Deep brain stimulation (DBS) may offer a viable treatment strategy for children with autism spectrum disorder (ASD) who display severe self-injurious behavior.

In a first-in-human pilot trial of children with ASD and severe, refractory self-injurious behavior (SIB), targeted DBS of the nucleus accumbens (NAcc) led to clinically meaningful reductions in self-harming in five of six children.

These behavioral improvements were accompanied by structural and volumetric changes in neural circuitry implicated in the emergence of SIB in ASD.

“Our findings advance mechanistic insight into the neural circuitry underlying SIB and delineate how targeted modulation of corticostriatal networks drive SIB improvement in ASD,” the researchers wrote.

“Together, our findings begin to explain how DBS may be reshaping the underlying neurocircuitry,” George Ibrahim, MD, PhD, with The Hospital for Sick Children and University of Toronto, both in Toronto, Ontario, Canada, who worked on the study, told Medscape Medical News.

The study was published online on April 10 in Science Advances.

Clinically Meaningful Improvement

SIB represents one of the most debilitating manifestations of ASD, characterized by repetitive behaviors that cause physical harm without suicidal intent and often persist despite behavioral and pharmacologic interventions. Roughly 20%-25% of cases are refractory, highlighting the need for biologically targeted therapies.

Converging evidence has implicated dysfunction in corticostriatal circuitry — particularly imbalances between limbic and prefrontal inputs to the striatum — in the pathogenesis of repetitive and self-injurious behaviors.

Yet the lack of causal evidence linking modulation of corticostriatal networks to changes in SIB has limited advancement of effective, targeted therapies.

In this phase 1 study, six children, ranging in age from 7 to 14 years, with severe refractory SIB underwent bilateral DBS of the NAcc and were followed for 1 year. Clinical outcomes were assessed using the Repetitive Behavior Scale-Revised (RBS-R) SIB subscore and quality-of-life measures.

The DBS procedure was “generally safe and feasible,” said Ibrahim and five of the six children experienced clinically meaningful improvement self-harming behaviors.

Across the cohort, SIB decreased by an average of 44.3% at 1 year, accompanied by significant improvement in health-related quality of life.

Individual responses varied, with reductions in RBS-R SIB scores ranging from 45.5% to 72.7% in responders. One child who had significant baseline neurodevelopmental abnormalities showed no improvement.

These clinical gains were associated with engagement of a distributed network encompassing sensorimotor, limbic, and striatal regions, identified through connectivity mapping of stimulation sites.

Neuroimaging further demonstrated longitudinal structural changes following DBS.

After 1 year, compared to baseline, participants exhibited reduced volume in regions implicated in reward and emotional processing, including the caudate nucleus, putamen, anterior cingulate cortex, and hippocampus, alongside volume increases in areas involved in motor control, social-emotional processing, and pain modulation.

The findings were supported by findings in a well-established mouse model of ASD characterized by excessive self-grooming that results in tissue injury.

Mice receiving NAcc stimulation showed significant reductions in repetitive and injurious self-grooming, and these behavioral shifts were associated with similar neural circuitry changes as seen in the children.

“Building on the promising clinical results from our pilot study, as well as the mechanistic insights from this new work, we are now conducting a double-blind, randomized, sham-controlled clinical trial,” Ibrahim said.

“This next phase is essential to rigorously evaluate both the safety and effectiveness of DBS in this vulnerable population. At the same time, we are continuing translational research to better understand the biological basis of these behaviors and how DBS affects the underlying brain circuits,” he added.

He also said researchers are working closely with affected families to gain insight into their experiences and the factors guiding treatment decisions, adding that additional evidence will be needed before the strategy can be adopted more broadly in clinical practice.

Strong Proof-of-Principle Research

Reached for comment, Shaheen Lakhan, MD, PhD, neurologist, and researcher based in Miami, said what makes this study “powerful” is not just symptom reduction but also that it demonstrates “causal, circuit-level modulation of a defined fronto-limbic-striatal network, with measurable structural remodeling across species.”

“Self-injurious behavior is not just a symptom, it is a network state, and this shows that changing the state changes the outcome. This is not correlation, it is intervention at the level of brain circuitry, which is a fundamentally different therapeutic paradigm,” said Lakhan, who wasn’t involved in the study.

He cautioned, however, that physicians should be “clear-eyed” about the limitations — a six-patient phase 1 “signal” in a highly refractory population and DBS is “invasive and not scalable in its current form.”

“But conceptually, this is a proof-of-principle that precision neuromodulation works. DBS may be the first-generation hardware, but the future is adaptive, intelligent, digital medicine built on exactly this kind of brain circuit insight,” Lakhan told Medscape Medical News.

This study had no commercial funding. Ibrahim and Lakhan had no conflicts of interest.