If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Common and Distinct Neural Patterns of Attention-Deficit/Hyperactivity Disorder and Borderline Personality Disorder: A Multimodal Functional and Structural Meta-analysis
Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Chengdu, ChinaFunctional and Molecular Imaging Key Laboratory of Sichuan Province, West China Hospital of Sichuan University, Chengdu, ChinaResearch Unit of Psychoradiology, Chinese Academy of Medical Sciences, Chengdu, China
Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Chengdu, ChinaFunctional and Molecular Imaging Key Laboratory of Sichuan Province, West China Hospital of Sichuan University, Chengdu, ChinaResearch Unit of Psychoradiology, Chinese Academy of Medical Sciences, Chengdu, China
Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Chengdu, ChinaFunctional and Molecular Imaging Key Laboratory of Sichuan Province, West China Hospital of Sichuan University, Chengdu, ChinaResearch Unit of Psychoradiology, Chinese Academy of Medical Sciences, Chengdu, China
Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Chengdu, ChinaFunctional and Molecular Imaging Key Laboratory of Sichuan Province, West China Hospital of Sichuan University, Chengdu, ChinaResearch Unit of Psychoradiology, Chinese Academy of Medical Sciences, Chengdu, China
Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Chengdu, ChinaFunctional and Molecular Imaging Key Laboratory of Sichuan Province, West China Hospital of Sichuan University, Chengdu, ChinaResearch Unit of Psychoradiology, Chinese Academy of Medical Sciences, Chengdu, China
Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Chengdu, ChinaFunctional and Molecular Imaging Key Laboratory of Sichuan Province, West China Hospital of Sichuan University, Chengdu, ChinaResearch Unit of Psychoradiology, Chinese Academy of Medical Sciences, Chengdu, China
Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Chengdu, ChinaFunctional and Molecular Imaging Key Laboratory of Sichuan Province, West China Hospital of Sichuan University, Chengdu, ChinaResearch Unit of Psychoradiology, Chinese Academy of Medical Sciences, Chengdu, China
Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Chengdu, ChinaFunctional and Molecular Imaging Key Laboratory of Sichuan Province, West China Hospital of Sichuan University, Chengdu, ChinaResearch Unit of Psychoradiology, Chinese Academy of Medical Sciences, Chengdu, China
Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Chengdu, ChinaFunctional and Molecular Imaging Key Laboratory of Sichuan Province, West China Hospital of Sichuan University, Chengdu, ChinaResearch Unit of Psychoradiology, Chinese Academy of Medical Sciences, Chengdu, China
Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Chengdu, ChinaFunctional and Molecular Imaging Key Laboratory of Sichuan Province, West China Hospital of Sichuan University, Chengdu, ChinaResearch Unit of Psychoradiology, Chinese Academy of Medical Sciences, Chengdu, China
Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Chengdu, ChinaFunctional and Molecular Imaging Key Laboratory of Sichuan Province, West China Hospital of Sichuan University, Chengdu, ChinaResearch Unit of Psychoradiology, Chinese Academy of Medical Sciences, Chengdu, China
Imaging of Mood- and Anxiety-Related Disorders Group, Institut d'Investigacions Biomèdiques August Pi i Sunyer, Centro de Investigación Biomédica en Red de Salud Mental, Barcelona, SpainDepartment of Psychosis Studies, Institute of Psychiatry, Psychology, and Neuroscience, King's College London, London, United Kingdom
Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Chengdu, ChinaDepartment of Psychiatry, University of Cincinnati, Cincinnati, Ohio
Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Chengdu, ChinaDepartment of Radiology, West China Xiamen Hospital of Sichuan University, Xiamen, China
Attention-deficit/hyperactivity disorder (ADHD) and borderline personality disorder (BPD) have partially overlapping symptom profiles and are highly comorbid in adults. However, whether the behavioral similarities correspond to shared neurobiological substrates is not known.
Methods
An overlapping meta-analysis of 58 ADHD and 66 BPD whole-brain articles incorporating observations from 3401 adult patients and 3238 healthy participants was performed by seed-based d mapping. Brain maps were subjected to meta-analytic connectivity modeling and data-driven functional decoding analyses to identify associated neural circuit alterations and relations to behavioral dimensions.
Results
Both groups exhibited hypoactivated abnormalities in the left inferior parietal lobule, and altered clusters of the bilateral superior temporal gyrus were disjunctive in ADHD and BPD. No overlapping structural abnormalities were found. Multimodal alterations of ADHD were located in the right putamen and of BPD in the left orbitofrontal cortex.
Conclusions
The transdiagnostic neural bases of ADHD and BPD in temporoparietal circuitry may underlie overlapping problems of behavioral control, while disorder-specific substrates in frontostriatal circuitry may account for their distinguishing features in motor and emotion domains, respectively.
). Borderline personality disorder (BPD) is also characterized by behavioral impulsivity and affective dysregulation that contributes to interpersonal instability and severe stigma (
Do borderline personality disorder and attention-deficit/hyperactivity disorder co-aggregate in families? A population-based study of 2 million Swedes.
The lifetime impact of attention deficit hyperactivity disorder: Results from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC).
). Behavioral disinhibition has been recognized as a shared functional impairment in ADHD and BPD, in which impulsivity is a prominent psychopathological dimension characterized by premature and ill-considered actions in both conditions (
A neuroimaging meta-analysis of ADHD has suggested that volumetric reduction of the orbitofrontal and temporal cortices and blunted activation of inferior frontal gyrus (IFG), temporal, and striatal substrates have relevance to impairments in executive functions in adulthood (
Comparative meta-analyses of brain structural and functional abnormalities during cognitive control in attention-deficit/hyperactivity disorder and autism spectrum disorder.
). Thus, ADHD and BPD may have shared neural substrates apart from those behavioral similarities, indicating a transdiagnostic phenotype following a common psychopathological pathway (
). Individuals with common neural markers may confer high risks for the comorbid condition, and additional attention is warranted to achieve a more effective treatment response in clinical management (
Borderline personality disorder (BPD) and attention deficit hyperactivity disorder (ADHD) revisited – a review-update on common grounds and subtle distinctions.
Borderline Personal Disord Emot Dysregulation.2021; 8: 22
). Identifying their disjunctive neural mechanisms may potentially facilitate differential diagnosis. Regarding insufficient studies investigating their comorbidities and making comparisons directly (
Inferior frontal white matter microstructure and patterns of psychopathology in women with borderline personality disorder and comorbid attention-deficit hyperactivity disorder.
), a systematic understanding of the conjunctive and disjunctive neural alterations with these disorders remains to be established. Overlapping and comparative meta-analytic approaches offer a promising approach for addressing this issue.
Herein, we performed a multimodal meta-analysis of functional and structural neuroimaging studies to map the common and disorder-specific brain abnormalities in adults with ADHD and BPD (
Structural and functional brain abnormalities in attention-deficit/hyperactivity disorder and obsessive-compulsive disorder: A comparative meta-analysis.
). Obtained neural markers were then subjected to meta-analytic connectivity modeling (MACM) and data-driven functional decoding analyses to identify the associated neural circuit alterations and their relation to behavioral dimensions (
Prior to obtaining any dataset, we preregistered our meta-analytic plan on the Open Science Framework (registration doi: 10.17605/OSF.IO/KFZQN). We performed a comprehensive literature search in PubMed, Web of Science, and Embase for whole-brain functional magnetic resonance imaging (fMRI) or voxel-based morphometry (VBM) studies based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) criteria before March 2, 2021 (
). Studies were included in the meta-analysis if they compared adult patients with ADHD or BPD (ages 18–65 years) with healthy individuals on blood oxygen level–dependent signals or gray matter volume (GMV) (details of the search and article eligibility criteria in the Supplemental Methods).
A total of 124 whole-brain articles were included in the current meta-analysis (ADHD-fMRI, 45 articles [proportion of cognitive/emotion experiments = 31/16]; BPD-fMRI, 52 articles [proportion of cognitive/emotion experiments = 19/33]; ADHD-VBM, 14 articles; BPD-VBM, 14 articles; one article reported parallel anatomic and functional findings and three articles enrolling duplicated samples were excluded in further analysis). These articles incorporated eligible observations from 3401 adult patients (ADHD, 1816; BPD, 1585) and 3238 healthy participants (procedures of literature search and included articles in Figure 1 and Table S1). All included studies were evaluated for quality and limitation to infer the importance of these findings with a 12-point Imaging Methodology Quality Assessment Checklist (
Figure 1Flowcharts of the literature search and selection criteria for articles on attention-deficit/hyperactivity disorder (ADHD) and borderline personality disorder (BPD) in the meta-analysis. Panel (A) shows the literature search for articles on fMRI, and panel (B) for article on the VBM method. DTI, diffusion tensor imaging; fMRI, functional magnetic resonance imaging; ROI, region of interest; rs-fMRI, resting-state functional magnetic resonance imaging; SBM, surface-based morphometry; VBM, voxel-based morphometry.
We extracted peak coordinates for brain functional or structural abnormalities and corresponding t values from those articles. Then, we coded each study with sample size, mean age, and sex ratio, as well as for parameters of scanner (i.e., tesla and slice thickness), statistical approach (i.e., kernel smoothing and multiple corrections), and reported clusters. For patient groups, the proportion receiving psychiatric medication and comorbid conditions were recorded. The task and corresponding Research Domain Criteria construct and domain were labeled for included fMRI studies (details of database construction in the Supplemental Methods and Table S2) (
), indicating that not only one-to-many, but also many-to-one theory reflects the nature of the relationship between task paradigms and neural circuits. Pooling findings across experiments might be an objective approach that facilitates the comprehensive investigation of brain functional abnormalities of clinical populations (
To investigate brain functional and structural abnormalities of individuals with ADHD and BPD, we conducted a multimodal meta-analysis with Seed-based d Mapping with Permutation of Subject Images toolbox (SDM-PSI, version 6.21; https://www.sdmproject.com/). Voxelwise effect-size brain maps were reconstructed accommodating the peak coordinates and corresponding statistical values in the preprocessing (
Structural and functional brain abnormalities in attention-deficit/hyperactivity disorder and obsessive-compulsive disorder: A comparative meta-analysis.
). We created mean effect-size maps across studies and experiments with covarying age and sex in random-effect general linear models separately for modalities and disorders to estimate the meta-analytical case-control difference (
To evaluate the converging neural substrates within and across disorders and imaging modalities, we performed the overlapping analysis following separate disorder- and modality-specific analyses to obtain conjunctive and comparative findings (
). A comparative coordinate-based meta-analytic approach could identify both common and distinct patterns of two disorders, and it has the potential to inform the emergence of diagnostic biomarkers for comorbid conditions (
Structural and functional brain abnormalities in attention-deficit/hyperactivity disorder and obsessive-compulsive disorder: A comparative meta-analysis.
). To avoid the likelihood that the false positive rate is higher than a preferred degree in the worst-case scenario, the SDM adjusts the raw union of probabilities to the desired threshold (
). To threshold the statistical maps of neural abnormalities of ADHD and BPD, we applied the familywise error rate correction with threshold-free cluster enhancement statistics at p < .05 and cluster size >10 voxels to control for multiple comparisons and obtain robust findings (detailed introduction of SDM-PSI and procedures of overlapping analysis in the Supplemental Methods) (
Meta-analytic coactivation analysis examines regions that typically are coactive in prior imaging research with those found to be abnormal in a clinical or psychophysiological study (
), allowing for a practical interpretation of the interaction across the whole brain and adding complementary information to task-related and resting-state functional connectivity studies (
A domain-general brain network underlying emotional and cognitive interference processing: Evidence from coordinate-based and functional connectivity meta-analyses.
). We delineated the coactivated neural circuits of identified peaks by extracting the peak coordinates and integrated findings of articles that reported activation in the seed regions to yield a convergent coactivation brain map in the Neurosynth package (http://www.neurosynth.org) (
). A 5-mm sphere was created for peaks of the identified clusters to search the candidate coactivated brain regions with the false discovery rate correction (p < .01) to reduce the false positive connectomes (
). To characterize neural coactivation at the large-scale network level, we overlaid the coactivation patterns into the default mode network (DMN), central executive network (CEN), dorsal attention network, ventral attention network (VAN), somatomotor network, visual network, and cortical affective network (
) and calculated the similarity of coactivation patterns to these networks (introduction of Neurosynth and details of MACM and related analysis in the Supplemental Methods) (
Functional decoding analysis allows for a data-driven understanding to bridge the gap between psychophysiological functioning and MRI-derived brain alterations (
). We used Neurosynth to decode the pooled neural anomalies, as it currently contains a large-scale neuroimaging database and encodes articles with interactive psychological terms (
). Furthermore, we gathered the correlation coefficients of five behavioral domains (i.e., action, cognition, emotion, interoception, and perception) derived from a paradigm taxonomy and presented with their relative proportions (
) to determine the most prominent behavioral domain related to suprathreshold brain regions (details of functional decoding and behavioral domains in the Supplemental Methods and Table S3).
Ancillary Analyses
We conducted separate disorder- and modality-specific analyses to supplement with overlapping analysis. Meta-regression analysis was employed to examine the potential role of demographic factors and medication status at the whole-brain level (
Comparative meta-analyses of brain structural and functional abnormalities during cognitive control in attention-deficit/hyperactivity disorder and autism spectrum disorder.
). Following the main meta-analyses, we extracted the Hedges’ g imputations for peak coordinates of clusters and performed Spearman’s correlation, Mann-Whitney U, and Kruskal-Wallis H tests to explore the modulatory effects of confounding variables (
). For fMRI studies, we calculated the contribution of included experiments attributed to each Research Domain Criteria domain for identified clusters based on Hedges’ g imputations. Additionally, we conducted cognition and emotion subgroup analyses to investigate specific brain functional profiles. Publication bias and between-study heterogeneity were assessed via funnel plots, Egger’s test, and I2 statistics (details of ancillary analyses in the Supplemental Methods) (
Among 122 eligible studies in the meta-analysis, no statistical difference between patient groups was noted in the sample size-weighted t tests in age (fMRI: t91 = 0.510, p = .611; VBM: t25 = 0.052, p = .959), but BPD studies consisted of a larger proportion of females than ADHD studies in both modalities (fMRI: t91 = 11.027, p < .001; VBM: t25 = 5.621, p < .001). The proportion of medicated patients did not differ by diagnosis (χ21 = 0.020, p = .889) (Table 1).
For the proportion of medicated patients 2 weeks before scanning.
38.2%
–
27.8%
–
Characteristics are reported based on eligible studies excluding three articles with duplicated samples, and one article on ADHD reported parallel anatomic and functional findings.
ADHD, attention-deficit/hyperactivity disorder; BPD, borderline personality disorder; HC, healthy control; fMRI, functional magnetic resonance imaging; sMRI, structural magnetic resonance imaging.
a For the proportion of females in whole sample.
b For the proportion of medicated patients 2 weeks before scanning.
The overlapping analysis indicated that functional abnormalities of the left inferior parietal lobule (IPL) were conjunctive in ADHD and BPD cases (peak coordinates: −40, −50, 52; Z = 4.454; k = 23; shared hypoactivation pattern), and those of the left anterior superior temporal gyrus (STG)/temporal pole (peak coordinates: −48, −2, −14; Z = 3.688; k = 82) and right posterior STG (peak coordinates: 52, −38, 18; Z = 4.255; k = 44) (Table 2 and Figure 2) were disjunctive (reduced response in ADHD but hyperactivation in BPD). The meta-analytic connectivity of the left IPL mainly overlaid on the CEN (% relative distribution [%RD]: 12.13%), the left STG on the DMN (%RD: 15.49%), and the right STG on the VAN (%RD: 10.60%) (Figure 2; Table S5). The convergent findings of the left IPL and right STG were predominantly correlated with the action domain (% correlation coefficient: 47.33% and 27.61%, respectively), while the left STG/temporal pole was linked to the cognition domain (% correlation coefficient: 39.82%) (Figure 2; Table S6). No conjunctive or disjunctive brain structural abnormalities were found between ADHD and BPD, though both structural patterns involved a part of the left orbitofrontal cortex (OFC) that had different specific anatomic locations (Table 3).
Table 2Common and Distinct Brain Abnormalities of ADHD and BPD
Contrast/Brain Region
BA
MNI Coordinates (x, y, z)
Z
p Value
Cluster Size
(ADHD vs. HC) vs. (BPD vs. HC) in fMRI
Conjunction: L inferior parietal lobule
40
−40, −50, 52
4.454
.026
23
Disjunction: L anterior superior temporal gyrus/temporal pole
21
−48, −2, −14
3.688
.001
82
Disjunction: R posterior superior temporal gyrus
42
52, −38, 18
4.255
.002
44
(ADHD vs. HC) vs. (BPD vs. HC) in sMRI
Disjunction or conjunction: None
fMRI vs. sMRI in ADHD
Conjunction: R putamen
48
30, 10, 4
−4.585
.003
145
fMRI vs. sMRI in BPD
Conjunction: L orbitofrontal/anterior cingulate cortex
11
−4, 34, −10
−4.044
.016
56
Suprathreshold clusters were identified at FWE rate correction with pFWE < .05 and cluster size >10 voxels. Breakdown clusters are shown in Table S4.
Figure 2Overlapping functional patterns across attention-deficit/hyperactivity disorder and borderline personality disorder. (A) Transdiagnostic clusters in the left inferior parietal lobule (L. IPL) (conjunctive findings) and bilateral superior temporal gyrus (STG) (left STG [L. STG] and right STG [R. STG]) (disjunctive findings). Statistical brain maps are available online at https://osf.io/jsdgq/files/. (B) Similarity of coactivation pattern to large-scale network. Additional details in Table S5. (C) Contribution of each behavioral domain to each cluster in functional decoding. Additional details in Table S6. AFN, cortical affective network; CEN, central executive network; DAN, dorsal attention network; DMN, default mode network; SMN, somatomotor network; VAN, ventral attention network; VN, visual network.
In ADHD, multimodal conjunction analysis revealed that the cluster in the right putamen (peak coordinates: 30, 10, 4; Z = 4.585; k = 145) exhibited both blunted functional activation and decreased GMV in contrast with control subjects, while those of the left OFC/anterior cingulate cortex (ACC) converged in multimodality in individuals with BPD (peak coordinates: −4, 34, −10; Z = 4.044; k = 56) (Table 2 and Figure 3). When overlaying the findings of MACM analysis, the coactivation pattern of the right putamen was largely located in the VAN and that of the left OFC in the DMN (%RD: 11.79% and 23.15%, respectively) (Figure 3; Table S5). The multimodal abnormality of ADHD was predominantly associated with the action domain, while that of BPD was associated with the perception domain (% correlation coefficient: 48.75% and 36.53%, respectively) (Figure 3; Table S6). No disjunctive findings in multimodality were found in either ADHD or BPD.
Figure 3Divergent multimodal abnormalities of attention-deficit/hyperactivity disorder (ADHD) and borderline personality disorder (BPD). (A) ADHD in multimodal clusters of the right putamen (R. putamen) and BPD in the left orbitofrontal cortex (L. OFC). Statistical brain maps are available online at https://osf.io/jsdgq/files/. (B) Similarity of coactivated pattern to large-scale network. Additional details in Table S5. (C) Contribution of each behavioral domain to each cluster in functional decoding. Additional details in Table S6. AFN, cortical affective network; CEN, central executive network; DAN, dorsal attention network; DMN, default mode network; SMN, somatomotor network; VAN, ventral attention network; VN, visual network.
Separate analyses of ADHD cases alone compared with healthy individuals demonstrated hypoactivation in regions including the bilateral IFG/insula/STG, bilateral IPL, left supplementary motor area, right cerebellum, and left OFC and GMV reduction in the right putamen, left OFC, and right supramarginal gyrus (SMG). Patients with BPD exhibited hyperactivation in the bilateral STG, right precuneus, left superior frontal gyrus, and right parahippocampal gyrus/amygdala; reduced activation in left superior parietal lobule/IPL and ventral ACC; and GMV reduction in the left OFC and left middle occipital gyrus relative to control subjects (Table 3). Coactivation patterns in MACM analysis are shown in Table S5 and Figure S1, as anomalies of ADHD mainly coactivated with the CEN, while those of BPD linked with the DMN. Functional decoding analysis revealed that most of the identified clusters of ADHD were primarily associated with the action domain, while those of BPD were diverse (Table S6 and Figure S2). The top 10 psychophysical terms correlated with those brain abnormalities to infer its main functioning are shown in Table S7, as the left IPL of overlapping analysis predominately correlated with “retention” and “execution,” the left STG correlated with “music” and “speak,” the right STG correlated with “pain” and “listening,” and the right putamen, from distinct multimodal analysis, correlated with “sensation” and “motor” and left OFC with “decision” and “value.”
Meta-regression analysis revealed that sex modulated the neuropathological processing in the right putamen in ADHD and that in the left cerebellum and right ACC in BPD. Additionally, the medication status modulated structural alterations of the right putamen, SMG, and left OFC in ADHD, but no significant clusters were identified in the fMRI modality. In BPD, medication status was associated with structural patterns in the right insula and functional patterns in the left amygdala (details and findings of age in Table S8). The modulatory effects of medication status on brain structural patterns of ADHD and BPD overlapped in the right putamen/insula. Among confounding variables of interests, sex and age played critical roles in identified clusters (20 out of 25 clusters and 7 out of 25 clusters, respectively) (Table S9). The Research Domain Criteria domain of cognitive systems was prevalent in functional profiles of ADHD (all contributions >62%) and in that of negative valence systems in profiles of BPD (all contributions >38%), and the differences of contributions did not reach statistical significance within disorders for any clusters (Table S10). In subgroup analysis for fMRI studies, ADHD samples exhibited hypoactivations in the right cerebellum, bilateral insula, left supplementary motor area, left superior frontal gyrus, right IPL, and right striatum relative to healthy control subjects in the cognition category, while BPD cases showed functional abnormalities in the right parahippocampal gyrus, right STG, right inferior temporal gyrus, and left ACC in the emotion category (Table S11). Interstudy heterogeneity was not significant among all identified clusters according to I2 statistics (all I2 < 10%) (details and publication bias results in Table S12).
Discussion
The current multimodal meta-analysis provides a comprehensive delineation of the similarities and differences in the neural bases of ADHD and BPD in adulthood underlying their shared and distinct clinical features. Following up on overlapping functional patterns centered on the left IPL and bilateral STG, the identified clusters coactivated with the CEN, DMN, and VAN in the large-scale MACM analysis and data-driven functional decoding indicated that these patterns mainly corresponded to the action and cognition domains. Structural abnormalities of both disorders emerged in the left OFC, yet without shared profiles. The multimodal meta-analysis demonstrated that converging brain functional and structural alterations were located in the right putamen in ADHD and in the left OFC in BPD, with linkage to the VAN and DMN and predominant involvement of the action and perception domains, respectively. These findings shed light on the similarities and divergences of brain alterations associated with ADHD and BPD in multimodal dimensions, which may facilitate mechanistic understanding of the two disorders, provide potential information about the basis for their shared behavioral features, and in the longer term, potentially facilitate differential diagnosis and provide interventional targets for treatment discovery programs.
Common Neural Abnormalities in ADHD and BPD
The overlapping functional abnormality in the left IPL may be responsible for the psychopathological model of the comorbid condition of ADHD and BPD in clinical settings. The prior model was constructed based on the view that behavioral disinhibition derived from alterations in the frontostriatal circuitry and attention deficits were triggered by those in frontoparietal systems (
). However, considering symptomatic overlaps of ADHD and BPD, we rephrased the model given that the reduced activation of the left IPL is linked with deficits in attention (
Meta-analysis of functional magnetic resonance imaging studies of inhibition and attention in attention-deficit/hyperactivity disorder: Exploring task-specific, stimulant medication, and age effects.
). Abnormalities of the IPL may indicate the executive network disruption in the neuropathological processing underlying cognitive control and decision making (
). Findings of the left IPL in a transdiagnostic approach may contribute to problems of reduced attentional control and behavioral disinhibition that are associated with both ADHD and BPD (
The overlapping but disjunctive functional patterns in the bilateral STG have relevance to the cognition and action domains, and their coactivated brain regions are largely instantiated in VAN and DMN. A hypoactivated STG/temporal pole in adults with ADHD impinged on cognition as a result of dysfunctional temporal processing (
). As above, although functional abnormalities of ADHD and BPD overlapped in the STG and were anchored to the cognition and action realms, the inverse activation direction suggests diverse psychopathological mechanisms in the two disorders (
Both ADHD and BPD cases exhibited a GMV decrease in the medial OFC that coactivated with the DMN, yielding different locations (the gyrus rectus and part of the ACC, respectively) and functional decoding (interoception and emotion domains). One role of the OFC is to flexibly adjust adaptive behaviors through reward processing to optimize goal-directed action (
). For correlated psychophysical terms in Neurosynth, the OFC cluster of ADHD was mainly associated with “craving” and that of BPD with “value” and “decision.” Thus, while the OFC is impacted in both disorders, somewhat different regions are involved with potential differences in behavioral aspect (
), regarding the fact that neuroimaging researchers selectively and exclusively enroll ADHD and BPD cases to emphasize case-control differences at the expense of the representativeness (
Comparative meta-analyses of brain structural and functional abnormalities during cognitive control in attention-deficit/hyperactivity disorder and autism spectrum disorder.
Distinct Multimodal Abnormalities in ADHD Versus BPD
The divergent brain functional and structural alterations of ADHD and BPD were located in the right putamen and left OFC/ACC, respectively. The connectivity between the OFC and striatum accounts for reinforcement learning and behavioral flexibility (
). Notably, high trait impulsivity, as a transdiagnostic symptom shared by ADHD and BPD, has been associated with decreased functional activation and integrity of myelination in circuitry including the medial OFC/ACC and striatum (
Impulsive personality and the ability to resist immediate reward: An fMRI study examining interindividual differences in the neural mechanisms underlying self-control.
). In the MACM analysis, the coactivated spatial pattern of the right putamen for ADHD cases is closely aligned with the VAN, which partially explains their maladaptive cognitive and emotional processing of external stimuli and dysfunctions of inhibitory control in ADHD (
). The coactivated network of the left OFC for BPD individuals was most similar to DMN, which may account for the associated loss of dynamic control for attention reorienting between internal and external events (
). Functional decoding indicated that the cluster of the putamen was anchored on the action domain and the OFC on the perception domain. Striatal alterations in individuals with ADHD may contribute to their dysfunction in the bottom-up processing of behavioral control regarding its role in the action realm (
Structural and functional brain abnormalities in attention-deficit/hyperactivity disorder and obsessive-compulsive disorder: A comparative meta-analysis.
), while the abnormal OFC pattern in individuals with BPD underlies their processing emotional events, alexithymia, and integrating emotional experience with adaptive action planning—potentially contributing to their increased risk for impulsive self-injury (
). Notably, our disjunctive findings should be interpreted with caution regarding the possibility that differences in task paradigms restricted us from a comprehensive understanding toward neural substrates of ADHD and BPD, and regarding the possibility that those distinct abnormalities would be identified in both disorders as more unbiased evidence emerged.
Disorder-Specific Abnormalities in ADHD and BPD
In functional alteration patterns, individuals with ADHD showed hypoactivation in the bilateral IFG/insula/STG. Psychological functions of the IFG have been generally recognized as a brake on affective processing and behavioral actions (
). Among individuals with ADHD and comorbid BPD, microstructural abnormalities of IFG co-occurring with limbic alterations were linked to dysfunctional emotion regulation and other behavioral features (
Inferior frontal white matter microstructure and patterns of psychopathology in women with borderline personality disorder and comorbid attention-deficit hyperactivity disorder.
). In BPD cases, abnormalities were found in the left superior frontal gyrus, superior parietal lobule, and right parahippocampal gyrus/amygdala. Psychological functions of frontal components correspond to top-down cognitive control and emotion regulation, of parietal ones to behavioral integration and social mentalizing, and of limbic substrates to salience detection and emotion processing (
). In addition, patterns of GMV reduction were also found in brain regions of the right putamen and SMG in ADHD and the left middle occipital gyrus in BPD. Given the predominant linkage with the action domain, the SMG contributes to action reprogramming for rapid adaptation to external alterations and to maintaining the memory flow of serial orders (
). Notably, large-scale mega-analyses revealed that reduced volume of subcortical regions was not found in adults, but rather was found only in children with ADHD compared with typically developing control subjects, suggesting a neuropathological model of altered subcortical trajectories (
Subcortical brain volume differences in participants with attention deficit hyperactivity disorder in children and adults: A cross-sectional mega-analysis.
The meta-regression analysis revealed that modulated brain structural alterations of ADHD and BPD overlapped in the right putamen/insula pertaining to current medication treatment. The putamen is a potential therapeutic target for ADHD given its role in behavioral inhibition and control (
). While individual relation of imaging to particular medications is not possible for a meta-analysis such as ours, these findings highlight the potential value of studying how treatments typically provided to patients with ADHD and patients with BPD may modulate these particular brain regions to reduce behavioral disturbances.
Limitations
First, our findings may not generalize to pediatric or geriatric populations. Because pediatric BPD is not well studied compared with pediatric ADHD, only articles recruiting adult participants were included in the meta-analysis. Second, the representativeness of our transdiagnostic findings might be weakened due to study issues such as differences in sex ratios or differences in the use of emotional and cognitive fMRI task paradigms between the two disorders of interest. Differences in sex ratios of included samples result from different susceptibility of sex populations in community samples (
). Adding sex to covariates in the meta-analysis may reconcile the concern, but further studies accounting for the general clinical presentation of those disorders might be of great interest. Finally, data to permit direct association of brain alterations and behavior at the individual level were not available, so we used data-driven approaches to provide suggestive behavioral linkages with our regional brain findings.
Conclusions
The current meta-analysis, to our knowledge, is the first to investigate the multimodal transdiagnostic neural patterns of ADHD and BPD, and it includes data-driven functional decoding and meta-analytic connectivity approaches. Our study depicts a comprehensive presentation of common and distinct neural bases of ADHD and BPD, which indicates shared abnormal patterns in the temporoparietal profile and distinct patterns in frontostriatal circuitry. These findings provide novel transdiagnostic information regarding two psychiatric conditions that share clinical features of behavioral disinhibition and affective reactivity (
). Understanding of the distinct disrupted neurocircuits with each disorder could help establish better pathophysiological models and facilitate differential diagnosis. Further, the transdiagnostic and disorder-specific neural bases of illness have the potential to serve as therapeutic targets in the development of novel treatments for ADHD and BPD, and to better understand mechanisms for their high comorbidity rates and overlapping symptom profiles (
This work was supported by the National Natural Science Foundation of China (Grant Nos. 81621003 [to QG], 82027808 [to QG], and 31800963 [to SW]), the National Natural Science Foundation (Grant No. 81820108018 [to JAS and QG]), and the Key Research and Development Project of Science and Technology at Department of Sichuan Province (Grant No. 2021YFS0242 [to LL]).
We deeply appreciate all the authors of the included studies who responded to our requests for further information.
JAS is a consultant for VeraSci. All other authors report no biomedical financial interests or potential conflicts of interest.
Do borderline personality disorder and attention-deficit/hyperactivity disorder co-aggregate in families? A population-based study of 2 million Swedes.
The lifetime impact of attention deficit hyperactivity disorder: Results from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC).
Comparative meta-analyses of brain structural and functional abnormalities during cognitive control in attention-deficit/hyperactivity disorder and autism spectrum disorder.
Borderline personality disorder (BPD) and attention deficit hyperactivity disorder (ADHD) revisited – a review-update on common grounds and subtle distinctions.
Borderline Personal Disord Emot Dysregulation.2021; 8: 22
Inferior frontal white matter microstructure and patterns of psychopathology in women with borderline personality disorder and comorbid attention-deficit hyperactivity disorder.
Structural and functional brain abnormalities in attention-deficit/hyperactivity disorder and obsessive-compulsive disorder: A comparative meta-analysis.
A domain-general brain network underlying emotional and cognitive interference processing: Evidence from coordinate-based and functional connectivity meta-analyses.
Meta-analysis of functional magnetic resonance imaging studies of inhibition and attention in attention-deficit/hyperactivity disorder: Exploring task-specific, stimulant medication, and age effects.
Impulsive personality and the ability to resist immediate reward: An fMRI study examining interindividual differences in the neural mechanisms underlying self-control.
Subcortical brain volume differences in participants with attention deficit hyperactivity disorder in children and adults: A cross-sectional mega-analysis.
00147-1&id=gr1.jpg){kind=link}
00147-1&id=gr2.jpg){kind=link}
00147-1&id=gr3.jpg){kind=link}