How many glial cells are there in the brain?

One of the fundamental principles of scientific thinking is skepticism. A good scientist refuses to accept anything blindly, instead scrutinizing every purported statement of fact to make sure the evidence backs it up.

What are glial cells? Watch this 2-minute video to find out.

Because this mindset is so pervasive in scientific disciplines, it’s difficult to understand how unsubstantiated claims can be accepted as fact in science. But this does happen occasionally. Some unproven assertions have even found their way into the vaunted territory of common knowledge— a designation that means something is so well-established as true that you don’t even need a source to back it up anymore. For the last several decades, this has been the status of claims that there are many more glial cells than neurons in the brain.

Neuroscientists have been interested in finding accurate estimates of the number of neurons and glia in the brain for at least a century and a half. While figuring these numbers out is a prodigious feat no matter how you cut it, determining glial cell counts has been particularly challenging due to the small size of glia and the difficulty in telling them apart from other small cells. Still, cell counting methods have improved drastically over the the last several decades, and there’s reason to believe that we finally have some valid estimates of both neurons and glial cells.

Methods for counting brain cells

Despite technical limitations like poor microscope resolution and undeveloped approaches to staining cells, early neuroscientists sometimes still managed to arrive at credible counts of neurons in the brain. Helen Bradford Thompson, for example, published an estimate in 1899 of the number of neurons in the cerebral cortex (about 9 billion) that matches up well with current estimates of about 10-20 billion.

Early neuroscientists like Helen Bradford Thompson arrived at neuronal numbers by actually counting neurons. In fact, this approach is still used today, just in a more refined manner. But the overall idea is the same: count the number of cells in various samples of brain tissue and extrapolate the numbers obtained to a larger brain region, or the whole brain.

A more recently developed method of cell counting uses some additional steps to make the process a bit easier and more precise. It involves taking a sample of brain tissue and homogenizing it—destroying the cell membranes, leaving the nuclei intact, and creating a soup-like mixture of liquefied brain. The nuclei can be stained with a fluorescent dye, antibodies can be used to differentiate between neural and non-neuronal cells, and then the nuclei can be counted.

This process is called isotropic fractionation. Isotropy is uniformity, and refers to the mixture formed after homogenization of the brain tissue. And fractionation indicates that cells are counted in a fraction of the whole tissue, and then the results are used to infer numbers in the rest of the brain region.

Gial cell estimates

Isotropic fractionation is a relatively new method. Before it was developed, finding accurate cell numbers in the brain was more painstaking and susceptible to errors. And, as mentioned above, glial cells were especially problematic.

This difficulty in counting glia was represented in some of the uncertainty researchers expressed about the number of glial cells in the brain before the 1980s. Although it was widely believed that the tiny glia outnumbered neurons, there was not a lot of hard evidence to prove this was the case. So it wasn’t uncommon to find scientists using qualifiers like "perhaps" when they made statements about glial cell counts. A common estimate at this time was that there were "perhaps" ten times as many glial cells as neurons.

But there were some who made more definitive statements. Holger Hyden, for example—a reputable biochemist and neuroscientist—stated more decisively in the 1960s that there are ten times as many glial cells as neurons. It’s probably the case that Hyden based his proclamations on specific regions of the brainstem he was studying where glia really do outnumber neurons significantly. But the extrapolation to the entire brain was nevertheless speculative, even though it was stated conclusively.

As can happen in scientific writing, researchers found Hyden’s declarations and others like them and cited them when writing journal articles or textbooks. Over time, this happened enough that the statements, which should never have been definitive, became common knowledge.

By the 1980s, even the most reputable sources in neuroscience were asserting that there are at least ten times as many glia as neurons in the brain. For example, in the 1985 edition of the famous neuroscience textbook, Principles of Neural Science (sometimes called the “bible of neuroscience”), it’s stated that glia outnumber neurons anywhere from 10 to 50 times. Because the text also estimated the number of neurons in the brain to be around 1 trillion (now considered a huge overestimate), the number of glia are implied to be somewhere between 10 and 50 trillion.

But again these numbers were all speculative, and they didn’t match up with the data researchers who were actually counting cells had obtained. For example, a respectable appraisal of the number of neurons and glia in the brain was published in 1986, and it suggested there are about 70-80 billion neurons and 40-50 billion glial cells. The largest number of glial cells reported in a primary research report was 130 billion in 1968.

These seemingly more accurate estimates, however, were largely ignored. It wasn’t until the late 2000s, when researchers began publishing data using isotropic fractionation, that the field took note of the discrepancies.

Debunking the myth

The groundbreaking paper in this respect was published by the Brazilian neuroscientist Suzana Herculano-Houzel and her colleagues in 2009. They used the isotropic fractionation method to count neurons and glia in the brain, and ended up with estimates of 86 billion neurons and 85 billion non-neuronal cells (which included glia and other cells, like endothelial cells). This suggested that there were actually fewer glial cells than neurons, which agreed with some of the data obtained earlier.

There was a bit of resistance to accepting these numbers at first, as some argued that isotropic fractionation had not yet been validated by comparing its results with those obtained through more well-known cell-counting methods. These validation studies came with time, however, and subsequent studies supported the numbers of Herculano-Houzel’s group. Today, most researchers have accepted the data obtained with isotropic fractionation, and the preponderance of the evidence supports the idea that the ratio of glia to neurons is about 1:1.

Of course, this doesn’t diminish the importance of glia. Historically, they have not been given due credit for the integral roles they play in the brain. That seems to be changing in recent years, however, as we learn more about the functions of glia. And as we get a more accurate view of all that glia do, we also seem to be letting go of inaccurate estimates of their numbers.

Perhaps more than anything, the history of glial cell counting teaches us that we should be skeptical of any sources that make claims that aren’t directly supported by primary research. Just because an authoritative source says something definitively, it doesn’t necessarily mean that it’s true or even that the research backs it up. It’s important, especially in this age of extreme information availability, that we be highly critical of the information we consume.

Reference (in addition to linked text above):

von Bartheld CS, Bahney J, Herculano-Houzel S. The search for true numbers of neurons and glial cells in the human brain: A review of 150 years of cell counting. J Comp Neurol. 2016 Dec 15;524(18):3865-3895. doi: 10.1002/cne.24040. Epub 2016 Jun 16.

Know Your Brain: Fatal Insomnia

In 1991, Michel A. Corke was enjoying the summer break from his position as a music teacher in a Chicago high school when he started to develop sleeping problems. He had recently turned 40, and seemed to be in good health, but it was soon very obvious that he was suffering from more than just your run-of-the-mill insomnia. It wasn't just taking him longer to fall asleep than normal. Nor was he suffering from the common problem of waking up frequently over the course of the night. He wasn't sleeping at all.

Within a few months, Corke's lack of sleep was causing obvious physical and mental deterioration. He developed problems with balance and had trouble walking. He began to display signs of dementia and there were times when he appeared to lose touch with reality. Sometimes these episodes involved hallucinations.

After Christmas, he was admitted to the hospital. At that point, Corke was unable to communicate. He had become completely dependent on his family to help him perform even the simplest tasks like showering and getting dressed. His decline had been rapid and pervasive.

At first, doctors could not figure out was wrong with Corke. They diagnosed him with multiple sclerosis, despite the fact that he didn't really have the symptom profile of someone with the disease.

But doctors soon recognized that something was very unusual about Corke's "sleep." Even though Corke would often close his eyes and appear to be sleeping, measurements of his brain activity found that his brain never did actually fall asleep. This helped doctors to realize that he was suffering from a disease that had only been recognized within the previous decade, called fatal familial insomnia (now often called fatal insomnia because not all cases seem to be hereditary).

In the disease, sleep becomes progressively disrupted until patients exhibit little to no sleep. Eventually, death is inevitable.

Michel Corke's case was no exception. By the time he had been admitted to the hospital, about 130 days had passed with minimal sleep. When he died, he had essentially been awake for 6 months.

What are the symptoms of fatal insomnia?

Fatal insomnia is a rare disease that usually develops in middle age or later (the average age of onset is 51 years), and begins with complaints of trouble sleeping or excessive fatigue during the day. At first, due to this extreme daytime sleepiness, the assumption might be that the patient is plagued by a condition that is making him too sleepy.

Sometimes there are other abnormal signs early on, like double vision, impotence, hypertension, and increased perspiration, lacrimation (i.e. tear production), and/or salivation.

As the disease progresses, patients lose their ability to sleep altogether. A variety of movement problems appear, including difficulties with balance and coordination and abnormalities in gait. Patients also will sometimes become delusional and display unusual behavior that resembles dream-enactment, which involves unconsciously making movements related to what's going on in a dream.

Later in the disease, after a patient has been deprived of sleep for some time, he begins to spend more and more time in a stupor from which it's difficult to rouse him. He may experience sudden, spasmodic movements, but voluntary movement like standing and walking often become difficult to impossible. He may lose the ability to speak, have trouble swallowing, and fall into a vegetative state.

Death can occur at any time throughout these phases of the disease, but if the patient survives long enough often he will fall into a coma, which will lead to death. The duration of the disease ranges from 8 to 72 months, with an average disease course lasting just over 18 months.

What happens in the brain in fatal insomnia?

Brain activity during "sleep"

One way to verify the sleep disturbances occurring in fatal insomnia is to measure sleep activity over the course of a night using a technique known as polysomnography. Polysomnography measures the electrical activity in the brain (using an electroencephalogram, or EEG) along with a number of other physiological changes that occur during sleep like eye movement, muscle activity, and the electrical activity of the heart.

Polysomnography is often used to verify a case of fatal insomnia because patients may appear to spend periods of the night sleeping, as they have their eyes closed and aren't moving. Polysomnography reveals, however, that their brain activity doesn't resemble a pattern of normal sleep.

In a healthy person, during sleep the brain cycles from relatively light sleep into a period of deep sleep into a period a rapid eye movement (REM) sleep. A full cycle takes about 90-120 minutes and is repeated 4-6 times per night. These different stages of sleep have characteristic electrical activity that can be measured with an EEG.

Patients with fatal insomnia will, of course, display drastically reduced total sleep time. But even at times when it appears as if they are asleep, the EEG still won't show these characteristics of healthy sleep. Instead, their brain activity generally indicates wakefulness for most of the night, with very brief periods of light sleep (i.e. stage 1 or stage 2 sleep) and occasional sudden episodes of REM sleep that only last seconds or minutes. Deep sleep mostly disappears, and as the disease progresses all traces of REM sleep may disappear as well.

The thalami are the orange, oval-shaped structures in the image. They are the site of the most significant neurodegeneration in fatal insomnia.


Fatal insomnia is associated with severe neurodegeneration of the thalamus, which is thought to be a critically important structure in sleep regulation. The thalamus is believed to play an especially important role in the generation of a type of deep sleep known as slow-wave sleep. Over the course of the disease, up to 80% of neurons are lost in certain nuclei of the thalamus.

The inferior olivary nucleus, a structure in the brainstem that is densely interconnected with the cerebellum, also suffers significant neurodegeneration in the disease, losing more than 50% of its neurons. The role of the inferior olivary nuclei in the symptoms of fatal insomnia is still unclear. It may be involved in generating movement-related symptoms like tremor and spasmodic muscle contractions, but some evidence suggests the olivary nuclei and cerebellum are involved with sleep as well.

The disease is also sometimes associated with the formation of large abnormal compartments, or vacuoles, within neurons. These "holes" in the brain can give the brain a sponge-like appearance. In fatal insomnia, this occurs primarily in the cerebral cortex.

Accumulation of prion protein

Fatal insomnia is considered a prion disease, and thus also involves the accumulation of abnormally-folded forms of prion protein in the brain. These misfolded proteins have a tendency to accumulate into clusters that are resistant to being broken down by brain enzymes. The implications of these protein clusters forming in the brain is unclear, although they are often linked to pathological changes in the brain. (For a short primer on prion diseases, read this article.)

Prion proteins also are capable of passing their misfolded state on to other healthy proteins. Thus, they can spread within the brain of an infected patient, gradually increasing the number of misfolded prion proteins. Interestingly, their "infectious" quality also allows prions to cause disease if transmitted from one host to another. While it isn't thought that fatal insomnia is spread among people in this way, the disease has been transmitted to mice by injecting them with a liquefied piece of brain tissue from a human patient who had the disease.

In fatal insomnia, however, there are relatively few clusters of prion protein in the brain as compared to other prion diseases. And, while deposits in some areas of the brain increase in number as the disease progresses, this isn't true for the areas that experience the most neurodegeneration---like the thalamus. Thus, it's still unclear what exactly causes the neurodegeneration that produces the symptoms of fatal insomnia.

What causes fatal insomnia?

Most of the cases of fatal insomnia identified to date are considered genetic diseases, attributable to a genetic mutation in the PRNP, or PRionN Protein, gene---a gene that's implicated in other prion diseases as well. The mutation is an autosomal dominant mutation, which means that if someone with the mutation has a child, the child has a 50% chance of ending up with the same mutation. When fatal insomnia is inherited, it is generally referred to as fatal familial insomnia.

To date, just over 200 individuals worldwide are known to carry the mutation associated with fatal familial insomnia. Due to the global distribution of the disease, some researchers have suggested it is caused by a recurrent mutation that has happened independently in a number of families.

In 1999, the first cases of what seems to be non-hereditary fatal insomnia appeared. Non-hereditary fatal insomnia is commonly referred to as sporadic fatal insomnia, and to date 32 cases have been identified. These patients display most of the same symptoms and pathology as fatal familial insomnia patients, but they have no family history of the disease and do not have the mutation of the PRNP gene seen in fatal familial insomnia patients.

What is the treatment for fatal insomnia?

We are severely limited in our ability to treat fatal insomnia patients. Even the strongest sedatives (e.g. barbiturates, benzodiazepines) do not cause patients of the disorder to sleep. Thus, treatment focuses on relieving the symptoms of the disorder as much as possible (which alone is a challenge).

References (in addition to linked text above):

Cracco L, Appleby BS, Gambetti P. Fatal familial insomnia and sporadic fatal insomnia. Handb Clin Neurol. 2018;153:271-299. doi: 10.1016/B978-0-444-63945-5.00015-5.

Montagna P. Fatal familial insomnia and the role of the thalamus in sleep regulation. Handb Clin Neurol. 2011;99:981-96. doi: 10.1016/B978-0-444-52007-4.00018-7.

Montagna P, Gambetti P, Cortelli P, Lugaresi E. Familial and sporadic fatal insomnia. Lancet Neurol. 2003 Mar;2(3):167-76. (in