Your brain is carrying a massive reserve of wiring it has never switched on. Researchers at MIT have found that roughly 30 percent of all synapses in the adult mouse cortex are functionally dormant, lacking the molecular hardware needed to transmit signals under normal conditions. These “silent synapses,” long assumed to disappear after early childhood development, instead persist in enormous numbers into adulthood, sitting idle until the brain needs to encode something new.
The discovery, published in Nature and expanded in a 2024 review in Annual Review of Neuroscience, reframes how scientists think about learning capacity in mature brains. It also raises pointed questions about how this hidden reserve might be activated on demand or lost in disease.
What the MIT team actually found
The central finding comes from a team at MIT’s Picower Institute for Learning and Memory, led by neuroscientists Dimitra Vardalaki, Matthew Bhatt, and Mark Harnett. Using super-resolution protein imaging and electrophysiology in the adult mouse visual cortex, the group showed that a large fraction of synapses lack AMPA receptors. Those receptors are the molecular gatekeepers that allow a synapse to fire in response to routine neural signals. Without them, a connection relies solely on NMDA receptors, which activate only under specific voltage conditions. The result: a synapse that exists structurally but stays electrically quiet during ordinary brain activity.
The team traced these silent synapses to thin, finger-like protrusions called dendritic filopodia. These structures were previously considered temporary features of developing neurons, not something you would expect to find in abundance on adult brain cells. But there they were.
Critically, the researchers also showed that silent synapses can wake up. When neurons receive the right pattern of coordinated stimulation, the dormant connections recruit AMPA receptors and become functional. This “unsilencing” mirrors the mechanism behind long-term potentiation (LTP), the best-studied cellular process for strengthening connections during learning.
The 30 percent figure comes from MIT’s institutional reporting on the study, where Vardalaki described these dormant connections as a large reservoir of plasticity that could be recruited to form new memories. A follow-up review by Vardalaki, Yaeger, and Harnett, drawing on 186 citations, synthesized the broader evidence and proposed that silent synapses help the brain solve one of its oldest engineering problems: holding onto existing memories while staying flexible enough to store new ones.
Why neuroscientists think this matters
That tension between stability and flexibility has nagged neuroscience for decades. Theoretical work by Stefano Fusi and L.F. Abbott demonstrated mathematically that ongoing synaptic plasticity can overwrite stored memories, and that physical limits on how strong or weak any synapse can become constrain how long a memory survives. In plain terms, every time you learn something new, you risk degrading something you already know.
Silent synapses offer a potential escape from that bind. Because they sit outside the active circuit until recruited, they could absorb new learning without disturbing the connections already encoding established knowledge. Think of it as the brain keeping a stack of blank pages bound into a notebook that is already full of writing. New notes go on fresh pages instead of overwriting old ones.
If that model holds, it would help explain why adult brains retain significant learning capacity even as their neural architecture becomes more fixed with age. It would also suggest that the brain’s potential for adaptation is far greater than its moment-to-moment activity reveals.
What has not been proven yet
All of the primary data on silent synapse prevalence in adults comes from one region of the mouse brain: the visual cortex. Whether the 30 percent figure holds across other cortical areas, or in the human brain, has not been directly tested. The filopodia study focused on a single sensory area, and while earlier developmental work on thalamocortical inputs showed silent-to-active conversion in young animals, no published experiment has mapped adult silent synapse density across the full cortex or in non-rodent species.
The functional story also remains partly theoretical. The MIT group proposes that silent synapses serve as a ready-made substrate for new learning, but no longitudinal study has tracked individual silent synapses being recruited during a real learning task in a living animal over days or weeks. The unsilencing demonstrated in the Nature paper occurred under controlled stimulation protocols, not during naturalistic behavior. How the brain decides which silent synapses to activate, and whether certain types of experience preferentially recruit certain connections, is still an open question as of June 2026.
Some secondary coverage has linked impaired unsilencing to neurodegenerative conditions such as Alzheimer’s disease, but the primary literature from the MIT group does not include experimental data on disease states. That connection remains a hypothesis without direct evidence from these studies.
What this could change if confirmed in humans
The strongest evidence here is the imaging and electrophysiology data from the Nature paper, which measured receptor presence and synaptic function at the single-synapse level. That is primary, peer-reviewed work with a clear experimental design. The 30 percent estimate and the localization to filopodia are both grounded in that dataset.
The theoretical models from Fusi and Abbott provide a useful framework for why silent synapses might matter, but they were not designed to predict or explain the specific biology the MIT team found. Readers should treat the stability-plasticity framing as a plausible interpretation, not an established mechanism.
For the practical implications, the gap between mouse cortex data and human application is real. Bridging it will require new imaging tools capable of resolving individual synapses in living human tissue. That technology does not yet exist at the required scale, though emerging approaches like expansion microscopy and advanced molecular imaging are pushing closer.
If the findings do translate, the downstream effects could be substantial. Clinicians working on rehabilitation after stroke or traumatic brain injury would have a new target: dormant connections that might be coaxed back online. Educators and cognitive scientists would need to reconsider assumptions about the limits of adult learning. And drug developers targeting synaptic plasticity would have a far more specific mechanism to work with than the broad-spectrum approaches that have dominated the field.
A brain with more capacity than it lets on
The MIT findings challenge a decades-old assumption that silent synapses are a developmental leftover, relevant only in the first years of life. Instead, they point to a brain that maintains a standing inventory of dormant connections, ready to be switched on when circumstances demand it. Nearly a third of the cortical wiring in adult mice appears to be held in reserve, quiet but structurally intact.
The science is solid within its scope. The full story of what silent synapses do in human cognition is still being assembled, one experiment at a time. But the picture emerging from MIT’s work is striking: the adult brain may be far less fixed, and far more prepared for the unexpected, than neuroscience assumed for most of the last century.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
