When Maya's parents watched her lose the ability to say words she had been saying for months, they had no explanation. For families facing what researchers call "regression" in autism — a loss of previously acquired social, communication, or motor skills — the experience can be bewildering and painful. Now, a team at Yale University School of Medicine may have found one reason why this happens.

Researchers led by Sheng-Nan Qiao and Sung Eun Wang have discovered that neuroinflammation can trigger behavioral regression in individuals with SHANK3 haploinsufficiency, a genetic form of autism. Their study, published in Molecular Psychiatry, found that roughly 40 percent of autistic people with this genetic profile experience such regression over time. The new research suggests that subclinical infections — ones too mild to cause obvious illness — may activate an immune response in vulnerable brains, setting off a cascade that erodes newly learned skills.

To understand this connection, the team studied mice with a partial loss of the SHANK3 gene. These mice appeared completely normal at first — no behavioral impairments whatsoever. But when the researchers injected them with lipopolysaccharides, a molecule that triggers immune responses, something striking happened. Two weeks later, while ordinary mice recovered fully, the SHANK3-deficient mice developed motor impairments, anxiety-like behaviors, and repetitive grooming patterns similar to mice with complete SHANK3 loss. The mice had regressed.

The findings point to a critical intersection between genetic vulnerability and immune activation. The researchers discovered that in the affected mice, genes related to neuroinflammation were upregulated while synaptic function genes were downregulated. Microglia — the brain's immune cells — became overactive, beginning to engulf synapses and disrupt the connections between neurons. This overactivation appeared to be driven in part by elevated toll-like receptor 4 (TLR4) in the SHANK3-deficient mice.

But the study also carried a message of hope: when the researchers gave the mice anti-inflammatory treatment after the immune challenge, the behavioral changes partially reversed. The repetitive grooming decreased, anxiety lessened, and motor skills improved. This suggests that neuroinflammation could be a therapeutic target — that intervening after a triggering event might help restore lost function.

The team had previously observed this pattern in a small cohort of human patients, noting that behavioral regression triggered by subclinical infections responded to immunomodulator treatments. The new mouse model now provides a way to study the mechanism in detail and test new interventions. Autism spectrum disorder affects approximately 1 in 100 children worldwide, and for the subset with SHANK3 haploinsufficiency, understanding why regression occurs — and how to stop it — could be life-changing.