The history of autism research has long been haunted by a discouraging assumption: if the brain is wired differently during early development, then by adulthood those circuits are essentially locked in place. You cannot rebuild a foundation that was laid wrong in the first years of life. Therapy, under that view, can only teach compensation strategies or manage symptoms, not change the underlying biology.
A study published July 23 in Nature Communications by researchers at the University of California, Los Angeles, now challenges that assumption head-on. The team, led by Dr. Janel Le Belle, Dr. Harley Kornblum, and Dr. Neil Harris of UCLA Neurosurgery, showed that in a mouse model of autism, adult brain function can be normalized within hours, even while structural differences from early development remain untouched.
The finding points toward a new way of thinking about treatment for autism and related neurodevelopmental conditions. Rather than trying to undo the structural architecture of the developing brain, it may be possible to target the dynamic signaling pathways that control moment-to-moment neural activity, effectively rewiring the brain’s function without altering its physical form.
A Model Built on Inflammation
To understand what the UCLA team discovered, it helps to know how they built their experimental system. The researchers exposed pregnant mice to a mild immune challenge, a synthetic mimic of viral infection, during a critical window of fetal brain development. The offspring of those pregnancies grew up displaying a constellation of traits that closely model core features of autism in humans: persistent brain overactivity measured by electroencephalography, heightened sensitivity to sensory stimuli such as sound, repetitive behaviors, and a lowered threshold for seizures.
These are not superficial behavioral quirks. The mice showed measurable differences in their brain structure and gene expression patterns, the kind of changes neuroscientists believe arise when the developing brain responds to prenatal immune activation. A mild immune challenge during pregnancy altered the course of brain development and produced lifelong differences in behavior and neural function.
The mTOR Pathway as a Cellular Thermostat
The team then turned to a molecule called rapamycin, a drug discovered in soil bacteria on Easter Island and best known in medicine as an immunosuppressant and anti-cancer agent. Rapamycin works by inhibiting a cellular signaling cascade called the mTOR pathway, short for mechanistic target of rapamycin.
Think of the mTOR pathway as a thermostat for a cell’s growth and metabolism. When it is set too high, cells grow too fast, consume too much energy, and become hyperactive. When it is set too low, cells become sluggish and fail to respond to their environment. In many neurodevelopmental conditions, including autism, the mTOR thermostat appears to be stuck in the high position in certain populations of neurons, driving excessive excitability that disrupts the brain’s delicate balance between excitation and inhibition.
Rapamycin, by turning down that thermostat, should in theory bring overactive neurons back toward a normal resting state. The UCLA team wanted to know whether doing so in adult mice, animals whose brains had been developing differently since before they were born, could reverse the functional consequences of that early-life wiring.
The answer was striking. A single dose of rapamycin, delivered to adult mice, reversed nearly all of the autism-like traits within roughly two hours. Brain overactivity quieted down. Sensory hypersensitivity diminished. Seizure vulnerability dropped. Repetitive behaviors eased.
Function Without Form
But here is where the story departs from the headline. When the researchers examined the animals’ brains at the molecular level, they found that rapamycin had not repaired the structural differences established during development. The abnormal gene expression patterns, the altered synaptic architecture, the fundamental wiring changes that had been present since fetal life: none of these were corrected by the drug.
What rapamycin did instead was more subtle, and arguably more important. The team performed detailed gene activity analysis and found that the drug reversed the abnormal expression of genes linked to autism, epilepsy, and ion channel function, specifically in excitatory neurons, the brain cells responsible for transmitting activating signals. Within hours of a single dose, the molecular machinery controlling how excitable those neurons were had been recalibrated. The neurons still sat in a brain whose physical structure was different from typical development, but their moment-to-moment firing patterns had returned to a normal range.
This distinction matters deeply. It suggests that the adult brain retains a degree of functional plasticity, the ability to alter its dynamic behavior without altering its hard-wired architecture, that has not been fully appreciated. The UCLA team demonstrated that a signaling pathway central to cellular metabolism (mTOR) can act as a master regulator of neuronal excitability, and that manipulating that pathway can restore normal brain function even in a nervous system whose development went off course months earlier.
The Catch: Temporary Relief and Rapid Tolerance
The findings come with significant caveats. The improvement after a single dose lasted only as long as the drug remained active in the animals’ systems. When the researchers attempted repeated dosing, the animals quickly developed tolerance; the same dose no longer produced the same effect. This makes rapamycin itself an unsuitable candidate for human therapy, quite apart from its well-known toxicity profile as an immunosuppressant.
The UCLA group does not recommend rapamycin as a treatment for autism, and they are explicit about that in the paper. The value of the study is not in the drug but in what the drug reveals about the brain’s hidden capacity for change.
Toward a New Class of Targets
If the brain’s function can be normalized without correcting its structure, then the targets for therapy shift. Instead of trying to intervene during narrow developmental windows to prevent atypical wiring from occurring, a difficult proposition in humans where diagnosis often comes years after the critical periods have closed, researchers can look for ways to modulate the signaling pathways that control moment-to-moment neural activity.
The mTOR pathway is one such target. But there are others. Ion channels, the proteins that control the flow of charged particles across neuron membranes, were among the genes whose abnormal expression was reversed by rapamycin in the UCLA study. These channels determine how easily a neuron fires, how quickly it recovers after firing, and how it integrates signals from thousands of synaptic inputs. Molecules that fine-tune ion channel function, or that modulate the broader excitation-inhibition balance in neural circuits, could achieve the same kind of functional normalization without the toxicity and tolerance problems of rapamycin.
There is also a deeper conceptual implication. If the adult brain remains functionally modifiable even after early developmental differences have been laid down, then the therapeutic window for intervention may be far wider than previously assumed. A person diagnosed with autism at age five or fifteen or fifty is not necessarily locked into a fixed neural trajectory. The circuits are there, built differently, but they may be continuously adjustable through the right molecular levers.
Limits and Next Steps
This is, of course, a mouse study. The gap between rodent models and human clinical application is vast, and the history of neuroscience is littered with treatments that worked beautifully in mice and failed in people. The tolerance problem, the fact that the brain adapts to repeated mTOR inhibition, is a serious barrier that will require new strategies, perhaps intermittent dosing regimens or combination therapies that target multiple pathways simultaneously.
Nor should the finding be interpreted as suggesting that autism is simply a condition of neuronal hyperexcitability that can be fixed with a pill. Autism is a heterogeneous set of conditions with diverse genetic and environmental origins, and what holds true in a mouse model of prenatal immune activation may not hold true for other subtypes. The behavioral and cognitive differences associated with autism are not necessarily deficits that require correction; many autistic people view their neurotype as a valid form of human variation rather than a disorder to be treated.
What the study does offer is a concrete biological proof of concept: the adult brain’s functional state is more plastic, and more accessible to intervention, than many researchers assumed. And it identifies a tractable set of molecular targets, including the mTOR pathway, ion channel regulation, and excitation-inhibition balance, that merit deeper investigation.
The research was funded by the Adelson Medical Research Foundation, the UCLA Brain Injury Research Center, the National Institutes of Health, the Simons Foundation, Autism Speaks, and the UCLA Center for Autism Research and Treatment.
References
Le Belle, J. E., Harris, N. G., & Kornblum, H. I. (2026). Rapamycin reverses functional deficits in a maternal immune activation model of autism by normalizing neuronal excitability. Nature Communications. DOI: 10.1038/s41467-026-74958-1