Last month, I wrote a diary detailing the accumulating scientific evidence, in scattered academic papers and reports, showing that neuroinflammation from COVID-19 could lead to deficits in the prefrontal cortex of the brain. If these deficits were to occur, they could lead to what is known as executive dysfunction, a condition where certain sectors of the brain are simply not consulted before activity takes place, or some level of disregulation, disorganization or disorder could occur. As I stated there, the prefrontal cortex can be seen as the living room of cognition. It is the seat of self.
This month, Nature Magazine has accelerated the publication of research from England showing the best evidence yet that this is actually happening. Brain tissue is being directly damaged by COVID-19, with measurable, significant differences in brain volume and well as in cerebrospinal fluid, the liquid in which the brain floats and bathes. There is a reduction in the former, an increase of the latter.
We know this—that Covid causes such damage—more definitively than ever, because the study uses subjects who submitted to brain scans before contracting Covid as well as after. Also, the Covid cohort of the study were well-matched to their healthy controls. The study is longitudinal, following these subjects over time. “To our knowledge,” the authors of the study write, “this is the first longitudinal imaging study of SARS-CoV-2 where participants were initially scanned before they had been infected.”
The Guardian reports:
In the brain study, researchers at the
University of Oxford studied 785 people aged between 51 and 81 who had received brain scans before and during the pandemic as part of the UK Biobank study. More than half of them tested positive for Covid between the two scans.
Compared with 384 uninfected control subjects, those who tested positive for Covid had greater overall brain shrinkage and more grey matter shrinkage, particularly in areas linked to smell. For example, those who had Covid lost an additional 1.8% of the parahippocampal gyrus, a key region for smell, and an additional 0.8% of the cerebellum, compared with control subjects.
In fact, the study authors detail the top ten areas affected by the virus, according to the imaging. They include “the anterior cingulate cortex, orbitofrontal cortex and insula, as well as the ventral striatum, amygdala, hippocampus and parahippocampus gyrus[…], the frontal and temporal piriform cortex, as well as the olfactory tubercle and anterior olfactory nucleus” (p. 2). The anterior cingulate gyrus is often implicated in addiction studies, as it is recognized as part of the reward circuit. The hippocampus is crucial for episodic memory and for moving memories into long-term storage. The amygdala is intimately involved with hippocampal function and is, among other tasks, heavily involved in the regulation of fear.
The orbitofrontal cortex (OFC) is situated directly above the sockets (the orbits) of the eyes. It can be seen as the very outer ridge or ledge of the prefrontal cortex.
It is right next door to the ventromedial prefrontal cortex (vmPFC), the region I detailed in my earlier diary. Both regions “are thought to be involved in assessing information about outcome values during decision processes,” according to previous work published in 2010. The OFC appears to be more involved with evaluating external aspects of reward. Incidentally, the ventral striatum connects to the vmPFC and enables it to communicate with other brain regions.
The insula, a hidden structure in the brain, is one of the oldest and helps regulate aspects of olfactory (smell) and gustatory (taste) sensations. It also houses our “ideas” (really, “gut” reactions) of disgust, notably. This has wide implications. In 2017, a team of researchers published a paper linking the insula to political behavior, specifically detailing how the insula may relate to attitudes about immigrants. Those researchers term the insula “the behavioral immune system.”
In the current research, conducted in relation to the UK Biobank, the authors tell us that “average percentage change differences between the two groups [the Covid-positive group and the healthy controls] was moderate, ranging from -0.2 to -2% (except for two diffusion measures in the fimbria at >6%, due to the very small size of these regions-of-interest), with the largest differences seen in the volume of the parahippocampal gyrus and caudal anterior cingulate cortex” (p. 3).
I quote the study in technical detail as the popular press has a tendency to reduce the information and couch it in acceptable terms. For instance, the New York Times, which also ran this story, found two neuroscientists not involved with the study. Ostensibly, this was done to place the study in context; however, those outside experts continually downplayed the information. One even made it seem as though the implications were so unusual that the results may not be intelligible. “None of them got thorough enough cognitive testing to know if they had significant deficits in these many regions where they found these changes in volume… We don’t know that it actually means anything for the patient’s quality of life or function.”
The other outside expert plainly stated, “We don’t want to scare the public and have them think, ‘Oh, this is proof that everyone’s going to have brain damage and not be able to function.’” This is of course important, that the public understands that the results are not necessarily able to be extrapolated to the individual level. But what stands out is that she is more concerned about tamping down a fear response than in fully communicating the results of the paper.
The research team of the UK Biobank study, led by Dr. Gwenaëlle Douaud, found that “significantly greater cognitive decline, which persisted even after excluding the [15] hospitalized patients, was seen in the SARS-CoV-2 positive group between the two timepoints, and this decline was associated with greater atrophy of crus II, a cognitve lobule of the cerebellum” (p. 4). In response to this, one of the skeptical experts stated “[It] is not the first brain structure you think of” to explain the discrepancies demonstrated in a well-known cognitive test called the Trail-making test. The comment sounds almost off-hand and has the effect of being dismissive. In fact, this may be one of the more important findings.
The trail-making test is a simple measure of being able to follow a sequence. The first part of the test has the subject follow various bubbled numbers with a pencil or stylus to draw a path between them; this runs from one to twenty. The second test is similar to the first, but instead the subject must intersperse the numbers with letters (an alphanumeric test); so in this case, they would need to go from 1 to A, then 2 to B, and so forth. This is a standard test and has reliably been used to measure the slowing of cognitive function.
The Covid-infected cohort had significantly slower response times for task completion. When viewed in conjunction with the brain imaging scans and other data analysis of the study, it is clear that this test is but one dimension of demonstration of diminishment of cognitive function. But the expert at NYT does not acknowledge this in so many words but instead appears to want to find many caveats.
In fact, Douaud et al., clearly state that “the exploratory analysis revealed a more pronounced loss of grey matter in crus II, part of the cognitive, and olfactory-related lobule VII of the cerebellum.” In a separate study, published in 2018, it was shown that “Crus I was functionally connected with anterior PFC, Crus II with middle PFC” and that “the suggestion that the cerebellum may play a role in non-motor behavior is not a new idea.”
That the cerebellum may be implicated deserves further study. This region is notoriously difficult to study due to the structure of the neuronal matrix and the way the neurons reach into other regions of the brain. They simply are difficult to distinguish. Additionally, that portion of the brain is not often included in standard imaging. Like when researchers study the brainstem, they must plan carefully in advance which structures will be highlighted; otherwise, the methods and materials used will not measure or highlight detail. (Certain staining techniques, for example, must be used when examining brainstem components.)
The 2018 study states in its conclusion, “While the exact role of the cerebellum in non-motor behavior remains unknown, it has been suggested that the structure acts to process internal models of thought …. With damage and disease, one would therefore experience deficits in thought and processing, such as those seen in patients with schizophrenia or those with cerebellar infarct.” Would that the NYT find time to interview those authors.
Douaud et al. inform us as to possible future implications:
The overlapping olfactory- and memory-related functions of the regions shown to alter significantly over time in SARS-CoV-2, including the parahippocampal gyrus/perirhinal cortex, entorhinal cortex and hippocampus in particular, raise the possibility that longer-term consequences of SARS-CoV-2 infection might in time contribute to Alzheimer’s disease or other forms of dementia. [...] In our sample of infected participants with mainly mild symptoms, we found no signs of memory impairment. However, these SARS-CoV-2 positive participants showed a worsening of executive function, taking a significantly greater time to complete trail A and particularly trail B of the Trail Making Test. These findings remained significant after excluding the few hospitalised cases. (p. 6, emphasis added)
Also,
[T]he two trails [are] two measures known to be sensitive to detect impairment of executive function and attention, for instance in affective disorders and in schizophrenia, and to discriminate mild cognitive impairment and dementia from healthy ageing. (p. 6)
The research team acknowledges that the main drawback to their study is that it was not randomized. The participants had been scanned prior to the pandemic and were later matched to others in the biobank after their Covid-positive status became known (either through public records or by two separate positive Covid antigen tests). The authors recommend further longitudinal studies structured specifically to address this drawback in design.
That being said, they also note that “while the two groups were not significantly different across major demographic and risk factor variables, we identified a subtle pattern of lower cognitive abilities in the participants who went on to be infected, but this could not explain away our longitudinal findings. The future positive cases also showed lower subcortical volume; and higher diffusion abnormalities at baseline compared with the controls in brain regions not overlapping with our longitudinal results.” In other words, the Covid-infected patients were well-matched and that the existing differences in brain architecture in individuals did not answer the findings of the study.
One of the major critiques of my diary from February had to do with the fact that the studies I cited were not longitudinal. Also, the lack of imaging was a drawback. We now have these in one elegant study. While the study could be even more robust with alterations in structural design, the results are impressive and go far to demonstrating exactly what type of damage Covid can do in even mild or asymptomatic cases.