Why you can’t learn about the human brain from monkeys, rats and mice
Much of neuroscience research spending and published output relates to studies done in rats, mice and monkeys. The results of these are generally presented as universally representative of brain function across species. To be sure all mammalian brains do bear some structural and functional similarities, and these systems do provide insights into the nature of cellular processes. Furthermore, at least morphologically, the neurons look similar. All this can be tremendously useful in experimenting with techniques and understanding general concepts. But extrapolating specific findings from these creatures to humans can be misleading.
It is obvious to everyone that the human brain produces more unique capabilities than other species. The question is at what level these differences manifest and where we must draw the boundaries of extrapolation.
Cellular level approaches to understanding brain function
There are three major areas of study of brain function at the cellular level – gene expression, which dictates the molecular toolkit and components that the cell builds and uses to form connections and modulate its activity, synaptic plasticity, which looks at how neurons wire up and alter the nature of their connections in response to stimuli, and spiking activity– the kinds of temporal patterns it produces in response to a stimulus or in association with behaviors.
Studying synaptic plasticity and the patterns of spiking of neurons are tremendously invasive and require direct access to tissue. Most experiments are done by extracting tissue from the brains of mice and rats and growing it in a dish. These are called in vitro experiments, i.e. done in a dish, and require sacrifice of the animal. Another approach involves opening up or drilling holes in the skulls of mice, rats or monkeys to place electrodes or deliver genetic material that produce molecules (for example, ones that fluoresce) that enable cell imaging. In this case you can study the activity of cells and tissue in vivo or while the animal is awake and behaving. However, neither synaptic activity nor neuronal spiking can be easily studied in humans unless they are undergoing brain surgery for some other purpose and consent to some experimental probing that naturally comes with risks. Gene expression, on the other hand, is more easily studied across species using post mortem tissue samples and is closely linked to synaptic plasticity and spike firing.
Gene Expression in the human brain relative to other species
The human brain expresses a much larger repertoire of genes than any other species. One study shows as many as 2014 genes that are differentially expressed in the human brain relative to chimpanzee brains (Khaitovich et al). Furthermore these genes are spread out across all known cellular processes from neuronal differentiation and development to synaptic transmission and signal transduction and even metabolic processes. This is a strong indication that the cells have some pretty global differences in how they function. Also of note are results from the Allen Institute on differences in gene expression between mouse and humans. They find that patterns of gene expression associated with various diseases in humans such as Alzheimer’s are not observed in the mouse suggesting that this is one reason why drugs that work for mouse brains often don’t work in humans.
Interestingly, Khaitovich et al (Svante Paabo’s group) also found quite some diversity among humans stating that differences between brain regions within any individual were far less than the differences between individuals. Another more recent study (read more about it here) shows that individual neurons actually edit their genes and can therefore have widely different expression patterns from one cell to the other. Finally, behavior or environmental context can change gene expression patterns (read more here). These many sources of cellular diversity and the generally greater diversity of experiences among people relative to other species adds another layer of complexity.
All this means that the brain is fundamentally different and wildly diverse in humans relative to other species. What works for mice and monkeys are not likely to work for humans and what works for one person may not even work for another.
The story of glia
The big story that might be emerging, however, is that the biggest differences are not in neurons, which are the cells involved in fast electrical signaling, but in glia. Glial cells outnumber neurons in the human brain by almost 1.4x but were long thought to play just a housekeeping role. A study from the Nedergaard lab however has shown that human astrocytes (a type of glial cell) are 2.6-fold larger in diameter, extend 10-fold more processes in humans and come in way more complex sub types than in mice. They also propagate Ca2+ waves 4x as fast as the ones in mice. Studies from Ben Barres’ lab show that glia play a more significant and active role in modulating the behavior of neurons than previously thought. Most interesting of these findings, however, is that human astrocytes respond far more robustly to glutamate relative to mice (Zhang et al) . Glutamate is the most ubiquitous neuronal signaling molecule in the brain which suggests that glia may have a bigger role to play in shaping the electrical dynamics of the human brain.
Given these differences between human and mouse glia, it is perhaps not surprising that where gene expression differs most between mouse and humans is in glia and not neurons. A study by Hawrylycz et al from the Allen Institute shows bigger differences between non neuronal gene expression, Zhang et al (Barres lab) have shown large differences in the gene expression of astrocytes in mouse and humans , and the Pavlidis lab observed that gene expression patterns between neurons and oligodendrocytes (another type of glial cell) were actually swapped.
These results suggest that human brain cells can behave totally differently. One could imagine the analogy that studying one kind of atom to infer the properties of another may not be quite the best approach, and by analogy, studying neurons from one kind of animal to infer something about another. This makes a bottom up approach extraordinarily complicated and casts tremendous doubt on the relevance to humans of most any result of physiology obtained from other species .
Where do we go from here?
Much of animal research has its roots in a reductionist framework that posits that to understand a system, one must know the function and behavior of its elements. One can then infer the overall behavior of the system. If we were to subscribe to such as framework then given what we know about the differences between animals and humans we would have to undertake the mammoth task of not just understanding one neuron in one human but billions of neurons and glia that are potentially genetically unique in each human. The task is technically mammoth.
And of course, there is the question of whether a bottom up approach is even the right path for what we want to accomplish in our understanding of the human brain. Consider that when we want to understand the properties of a material that describe its nature and utility, what matters are its macro features such as color and hardness. We don’t seek to predict these through measurement of the electron spins of each atom and their bond energies (which could work well some of the time but would be both inordinately expensive and technically challenging) but rather seek different system level measurement tools that are far easier and in the end more practically informative. Fundamentally it comes down to what we are trying to accomplish and how we want science to contribute to humanity’s understanding and care of the human brain.
A great and useful statement, and very clear. I am glad I somehow appeared on your mailing list.