When you choose to publish with PLOS, your research makes an impact. Make your work accessible to all, without restrictions, and accelerate scientific discovery with options like preprints and published peer review that make your work more Open.

PLOS BLOGS The Official PLOS Blog

“Neuroscience at Scale: Synaptomics tries to make sense of the brain at a stupefying scale” by Samuel Rose

Even though it’s smaller than a basketball, the scale of the human brain can feel astronomical. “There are as many neurons in the human brain as there are stars (100 billion) in the Milky Way” is commonly stated in musings on the brain. And while this figure isn’t quite accurate (it’s closer to 86 billion neurons/brain), the comparison is staggering.1 Neurons, though, aren’t useful without synapses – the point of contact between neurons where chemical messages are conveyed. It’s at synapses where memories are formed and skills like dribbling a basketball are honed. And it’s at the level of the synapse where the human brain scales up to numbers that dwarf our tiny galaxy – 100 trillion synapses is the crude estimate. Any effort to catalog and characterize the brain’s population of synapses probably sounds as futile as cataloging all the stars in the universe. A new field called synaptomics, however, has emerged to try and do just that: map the molecular composition and spatial distribution of every single synapse in the brain.

The scale of synapses is so staggering, there actually isn’t a consensus on how many there are. The number 100 trillion appears in popular articles and teaching tools, but it’s not typically referenced in peer-reviewed journals. 100 trillion is calculated by multiplying 100 billion neurons by 1,000 connections each, which belies the diversity in neuronal shape and size.

A sampling of the dendritic arbors of different neuron types, each making a different number of synapses.
A sampling of the dendritic arbors of different neuron types, each making a different number of synapses.

Neurons, like “E” in the figure above, may only form a dozen synapses on its simple dendritic arbor. A neuron with a dense network of dendrites, like in “B” or “F”, above, may form 10,000 to 100,000 synapses. With 86 billion neurons spread through roughly 200 unique regions2 – each region containing unique combinations of neuron types – accurate estimation becomes close to impossible. This is not to say that finding the precise number of synapses in a human brain is the most pressing question in neuroscience. But framing this question does show how difficult it is to answer even the most basic questions at this scale. In other words, if figuring out the number of synapses isn’t feasible, how does the field begin to answer more complex questions about the brain’s catalog of synapses?

The approach most neuroscientists take is to compartmentalize. How do you eat an elephant? One bite at a time, according to proverb. Neuroscientists take their fork and steak knife and start sawing away at one brain region or circuit and learn the rules that govern the set of synapses there. For example, there have been over 11,500 Pubmed cataloged publications on “hippocampal plasticity” – studies that furthered our understanding of synapses in this memory-formation region. But how do the synapses in this region compare to the rest of the brain? Will it take another 2.3 million bites (11,500 x 200 brain regions) of elephant meat to figure this out?

A group led by Seth Grant at University of Edinburgh is attempting to define some simple principles that distinguish and group the brain’s trillions (?) of synapses, in an effort they call the Genes to Cognition Synaptome Mapping pipeline or, more simply, synaptomics. They started with the mouse brain in a recent issue of Neuron.3 The group genetically engineered mice to express two different fluorescently “tagged” proteins: PSD95 attached to a green marker and SAP102 to an orange marker. PSD95 and SAP102 are found exclusively in synapses, so under a microscope the brains of these mice are a canvas of green and orange dots, each dot a synapse. With the naked eye, Grant and colleagues could distinguish three groups of synapses: those with PSD95, those with SAP102, and those with both. But they also measured the brightness, size, and relative density of each dot throughout the brain. Including these factors, they used machine learning to characterize the ~1 billion glowing synapses in the mouse brain. Clustering algorithms categorized 37 subtypes of synapses, each subtype with its own unique signature of brightness, size and density. The way these 37 subtypes are distributed throughout the mouse brain creates the first synaptome “signature” of a brain.

At this point, you could argue that an attempt to catalog 1 billion of anything would find clusters. If one were to catalog a billion cracks in the surface of a highway, groupings would likely arise. Do the synapse subtypes that Grant and colleagues observed have any significance to brain function? The fact that different synapse subtypes were enriched in certain brain regions argues that they do. For example, subtype “24” was abundant in the cerebral cortex and absent in the cerebellum. This makes sense since cerebral cortex and cerebellum develop differently and perform different functions.

Grant and colleagues also show that the synaptome signature is sensitive to insults in brain development. It has long been known that neuropsychiatric diseases like autism and schizophrenia are associated with disrupted brain connectivity. The authors saw changes in the synaptome signature of mutant mice with behavioral abnormalities resembling these neuropsychiatric diseases. These data show that the synaptome is sensitive to developmental disturbances, and changes in it may be a marker for psychiatric disease.

This study is more of a starting point for synaptomics than a presentation of its fruition. For one, this study builds its synaptome map with only two of the synapse’s thousands of proteins. The stated goal of the Genes to Cognition Synaptome Mapping pipeline is to map as many of these thousands of proteins as possible and incorporate them into a global catalog. What will neuroscience gain from this this catalog? Will future neuroscientists be able to see a memory form or a disease develop by simply “perusing” the catalog? These questions aren’t possible to answer right now, but hopefully this new field will provide some general principles that transcend the astronomical.


1          Azevedo, F. A. et al. Equal numbers of neuronal and nonneuronal cells make the human brain an isometrically scaled-up primate brain. The Journal of comparative neurology 513, 532-541, doi:10.1002/cne.21974 (2009).

2          Glasser, M. F. et al. A multi-modal parcellation of human cerebral cortex. Nature 536, 171-178, doi:10.1038/nature18933 (2016).

3          Zhu, F. et al. Architecture of the Mouse Brain Synaptome. Neuron 99, 781-799 e710, doi:10.1016/j.neuron.2018.07.007 (2018).

Any views expressed are those of the author, and do not necessarily reflect those of PLOS. 

Sam Rose received his PhD in Neuroscience from Emory University. He is currently a postdoctoral scholar at Duke University (Durham, North Carolina). His research focuses on the role of G-protein coupled receptors in various neuropsychiatric disorders. You can email him at samueljosephrose@gmail.com



Leave a Reply

Your email address will not be published. Required fields are marked *

Add your ORCID here. (e.g. 0000-0002-7299-680X)

Back to top