Deep brain stimulation of the nucleus accumbens and anterior limb of internal capsule in a paediatric patient with autism spectrum disorder: A case report and literature review

INTRODUCTION

Autism spectrum disorder (ASD) is a congenital neurodevelopmental psychiatric disorder characterised by persistent deficits in social communication and social interaction, together with restricted and repetitive patterns of behaviour, interests or activities. Epidemiological studies indicate a continuously increasing prevalence of ASD worldwide. ¹ Currently, treatment strategies for ASD primarily rely on pharmacological therapy and behavioural interventions. However, neuromodulatory approaches—particularly transcranial magnetic stimulation (TMS) and deep brain stimulation (DBS)—have increasingly been explored for the clinical management of patients across the ASD spectrum. ² Here, we report the case of a 12‐year‐old girl with ASD complicated by severe mood disturbance and self‐injurious and aggressive behaviours, who was treated with DBS targeting the bilateral anterior limb of the internal capsule (ALIC) and nucleus accumbens (NAc). To our knowledge, the present case represents the first reported application of this dual‐target DBS strategy in a patient with ASD.

CASE SUMMARY

The patient had a normal birth and early development until age two; then, at around two and a half years, she exhibited language regression and stagnation of cognitive development. At 5 years of age, she developed prominent stereotyped behaviours, including repetitive pouring of water between containers, echolalia, butterfly‐like hand movements, and impulsive circular running. These symptoms progressively worsened, and by 7 years of age she began to display severe self‐injurious and aggressive behaviours, manifested by repetitive hitting of her head, face and limbs, alongside intermittent physical aggression towards her guardians (figure 1A).

FIGURE 1.

Clinical symptoms and treatment evaluation. (A) The injured arms of the patient and her guardian. (B) Schematic reconstruction of the stimulating electrode placement. The nucleus accumbens is represented by the light green zone, and the anterior limb of the internal capsule by the blue area. Volume of tissue activated spheres for the nucleus accumbens and anterior limb of the internal capsule electrodes are shown in orange and red, respectively. (C) Changes in disease severity scores over 1 year of deep brain stimulation. (D) Morphometric changes over 1 year of deep brain stimulation. ALIC, anterior limb of the internal capsule; CARS, Childhood Autism Rating Scale; DBS, deep brain stimulation; M, month; NAc, nucleus accumbens.

The patient underwent behavioural intervention training at multiple institutions without significant clinical benefit. Pharmacological treatment with antipsychotic medications, including risperidone and quetiapine, also failed to adequately control her mood instability or aggressive behaviours. A 3.0‐T brain magnetic resonance imaging (MRI) scan revealed no overt structural abnormalities. She was formally diagnosed with autism spectrum disorder at the age of 12 years.

This case represents the first paediatric patient enrolled in our clinical registry study (an exploratory clinical trial on the effectiveness and safety of a dual‐targeted implanted neural stimulation system for treating autism, ChiCTR2400084724). This study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board of the hospital (Approval No. 2023–1025). Written informed consent was obtained from her legal guardians.

Following comprehensive multidisciplinary evaluation and detailed discussion, the dual‐target DBS of the NAc and ALIC was implemented. Preoperative planning was performed using iPlan 3.0 software (Brainlab, Germany). The patient's head was fixed using a Leksell G–frame (Elekta, Sweden), followed by computed tomography (CT) scanning, which was fused with preoperative MRI images. Bilateral trajectories targeting the NAc and ALIC were planned for the two electrodes. SR1212–S eight‐contact leads and an SR–1103 implantable pulse generator (contact length 1.5 mm, inter‐contact spacing 1.5 mm; SceneRay, Suzhou, China) were implanted, positioning three ventral contacts within the NAc and five dorsal contacts within the ALIC. Postoperative CT scanning (slice thickness of 1 mm) was performed immediately to verify accuracy (figure 1B). The methods of DBS electrode localisation and stimulation volume of tissue activated (VTA) modelling can be found in the Supporting Information S1. The patient recovered uneventfully and was discharged on postoperative day six. DBS stimulation was activated 4 weeks post‐implantation, followed by monthly follow‐ups for 6 months. Thereafter, visits were conducted every 3 months, and the interval was eventually extended to 12 months. Standard behavioural and cognitive rehabilitation therapies were consistently maintained post‐surgery.

Routine programming sessions were conducted biweekly using a hybrid approach that combined in‐person clinic visits with remote telemedicine, supplemented by unscheduled adjustments to manage symptom fluctuations. Notably, all programming sessions and behavioural evaluations were successfully completed while the patient was fully awake and cooperative, without the need for sedation.

Clinical outcomes and behavioural assessments

During the 12‐month follow–up period after the DBS activation, the child did not experience any complications such as cerebral haemorrhage, epileptic seizures or hardware failure.

In this patient, we used constant–voltage stimulation with a frequency of 130 Hz and a pulse width of 120 μs. During the first 6 months after DBS activation, the stimulation amplitude at the ALIC site was gradually increased from 0.0 to 5.0 V, and at the NAc site from 0.0 to 3.0 V. As shown in figure 1C, the patient's symptom severity scores on the Childhood Autism Rating Scale (CARS) showed an overall decline, marking a clinical transition from severe (CARS > 36) to mild (CARS 30–36) within 6 months. ³ Furthermore, other related clinical scales—including the Autism Behaviour Checklist (ABC), the Repetitive Behaviour Scale‐Revised (RBS–R) and the Autism Treatment Evaluation Checklist (ATEC)—demonstrated significant multidimensional improvements (e.g., in social interaction and self–injurious behaviour) compared to the preoperative baseline (figure S1). ⁴ , ⁵ , ⁶

To further evaluate the specificity of the therapeutic effects of DBS and rule out potential placebo responses, a controlled single‐blind DBS–OFF period was implemented 6 months post‐activation. Specifically, the patient and the clinical assessors were blinded to the stimulation status (DBS‐ON vs. DBS‐OFF). However, the treating physicians and the patient's parents were not blinded. During the DBS‐OFF period, the patient developed emotional lability with irritability and angry outbursts and gradually resumed self‐injurious and aggressive behaviours. Although we had planned a 3‐month off period, the family requested earlier reactivation at 2 months after DBS cessation to control mood and behavioural symptoms. Notably, when we gradually reinstated the stimulation parameters after the OFF period, the patient's condition did not immediately return to the baseline level prior to the OFF period. During periods of clinical fluctuations, we temporarily adjusted the amplitude by 0.1–0.3 V and then gradually returned it to the baseline (ALIC 5.0 V, NAc 3.0 V) within 2 weeks. After 2 months of continuous re–stimulation, the patient's condition was finally stabilised at a mild level (figure 1C).

Neuroimaging findings

The patient underwent fluorodeoxyglucose positron emission tomography/computed tomography (FDG–PET/CT) and MRI imaging preoperatively and at 1 year postoperatively. To prevent motion artefacts and ensure safety, an anaesthesiologist administered minimal‐dose sedatives (midazolam and dexmedetomidine) alongside continuous oxygen supplementation and vital sign monitoring during the imaging examination. PET images were normalised to cerebellar uptake values. One year after DBS implantation, compared with the baseline, global cerebral metabolic activity was reduced, with a more pronounced decrease in the bilateral frontal, temporal and parietal lobes (figure S2). The specific change values can be found in the Supporting Information S2.

Voxel‐based morphometry (VBM) analysis revealed a reduction in cortical grey matter density, with relative preservation of white matter density, at 1 year postoperatively. Compared with baseline measurements, deep brain nuclei showed reduced volumes in the bilateral nucleus accumbens (associated with the directly stimulated regions), as well as in the adjacent caudate nucleus and putamen. The bilateral amygdala, thalamus and hippocampus also showed volume decreases. In terms of cortical regions, the right supramarginal gyrus decreased in volume, whereas the bilateral occipital cortex increased. Overall, the cortical volumes exhibited heterogeneous changes; the detailed changes can be found in the Supporting Information S3.

DISCUSSION

ASD is frequently accompanied by diverse psychiatric comorbidities, including attention deficit hyperactivity disorder, anxiety and mood disorders. ⁷ Although standard interventions—including pharmacotherapy and intensive behavioural therapies such as applied behaviour analysis (ABA)—offer symptomatic relief, their efficacy remains suboptimal for core deficits, particularly in severe treatment‐refractory cases. Consequently, given the limitations of these conventional approaches and their associated long‐term risks, there is a compelling clinical rationale to explore neuromodulation as a targeted therapeutic option for carefully selected patients. ⁸

Neuromodulation for ASD is categorised into non‐invasive and invasive approaches. Non–invasive methods, such as TMS and transcranial direct current stimulation (tDCS), target cortical excitability in regions such as the dorsolateral prefrontal cortex (DLPFC) to improve social and executive functions. However, their clinical utility is limited by heterogeneous outcomes, modest effect sizes, and a lack of long‐term stability. ⁹ These limitations underscore the need for invasive strategies such as DBS, which offers stable, circuit‐specific modulation for treatment‐refractory cases.

A summary of published clinical studies investigating DBS in ASD is provided in table 1. However, substantial heterogeneity in patient selection and parameters necessitates further systematic evaluation to establish optimal stimulation strategies.

TABLE 1.

Deep brain stimulation targets and stimulation parameters in autism spectrum disorder

Authors	Age (years), sex	Presentation	DBS target	DBS parameters	Outcome
Sturm et al. 10
Case 1	13, M	Life‐threatening SIB, nonverbal phenotype, & profound social cognition deficits	BLA	2–6.5 V, 120 μs, 130 Hz	Decrease in SIB & improvement in social cognition.
Stocco et al. 11
Case 1	19, F	Refractory SIB, motor stereotypies, & dystonia due to antipsychotics	GPi	3.3 V, 120 μs, 80 Hz	Decreased stereotypies & decreased dystonia.
Case 2	17, M	Refractory SIB, motor stereotypies, & intellectual disability	GPi & ALIC	2.5 V, 210 μs, 100 Hz	Transient decrease in SIB & stereotypies, with a return to the initial baseline after 6 months.
Park et al. 12
Case 1	14, M	Refractory SIB	NAc	3–5 V, 130 Hz, 90 μs	Decrease in SIB.
Segar et al. 13
Case 1	24, F	Biting hands, picking skin	NAc	2–8 V, 130 Hz, 90 μs	Improvements in compulsive behaviours, coprolalia, language and social interaction.
Doshi et al. 14
Case 1	42, F	OCD, aggression	NAc	5–6.5 V, 100 μs, 130 Hz	Marked improvements in OCD symptoms, aggressive behaviour, eye contact and appropriate laughter.
Davis et al. 15
Case 1	44, M	OCD, aggression	Ventral capsule/ventral striatum	5–6.5 V, 100 μs, 130 Hz	The clinical improvements were maintained, albeit with fluctuations, after 3 years. No effect on core symptoms of ASD.
Graat et al. 16
Case 1	39, F	OCD	Ventral ALIC	3–5.5 V, 130 Hz, 90 μs	50% reduction of OCD symptoms following DBS, especially obsessions.
Case 2	54, F	OCD	Ventral ALIC & MFB		Initially did not benefit from DBS. Thereafter, OCD symptoms improved and decreased by more than 50%.
Case 3	32, M	OCD, aggressive intrusions	Ventral ALIC		Partial responder, probably due to several transient side effects of DBS.
Case 4	31, F	OCD	Ventral ALIC		Only some subjective improvements.
Case 5	51, M	OCD	MFB		Obsessive‐compulsive symptoms disappeared entirely. Improved confidence and less social shyness.
Case 6	30, F	OCD	MFB		35% reduction of OCD symptoms following DBS.
Benedetti et al. 17
Case 1	27, M	Aggressive behaviour towards self	pHyp	2.7–2.8 V, 185 Hz, 90 μs	Improvements in seizures, in aggressive behaviour, in quality of life, in daily living skills.
Case 2	16, M	Self‐aggression			Aggressive behaviour was controlled for a month. After 2 months, it reappeared as before surgery.
Gorodetsky et al. 18
Case 1	8, F	SIB: Head banging, self‐punching, self‐slapping, self‐biting, self‐gagging	NAc	2–2.5 mA, 130 Hz, 60 µs	Marked improvement in the SIB.
Case 2	14, F	SIB: Head banging, self‐punching, self‐scratching
Case 3	13, M	SIB: Self‐punching, self‐slapping, self‐gagging, self‐biting, eye poking
Case 4	12, F	SIB: Head banging, self‐slapping, self‐biting
Case 5	9, M	SIB: Self‐punching, self‐slapping, eye poking
Case 6	7, M	SIB: Head banging, self‐punching, self‐slapping

Abbreviations: ALIC, anterior limb of the internal capsule; ASD, autism spectrum disorder; BLA, basolateral amygdala; DBS, deep brain stimulation; F, female; GPi, globus pallidus internus; M, male; MFB, medial forebrain bundle; NAc, nucleus accumbens; OCD, obsessive–compulsive disorder; pHyp, posteromedial hypothalamus; SIB, self‐injurious behaviour.

Both the ALIC and NAc have individually been reported as targets in DBS studies for ASD patients. Emerging evidence, particularly from studies conducted in China, suggests that simultaneous modulation of these interconnected targets may provide enhanced therapeutic effects in severe psychiatric disorders, such as obsessive‐compulsive disorder (OCD) and substance use disorders (SUD). ¹³ , ¹⁹ In the present study, we employed combined stimulation of the ALIC and NAc, with the aim of alleviating severe mood dysregulation and self‐injurious behaviour while potentially enhancing higher‐order emotional and cognitive processing.

The ALIC is a critical white matter conduit connecting the prefrontal cortex to subcortical structures, integrating circuits essential for cognitive flexibility and emotional regulation. Given the phenomenological overlap between the repetitive behaviour of ASD and OCD, where ALIC–DBS is already established, modulating this target may directly address restrictive behavioural dimensions. ¹⁶ Simultaneously, the NAc serves as a central hub for reward processing; its dysfunction is closely linked to the social motivation deficits characteristic of ASD. ²⁰

In this case, we sought to exclude placebo influences by conducting a controlled DBS‐OFF phase. The resulting data strongly argue against a mere placebo effect; rather, the delayed, progressive recovery observed after turning the stimulation back on points to a genuine neuromodulatory process involving long‐term plasticity.

Our rationale for dual targeting is supported by emerging evidence suggesting that simultaneous modulation of these interconnected nodes may yield enhanced therapeutic synergy in severe psychiatric conditions. ¹⁸ This approach aims to address the distributed network dysfunction central to ASD models, particularly those involving social motivation and cognitive control. Notably, our post‐operative findings revealed reduced glucose metabolism and altered cortical T1 density in frontal and motor regions, suggesting that ALIC/NAc–DBS does not globally activate the brain but rather suppresses specific hyperactive neural circuits via high‐frequency stimulation. In ASD patients with severe self‐injurious behaviour, ALIC/NAc DBS may improve symptoms by inducing local functional inhibition and long‐term neuroplasticity.

The study is based on a single case, which limits the generalisability of the findings. Looking forward, large‐scale datasets will be essential for constructing stimulus–response models and standardising personalised stimulation paradigms. Integrating intelligent assessment systems with DBS could enable precise monitoring of rehabilitation trajectories. Such advancements will not only refine individualised treatment for refractory ASD but also deepen our understanding of the mechanisms underlying neurorehabilitation and circuit‐based recovery in this population.

LEARNING POINTS

• Dual‐target DBS of the NAc and ALIC can alleviate self‐injurious and repetitive behaviours in paediatric ASD patients.

• Radiological findings suggest that DBS may induce adaptive structural and functional remodelling in the brain of the patient with ASD.

AUTHOR CONTRIBUTIONS

Liang Chen, Ying Mao and Zhen Wang: study conception and design; Liang Chen, Xin Zhang and Liqin Lang: surgical procedure; Jingchuan Wu, Zhiyuan Sheng and Weiliang Hou: data collection; Xin Zhang, Ankang Hu and Jingchuan Wu: analysis and interpretation of results; Zhimin Zhao: psychiatric evaluation; Min Shen, Chunxin Xu, Mei Lv and Yangyang Lu: rehabilitation evaluation; Xin Zhang, Jingchuan Wu: draft manuscript preparation.

FUNDING

This study was supported by Brain Science and Brain‐like Intelligence Technology‐National Science and Technology Major Project (2025ZD0215100), National Natural Science Foundation of China (82272116) and Shanghai Hospital Development Centre Foundation (SHDC12024131).

CONFLICT OF INTEREST STATEMENT

Zhen Wang is an associate editor/editorial board member of General Psychiatry and a co‐author of this article. He was excluded from editorial decision‐making related to the acceptance of this article for publication in the journal. The other authors declare no conflicts of interest.

ETHICS STATEMENT

This study involves human participants. All procedures were approved by the ethics committee of Huashan Hospital (Approval number: 2023‐1025). For this patient, written informed consent was obtained from her legal guardians before participation in the study. This study is registered with Chictr.org.cn, number: ChiCTR2400084724.

Supporting information

Supporting Information S1

Supporting Information S2

Supporting Information S3

Figure S1

Figure S2

Biography

Xin Zhang graduated from Shanghai Medical College, Fudan University in China with an MD degree in 2012. He joined the Department of Neurosurgery at Huashan Hospital, Fudan University in 2012, where he completed his residency and clinical fellowship training, and has since served as an attending physician. From 2014 to 2015, he completed specialised training in endoscopic skull base surgery under the mentorship of Dr Daniel F. Kelly at Saint John's Health Centre in Los Angeles, USA. Dr Zhang currently serves as a committee member of the Neuromodulation Branch of the Chinese Neuroscience Society. His current main research activities and clinical interests focus on functional neurosurgery and on the translation and clinical trials of invasive neuromodulation technology for movement disorders and refractory psychiatric disorders.

DATA AVAILABILITY STATEMENT

Data are available upon reasonable request.