When epidemics and pandemics washed over humanity through the ages, watchful doctors noticed that in addition to the usual, mostly respiratory ailments, the illnesses also seemed to trigger neurological symptoms. One British throat specialist observed in the late 1800s that influenza appeared to “run up and down the nervous keyboard stirring up disorder and pain in different parts of the body with what almost seems malicious caprice.” Indeed, some patients during the 1889–92 influenza pandemic reportedly became afflicted with psychoses, paranoia, stabbing pains, and nerve damage. Similarly, scholars have linked the 1918 flu pandemic to parkinsonism, neuropsychiatric disorders, and a broadly coinciding outbreak of the “sleeping sickness” encephalitis lethargica, which would often arrest patients in a coma-like state—although researchers still debate whether the two are causally connected.

That SARS-CoV-2, the culprit of the COVID-19 pandemic, is also associated with neurological symptoms isn’t entirely surprising, given some evidence that its close relatives, MERS-CoV and SARS-CoV-1, have been associated with neurological symptoms too. But the proportion of patients developing such symptoms—and their mounting collective numbers—has startled some scientists. When the news broke early last year that some 36% of COVID-19 patients in Wuhan hospitals were developing impaired consciousness, seizures, sensory impairments, and other neurological symptoms, “that floored me,” remarks Shibani Mukerji, a neuroinfectious disease specialist at Massachusetts General Hospital.

Now, entire clinics are being created specializing in subsets of patients with neurological symptoms, including one in London co-led by Patricia McNamara, a neurologist at the National Hospital for Neurology and Neurosurgery. She broadly separates neurological symptoms into two diverse groups. The first—described in several reports last year—comprises acute symptoms often afflicting hospitalized patients with severe disease. These can manifest as a confused, delirious state known as encephalopathy, or as strokes, peripheral nerve damage, or encephalitis, an inflammation of the brain. The second group represents long-term symptoms usually following milder infections, ranging from headaches, fatigue, sensations of numbness or tingling, and cognitive difficulties to occasionally seizures and inflammation of the heart, McNamara says. “The people I’ve seen so far [in this group] are people who slowly improve, but it’s certainly a very slow improvement.”

As this cohort grows, so too does the global effort to investigate how SARS-CoV-2 causes such symptoms. The picture so far remains somewhat puzzling. Autopsy studies—which have been limited—have found clear signs of damage in dozens of brains of COVID-19 patients. Although traces of the virus have been reported in some brain specimens, in many cases it is nowhere to be found. While the question of whether SARS-CoV-2 directly infects the brain remains unresolved, researchers are exploring other mechanisms whereby it could affect the human brain.

“I think all of us probably . . . would agree that there is no overwhelming infection of the brain,” notes Avindra Nath, a neurovirologist at the National Institute of Neurological Disorders and Stroke. “If there is, it’s a very, very miniscule amount. That cannot explain the pathology that we see. It has to be something more than that.”

Does SARS-CoV-2 infect the human brain?

Early in the pandemic, some researchers worried that SARS-CoV-2 could be gaining access to the brain and represent a “neurotropic” virus, McNamara says. That would offer an obvious hypothesis for some of the neurological symptoms observed, while also posing complex questions around how to therapeutically target pathological processes in the brain. And after all, the genetic material of SARS-CoV-1 and MERS-CoV have been spotted inside human brains, and even common cold-causing coronavirus proteins are—curiously, given that they rarely lead to neurological symptoms—common in autopsied brains, adds Helena Radbruch, a neuropathologist at the Charité hospital in Berlin.

Yet the first brain autopsies of COVID-19 patients didn’t find much SARS-CoV-2 RNA, let alone viral protein. In a September study of 18 COVID-19 patients with neurological symptoms who died in hospitals last April, Mukerji and her colleagues found very low levels of RNA—the source of which is a mystery—in only five of the patient brains, and no signs of viral protein. Because the low RNA concentration “seems out of proportion to the profound deficits that people are experiencing,” she says, “I’d be extremely surprised [if] the majority of cases where people are having neurological symptoms are due to direct viral invasion.”

In a more recent postmortem analysis of 18 patients published in the New England Journal of Medicine, Nath and his colleagues couldn’t find any evidence of the virus in the brain. However, an earlier study from researchers in Germany reported viral protein in the cranial nerves and isolated brainstem cells in 21 of 40 patients examined postmortem. Viral protein that has somehow reached the central nervous system typically elicits an immune response, Mukerji notes, but its presence in the German study didn’t correlate with the severity of neural inflammation the team observed.

Whatever viral protein is doing inside the brain, Radbruch and her colleagues assert that it is indeed finding its way there. In an autopsy study published in November in Nature Neuroscience, they propose that the virus could get into the brainstem through the nose. Based on detailed autopsy analyses from 33 COVID-19 patients, they discovered intact coronavirus particles in supporting cells of the olfactory mucosa at the roof of the nose, along with evidence of active replication in the tissue. Perhaps viral replication destroys those cells, and/or induces inflammation, which could help explain the frequent loss of taste and smell at the start of SARS-CoV-2 infections, notes Charité neuropathologist Frank Heppner, a coauthor of the study. The virus could then work its way into the olfactory bulb, a hub for processing sensory information, and via specific cranial nerves into the brain. Indeed, they observed evidence of viral RNA in these tissues as well as viral RNA and protein inside cells in the medulla oblongata in the brainstem, and in other structures such as the cerebellum, hinting that the virus could be using multiple routes of entry into the brain.

Nevertheless, Radbruch and Heppner agree that the extent of SARS-CoV-2’s infection pales in comparison to that of fellow RNA viruses such as rabies, which is devoted to infecting brain tissue. SARS-CoV-2 is more of an “incidentally neurotropic” virus, likely getting into the brain by accident, Heppner says. Notably, in the brainstem they only found viral protein inside endothelial cells that make up the lining of blood vessels—the blood-brain barrier—and not inside neurons. That could be due to difficulties detecting the virus inside neurons, or could indicate that it doesn’t infect neurons. In contrast to endothelial cells, which have an abundance of ACE2, the molecular doorways SARS-CoV-2 uses to enter cells, neurons tend not to have ACE2 receptors.

Still, the jury’s still out on whether SARS-CoV-2 infects the brain, notes Eric Song, an immunobiologist at Yale University who recently completed his PhD in immunologist Akiko Iwasaki’s lab there. He points to their recently published study on brain organoid models, which suggests that it’s possible for SARS-CoV-2 to infect and exert pathological effects in neural tissue. Plus, postmortem samples are limited in number, only show the final picture of a disease, and, by necessity, reflect what happens in just the sickest patients. By the time COVID-19 patients die, most virus in the lung has often already been cleared, and that might be the case for the brain too, Song says.

In a recent preprint, he and his colleagues couldn’t find viral RNA in the brain-engulfing cerebrospinal fluid (CSF) of six living COVID-19 patients with neurological symptoms. But interestingly, they did find B cells and antibodies in the CSF, and not just ones that target SARS-CoV-2, but also ones that target the body’s own proteins, including components of neurons. What this finding means is unclear, but it adds to a string of findings of body-attacking immune machinery in COVID-19. The looming question now, Song says, “is whether this is an enriched process in [COVID-19] or if it’s a process that occurs with the same magnitude in other viral diseases.”

Indirect effects on the brain

While the evidence of SARS-CoV-2 inside human brains remains murky, autopsy studies have revealed a range of recurrent neuropathological features in hospitalized COVID-19 patients. Pathologists have frequently found localized hypoxic damage caused by a lack of oxygen and associated infarcts (or ischemic strokes), and less commonly, signs of bleeding in the brain, as well as some inflammation—in some cases a severe inflammation known as acute disseminated encephalomyelitis. What exactly causes these pathologies is unclear, but scientists have some suspicions about contributing factors.

For instance, blood clots—which COVID-19 patients are prone to develop—could create blockages that restrict blood supply to neural tissue, explaining the observed “mini strokes,” Song says. The difficulty in getting enough oxygen through damaged lungs probably also makes patients more vulnerable to hypoxic brain damage. Systemic hypoxia, in turn, could result in neurological symptoms, but in general, it’s hard to say how pathologies discovered in autopsies are connected to clinical symptoms observed in patients, Mukerji adds in an email, as neurological exams aren’t often conducted when patients are hospitalized.

The immune response to the virus could also explain some neurological complications. Perhaps the exuberant flush of proinflammatory cytokines in the body’s periphery could “cause inflammatory cytokines within the brain to be more active and cause inflammation [there],” McNamara suggests. Interestingly, the Charité neuropathologists’ recent paper also found signs of inflammation and hordes of highly activated microglial cells—brain-resident immunological defenders—that could feasibly be the result of a local or peripheral immune overactivation, Heppner hypothesizes. In turn, this could perturb neuronal function, and help explain certain COVID-19 encephalopathy symptoms, such as agitation, confusion, excessive sleepiness, and comas. Some patients, for instance, “have problems weaning from ventilation—they do not wake up, and so far, we do not really understand why,” Radbruch says.

Curiously, some patients also exhibit signs of microscopic damage to the small blood vessels in their brains, as evidenced in Nath and his colleagues’ recent brain autopsy study. The SARS-CoV-2–positive subjects—who were obtained from New York City’s chief medical examiner’s office—represent a unique patient cohort, most of them having died suddenly. Some had been found dead in nursing homes or in the subway, Nath says. In 10 of 13 patient brains examined under a high-resolution MRI scanner, the team noticed patches that appeared unusually bright or dark—the latter likely signifying signs of bleeding. Scrutinizing those areas, which sometimes included the olfactory bulb, the team found evidence of damaged blood vessel walls. And wherever blood vessels were damaged, protein staining revealed leakage of fibrinogen—a blood coagulation protein whose presence in the brain is associated with various neurological disorders—into neural tissue. Clustering around those sites they found macrophages and sometimes also activated microglia and T cells.

A plausible hypothesis is that some of those immune cells attack the endothelial cells that line the blood vessels, a cell type that SARS-CoV-2 is known to infect. Nath’s study didn’t find any evidence of that, but “at least that’s one way to explain it,” he says. “And maybe that’s what’s causing the leakage. Then once you get the fibrinogen in the brain, you can incite more inflammation there . . . so it keeps building up a spiral.”

He and his team also discovered signs that shine light on the patients’ sudden deaths. Assuming the patients died of cardiac dysfunction, they examined the nuclei in the base of the brainstem involved in the control of breathing. There, they spotted macrophages clustering around those neurons, a possible sign of neuronophagia, in which phagocytes devour unhealthy neurons. “These respiratory centers and some of these other brainstem nuclei are impaired in their function,” Nath says. It’s hard to say how generalizable these findings are among COVID-19 patients, but “I think it’s quite possible that some of these long-term symptoms that people have may be related to some of these things.” Along with other researchers, he’s working on a series of studies in “long hauler” patients to find out.

It’s unlikely that one mechanism of neurological effects will fit all patients. Finding out what causative processes are unfolding in individuals will require integrating different techniques and many more brain autopsies, of which there have been only around a few hundred reported so far, largely due to a shortage of specialized equipment and labs to conduct them, Nath says. It will also require some sorting of what is currently a hodgepodge of neurological symptoms—for example, by establishing diagnostic criteria of post-COVID-19 syndrome and defining what acute versus chronic symptoms are, Mukerji adds.

She says she hopes that these studies will also help those with neurological conditions caused by other infections. “There are a variety of people who’ve had [viral] infections [such as] Ebola or West Nile, who have said that they have cognitive complaints, brain fog, and that they’re disabled—and it’s largely fallen on some deaf ears,” she says. “I would be surprised if the world does not take up this discussion much more scientifically than it has done in the past [and] try to understand if we can develop . . . some sort of diagnostic, and then some sort of therapeutic agent to help what is now going to be a large percentage of the world’s population.”