOVER-TLC is currently moving forward with 3 specific therapeutic approaches. So, it’s the low-dose naltrexone trial, a GLP-1 [glucagon-like peptide-1] trial, and a stellate ganglion block trial, all looking at cognition as a secondary outcome. And one of the major benefits of falling under the same RECOVER umbrella is that we’ve been able to at least somewhat harmonize these measures across at least the adult trials so that we can later compare results within and between the trials and the observational cohort as well to gather even more information on what works and what doesn’t.

And briefly also under the RECOVER umbrellas, the REVERSE-LC trial, which is a Phase 3 trial that’s testing another pharmaceutical called baricitinib. And for this, the primary outcome is specifically neurocognitive function and physical impairment. So, just to quickly bring it back to where we started, I talked about cognition as sort of the tip of the iceberg, the thing that patients see and feel, but what I hope I’ve shown here today is that when we measure it rigorously, cognition becomes more than just the visible part, but rather it can be one of the few tools we have that can tell us what’s going on underneath the symptom layer. And so, RECOVER Cycle 2 is really an opportunity to move forward from just describing these symptoms to understand what is driving them and how we can treat them. And so with that, just want to quickly acknowledge RECOVER leadership, the Sinai PIs, the in-person working group members that were so thoughtful in thinking about Cycle 2’s cognitive battery, and the RECOVER staff and participants, and of course, the Patient Representatives who really bring such critical perspectives to this work.

Thank you so much, everybody.

Quinn Barnette: 

All right. Thank you so much, Dr. Becker. I think we will now turn it over to Dr. Frontera to give us an overview of the state of the science with Long COVID and cognitive impairment. Dr. Frontera?

Dr. Jennifer Frontera:

Okay, great. Thank you. I’m just going to share my screen. Okay. Okay, can you see my screen okay?

Quinn Barnette: 

Yes, we can.

Dr. Jennifer Frontera:

Great. Okay. So, I’m going to pick up where Dr. Becker left off and talk a little bit more about possible mechanisms that might underlie the cognitive findings that she discussed. Let me just see. And so, I want to tackle this kind of elephant in the room here is what is causing this kind of cognitive impairment, and could it actually be a neurodegenerative process? And neurodegenerative is sort of code for Alzheimer’s type of neuropathology or Alzheimer’s disease. So, post-viral Alzheimer’s pathology is not a new idea. It’s been thought of for decades, mostly linking herpes virus–type infections like HSV [herpes simplex virus], EBV [Epstein-Barr virus], to neurodegenerative pathways. And so, the inflammatory pathway leading to buildup of typical Alzheimer’s-type neuropathological substances has been looked at for many decades.

In the context of SARS-CoV-2, particularly the first strains and up through Delta, we had this sort of perfect storm where these patients were hypoxic and had a cytokine storm, so a hyperinflammatory state. And these things can conspire together to really lead to blood–brain barrier permeability, particularly in the context of systemic hypoxia. And when this happens, when you have a leaky blood–brain barrier, you can activate these cells inside the brain called microglia. Microglia then secrete their own cytokines, so they perpetuate this sort of inflammatory process in the brain. Downstream from these cytokines, you can really activate a variety of enzymes that can lead to increased amyloid beta production, decreased amyloid beta breakdown, and increased neurofibrillary tangle formation. And these are all classic neuropathological components of Alzheimer’s disease pathology.

Just to show you a few different examples, there’s been a lot of big data using electronic health records looking at correlations between having COVID and increased diagnoses of dementia using ICD-10 [International Classification of Diseases, 10th revision] codes. And so, I’m just showing you 3 papers across different spans from index SARS-CoV-2 infection at 6 months, 12 months, and 2 years. And you can see the red line is the COVID-19 patients. The blue line is other types of respiratory tract infections. And there were higher frequency of diagnoses of dementia in the COVID-19 population compared to other types of respiratory infections at 6 months, at 12 months, similar kind of finding increased new dementia, specifically Alzheimer’s disease diagnoses in the COVID-19 people compared to COVID-negative cohorts. And at 2 years, similar findings, where here again, the red line is COVID-19 and the blue is other respiratory tract infections where there’s increased dementia and cognitive deficit diagnoses.

Now, this is all very interesting, but there’s a lot of limitations related to ICD-10 coding. First of all, it’s not necessarily accurate, but also there may be a component of referral bias. So, obviously the pandemic was highly visible. There was a lot of press related to brain fog and cognitive impairment. It’s possible that people that had preexisting or pre-COVID cognitive impairment or dementia or mild cognitive impairment sought medical attention because of this and there’s just more diagnosis happening. It’s also possible that people that have dementia or mild cognitive impairment are more predisposed to having COVID and more likely to present to the hospital. So, there’s issues with this kind of data.

I’m going to show you some of our work. This is NIH NIA [National Institute on Aging] funded. We basically looked at longitudinal cohorts of patients that were COVID negative, meaning they were nucleocapsid antibody negative and had no clinical history or episodes that resembled SARS-CoV-2 infection. We looked at COVID positive folks and then people with Long COVID, meaning new symptoms following index infection that lasted at least 3 months. And we followed these people over time and subjected them to not only comprehensive neuropsych testing, but also interviews with clinicians specializing in cognitive impairment, as well as informant interviews. And then taking into account the neuropsych findings, as well as the interviews, as well as other comorbidities, there is an adjudication process that occurred with a panel of experts, and patients were diagnosed with a variety of things that could be cognitively normal or have mild cognitive impairment [MCI] or dementia.

And so over time, when we looked at these 3 groups, the cumulative incidence of MCI or dementia over 4 years was significantly higher in patients that had Long COVID compared to those that had COVID but recovered and compared to those who are COVID negative. And just to give you a sense, the background rate of mild cognitive impairment diagnoses, and you can see the age groups here; the Long COVID people, if anything, were somewhat younger, and more than 50% were female but less a frequency of female predominance compared to the other 2 groups. So, they had less risk overall of MCI or Alzheimer’s-type dementia, and yet they still had a higher incidence over time. The background incidence for this age group is about 2% to 3% per year. So, the cumulative incidents, as we would expect, would be 8% to 12% over this 4-year period. And so, what we’re seeing here is substantially higher for the Long COVID patients.

When we sort of break this down, mild cognitive impairment has several types. What we’re interested in is the Alzheimer’s type, meaning that there’s memory impairment, which could be single domain or multi-domain. There could be other types of impairment also, but there has to be a predominant memory component to it. When we looked at this, Alzheimer’s type of MCI was also significantly higher in patients that had Long COVID than those who did not. MCI or mild cognitive impairment related to psychiatric illness was not significantly different across these groups. And MCI related to systemic illness showed some significance. So, some of this is related to being chronically ill, but notably, the MCI Alzheimer’s type was certainly substantially significant. And this is after adjusting for other covariates.

We also looked at the biomarker phosphorylated tau 217. This is really turning out to be probably one of the most important biomarkers related to Alzheimer’s-type pathology, particularly more so than pTau 181 or 231, some of the other types of pTau that have been looked at in the past. And there was a significantly higher pTau217 in those that were diagnosed with Alzheimer’s-type MCI compared to those who were not, sort of substantiating these diagnoses. We did take out patients that were hospitalized just to see if some of this cognitive impairment was due to being critically ill, and the results are still robust and significant, suggesting this is not just related to the sickest of the COVID patients who might have been ventilated or hospitalized.

Looking a little bit deeper at pTau 217 and Long COVID, this is a slightly larger cohort than what I just showed you, but similarly, there’s Long COVID, recovered COVID, and COVID-negative controls, and pTau 217 was substantially higher, significantly higher, in Long COVID patients compared to other groups. I’m just showing you compared to recovered COVID and compared to COVID-negative folks. Interestingly, if we looked at people with just brain fog versus no brain fog, there was not a significant difference in pTau 217. And this goes back to some of the things that Dr. Becker was referring to earlier about symptoms not necessarily correlating with objective pathology. So, in the context of Alzheimer’s disease, having what we call an anosognosia, which means you don’t really recognize that you have a problem, is a prominent finding. And so, there may be a component or patients that don’t think they have brain fog or don’t mention it but are still having a potentially neurodegenerative process. And so, I think it’s important to note that we don’t want to just look at people who acknowledge that they have this symptom, but some patients that don’t even have this symptom may be at risk.

Part of our study looks also at MRI, and I’m just going to present you with some of this, because we’re still analyzing all of our data, but we looked at the choroid plexus. Now, choroid plexus is interesting, because it’s part of the glymphatic system, which is sort of like brain lymphatics. It helps clear waste products from interstitial space, which is one idea that if the glymphatics aren’t working, you can get reduced clearance of basically cellular garbage such as neurofibrillary tangles or amyloid beta waste products. And when the glymphatic system is not working quite well, you actually see enlargement of the choroid plexus. It’s almost like a fibrotic change in the choroid plexus. And this correlates very strongly with Alzheimer’s, dementia, as well as elevated pTau 217 in the literature apart from in the context of COVID. And we’ve broken down our patients again into Long COVID recovered and healthy controls here.

You can see there’s a significantly larger volume of choroid plexus in the Long COVID patients compared to recovered COVID and healthy controls. There’s also reduced cerebral blood flow in the Long COVID patients compared to the recovered patients and healthy controls. And so, sort of correlating with tissue that’s not getting as much cerebral blood flow and is enlarged, so dysfunctional. Choroid plexus volume also increased volume inversely correlated with things like mini mental status and some of the other neuropsych testing that we did. pTau 217 is also significantly correlated with volume. So, higher chord plexus volume was associated with higher pTau 217 levels and inversely correlated with cerebral blood flow. So, lower cerebral blood flow was related to higher pTau 217 levels.

Now, just as a sort of last word, the way that Alzheimer’s is diagnosed in terms of being standard is neuropathology. And so, what we would like to really confirm is if we’re seeing these neuropathological changes in Long COVID patients. If this is a possible etiology of brain fog, then we really need to look at the neuropathology. We did this in a very small group of patients that were COVID positive and then 12 COVID negatives before 2020, as well as people who died from hypoxia. And just to show quickly the amyloid beta deposits, which are these dark kind of reddish brown, you can see that these are significantly more commonly occurring in the COVID-19 patients compared to controls. Similarly, also COVID-negative hypoxia patients had these similar findings, so there’s a question of whether hypoxia was driving a lot of this pathology to begin with. Phosphorylated tau, which is another hallmark of Alzheimer’s-type pathology, is also significantly higher in the COVID-19 patients. And then we saw neuroinflammatory markers like activated microglia and activated astrocytes occurring more frequently in COVID-19 cortical regions compared to control patients.

This is a small subset of patients; we really need more neuropathological data to sort of understand if what we’re observing, at least in our cohort, is indeed Alzheimer’s-type pathology. The RECOVER autopsy cohort has a really rich body of evidence related to brain specimens, half of which have undergone MRI with 7T imaging. So, you could actually correlate neuropathological findings to MRI findings, which would be a huge advance in our understanding of if there is actually neurodegeneration occurring and what it might look like on MRI. So, a lot more to go, and this is just a sort of view into where we’re at right now, I think neuropathologically speaking. I’m going to stop here and turn it over to our next speaker.

Quinn Barnette: 

All right. Thank you very much, Dr. Frontera. I think now we’ll turn it over to Dr. VanElzakker.

Dr. Michael VanElzakker:

 Okay. Hi, everybody. Thanks so much for the invitation and for letting me talk about this today. I work in a lab that’s been focused on infection-associated chronic illness for a while now. And so, we have sort of a research program that is ongoing, and we are including people with Long COVID. I think when Long COVID hit, a lot of people were a little bit surprised, at least a lot of reports showed some surprise that some people weren’t feeling better after weeks or months, or even in some cases, years. And so, we decided early on to write this paper, Dr. Proal and I, with whom I co-founded PolyBio, where we laid out some research priorities that we thought were important for studying Long COVID, including taking the pathogen itself seriously. And one of the things that we focused on in this paper was neuroinflammation. We’ve heard some great background from Drs. Becker and Frontera about this topic.

Neuroinflammation is the activation of the central nervous’s glial immune cells. So, when you look at a brain, the majority of the cells that you are looking at are not neurons; they’re actually glia, which is a type of neuroimmune cell. Glia actually means “glue” in Greek, because it used to be thought that they just simply held together neurons, but it turns out that they actually do a lot, and they activate upon inflammatory signaling and they produce their own inflammatory signaling.

Something like 40 years ago maybe, we could measure neuroinflammation by doing a lumbar puncture and then just measuring inflammatory molecules in the cerebral spinal fluid, and that’s what you had to do. That’s still what happens in clinics. And you can measure neuroinflammation in very severe neuroinflammatory conditions like multiple sclerosis. What we’re talking about here is not that level of neuroinflammation; it’s something a little bit more subtle. And with modern research techniques, we can use sensitive neuroinflammation imaging to try and understand if that’s an important process. We use positron emission tomography brain imaging, so PET scan imaging, where you’re actually injected with something that binds to the immune cells that are activated in the brain or anywhere else in the body, and you can image that. And we’re lucky enough to have a scanner that has PET scanning and MRI scanning in the same scanner. So, we can do concurrent MRS [magnetic resonance spectroscopy] scanning, which measures the concentration of certain chemicals.

I’m not going to talk about that today, but what I will talk about is at the same time that we did the PET scanning for neuroinflammation, we did some cognitive tests with an MRI scanner to look at fMRI [functional magnetic resonance imaging] responses. And so, this is all taking an existing PET/fMRI program that we had going with patients who were diagnosed with ME/CFS before the pandemic, and we turned it over to Long COVID. A lot of this work is based on this paper that we wrote before the pandemic about how to study neuroinflammation and some of the mechanisms that may be important.

So—glia, there are many different types of glial cells. The main one is microglia, which are the macrophages of the brain. And just like other immune cells in the body, when they come in contact with something from a pathogen, including something like a spike protein or RNA, or when they come in contact with inflammatory molecules like pro-inflammatory cytokines, they actually change their function and shape a little bit. They pull in their long, skinny arms; they kind of get thick and fat, and then they release a whole bunch of stuff. They release immune signaling, so like cytokines, which then can turn on other neighboring immune cells like other glia, and they release a lot of pro-excitatory modulators like glutamate, which is the main neuroexcitatory neurotransmitter, ATP [adenosine triphosphate], prostaglandins, lots of other things. This is really important for keeping neighboring neurons alive and robust in the short term. But then if this keeps happening long term, it can drive problems, because the brain normally operates on a very sensitive signal-to-noise functionality, and this additional excitation cannot help.

One of the things that these activated glial brain cells release or produce is called the translocator protein. And this is the thing that we can detect using PET scanning, and this is what we’ve done. In the first paper that we did, these scans are really expensive. NIH RECOVER was able to pay for about 20 of the scans, and the rest we paid for with PolyBio; blood tests and saliva tests, et cetera, was paid by PolyBio. In this study, we scanned 12 people with Long COVID, and we used 43 controls from people who were scanned before the pandemic, so we know that they didn’t have SARS-CoV-2. And the patients that we scanned got sick before Omicron. Most of them were not hospitalized, and our recruitment materials prioritized neuropsych symptoms, and we tried to make sure that they had sort of a broad phenotype symptom expression. So, we used the ME/CFS-ICC [International Consensus Criteria] criteria to make sure that they qualified for that. And then later on, I’ll show you post-COVID controls.

This is all just showing that our patients and our controls were reasonably well matched in other things that might explain the differences that we detected. We used sort of a standard PET scan analysis normalizing to cerebellum within each person and then comparing across groups. And this is basically what we found. What we’re seeing here is a heat map with the signal for the PET ligand that binds to excited or inflamed glial cells showing up in red and yellow. And it happens across a wide swath of brain regions. Interestingly, as Dr. Frontera mentioned, we did see signal in areas where there seemed to be normally and naturally a diminished blood–brain barrier in the brain. And in the middle here where there’s some areas which informed some questions that we answered next. We also saw activation in this region called the anterior cingulate cortex, which is important for cognition, and I’ll talk about that in a bit.

This is just pulling out data from each individual and plotting it across different brain regions. The point here is to show that there isn’t really one brain region; this is the Long COVID brain region. It doesn’t really work like that. And also, it’s probably a little small on your screen, but I put the hospitalized patients in an open circle, and the point there is to show that this isn’t an effect that was just driven by hospitalized patients. So basically, the patient showed more than the controls, even though there is a natural and normal variability within patients. Because we saw signal in areas near the blood–brain barrier, we thought that we should take some blood that we’d collected right before they got into the scanner and see if that predicted anything in the PET signal that we saw. And we did find several markers in the blood that are related to vascular dysfunction correlated pretty well with the PET signal in the brain.

This is showing sL-selectin, which is a kind of immune signaling molecule. Pentraxin 2 regulates response to tissue damage, and a correlation of 0.7/0.6 is moderate to strong. Here we’re showing fibrinogen, which is part of the clotting cascade and alpha-2-macroglobulin, which is a marker of endothelial dysfunction, vascular dysfunction, and a risk factor for cardiovascular disease. This is to show that we think maybe there’s a vascular component here. And there’s ongoing work looking at fiber and amyloid microclots and also looking to see if we can detect spike protein, viral RNA, et cetera, that might be driving the neuroinflammation, because we don’t really think it’s sort of coming from nowhere. All that is described in this paper here, if you’d like to go look it up.

The next step in our program was to look at COVID-recovered controls to see if there’s any difference there. And same thing here where we’re showing that the controls and the cases are pretty equally matched. We went out of our way to try and get the same ratio of sex, the same genotype that matters for the translocator protein, body mass index, et cetera. And so, what we’re showing here is, again, a heat map where areas of neuroinflammation are increased in the patients versus the controls. You can see the bar graph over here—each dot is an individual. And again, it’s not the case that I could just look at one individual and say, “Ah, there it is. There’s the Long COVID.” But rather when we take a group of patients and compare to a group of controls, they show up statistically different. And most of the patients were above average in terms of the control average. This here is just highlighting brainstem, which we think is an important issue, but there are other brain regions that I’ll talk about in a moment.

We had a great intro from Dr. Becker about brain fog, that it’s sort of this general nonspecific symptom that can have a wide variety of severities. And we even see patients with Long COVID that don’t really report brain fog. They say, “I can’t really make it up the stairs without breathing hard, but I feel pretty sharp still.” So, we deliberately recruit for patients with these sorts of symptoms, but in order to study it, we have to be able to operationalize it. How do we turn this into something that we can actually study in a lab setting inside of a scanner? So, it can’t sort of mimic everyday life. We have to sort of boil it down. And what we can do is to conduct cognitive tasks right inside the scanner and then try to understand whether performance on those tasks might be related to the things we’re measuring.

There’s a fundamental aspect of cognition called top-down cognitive control. And this is a really important aspect of being able to stay focused and being able to stay on task, et cetera. Let’s say I’m looking through a crowd for my friend with a red shirt. It’s a crowd, it’s busy, there’s lots of things going on. Being able to focus on my task of looking for my friend is a type of cognitive control. And we understand how this works pretty well thanks to research that’s been around for a long time. So, I’m focused on my friend with the red shirt. Let’s say all of a sudden, a balloon floats up and it grabs my attention. What that will do is to bring my attention to this thing that I wasn’t really even caring about, and I’ll lose attention for the thing that I was caring about.

And we think that this is probably an important function in cognition in patients that sort of informs cognition across multiple domains. And it turns out that inflammation affects some of the circuitry that’s involved. Here you’re seeing a paper that looked at a lot of different neuron inflammation studies, and they’re showing that cingulate cortex is one of the areas that gets activated by inflammation.

Well, it turns out that we can study that pretty well with lots of different tasks—one called the Multi-Source Interference Task, which is what we gave to our patients inside of the scanner. They hold onto a button box inside of the scanner that’s magnet safe, and they have to choose the numeral that is different than the other numeral. So, if you put your right hand with the palm facing the screen, here you press your index finger because the answer is 1. That’s the different numeral. Here, the answer is 2, here the answer is 3. Here, the answer is 3. Now, it’s actually kind of hard, because it’s in the space of 1, but it’s in 3. And that is part of overriding that sort of red balloon of distraction. So, this is what patients experience in the scanner.

Two, 1, easy, 2, 1, 3, hard, 2, 3, 1. So, when we give this task to patients inside of the scanner, we sort of expect that patients will have to work a little bit harder than controls, and that’s what we see. So, the Long COVID patients, this here is actually a subset of what I showed you before. Right in the area called the anterior mid-cingulate cortex, which is the area that I showed you was affected by inflammation. They require more brain activity to do the task. They can do the task. It just takes a little bit more brain effort, it looks like, in the patients they control.

If I take the PET signal—what you’re looking at now is the neuroinflammation signal right here in this area that was predicted to be more affected by inflammation—and I superimpose it with the MRI of the cognitive task, it looks like this. We think that this is part of what is happening with brain fog and trouble concentrating is that the inflammatory signaling, which again, may come from spiked protein, from viral RNA, it may be vascular damage, it may be sort of ongoing inflammation, immune cells can become activated and stay activated. It could come from different sources, but you end up with this scenario where cognition is affected by neuroinflammation.

With that, I just want to thank my team. They do a great job. Patients have a tough time in the medical system often, and my team does a great job. Everybody reports good experiences. We’re standing next to our PCR [polymerase chain reaction] test where we test ourselves. We always wear masks when the patients come and visit us, and just thank you to everybody that was involved.

Quinn Barnette: 

All right. Thank you so much, Dr. VanElzakker. I think we’re going to turn it over now to a panel discussion with our Patient Representatives, Liza and Mike, and I’m going to turn that over to Dr. Frontera to lead that discussion.

Dr. Jennifer Frontera: 

Great. Thanks everyone for your talks. They’re wonderful, very provocative. And so, I’d like to start with Liza and Mike, and maybe let’s take Mike first, and maybe you could tell us just briefly about your personal experience with COVID, and maybe specifically as it pertains to brain fog or cognitive issues.

Mike Zissis:

Hey, sure. Doing my favorite thing, public speaking. My COVID started January 7, 2021, and on the third day was the first strange thing that made it feel like something weird was going on in my brain. I had this bicep that I’d torn 2 months earlier, and it had been busted, repaired. I woke up third day of COVID, and it felt the same way that it felt when I busted it for 2 days and then it went away. I woke up on day 8 with numb shins and numb feet, bilateral unequal, and it was just a really weird presentation where I’m going like, “This doesn’t make any sense.” And then multitasking—that’s the first thing I noticed was that I couldn’t multitask. And then I started noticing that I couldn’t... My short-term memory was extremely... I call it extreme short-term memory loss, which has come back a little in the last couple of weeks.

I did a neurocognitive exam, not the one with RECOVER. It was recommended by a neurologist that I was seeing. It was 4 hours long, longer than the one I’ve done with RECOVER. And it crashed me. That was the first indication that cognitive stuff could crash me. That’s the first time it happened in my life, and that put me in a 4- or 5-week crash. Now I’ve got ADHD-type behavior, inattention, trouble focusing, staying on task, and easily distracted. And then I wonder about neurodegeneration. So, I’ve had a couple of brain MRIs. The second one I did, the hippocampal volume is at fifth percentile left, seventh on the right, which all things are measured differently, but with all the other symptoms, I’d like to know what’s going on. I believe that the brain is where everything starts. To me, it was always like something was wrong in my brain. The signals going to it weren’t always right. The signals coming from it weren’t always right. And I don’t have neuroinflammation like some people where they can feel it, but there are people that say their brains hurt. So, that’s enough. I rambled a little.

Dr. Jennifer Frontera: 

Okay. Thanks, Mike. Yeah, I was hearing some post-exertional cognitive issues there, too, which is also something that we do hear from folks. Liza, do you want to share your experience?

Liza Fisher:

Sure. Thank you for having me, and thank you all for presenting your information today. It’s been very interesting and very relative. I had a little bit more of a severe acute COVID presentation but was kind of dismissed and unable to get care here in Houston, Texas, because it was just during that first wave in Houston in June of 2020. So, it was still difficult to get tested and whatnot. However, while I was home isolating, I had already started noticing neuro along with GI [gastrointestinal] and respiratory. But at that point, Mike said something similar where he could tell it was in his brain. I remember being on the phone with my mom and saying, “I don’t know what’s going on, but it’s in my head. Something is wrong with going on inside my head and I just don’t know what it is.” And then shortly after, I would be at the point where I would forget what I was talking about during conversation, along with some other neuro symptoms.

Five weeks later I got into ICU [intensive care unit], and that’s where things just kind of went “blah.” After that I went into a rehab hospital, and I had a lot of neurological effects. I was still having trouble speaking, everything like that, neuromuscular, but cognition, I was not doing very well on that MoCA [Montreal Cognitive Assessment] test. I can’t stand that thing. I couldn’t remember the animal, which you had to remember 5 minutes later, all that stuff. So, I was transitioned right into speech therapy. And then during my few months after that, I went into outpatient therapy, and that was based around a lot of people who have brain injuries. So, I got that type of treatment a little bit early on in that type of functional testing, which I think helped me in being able to get some progress back for me personally.

Along that line, I dealt with a lot of what you’re talking about, memory, attention—like Mike–I’m neurospicy now. Just like that balloon that Dr. VanElzakker showed, I’m not focusing on that red shirt at all. Everything is just balloons all the time, which affects my entire nervous system. And I do get different types of post-exertional malaise. The cognitive one is different as well. So, I will go ahead and stop before I start rambling as well. Thank you.

Dr. Jennifer Frontera: 

Okay. Neurospicy. Okay. I haven’t heard that, I love it. So Liza, since we’re on you, let’s just stick with you for a second, and maybe you could give us your thoughts about some of the stuff presented today and what would be important for clinicians or researchers to understand from your patient perspective as it pertains to ongoing studies of neurocognition post-COVID?

Liza Fisher: 

I think some of the most important parts that remain in the presentations today are how the symptoms both fluctuate and how difficult it can be to capture them in the current type of testing that we have for neurocognitive symptoms at this point. I think that is a huge point. And as I discovered in my own journey, the different types of testing that I got throughout the timeframe.

Dr. Jennifer Frontera:

I think-

Mike Zissis: 

Oh, no.

Dr. Jennifer Frontera: 

Liza’s freezing there. Mike, I might let you jump in while Liza’s coming back online.

Mike Zissis: 

Yeah, maybe she—

Dr. Jennifer Frontera: 

Same idea.

Mike Zissis:

... put off the camera. Let’s see.

Dr. Jennifer Frontera: 

Similar question, Mike. Just what do you think is important for researchers and clinicians to focus on, maybe moving forward as it pertains to cognitive symptoms after COVID?

Mike Zissis: 

Yeah, so there’s a lot there as far as I’m concerned. First of all, like I think Liza was saying, and we all say, we don’t always present on tests for what we’re feeling. So, to clinicians and researchers I say, but more to clinicians is, believe the patient. Even if they show up and you’re like, “I can’t find anything,” because that’s very important. There’s this lived experience conversation, and we have it amongst this community, but you would also have it with your doctor saying, if they say, “This isn’t showing anything, you’re the patient, you don’t have time to make things up.”

And it’s important in this community, because we’re the science, we are the science. Research can’t do anything without us, and we can’t do anything without research, but we’re the real-time learning right in front of you. So, we’d like to see things expedited, and we’re working on that. And then brain fog, just we got to come up with a better term. Or most of us with brain fog, I consider I have cognitive impairment or I have cognitive malfunction. Some people like to say they have brain damage. So, we would like to go with not necessarily brain fog, because if I say that to a neurologist that I’ve just met, that’s a lot wider term than saying, “I have cognitive impairment.” “Oh, okay, I understand what you’re saying.” That’s narrowed down a little bit. So, I think if Liza wants to come back, she can come back. But most importantly, believe the patients, because none of us is doing this for our own fun.

Dr. Jennifer Frontera: 

Thanks, Mike. Appreciate your insights. Liza, we lost you for a second. I don’t know if you want to finish your thoughts there about what you think researchers and clinicians should be focusing on in regard to post-COVID neurocognitive issues?

Liza Fisher: 

Yeah, sorry about that. And thank you for repeating the question because I already forgot where I was at. I think the points that you make in your presentation about how it is very difficult to... The testing that we have currently is sensitive and how it fluctuates at different times and phases throughout the illness per person, is something that people need to understand. And I think I caught a little bit of the tail end of Mike’s. I agree, the severity. I mean, it was not brain fog to me. It was definitely a deficit that was able to be calculated. So, that may not be what is necessarily presented on paper for everyone else, but it’s definitely what they live with every day.

Dr. Jennifer Frontera: 

Okay, thank you. I think we might start to take some questions from the audience. I don’t know if, Quinn, you want to take over for that?

Quinn Barnette: 

Sure. Yeah, we can turn over to the audience Q&A. There was a lot of interest. We got a lot of questions, so we’ll try to get to as many as possible. But just as a reminder for the audience, we will post a Q&A document on RECOVERcovid.org for any questions we can’t get to today. And also, just as a reminder, we won’t be able to answer any questions about individual clinical treatment. And the way I’ll do this is I’ll try to direct questions to the appropriate person that I think can answer the question, but I want this to be conversational, so please feel free to chime in if you have any input on any specific question. The first question I think I’ll direct towards Jennifer and Jackie, and it asks, “Do you see the same amount of degeneration in dementia in individuals who were much younger when they had SARS-CoV-2 and Long COVID?”

Dr. Jennifer Frontera: 

What I can say is that... I don’t want to say “dementia,” because mostly what we found was mild cognitive impairment, and that’s a diagnosis that’s extremely specific. And we had one patient in our cohort particularly that developed dementia, and that was a person that was working in finance pre-COVID, had COVID, and had a really precipitous decline and was not able to work anymore, et cetera. But the vast majority of people have mild cognitive impairment—they do tend to be younger than what we would see in a normal non-COVID population in terms of timing of presentation with mild cognitive impairment. That being said, usually we’re still looking at people in their 50s and 60s. Less often is it people in their 20s and 30s. They may have mild cognitive impairment, but not usually Alzheimer’s type, if that could be reassuring to people at all. But yes, we are seeing it sort of earlier than what you would typically see.

Quinn Barnette: 

Jackie, did you have anything you wanted to add to that?

Dr. Jacqueline Becker: 

Yeah, I agree with what Dr. Frontera said. And I think the dementia risk studies are sort of detecting early neurobiological signals over, say, 2 to 5 years. But I think within the broader dementia literature, we see that elevated dementia risk from serious infections can persist across follow-up periods of 25 to 30 years. So, it’s too soon, I think, for us to really pinpoint anything. But there was a recent Lancet study that showed that older adults had a higher absolute dementia rate because they started from a more vulnerable baseline, but younger patients actually showed a greater relative risk after SARS-CoV-2 than older patients. Again, this was just from a recent Lancet study.

Mike Zissis: 

If I could chime in real quick and anecdotal, I was 57 years old when it started, and so you’re all always assessing, okay, is this normal for this person? But our neurocog tests are done to your peer population, but we don’t know the person before they had that neurocog test. So, what I have going on, and a lot of people 20 years younger than me, my friends that have this, I’d say I know people in their 40s and 30s, late 30s that are having the same level and depth of this. So, it’s hitting younger people too as far as I’ve seen.

Quinn Barnette: 

All right, thank you. I think I’ll open our next question up to the full panel. It asks, “Can you talk about the loss of neurons in damaged brainstems that lead to breathing dysfunction?” And they’re citing a study from Avindra Nath, “Is the NIH RECOVER using MRIs to investigate the root causes of patients with neurological symptoms?”

Dr. Jacqueline Becker: 

Yeah. So, Avi Nath, who was at NINDS [National Institute of Neurological Disorders and Stroke], showed that in severe COVID, there can be a loss of neurons in brainstem regions, again, in severe COVID. And these brainstem regions are sort of in charge of controlling automatic breathing. So, it’s the reason that we know to inhale and exhale without having to think about it. But we don’t usually see that level of structural damage in the context of Long COVID necessarily, though it’s possible that the same sort of brainstem autonomic systems may be dysregulated in Long COVID. Maybe Dr. VanElzakker would like to comment, too, since it involves a lot of neuroinflammation.

Dr. Michael VanElzakker: 

Yeah, I think that’s exactly right, that there’s sort of the damage in the acute phase and then there’s sort of lingering inflammatory responses. So, a big area of inflammatory communication between the peripheral body and the brain is the brainstem. It’s where, for example, the vagus nerve, which detects inflammation out in the body, comes in and sends a signal to the brain that says, “Hey, your body is sick. Go to bed. You feel sick right now.” And that also triggers an inflammatory response. That cluster where the vagus nerve enters the brainstem is surrounded by autonomic nuclei, including nuclei that are important in breathing. It’s probably for most Long COVID patients not going to manifest in terms of like, “I have to remember to keep breathing, or I’ll forget.” But we think that maybe that’s part of why, for example, they may have lost some control over their heart rate when normally they could go from lying down to standing up.

Normally we all make these very subtle but important adjustments so that we don’t pass out when the blood rushes to our feet. And that sort of thing can be disrupted, because the inflammatory process adds some signal-to-noise disruption.

Dr. Jennifer Frontera:

And I can also add in terms of, are we looking at MRI? Yes. Cycle 1 has a large segment on MRI and much of that post-processing has been completed, and we’re moving into analysis stages of that. There are multiple other NIH-funded studies outside of RECOVER also that are looking at neuroimaging. So, there’s a lot more information to come, I think.

Dr. Jacqueline Becker: 

And Quinn, I noticed that online, there were a few questions actually about language disturbances that I thought I might tackle. People asked why are language disturbances like aphasia, paraphasias, anomias typically seen in Long COVID? So, I just wanted to sort of address that. Language disturbances are most commonly word finding difficulty, but these are typically influenced by things like executive functioning and processing speed. And so, remember how in my presentation, I mentioned that cognitive domains don’t work in isolation, right? This is the perfect example where executive functioning and processing speed deficits can sort of cascade into like slowed what we call semantic access, which is the retrieving words from our sort of semantic bank, so to speak. Phonemic paraphasias is when people are making specific errors when they’re saying words; they can happen, especially under cognitive load or fatigue. People who have verbal fluency impairment we find is often related to more of that initiation of speech, which again has prefrontal executive demands.

And also, there are some people who will say that they have difficulty sort of putting a sentence together and that that is really challenging, and that I sort of attribute to reduced working memory. And so, I think these are meaningful in that it’s different from what we see from like a primary language disorder, but rather it resembles sort of, again, these frontal temporal disruptions and maybe even in white matter pathways, which is different from the typical temporal profile that we see with patients with primary language disorders. So, I just wanted to emphasize that, because I saw 6 different questions about anomic aphasias and things like that. So again, it’s more, I think more granular language assessments are needed, but these are on our radar and certainly included for Cycle 2.

Liza Fisher: 

Can I jump in and talk about how that kind of combines, both the breathing and the language and the autonomic nervous system? In my lived experience that combines with not being able to speak in vocal cord dysfunction. So, they kind of played off of each other functionally in the lived experience.

Quinn Barnette: 

All right, great. Thank you.

Dr. Jennifer Frontera: 

I do want to just add that importantly, and not necessarily related to COVID at all, is that sleep apnea is definitely a risk factor for cognitive impairment and is easily treated. And so, if that is an issue, then that is something that folks should certainly consider and address.

Quinn Barnette: 

All right, thank you. Our next question I think I will direct to Dr. Becker as well. It asks, “Is there a difference in the prevalence of neurocognitive symptoms from previous variants of SARS-CoV-2 versus more recent dominant variants?”

Dr. Jacqueline Becker: 

Yeah. So, when I talk about variants, I typically like using variant eras, and I say that because the way that we really look at variants is by the date of infection. So, we don’t have a way really, a good way anyway, of knowing which specific variant someone was infected by. So, I think what we know is that brain fog specifically may actually be more prominent in the Omicron era Long COVID, even though the overall prevalence of Long COVID may have decreased. So, the pre-Omicron variants, so the original variants, were more strongly associated with things like shortness of breath, with anosmia and olfactory dysfunction, whereas Omicron was actually more associated with things like brain fog and fatigue. But again, these were eras. And so geographically, they varied substantially. So, vaccination may also have reduced overall Long COVID risk relative to the earlier variants. And so, I just want to highlight again that when we think about variants, we’re talking about specific eras, and we can’t say for sure which variant that people were infected with, but generally that seems to be the pattern.

Quinn Barnette: 

All right, thank you. We have time for maybe a couple more questions. This one is for Dr. Frontera, and it asks, “Is there an increase? Did you find an increase in choroid plexus volumes where the ventricles also increase in Long COVID participants?”

Dr. Jennifer Frontera: 

No. Ventricular size was not different in our particular cohort. By the time you get atrophy severe enough to cause ventricular size to increase, then you probably are moving more towards a dementia diagnosis. So no, we saw it independent of ventricular size. And in fact, in this particular cohort, we didn’t find differences in ventricular volume. Our patients are getting serial MRIs, so we’re still analyzing changes over time to try to get a sense if any of the volumetrics are changing over the time course of the study.

Quinn Barnette: 

Okay, thank you. And then I think we’ll have time for one more question, and I will also direct this one to Dr. Becker. It asks, “How is it determined that symptoms are Long COVID versus maybe a condition with overlapping symptoms such as perimenopause or menopause?”

Dr. Jacqueline Becker: 

Yeah, I think I answered this one actually in the chat, but there’s no single test that separates them cleanly. There’s a lot of symptom overlap between Long COVID specifically, and in speaking about menopause specifically. Women aged 45 to 54 tend to be disproportionately affected by Long COVID. And so, it’s right during that critical period that we’re talking about. Often, we do see though a differential cognitive pattern in Long COVID. So, what we do is when we do some of our gold standard neuropsychological tests, we look at the pattern of deficits and the degree of deficits relative to somebody’s baseline. And we know from many, many decades of literature what a typical profile looks like for someone who is in that having difficulty cognitively speaking because of menopause. And so, we’re able to see whether the deficits we’re seeing follow a similar pattern or if they’re in excess.

But with respect to other symptoms, those are always things that we say we need to try to rule out. So, there are a lot of contributing factors too outside of just Long COVID. So, it’s been mentioned here a whole lot that sleep very much impacts cognitive functioning. There are some medications and even supplements that people don’t think about that can impact cognitive functioning. Certainly, depression and anxiety can sort of reduce some of the attentional resources that you may need to be more cognitively efficient. And so, those things can definitely impact cognition as well. In a neuropsych assessment, we usually look at all of those and try to rule out other potential underlying etiologies.

Quinn Barnette: 

All right. Thank you so much. And I think that’s all the time we have for Q&A, but like I mentioned, there will be a Q&A document posted on RECOVERcovid.org for any questions we couldn’t get to during this time. So, thank you so much to our presenters, and thank you again to our audience for attending this seminar and engaging with the Q&A. For interested researchers, we would like to quickly remind you about the current Research Opportunity Announcement and notice for ancillary studies from NIH. These are some of the ways that researchers both within and outside of RECOVER can apply to use participant data similar to the data that was presented today in your own proposed research studies. The link to this opportunity is available in the chat window, and to learn more and apply, you can click on the link provided or go to RECOVERcovid.org/funding.

As a reminder, a recording of the seminar will be available on RECOVERcovid.org in about a week, and then we’ll also post a recap of the seminar, as well as the Q&A document that has responses to questions that we received today.

We’ll post information about upcoming seminars on the R3 Seminar series webpage on RECOVERcovid.org. You can also subscribe for R3 Seminar updates on this page. Additionally, you will see a short survey come up on your screen, which asks for your feedback on this seminar, and we’d appreciate it if you could take just a minute to fill out that brief survey. And with that, thank you all again to our presenters and panelists and to our audience for joining, and have a great rest of your day.

Dr. Jacqueline Becker: 

Bye, everybody.

Mike Zissis: 

Thanks, everyone.