Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by persistent deficits in social communication and interaction and restricted, repetitive behaviors, interests, and/or activities.1,2 Extreme behaviors in patients with ASD include self-injurious behavior (SIB) or persistent episodes of aggression.3,4 A recent meta-analysis estimated a pooled prevalence of self-injury of 42% among patients with ASD, noting poor communication and developmental delay as additional risk factors.5,6 Both SIB and aggression can lead to physical harm, psychological stress, and increased healthcare resource utilization, and therefore may adversely affect the quality of life of patients and their caregivers.7–9
The consequences of extreme behaviors of patients with ASD are further compounded by the limited efficacy of psychopharmacology and by challenges in access to behavioral treatment. Expert consensus and small-cohort studies have supported the use of second-generation antipsychotics—particularly risperidone and aripiprazole—to target aggression or the irritability that may underlie SIB.10–12 Applied behavioral analysis, including functional behavior assessment, has also met criteria as evidence-based practices for treating extreme behaviors by manipulating antecedents and reinforced consequences.13,14 However, the variation in efficacy and side effects of medications and the lack of accessibility of applied behavioral analysis limit the effectiveness of these treatments in real-world settings.15,16Alternative and innovative options are therefore increasingly sought, including deep brain stimulation (DBS).
DBS electrodes targeting the posterior hypothalamus (pHyp)17,18 and the nucleus accumbens (NAcc)19 have been used to treat aggression and SIB. However, the exact mechanism by which stimulation of these targets alters extreme behaviors in patients with ASD is not yet understood. It is likely that different DBS targets for extreme behaviors that are used in patients with ASD may engage common underlying neural circuitry. Recent connectivity modeling by Li et al.20 in patients with obsessive-compulsive disorder (OCD) demonstrated that optimal clinical response across several different DBS targets was associated with stimulation of common streamlines in a normative structural connectome. These streamlines could represent a singular indirect circuit that unifies disparate target regions and underpins at least some of their therapeutic effect.20 Network mapping approaches have also been used successfully to identify whole-brain networks incorporating both gray matter and white matter tracts that are associated with DBS treatment success from disparate targets in Parkinson’s disease and depression.21,22
The goals of this study were to 1) summarize current knowledge surrounding the effectiveness of DBS in treating extreme behaviors such as SIB and aggression in patients with ASD, and 2) evaluate neural circuitry engaged by DBS targets. The authors of all published studies of patients with extreme behaviors associated with ASD and treated with DBS were identified through a systematic review and invited to participate in the current study. A connectomic analysis of all DBS targets used to treat extreme behaviors in 4 patients with ASD was performed, and converged upon a common circuitry linking disparate targets. This multisite collaborative initiative advances understanding of DBS for the treatment of extreme behaviors in patients with ASD.
Methods
Systematic Review
The PRISMA search was performed by searching 3 electronic databases: PubMed, EMBASE, and Web of Science. This search included articles from inception to January 2020 to identify individual clinical data in patients with ASD who received DBS treatment for any indication. The following search terms were used either alone or in combination in the initial search: “autism,” “deep brain stimulation,” “neuromodulation,” and “ASD.” Inclusion criteria were articles reporting on patients with 1) a reported diagnosis of ASD; 2) DBS for the treatment of extreme refractory behaviors (specifically if the primary indication was not a movement disorder); and 3) individual patient data regarding demographics, DBS target, stimulation parameters, and reported outcome measures. Exclusion criteria included the following: 1) articles not in English; and 2) abstracts or conference proceedings. All articles reporting on DBS for patients with ASD were included, and that included case reports.
The initial title and abstract screening and subsequent full-text review were performed independently by 2 reviewers (L.M.E. and H.Y.), with disagreements resolved through conference with a senior author. Reviewed papers were cross-referenced to include missing articles. Corresponding authors of included articles were contacted for imaging sequences for connectomics.
Study Population
We performed a systematic review to identify all reports of patients with extreme behaviors associated with DBS published in the literature. This was completed according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.23 Data collected included the following (where available): age at surgery, sex, comorbidities, previous treatments, ASD severity scores, DBS target, stimulation parameters, complications, and reported follow-up.
The corresponding authors of all studies were invited to participate in the current analysis by sharing pre- and postoperative imaging, including MRI and CT. Two patients identified in the systematic review and 6 additional patients (of coauthors J.C.B.I. and C.V.T.) were added for imaging analysis, due to the data being either unpublished or found outside of the systematic review search dates.
Electrode Localization and Volume of Tissue Activated Modeling
Lead-DBS software (https://www.lead-dbs.org/) was used to perform DBS electrode localization and volume of tissue activated (VTA) modeling as previously described.24 Briefly, immediate postoperative and, where available, high-resolution preoperative structural MRI acquisitions were corrected for intensity inhomogeneity and rigidly coregistered using SPM12 software (https://www.fil.ion.ucl.ac.uk/spm/software/spm12).25 The images were then nonlinearly normalized to standard space (ICBM 2009b NLIN asymmetrical) by using "effective low variance" ANTs software (http://stnava.github.io/ANTs). To correct for postoperative brain shift, an additional subcortical affine transform was performed where necessary.26 The electrodes were manually localized using the postoperative acquisition as agreed on by 2 experienced users (A.L. and J.G.) and were then warped to standard space by using the previously generated transforms. Each patient’s bilateral VTAs associated with their individual settings were modeled by first constructing a 4-compartment volume conductor model (http://iso2mesh.sourceforge.net/cgi-bin/index.cgi) that segregated peri-electrode tissue by tissue type.27 The FieldTrip-SimBio finite element model pipeline was then used to simulate the potential electric field distribution around each active contact (https://www.mrt.uni-jena.de/simbio/index.php/; http://fieldtriptoolbox.org), and a binary VTA was generated by thresholding the gradient of this distribution at 0.2 V/mm.
Functional and Structural Connectivity Mapping
The brain-wide functional and structural connectivity of each patient’s bilateral VTAs was explored using high-quality normative connectomes as previously described.28–30 To assess functional connectivity, whole-brain connectivity maps of each VTA pair were computed. These maps are based on the resting-state functional MRI time course–dependent correlations of blood oxygen level–dependent signal within the seed (i.e., VTA pair) region and all the other voxels in the brain (in-house MATLAB script, version R2017b; The MathWorks Inc.) across 1000 healthy subjects from the Brain Genomics Superstruct Project data set (http://neuroinformatics.harvard.edu/gsp) (age range 18–35 years; 57% female).31
Structural connectivity was assessed by identifying all streamlines that touched the seed (i.e., VTA pair) within an approximately 12-million-fiber whole-brain tractography template (in-house MATLAB script). This template was assembled using generalized q-sampling imaging32 from a high-fidelity multishell diffusion-weighted MRI Human Connectome Project data set (http://www.humanconnectomeproject.org) of 985 healthy subjects (age range 22–35 years; 54% female).33
Statistical Analysis Identifying Common Network Components
R (https://www.r-project.org/, version 3.4.4) and RMINC (https://github.com/Mouse-Imaging-Centre/RMINC) software were used for statistical analysis. The 4 functional connectivity r-maps were converted to t-maps and, using the known p distribution, thresholded at t = 5.1 (pBonferroni corrected < 0.05) using the Bonferroni method to correct for multiple comparisons.34,35 Using this conservative approach, only voxels that showed significant functional connectivity to the seed VTAs in each individual map were considered. These 4 thresholded t-maps were binarized and then overlapped, creating a single sum map. A higher voxel value in the sum map indicated a brain area that exhibited significant functional connectivity to a greater number of VTA seeds. A similar approach was used to analyze the structural connectomes, in which individual streamline maps were also binarized (1 = touched, 0 = not touched by the seed). These individual streamline maps were summed and streamlines common to all 4 seeds were identified.
Results
Systematic Search
The initial search strategy identified 297 articles (171 after duplicates were removed) (Fig. 1). After initial title and abstract screening, a total of 18 articles underwent full-text review. An additional 2 articles not found in the database search were included after reviewing the references of the included studies. Of 20 articles undergoing full-text review, 7 articles reporting on 9 patients were included. Reasons for exclusion were due to lack of individual patient data, such as age, DBS target, stimulation parameters, or reported outcome.
PRISMA flowchart showing the search strategy and reasons for exclusion at each stage. Data added to the PRISMA template [from Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med. 6(7):e1000097] under the terms of the Creative Commons Attribution License. Figure is available in color online only.
Individual Participant Data
Data were extracted from 7 articles describing 9 patients who met eligibility criteria (age range 13–42 years). All patients had a diagnosis of ASD among several other comorbidities. The primary indication for DBS was SIB or aggression in 8 patients, whereas the last patient presented with OCD symptoms that progressed to SIB as well. All patients had undergone multiple attempts at pharmacological management prior to consideration of DBS. Patient demographic information can be found in Table 1.
Systematic review of patient demographics, symptoms, and DBS stimulation parameters
| Authors & Year | Age (yrs), Sex | Comorbidities | Behaviors | Indication for DBS | DBS Target & Stimulation Levels |
|---|---|---|---|---|---|
| Sturm et al., 200343 | 13, M | Kanner’s autism, cerebral palsy | Self-aggression | Self-injury | Basolat amygdala (120 µsec, 130 Hz, 2–6.5 V) |
| Stocco & Baizabal-Carvallo, 201438 | 19, F | Mental delay, monosomy 2q & trisomy 20p, autism | Self-picking | Self-injury & tardive dystonia | GPi (120 µsec, 80 Hz, 3.3 V) |
| Stocco & Baizabal-Carvallo, 201438 | 17, M | Mental delay, anxiety, autism | Punching of arms & legs, biting | Self-injury | GPi (120 µsec, 100 Hz, 2.5 V) + ALIC (210 µsec, 100 Hz, 2.0 V) |
| Benedetti-Isaac et al., 201517 | 27, M | TBI, epilepsy, autism | Aggressive behavior toward self | Aggression | pHyp (90 µsec, 185 Hz, 2.7 V) |
| Benedetti-Isaac et al., 201517 | 16, M | Epilepsy, autism, developmental delay | Self-aggression | Aggression | pHyp (90 µsec, 180 Hz, 3.0 V) |
| Segar et al., 201536 | 24, F | Developmental delay, OCD, TS, autism, Kleefstra syndrome | Biting hands, picking skin | OCD, TS | VC/VS (90 µsec, 130 Hz, 8 V) |
| Park et al., 201737 | 13, M | Developmental delay, autism | Self-mutilation, face-hitting causing fractures | Self-injury | NAcc (90 µsec, 130 Hz, 3–5 V) |
| Kakko et al., 201939 | 19, M | Mental retardation, autism, epilepsy, tardive dyskinesia | Aggression, self-mutilation, lacerations | Self-injury & tardive dyskinesia | GPi |
| Doshi et al., 201919 | 42, F | OCD, aggression, autism | Hitting, violent outbursts | OCD & aggression | NAcc (60 µsec, 130 Hz, 2.6 V) |
TBI = traumatic brain injury; TS = Tourette syndrome.
Two of the identified patients received DBS in the NAcc,19,36,37 2 in the globus pallidus internus (GPi),38,39 2 in the pHyp,17 1 in the ventral capsule/ventral striatum (VC/VS),36 1 patient had dual targets in the GPi and anterior limb of the internal capsule (ALIC),38 and 1 patient received DBS in the basolateral amygdala complex.40 All patients found in the systematic review received bilateral DBS. The GPi was chosen to improve both movement and behavioral symptoms in the 2 patients with tardive symptoms related to the pharmacological treatments targeting their SIB.
Clinical Outcomes
The primary indication for DBS among the included patients was refractory SIB or aggression. Several different outcome measurements were used to quantify behavior before and after surgery; the behaviors were measured with a number of different clinical grading scales between 2 months and 10 years. The effectiveness of DBS as measured with these scales is reported in Table 2.
Systematic review of patient treatment histories and perioperative outcomes
| Authors & Year | Age (yrs), Sex | Pharmacological Tx | Alternative Txs | Pre-DBS | Post-DBS (FU period) | Notes |
|---|---|---|---|---|---|---|
| Sturm et al., 200343 | 13, M | Risperidone, topiramate, aripiprazole, lorazepam | Permanent body restraint | Parental score of 6: restraint does not prevent skin lesions & life-threatening self-injury | Parental score of 2: restraint of the wrists suffices & is well tolerated (10 mos) | Battery depletion worsening restraint use to preop levels; battery replaced & restraint use decreased |
| Stocco & Baizabal-Carvallo, 201438 | 19, F | Risperidone | NR | JHMRS 46 | JHMRS 4 (13 mos) | NA |
| Stocco & Baizabal-Carvallo, 201438 | 17, M | Neuroleptics, SSRIs, anxiolytics, tricyclics, tetrabenazine, benzodiazepines | BT | JHMRS 67 | JHMRS 19 (6 mos) | Improvement in initial 3 mos, & regressed back to baseline |
| Benedetti-Isaac et al., 201517 | 27, M | AEDs (carbamazepine, valproic acid, clobazam, clozapine, phenobarbital); quetiapine | NR | OAS 9 | OAS 1 (4 yrs) | Infection, during which initial symptoms of seizure returned |
| Benedetti-Isaac et al., 201517 | 16, M | AEDs (valproic acid, phenytoin, carbamazepine, clonazepam, lorazepam); quetiapine; & clozapine | NR | OAS 8 | OAS 1 (2 mos) | Aggression controlled for 1 mo (OAS of 1), followed by regression to OAS 8, FU 6 yrs |
| Segar et al., 201536 | 24, F | Unnamed medications | BT | GAF 20 | GAF 50–60 (3 yrs) | DBS lead fracture, GAF returned to 20; replacement of lead, GAF 60 |
| Park et al., 201737 | 13, M | Quetiapine, valproate, phenytoin | BT, protective helmet | CGI-S 6; ABC 106; CY-BOCS 22; K-ARS 54; SRS 101 | CGI-S 4; ABC 40; CY-BOCS 7; K-ARS 36; SRS 98 (2 yrs) | Quetiapine weaned off after 12 mos, removal of headgear, SIB 40% improved |
| Kakko et al., 201939 | 19, M | Risperidone, melatonin, chlorprothixene, thioxanthene, valproate, quetiapine, haloperidol, clonazepam, diazepam | NR | NR | NR | Improvement in self-mutilation |
| Doshi et al., 201919 | 42, F | Aripiprazole, quetiapine, haloperidol, guanfacine hydrochloride, lamotrigine, alprazolam, propranolol, vitamin B6, & magnesium | Restraints while hospitalized | Y-BOCS 19, HDS 20, HAS 30, SCQ 26 | Y-BOCS 5, HDS 15, HAS 18, SCQ 16 (12 mos) | NA |
ABC = Antecedent, Behavior, Consequence; AEDs = antiepileptic drugs; BT = behavioral therapy; CGI-S = Clinical Global Impressions–Severity; CY-BOCS = Children’s Y-BOCS; FU = follow-up; GAF = Global Assessment of Functioning; HAS = Hamilton Anxiety Scale; HDS = Hamilton Depression Scale; K-ARS = Korea ADHD Rating Scale; NA = not available; NR = not reported; SCQ = Social Communication Questionnaire; SRS = Social Responsiveness Scale; SSRI = selective serotonin reuptake inhibitor; Tx = treatment; Y-BOCS = Yale-Brown Obsessive-Compulsive Scale.
All patients demonstrated improvement in their clinical symptoms; this improvement was usually sustained but was transient in 2 patients. The first of the patients with transient improvement was a 17-year-old boy, who demonstrated severe body rocking and repetitive punching to his legs and arms. He received DBS to the ALIC and the GPi.38 His symptoms showed improvement on the Johns Hopkins motor stereotypy rating scale (JHMRS)41 after 3 months (from 67 to 19), but reverted to 67 after 6 months with regression to baseline symptoms. The second was a 16-year-old boy with aggression, who received DBS to the pHyp and demonstrated improvement in hostility symptoms for 1 month with an Overt Aggression Scale (OAS) score of 1, but ultimately demonstrated remission of symptoms and return to an OAS of 8 for the last 6 years of follow-up.17,42
Symptoms transiently reoccurred in 2 other patients when they encountered a complication of DBS that interrupted therapy. A 13-year-old boy with ASD, cerebral palsy, and SIB had DBS of the amygdala; his symptoms worsened to preoperative levels with battery depletion but resolved with a new battery.43 A 24-year-old woman with developmental delay, OCD, and Tourette syndrome had biting and picking symptoms that improved with VC/VS DBS, regressed with a lead fracture, and improved again after revision surgery.36
Functional Connectivity
Two patients identified through the systematic review and 6 additional patients identified through our collaborative group had appropriate preoperative and postoperative imaging for connectomics analyses. The clinical data of these patients can be found in Table 3. The 4 patients found following the systematic review had DBS of the posterior medial hypothalamus for aggression.44 The patients in the 2 unpublished cases both had DBS of the pHyp for the treatment of SIB and aggression, measured with the OAS. Table 4 outlines the DBS targets, stimulation parameters, and percent clinical improvement of these patients, with data for connectomics analysis. Clinical improvement was measured as the percentage change between the pre- and postoperative outcome measures. Targeting coordinates for preoperative planning or postoperative confirmed placement are found in Table 5.
Clinical data for patients with ASD and SIB included in connectomic analysis
| Pt ID/ Authors | Age (yrs), Sex | Comorbidities | DBS Target | Pre-DBS | Post-DBS |
|---|---|---|---|---|---|
| A; Segar et al., 201536 | 24, F | Developmental delay, OCD, TS, autism, Kleefstra syndrome | Bilat VC/VS | GAF 20 | GAF 50–60 |
| B; Doshi et al., 201919 | 42, F | OCD, aggression, autism | Bilat NAcc | Y-BOCS 19 | Y-BOCS 5 |
| C; Torres et al., 202144 | 17, M | Developmental delay | Bilat pmHyp | ICAP (general): −43; ICAP (self): −49 | ICAP (general): −18; ICAP (self): −12 |
| D; Torres et al., 202144 | 22, M | Perinatal hypoxia, developmental delay | Rt pmHyp | ICAP (general): −32; ICAP (self): −16 | ICAP (general): −18; ICAP (self): +2 |
| E; Torres et al., 202144 | 23, F | Perinatal hypoxia, developmental delay | Bilat pmHyp | ICAP (general): −48; ICAP (self): −49 | ICAP (general): −24; ICAP (self): −49 |
| F; Torres et al., 202144 | 37, F | Epilepsy, developmental delay | Rt pmHyp | ICAP (general): −44; ICAP (self): −49 | ICAP (general): −3; ICAP (self): −12 |
| G; Benedetti-Isaac (unpub) | 31, M | Developmental delay | Bilat pHyp | OAS 9 | OAS 1 |
| H; Benedetti-Isaac (unpub) | 29, F | Developmental delay | Bilat pHyp | OAS 11 | OAS 9 |
ICAP = Inventory for Client and Agency Planning general aggressiveness score; pmHyp = posterior medial hypothalamus; pt = patient; unpub = unpublished.
Imaging data for patients with ASD and SIB included in connectomic analysis
| Pt ID | Target | MNI Coordinates (x, y, z) | Rt Contacts | Lt Contacts | Pulse Width (µsec) | Frequency (Hz) | Amplitude | Improvement (%) |
|---|---|---|---|---|---|---|---|---|
| A | Bilat VC/VS | Rt (7.4, 1.8, −6.9); lt (−10.1, 6.6, −7.3) | 1(−) c(+) | 1(−) c(+) | 90 | 130 | 6 V | 50 |
| B | Bilat NAcc | Rt (10.8, 6.1, −2.0); lt (−9.2, 6.9, −2.8) | 10(−) 11(+) | 2(−) 3(+) | 60 | 13− | 2.6 V | 74 |
| C | Bilat pmHyp | Rt (1.4, −15.7, −10.8); lt (−3.6, −10.8, −7.4) | 0(−) c(+) | 4(−) c(+) | 210 | 130 | 1.8 V | General: 58; self: 76 |
| D | Rt pmHyp | Rt (4.3, −5.4, −7.3) | 0(−) 3(+) | Not applicable | 90 | 185 | 1.8 V | General: 44; self: 113 |
| E | Bilat pmHyp | Rt (−0.1, −10.8, −8.8); lt (−6.7, −8.5, −3.5) | 0(−) c(+) | 6(−) c(+) | 60 | 185 | 2.5 V | General: 50; self: 73 |
| F | Rt pmHyp | Rt (9.9, −6.7, 9.7) | 2(−) c(+) | Not applicable | 450 | 15 | 1.3 V | General: 93; self: 76 |
| G | Bilat pHyp | Rt (4.3, −12.8, −6.3); lt (−0.3, −15.3, −8.9) | 8(−) c(+) | 0(−) c(+) | 90 | 180 | 3.5 V | 88 |
| H | Bilat pHyp | Rt (0.9, −14.9, −6.6); lt (−0.5, −14.5, −8.5) | 8(−) c(+) | 0(−) c(+) | 90 | 180 | 3.8 V | 18 |
MNI = Montreal Neurological Institute.
Coordinates of targeting method in DBS treatment for patients with ASD and extreme behavior
| Authors & Year | DBS Target | Targeting (in mm) |
|---|---|---|
| Sturm et al., 200343 | Basolat amygdala | NR |
| Stocco & Baizabal-Carvallo, 201438 | GPi | NR |
| Stocco & Baizabal-Carvallo, 201438 | GPi + ALIC | NR |
| Benedetti-Isaac et al., 201517 | pHyp | Plan of MRI-based direct targeting: 2.0 lateral to AC; 3.0 posterior to AC; 5.0 below AC |
| Benedetti-Isaac et al., 201517 | pHyp | Plan of MRI-based direct targeting: 2.0 lateral to AC; 3.0 posterior to AC; 5.0 below AC |
| Segar et al., 201536 | Bilat VC/VS | MRI-based direct targeting w/ resultant electrodes placed: 7.0 lateral to AC; rt electrode at the level of AC & lt electrode 1.0 anterior to AC; 3.0 below AC |
| Park et al., 201737 | NAcc | Plan of MRI-based direct targeting: 6.5 lateral to AC; 2.5 anterior to AC; 4.5 below AC |
| Kakko et al., 201939 | GPi | NR |
| Doshi et al., 201919 | NAcc | MRI-based direct targeting w/ resultant placement of rt & lt electrodes: 6.38 & 6.58 lateral to AC; 3.51 & 4.13 anterior to AC; 4.71 & 3.74 below AC |
| Torres et al., 202144 | pmHyp | Plan of MRI-based direct targeting: 2.0 lateral to wall of 3rd ventricle; 0.0 relative to MCP; 2.0 below MCP |
AC = anterior commissure; MCP = midcommissural point.
The functional connectivities of VTA pairs from different patients overlap in anatomical areas of shared circuitry. Within a high-fidelity brain template,45 Fig. 2 depicts the disparate DBS target locations of the 8 patients. Using functional network mapping, all 8 patients’ VTAs were found to be significantly functionally connected to the ALIC (Fig. 3). Additionally, VTA pairs in 4 of the 8 patients were significantly functionally connected to the pHyp, amygdala, insula, and anterior cingulate. Functional connectivity with distal brain regions could not be weighted according to percent clinical change due to the fact that all patients showed clinical improvement, but this was measured using different scales.
A 3D reconstruction showing the electrode placement for the 8 patients in the connectomic analyses, using a high-resolution, high-fidelity brain template.46 Renderings of the target structures are provided in 3D: NAcc in green27 and pHyp in blue.46 The DBS electrodes are identified in pairs, with the exception of patients D and F, who only received unilateral electrodes on the right. Note the considerable distance between the targets. Figure is available in color online only.
Functional network common to all 8 VTAs. A sum map showing the overlap of the 8 thresholded and binarized functional connectivity maps overlaid on coronal slices of a T1-weighted standard brain (Montreal Neurological Institute [MNI] 152). A higher voxel value in these summed images indicates a greater number of individuals, with thresholded connectivity maps overlapping in that particular area of the brain. Areas common to all 8 functional connectivity maps are shown in pink and areas covered by only 1 connectivity map are shown in light blue. Figure is available in color online only.
Structural Connectivity
There were no common streamlines identified among all 8 patients. Given that the ALIC was identified to be significantly functionally engaged in the VTA of all patients with implanted electrodes, we further probed this connectivity by using structural analyses. We assessed the extent of overlap in the structural connectivity of the ALIC in all participants with a previously published set of streamlines associated with successful treatment of OCD with DBS. The tract unifying the subthalamic nucleus and the ALIC/NAcc targets for patients treated by DBS for OCD20 was overlaid on our functional analysis (Fig. 4). Indeed, the area of perfect overlap of all 8 patients is very similar to the density projection presented by Li et al.,20 suggesting a commonly engaged structural network.
Functional network compared to patients with OCD and optimal DBS response. The summed map of the 8 functional connectivity maps is compared to the streamlines in red, which represent the common tract among patients with OCD who had optimal DBS response, identified by Li et al.20 This comparison is shown in sagittal (A), coronal (B), and axial (C) views. Figure is available in color online only.
Discussion
ASD is a complex neurodevelopmental disorder with heterogeneous symptoms and presentations. One of the most distressing manifestations is repetitive SIB causing physical harm to individuals and caregivers. DBS is an emerging option for these extreme behaviors in circumstances in which behavioral intervention and psychopharmacological treatments are insufficient.
The current study identified 9 patients with ASD reported in 7 studies, who received DBS for the treatment of extreme behaviors that mostly presented as SIB or aggression. The targets were the NAcc, GPi, ALIC, pHyp, VC/VS, and the basolateral amygdala. Through structural and functional network mapping, we show that at least 3 of these targets (NAcc, VC/VS, pHyp) in 4 patients engaged common neural circuitry.
Functional and Structural Connectivity
Connectomic analysis of 4 patients implanted with NAcc, VC/VS, or pHyp DBS electrodes provided insights into a common network underlying these targets for extreme behaviors. By mapping large-scale coordinated interactions, we were able to demonstrate that the 3 DBS target regions modulate common circuitry at the ALIC. Furthermore, some of these targets connect to a number of common distant regions such as the amygdala, insula, and anterior cingulate. Crucially, all of these areas have previously been implicated in the initiation, maintenance, or suppression of aggression and SIB.46–49 By evaluating challenging or provocative auditory, visual, and other sensory stimuli, the amygdala is responsible for the context-specific initiation and modulation of aggressive behavior.46,48 Through extensive connections with the hypothalamus and insular cortex, the hypothalamic-pituitary axis and autonomic brain centers are activated to prepare for an aggressive response.46,48 Furthermore, via reinforcement-based decision-making, the NAcc and the insular cortex, along with the ventromedial, orbitofrontal, and cingulate cortices, are responsible for the modulation and/or suppression of extreme behavior, as seen in patients with ASD.47,48 In acquired lesions leading to SIB, the anterior salience network, notably the anterior cingulate, has been identified as a possible mediator of pathological behavior.50 The current analysis suggests that DBS may exert a neuromodulatory effect on a common network.46,51,52
The functional connectivity of all 8 patients, which identified the ALIC, is in close proximity and follows a similar pathway as the previously published unified tract for DBS in OCD, with fibers originating from the ventrolateral prefrontal cortices terminating in the subthalamic nucleus.20,53,54 Episodes of SIB or aggression may share common neural correlates with compulsive behavior, all demonstrating voluntary and repetitive behaviors.55 Furthermore, extreme behaviors have been associated with a need to escape from demands or undesired sensory stimuli, which is analogous to compulsions, which serve as an escape from distress or anxiety (i.e., obsessions).13,56,57
Limitations
The current study provides a summary of the sparse literature pertaining to DBS as a treatment option for extreme behaviors in patients with ASD. Although only 9 reports of this treatment exist, and only 4 patient images were used in the connectomic analysis, this work presents the first multisite collaborative effect to use a network-based approach to understand the neural circuitry implicated in DBS for SIB and aggression. The selection of targets for imaging analysis was based on availability of high-quality pre- and postoperative MRI. Comparisons between targets could potentially be possible with more data in the future.
We conducted imaging analyses in large normative data sets, which may not be representative of the connectomic profiles of individuals with ASD. Neural networks in patients who demonstrate extreme behaviors may be associated with alterations in structure and function. Future studies in data sets acquired from specific patient populations may inform the findings of the current study. Furthermore, only 8 patients underwent connectomic analysis. The conservative strategy of stringently thresholding and binarizing connectivity masks was, however, optimized to identify robust patterns of shared connectivity among seed regions. It is expected that more widespread adoption of DBS for patients with extreme behaviors will facilitate better understanding of the neural circuitry engaged by stimulation.
Conclusions
This study performed a systematic review of all patients in the literature who have received DBS for extreme behaviors associated with ASD. The individual patient data reported an improvement in SIB and aggression, as measured by several different outcome scales. Furthermore, connectomic analysis in 8 patients revealed a common functional network upon which 3 of the DBS targets converge. Novel indications for DBS are emerging, with many case studies reported in the literature. The current report advances understanding of the neural circuitry engaged by DBS for extreme behaviors in individuals with ASD.
Disclosures
Dr. Fasano is a consultant for Abbott, Ceregate, Ipsen, Medtronic, Boston Scientific, and AbbVie Inc. He has received research support from Medtronic, Boston Scientific, University of Toronto, Michael J. Fox Foundation for Parkinson’s Research, and Dystonia Medical Research Foundation, and honoraria for serving as a speaker from Abbott, UCB, Medtronic, Paladin, Sunovion, Ipsen, Boston Scientific, and AbbVie Inc. Dr. Lozano is a consultant for Medtronic, Abbott, Boston Scientific, and Insightec. He is a scientific director at Functional Neuromodulation. Dr. Bennedetti-Isaac is a proctor for Medtronic.
Author Contributions
Conception and design: Ibrahim. Acquisition of data: Yan, Elkaim, Benedetti-Isaac, Doshi, Torres, Segar, Cosgrove. Analysis and interpretation of data: Yan, Huber, Germann, Loh, Elias, Boutet. Drafting the article: Yan, Elkaim, Venetucci Gouveia, Germann, Loh, Elias, Boutet. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Ibrahim. Statistical analysis: Yan, Germann, Loh. Administrative/technical/material support: Ibrahim, Fasano, Lozano, Kulkarni. Study supervision: Ibrahim.
References
- 2↑
Lobar SL. DSM-V changes for autism spectrum disorder (ASD): implications for diagnosis, management, and care coordination for children with ASDs. J Pediatr Health Care. 2016;30(4):359–365.
- 3↑
Adler BA, Wink LK, Early M, et al. Drug-refractory aggression, self-injurious behavior, and severe tantrums in autism spectrum disorders: a chart review study. Autism. 2015;19(1):102–106.
- 4↑
Raulston TJ, Hansen SG, Machalicek W, McIntyre LL, Carnett A. Interventions for repetitive behavior in young children with autism: a survey of behavioral practices. J Autism Dev Disord. 2019;49(8):3047–3059.
- 5↑
Arron K, Oliver C, Moss J, Berg K, Burbidge C. The prevalence and phenomenology of self-injurious and aggressive behaviour in genetic syndromes. J Intellect Disabil Res. 2011;55(2):109–120.
- 6↑
Steenfeldt-Kristensen C, Jones CA, Richards C. The prevalence of self-injurious behaviour in autism: a meta-analytic study. J Autism Dev Disord. 2020;50(11):3857–3873.
- 7↑
Boser K, Higgins S, Fetherston A, Preissler MA, Gordon B. Semantic fields in low-functioning autism. J Autism Dev Disord. 2002;32(6):563–582.
- 8
Bradley V, Hiersteiner D, Rotholz D, et al. Personal characteristics and outcomes of individuals with developmental disabilities who need support for self-injurious behaviour. J Intellect Disabil Res. 2018;62(12):1043–1057.
- 9↑
Shields MC, Akobirshoev I, Dembo RS, Mitra M. Self-injurious behavior among adults with ASD: hospitalizations, length of stay, and costs of resources to deliver care. Psychiatr Serv. 2019;70(6):457–464.
- 10↑
Malone RP, Waheed A. The role of antipsychotics in the management of behavioural symptoms in children and adolescents with autism. Drugs. 2009;69(5):535–548.
- 11
Sabus A, Feinstein J, Romani P, Goldson E, Blackmer A. Management of self-injurious behaviors in children with neurodevelopmental disorders: a pharmacotherapy overview. Pharmacotherapy. 2019;39(6):645–664.
- 12↑
Antonacci DJ, Manuel C, Davis E. Diagnosis and treatment of aggression in individuals with developmental disabilities. Psychiatr Q. 2008;79(3):225–247.
- 13↑
Iwata BA, Dorsey MF, Slifer KJ, Bauman KE, Richman GS. Toward a functional analysis of self-injury. J Appl Behav Anal. 1994;27(2):197–209.
- 14↑
Fitzpatrick SE, Srivorakiat L, Wink LK, Pedapati EV, Erickson CA. Aggression in autism spectrum disorder: presentation and treatment options. Neuropsychiatr Dis Treat. 2016;12:1525–1538.
- 15↑
Aman MG, De Smedt G, Derivan A, Lyons B, Findling RL. Double-blind, placebo-controlled study of risperidone for the treatment of disruptive behaviors in children with subaverage intelligence. Am J Psychiatry. 2002;159(8):1337–1346.
- 16↑
Dennison A, Lund EM, Brodhead MT, Mejia L, Armenta A, Leal J. Delivering home-supported applied behavior analysis therapies to culturally and linguistically diverse families. Behav Anal Pract. 2019;12(4):887–898.
- 17↑
Benedetti-Isaac JC, Torres-Zambrano M, Vargas-Toscano A, et al. Seizure frequency reduction after posteromedial hypothalamus deep brain stimulation in drug-resistant epilepsy associated with intractable aggressive behavior. Epilepsia. 2015;56(7):1152–1161.
- 18↑
Franzini A, Marras C, Ferroli P, Bugiani O, Broggi G. Stimulation of the posterior hypothalamus for medically intractable impulsive and violent behavior. Stereotact Funct Neurosurg. 2005;83(2-3):63–66.
- 19↑
Doshi PK, Hegde A, Desai A. Nucleus accumbens deep brain stimulation for obsessive-compulsive disorder and aggression in an autistic patient: a case report and hypothesis of the role of nucleus accumbens in autism and comorbid symptoms. World Neurosurg. 2019;125:387–391.
- 20↑
Li N, Baldermann JC, Kibleur A, et al. A unified connectomic target for deep brain stimulation in obsessive-compulsive disorder. Nat Commun. 2020;11(1):3364.
- 21↑
Schlaepfer TE, Bewernick BH, Kayser S, Hurlemann R, Coenen VA. Deep brain stimulation of the human reward system for major depression—rationale, outcomes and outlook. Neuropsychopharmacology. 2014;39(6):1303–1314.
- 22↑
Horn A, Reich M, Vorwerk J, et al. Connectivity predicts deep brain stimulation outcome in Parkinson disease. Ann Neurol. 2017;82(1):67–78.
- 23↑
Moher D, Shamseer L, Clarke M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev. 2015;4:1.
- 24↑
Horn A, Kühn AA. Lead-DBS: a toolbox for deep brain stimulation electrode localizations and visualizations. Neuroimage. 2015;107:127–135.
- 25↑
Friston K, Ashburner J, Kiebel S, Nichols T. Penny W, ed. Statistical Parametric Mapping: The Analysis of Functional Brain Images. Elsevier;2011.
- 26↑
Avants BB, Tustison NJ, Song G, Cook PA, Klein A, Gee JC. A reproducible evaluation of ANTs similarity metric performance in brain image registration. Neuroimage. 2011;54(3):2033–2044.
- 27↑
Ewert S, Plettig P, Li N, et al. Toward defining deep brain stimulation targets in MNI space: a subcortical atlas based on multimodal MRI, histology and structural connectivity. Neuroimage. 2018;170:271–282.
- 28↑
Elias GJB, Boutet A, Joel SE, et al. Probabilistic mapping of deep brain stimulation: insights from 15 years of therapy. Ann Neurol. 2021;89(3):426–443.
- 29
Elias GJB, Giacobbe P, Boutet A, et al. Probing the circuitry of panic with deep brain stimulation: connectomic analysis and review of the literature. Brain Stimul. 2020;13(1):10–14.
- 30↑
Coblentz A, Elias GJB, Boutet A, et al. Mapping efficacious deep brain stimulation for pediatric dystonia. J Neurosurg Pediatr. 2021;27(3):346–356.
- 31↑
Fox MD, Buckner RL, Liu H, Chakravarty MM, Lozano AM, Pascual-Leone A. Resting-state networks link invasive and noninvasive brain stimulation across diverse psychiatric and neurological diseases. Proc Natl Acad Sci U S A. 2014;111(41):E4367–E4375.
- 32↑
Yeh FC, Wedeen VJ, Tseng WY. Generalized q-sampling imaging. IEEE Trans Med Imaging. 2010;29(9):1626–1635.
- 33↑
Glasser MF, Sotiropoulos SN, Wilson JA, et al. The minimal preprocessing pipelines for the Human Connectome Project. Neuroimage. 2013;80:105–124.
- 34↑
Fox MD. Mapping symptoms to brain networks with the human connectome. N Engl J Med. 2018;379(23):2237–2245.
- 35↑
Mansouri AM, Germann J, Boutet A, et al. Identification of neural networks preferentially engaged by epileptogenic mass lesions through lesion network mapping analysis. Sci Rep. 2020;10(1):10989.
- 36↑
Segar DJ, Chodakiewitz YG, Torabi R, Cosgrove GR. Deep brain stimulation for the obsessive-compulsive and Tourette-like symptoms of Kleefstra syndrome. Neurosurg Focus. 2015;38(6):E12.
- 37↑
Park HR, Kim IH, Kang H, et al. Nucleus accumbens deep brain stimulation for a patient with self-injurious behavior and autism spectrum disorder: functional and structural changes of the brain: report of a case and review of literature. Acta Neurochir (Wien). 2017;159(1):137–143.
- 38↑
Stocco A, Baizabal-Carvallo JF. Deep brain stimulation for severe secondary stereotypies. Parkinsonism Relat Disord. 2014;20(9):1035–1036.
- 39↑
Kakko K, Bjelogrlic-Laakso N, Pihlakoski L, Lehtimäki K, Järventausta K. Tardive dyskinesia should not be overlooked. J Child Adolesc Psychopharmacol. 2019;29(1):72–74.
- 40↑
Sturm V, Fricke O, Bührle CP, et al. DBS in the basolateral amygdala improves symptoms of autism and related self-injurious behavior: a case report and hypothesis on the pathogenesis of the disorder. Front Hum Neurosci. 2013;6:341.
- 41↑
Miller JM, Singer HS, Bridges DD, Waranch HR. Behavioral therapy for treatment of stereotypic movements in nonautistic children. J Child Neurol. 2006;21(2):119–125.
- 42↑
Yudofsky SC, Silver JM, Jackson W, Endicott J, Williams D. The Overt Aggression Scale for the objective rating of verbal and physical aggression. Am J Psychiatry. 1986;143(1):35–39.
- 43↑
Sturm V, Lenartz D, Koulousakis A, et al. The nucleus accumbens: a target for deep brain stimulation in obsessive-compulsive- and anxiety-disorders. J Chem Neuroanat. 2003;26(4):293–299.
- 44↑
Torres CV, Blasco G, Navas García M, et al. Deep brain stimulation for aggressiveness: long-term follow-up and tractography study of the stimulated brain areas. J Neurosurg. 2021;134(2):366–375.
- 45↑
Neudorfer C, Germann J, Elias GJB, Gramer R, Boutet A, Lozano AM. A high-resolution in vivo magnetic resonance imaging atlas of the human hypothalamic region. Sci Data. 2020;7(1):305.
- 46↑
Gouveia FV, Hamani C, Fonoff ET, et al. Amygdala and hypothalamus: historical overview with focus on aggression. Neurosurgery. 2019;85(1):11–30.
- 47↑
Blair RJ. The neurobiology of impulsive aggression. J Child Adolesc Psychopharmacol. 2016;26(1):4–9.
- 49↑
Westlund Schreiner M, Klimes-Dougan B, Begnel ED, Cullen KR. Conceptualizing the neurobiology of non-suicidal self-injury from the perspective of the Research Domain Criteria Project. Neurosci Biobehav Rev. 2015;57:381–391.
- 50↑
Yan H, Elkaim L, Loh A, et al. Lesions causing self-injurious behavior engage putative networks modulated by deep brain stimulation. Brain Stimul. 2021;14(2):273–276.
- 51↑
Gouveia FV, Germann J, de Morais R, et al. Longitudinal changes after amygdala surgery for intractable aggressive behavior: clinical, imaging genetics, and deformation-based morphometry study—a case series. Neurosurgery. 2021;88(2):E158–E169.
- 52↑
Gouveia FV, Germann J, Devenyi GA, et al. Refractoriness of aggressive behaviour to pharmacological treatment: cortical thickness analysis in autism spectrum disorder. BJPsych Open. 2020;6(5):e85.
- 53↑
Coenen VA, Schlaepfer TE, Varkuti B, et al. Surgical decision making for deep brain stimulation should not be based on aggregated normative data mining. Brain Stimul. 2019;12(6):1345–1348.
- 54↑
Dougherty DD, Rezai AR, Carpenter LL, et al. A randomized sham-controlled trial of deep brain stimulation of the ventral capsule/ventral striatum for chronic treatment-resistant depression. Biol Psychiatry. 2015;78(4):240–248.
- 55↑
Barnard-Brak L, Rojahn J, Richman DM, Chesnut SR, Wei T. Stereotyped behaviors predicting self-injurious behavior in individuals with intellectual disabilities. Res Dev Disabil. 2015;36C:419-427.
- 56↑
Figee M, Pattij T, Willuhn I, et al. Compulsivity in obsessive-compulsive disorder and addictions. Eur Neuropsychopharmacol. 2016;26(5):856–868.
- 57↑
Starcevic V, Berle D, Brakoulias V, et al. Functions of compulsions in obsessive-compulsive disorder. Aust N Z J Psychiatry. 2011;45(6):449–457.
