Gray matter volume associations in youth with ADHD features of inattention and hyperactivity/impulsivity

Abstract

Background

Prior research has shown smaller cortical and subcortical gray matter volumes among individuals with attention‐deficit/hyperactivity disorder (ADHD). However, neuroimaging studies often do not differentiate between inattention and hyperactivity/impulsivity, which are distinct core features of ADHD. The present study uses an approach to disentangle overlapping variance to examine the neurostructural heterogeneity of inattention and hyperactivity/impulsivity dimensions.

Methods

We analyzed data from 10,692 9‐ to 10‐year‐old children from the Adolescent Brain Cognitive Development (ABCD) Study. Confirmatory factor analysis was used to derive factors representing inattentive and hyperactive/impulsive traits. We employed structural equation modeling to examine these factors' associations with gray matter volume while controlling for the shared variance between factors.

Results

Greater endorsement of inattentive traits was associated with smaller bilateral caudal anterior cingulate and left parahippocampal volumes. Greater endorsement of hyperactivity/impulsivity traits was associated with smaller bilateral caudate and left parahippocampal volumes. The results were similar when accounting for socioeconomic status, medication, and in‐scanner motion. The magnitude of these findings increased when accounting for overall volume and intracranial volume, supporting a focal effect in our results.

Conclusions

Inattentive and hyperactivity/impulsivity traits show common volume deficits in regions associated with visuospatial processing and memory while at the same time showing dissociable differences, with inattention showing differences in areas associated with attention and emotion regulation and hyperactivity/impulsivity associated with volume differences in motor activity regions. Uncovering such biological underpinnings within the broader disorder of ADHD allows us to refine our understanding of ADHD presentations.

The present study examines the neurostructural associations with ADHD core features—inattention and hyperactivity/impulsivity—represented as continuous factors. Findings reveal that inattention and hyperactivity/impulsivity display both shared and dissociable associations with smaller gray matter volumes in several regions.

Practitioner Points.

• The present study examines the neurostructural associations with ADHD core features—inattention and hyperactivity/impulsivity—represented as continuous factors.

• Findings reveal that inattention and hyperactivity/impulsivity display both shared and dissociable associations with smaller gray matter volumes in several regions.

• Our primary findings highlight neural specificity of volume deficits in core ADHD features and could refine our understanding of ADHD presentations.

1. INTRODUCTION

Attention‐deficit/hyperactivity disorder (ADHD) is one of the most common neurodevelopmental conditions in youth, with an estimated lifetime prevalence of 9.8% in children ages 3–17 years (Bitsko et al., 2022). The core features of ADHD include atypical levels of inattention, hyperactivity, and impulsivity (Regier et al., 2013). In addition to the diminished psychosocial, academic, and occupational well‐being that accompanies endorsement of these core features, estimates suggest that U.S. societal costs of ADHD are greater than 124 billion dollars (Erskine et al., 2016; Zhao et al., 2019). Economic and humanistic burdens underscore the importance of clarifying the neurobiological etiology of ADHD and its behavioral domains of inattention and hyperactivity/impulsivity.

Prior neuroimaging studies have reported diverging neural characteristics associated with ADHD (Durham et al., 2021; Reimann et al., 2022). Empirical studies and meta‐analyses have linked ADHD to both global and focal differences in gray matter volumes, including smaller volumes in cortical areas (e.g., temporal lobe, frontal lobe, precentral gyrus, postcentral gyrus, anterior cingulate cortex) and subcortical regions (e.g., caudate nucleus, putamen) in association with ADHD (Durham et al., 2021; Frodl & Skokauskas, 2012; Greven et al., 2015; Hoogman et al., 2017, 2019; Nakao et al., 2011; Valera et al., 2007). Not only have studies reported brain volume group differences between those with ADHD and typically developing individuals, but these regions have been inversely associated with the severity of ADHD symptoms as well (Castellanos et al., 2002).

Many studies have investigated the neuroanatomy of ADHD; however, these studies commonly treat ADHD as a homogeneous disorder without differentiating between its presentations—predominantly inattentive, predominantly hyperactive/impulsive, and a combined presentation. Examining inattention and hyperactivity/impulsivity together may clarify the neurobiology of ADHD's combined presentation, which is characterized by the presence of both core features (Regier et al., 2013). Yet this approach neglects that inattention and hyperactivity/impulsivity are considered distinct manifestations of the disorder in the Diagnostic and Statistical Manual of Mental Disorders, fifth edition, text revision (DSM‐5‐TR) and may have dissociable structural distinctions. Therefore, it is important to investigate whether cortical and subcortical gray matter volume differences represent a broad characteristic of ADHD or whether specific ADHD core features drive these structural differences.

The present study sought to parse the neurobiological heterogeneity of ADHD to examine mechanisms specific to the disorder's core features of inattention and hyperactivity/impulsivity using a large sample of children from the Adolescent Brain Cognitive Development^SM Study (ABCD Study®). We operationalized inattentive and hyperactive/impulsive traits as two separate continuous factors using a confirmatory factor analysis (CFA) on item‐level data from parent ratings of child behavior and emotions. We then related these factors to cortical and subcortical gray matter volume to analyze whether inattentive and hyperactive/impulsive features show distinct neurostructural properties. We hypothesized that the inattention and hyperactivity/impulsivity dimensions would show smaller gray matter volumes in regions relevant to these traits. In particular, we hypothesized that greater endorsement of hyperactivity/impulsivity traits would be associated with smaller volumes in movement‐related regions, such as the precentral gyrus, caudate, and putamen. We hypothesized that greater inattention features would be associated with smaller volumes in attention‐ and planning‐related regions, such as the anterior cingulate cortex and middle frontal gyrus.

2. MATERIALS AND METHODS

2.1. Participants

The present study used data from the ABCD Study Wave 1 (release 4.0). The ABCD Study obtained consent from all participants, and Vanderbilt University's Institutional Review Board approved the use of this deidentified dataset. Participants included 11,876 9‐ and 10‐year‐old children recruited at 21 sites across the United States. In the present study, we excluded participants based on missing data and failed neuroimaging quality assurance measures, resulting in a final sample size of 10,692. Specifics regarding the methods used for imaging data exclusions are described in the imaging acquisition, processing, and quality assurance section and in prior publications (Hagler et al., 2019). A summary of the demographic characteristics of the analyzed sample can be found in Table 1.

TABLE 1.

Demographics of the sample.

Variable	Mean	SD
Age (in years)	9.92	0.62
	N	(%)
Sex
Male	5562	52.0
Female	5137	48.0
Race/Ethnicity
White	5615	52.5
Hispanic	2197	20.5
Black	1551	14.5
Other	1334	12.5
Household Income
<$5000	354	3.3
$5000–$11,999	376	3.5
$12,000–$15,999	254	2.4
$16,000–$24,999	465	4.3
$25,000–$34,999	584	5.5
$35,000–$49,999	827	7.7
$50,000–$74,999	1350	12.6
$75,000–$99,999	1445	13.5
$100,000–$199,999	3006	28.1
≥$200,000	1134	10.6
Missing	904	8.5
Parent Education
No degree	541	5.0
High school/GED	1285	12.0
Some College	1750	16.4
Associate degree	1387	13.0
Bachelor's degree	3041	28.4
Master's degree	2056	19.2
Professional/Doctoral	639	6.0

Note: The “Other” Race/Ethnicity category includes those who were identified by their parent as American Indian/Native American, Alaska Native, Native Hawaiian, Guamanian, Samoan, Other Pacific Islander, Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, Other Asian, or Other Race.

Abbreviations: GED, General education development; SD, Standard deviation.

2.2. ABCD sample representativeness

Previous literature details the ABCD Study participant recruitment process (Garavan et al., 2018). In sum, each of the 21 ABCD Study‐designated sites across the United States had independent catchment areas. Within these catchment areas, researchers engaged in probability sampling of schools to recruit eligible children. Sociodemographic factors—including age, gender, race/ethnicity, socio‐economic status, and urbanicity—were considered in sample recruitment. Target numbers for representativeness in each of these factors were determined by the following sources: (1) the American Community Survey (ACS), a large annual survey by the U.S. Census Bureau, and (2) the National Center for Education Statistics' school enrollment data. Each site implemented the same unbiased recruitment. Additionally, post‐stratification weights were derived by the researchers of the ABCD Study and the present analyses utilized these weights to adjust the sample to be more representative of the U.S. population.

2.3. Family dependencies and nesting within site

The ABCD Study includes multiple children from the same family, such as siblings, twins, and triplets. Because the ABCD Study sample was collected at 21 sites across the U.S., families are nested within each site. To account for participant relatedness and nesting, we clustered by family ID using the CLUST command, and stratified by site using the STRAT command in Mplus. These approaches are recommended when working with complex survey data in the SEM framework in Mplus (Muthén & Muthén, 2017).

2.4. Measure of inattention and hyperactivity/impulsivity

The Child Behavior Checklist (CBCL) was used to assess symptoms of inattention and hyperactivity/impulsivity via a broad measure of parent‐reported emotional and behavioral problems (Achenbach & Rescorla, 2009). The CBCL is normed for children ages 6–18 and consists of 119 items (112 prompts, some with multiple items within a single prompt) related to various emotions and behaviors. Items are rated on a 3‐point scale: 0 = not true (as far as you know), 1 = somewhat or sometimes true, and 2 = very true or often true. Eighteen items reflecting ADHD symptoms were used to derive factors of inattention and hyperactivity/impulsivity.

2.5. Imaging acquisition

The ABCD Data Analysis and Informatics Center (DAIC) and the ABCD Imaging Acquisition Workgroup developed the imaging protocol used in the present study. These groups harmonized the imaging protocol for all scanner platforms. Imaging data was acquired across 21 sites using the following 3 tesla (3 T) scanner models: Siemens Prisma, Siemens Prisma Fit, General Electric Discovery MR750, Philips Achieva dStream, and Philips Ingenia. Imaging data collection occurred across one to two sessions. Further detail on these methods is documented elsewhere (Casey et al., 2018; Hagler et al., 2019). Each session included T1‐ and T2‐weighted images of brain structure. Details on the imaging parameters include the following: TR (repetition time) 2400–2500 ms; TE (echo time) 2–2.9 ms; FOV (field of view) 256 × 240 to 256; FOV phase of 93.75%–100%; matrix 256 × 256; 176–225 slices; TI (inversion delay) 1060 ms; flip angle of 8°; voxel resolution of 1 × 1 × 1 mm; total acquisition time from 5 min 38 s to 7 min 12 s.

2.6. Imaging processing and quality assurance

Previous work describes the processing and quality control of ABCD Study neuroimaging data (Casey et al., 2018; Hagler et al., 2019). Briefly, DAIC performed centralized processing and analysis of the structural data using the Multi‐Modal Processing Stream (MMPS) software; this software was developed and maintained at the University of California, San Diego's Center for Multimodal Imaging and Genetic (CMIG). Using this package, preprocessing included gradient nonlinearity distortion correction, intensity scaling and inhomogeneity correction, registration to an averaged reference brain in standard space, and manual quality control. Additionally, brain segmentation was performed using cortical surface reconstruction and subcortical segmentation performed based on automated, atlas‐based, segmentation procedures in FreeSurfer v5.3. Derivation of morphometric measures was conducted via volume calculation in each cortical parcel of the standard Desikan parcellation scheme and in each subcortical region via a subcortical segmentation procedure in FreeSurfer (Desikan et al., 2006; Fischl et al., 2002). Lastly, post‐processing quality control included manual review by trained technicians for motion artifacts, intensity inhomogeneity, white matter underestimation, pial overestimation, and magnetic susceptibility artifact.

2.7. Confirmatory factor analysis of inattention and hyperactivity/impulsivity traits

All analyses were conducted using Mplus version 8.4 (Muthén & Muthén, 2017). A correlated traits confirmatory factor analysis (CFA) was conducted in Mplus to evaluate continuous latent factors of inattention and hyperactivity/impulsivity (N = 11,876). CBCL items were selected for their relevance to the DSM‐5‐TR criteria for inattentive and hyperactive/impulsive presentations (e.g., inattentive or easily distracted, talks too much; Table 2). Hyperactivity and impulsivity items were grouped together given the junction of these features under the same behavioral domain in the DSM‐5‐TR (Regier et al., 2013). Items were retained if they had sufficient endorsement for each response option and if the correlation with other items did not exceed .87. In the case that model fit was inadequate, R's ICLUST function was used to assess for redundant CBCL items. We assessed the adequacy of model fit using the following widely accepted criteria: a Comparative Fit Index (CFI) ≥ .9, root mean square error of approximation (RMSEA) ≤ .06, and standardized root mean square residual (SRMR) ≤ .08 (Jackson et al., 2009; McDonald & Ho, 2002).

TABLE 2.

Child behavior checklist items contributing to the confirmatory factor analysis model.

Inattentive items	Hyperactive/impulsive items
Fails to finish things Inattentive or easily distracted Daydreams or gets lost in thoughts Poor schoolwork Stares blankly Confused or seems to be in a fog* Can't concentrate/pay attention for long*	Acts too young for age Can't sit still, restless, or hyperactive Disobedient at home Disobedient at school Gets hurt a lot, accident prone Impulsive or acts without thinking Nervous movements or twitching Unusually loud Showing off or clowning Talks too much Poorly coordinated or clumsy*

Note: *Indicates items that we attempted to include in the initial model but were removed due to redundancy with other items.

2.8. Statistical analysis

We employed structural equation modeling to examine the associations between gray matter volumes (68 cortical regions and 19 subcortical regions) and inattentive and hyperactive/impulsive factors obtained from the aforementioned CFA model. We included age, sex, race/ethnicity, and MRI scanner model as covariates. As mentioned previously, the data were weighted by the post‐stratification weights provided by the ABCD Study, stratified based on site, and clustered based on family membership. In addition, our prior work has shown that those who are excluded for failing to meet imaging quality metrics differ significantly from those who are included based on age, sex, race/ethnicity, and measures of socioeconomic status (Durham et al., 2021). Therefore, we generated and applied non‐participation weights to account for the non‐random nature of the exclusions. Given the overlapping variance between the inattention and hyperactivity/impulsivity factors (r = .95), we regressed out the effects of one from the other—inattentive factor variance regressed out of hyperactive/impulsive variance, and vice versa—to examine the unique association of either the inattention or hyperactivity/impulsivity factor with gray matter volume. Using this approach, our structural equation model ran a series of three analyses as follows:

1. Inattention = Hyperactivity/impulsivity (remove the effect of hyperactivity/impulsivity from inattention)

2. GMV = age + sex + race/ethnicity + MRI scanner model (remove the effects of standard covariates from GMV regions)

3. Inattention = GMV (test the unique association between inattention and GMV)

We then ran the same model with inattention and hyperactivity/impulsivity flipped to test the unique contributions of hyperactivity/impulsivity to GMV. We controlled for the false discovery rate (q < 0.05) in R version 3.6.1's ‘stats' package (http://www.r-project.org/). All cortical regions were tested in separate models, and false discovery rate corrections were applied across the 68 regions. The same procedure was followed for the 19 subcortical regions.

2.9. Sensitivity analyses

Additionally, we performed sensitivity analyses to determine whether associations between gray matter volume and inattention or hyperactivity/impulsivity persisted after accounting for the additional covariates of family income and current medication use, as well as in‐scanner motion. Separate sensitivity analyses were run to account for total intracranial volume (ICV) and total gray matter volume (total cortical gray matter volume and total subcortical gray matter volume for cortical and subcortical analyses, respectively; Table 5).

TABLE 5.

Associations between inattention and hyperactivity/impulsivity symptoms and gray matter volume of cortical and subcortical regions in four separate sensitivity analyses.

		Inattention			Hyperactivity/impulsivity
Brain region by sensitivity analysis		β	p fdr	R 2	β	p fdr	R 2
1. Income and medication use	Left caudal anterior cingulate	−0.058	<.001	0.04	‐	‐	‐
	Right caudal anterior cingulate	−0.046	<.001	0.04	‐	‐	‐
	Left parahippocampal	−0.045	<.001	0.06	−0.050	<.001	0.06
	Left caudate	‐	‐	‐	−0.062	<.001	0.08
	Right caudate	‐	‐	‐	−0.055	<.001	0.08
2. Total GMV	Left caudal anterior cingulatea	−0.081	<.001	0.19	‐	‐	‐
	Right caudal anterior cingulatea	−0.080	<.001	0.18	‐	‐	‐
	Left parahippocampala	−0.076	<.001	0.18	−0.093	<.001	0.18
	Left caudateb	‐	‐	‐	−0.117	<.001	0.32
	Right caudateb	‐	‐	‐	−0.112	<.001	0.34
3. Total ICV	Left caudal anterior cingulate	−0.078	<.001	0.15	‐	‐	‐
	Right caudal anterior cingulate	−0.077	<.001	0.15	‐	‐	‐
	Left parahippocampal	−0.077	<.001	0.17	−0.092	<.001	0.17
	Left caudate	‐	‐	‐	−0.119	<.001	0.33
	Right caudate	‐	‐	‐	−0.114	<.001	0.35
4. In‐scanner motion	Left caudal anterior cingulate	−0.045	.001	0.03	‐	‐	‐
	Right caudal anterior cingulate	−0.047	<.001	0.04	‐	‐	‐
	Left parahippocampal	−0.044	.001	0.05	−0.043	.001	0.05
	Left caudate	‐	‐	‐	−0.048	<.001	0.07
	Right caudate	‐	‐	‐	−0.044	.001	0.07

Note: Sensitivity analyses were performed only on regions found to be significant in the primary analyses. All sensitivity models controlled for age, sex, race/ethnicity, and MRI scanner model and added the additional covariates of (1) income and medication, (2) total gray matter volume (^atotal cortical gray matter volume; ^btotal subcortical gray matter volume), (3) total intracranial volume, or (4) in‐scanner motion.

Abbreviations: GMV, Gray matter volume; ICV, Intracranial volume.

2.10. Data and code availability

This study's data is from the ABCD Study. It is available through the National Institute of Mental Health Data Archive (https://nda.nih.gov/abcd). Analysis scripts and a corresponding wiki can be found at https://github.com/VU-BRAINS-lab/Reimann_ADHD_Volume.

3. RESULTS

3.1. A correlated traits confirmatory factor analysis defines inattention and hyperactivity/impulsivity factors

Prior to the CFA, it was determined that all 18 ADHD‐related CBCL items had sufficient endorsement across response options, and correlations among the items did not exceed .87. The initial CFA with all 18 items (7 items defining the inattention factor and 11 items defining the hyperactivity/impulsivity factor) did not achieve adequate fit (CFI = .87; RMSEA = .05; SRMR = .05). R's ICLUST function was used to identify variables forming composites. Three pairs of items were found to be redundant; (1) “Can't concentrate/pay attention for long” and “Inattentive or easily distracted,” (2) “Confused or seems to be in a fog” and “Stares blankly,” and (3) “Gets hurt a lot/accident prone” and “Poorly coordinated or clumsy.” The item within a redundant pair with greater variability in responses was retained in the model. As a result, three items were removed (Can't concentrate/pay attention for long, Confused or seems to be in a fog, Poorly coordinated or clumsy). The final CFA model included 15 items (5 items defining the inattention factor and 10 items defining the hyperactivity/impulsivity factor) and achieved adequate fit (CFI = .91; RMSEA = .04; SRMR = .04). Table 2 shows the CBCL items that contributed to the initial and final CFA models for each behavioral domain.

3.2. Inattention shows smaller volumes in distinct cortical structures

After extracting the variance associated with hyperactivity/impulsivity, the inattentive factor was inversely associated with left ($\beta$ = −0.045, p _fdr = .029) and right ($\beta$ = −0.046, p _fdr < .001) caudal anterior cingulate gray matter volumes (Table 3). Furthermore, inattention was inversely associated with left parahippocampal gray matter volume ($\beta$ = −0.044, p _fdr = 0.029). No other cortical regions showed significant associations with inattention. No significant associations were found for the inattention factor and any of the subcortical regions (Table 4; Figure 1).

TABLE 3.

Associations between inattention and hyperactivity/impulsivity symptoms and gray matter volume of 68 cortical regions.

Brain region	Inattention			Hyperactivity/impulsivity
	β	p fdr	R 2	β	p fdr	R 2
Left banks of superior temporal sulcus	−0.004	.923	0.07	0.006	.776	0.07
Left caudal anterior cingulate	−0.045	.029	0.03	−0.024	.364	0.03
Left caudal middle frontal	0.005	.908	0.10	0.004	.904	0.10
Left cuneus	0.014	.653	0.12	0.015	.526	0.12
Left entorhinal	−0.010	.736	0.11	0.023	.364	0.11
Left frontal pole	0.014	.653	0.11	0.015	.526	0.11
Left fusiform	−0.016	.626	0.18	−0.007	.770	0.18
Left inferior parietal	−0.013	.653	0.10	0.002	.956	0.10
Left inferior temporal	0.016	.626	0.21	0.022	.388	0.21
Left insula	−0.003	.946	0.18	0.011	.677	0.18
Left isthmus cingulate	0.011	.736	0.15	0.014	.526	0.11
Left lateral occipital	0.007	.906	0.22	0.015	.526	0.23
Left lateral orbitofrontal	−0.013	.684	0.19	−0.003	.948	0.19
Left lingual	−0.020	.626	0.11	−0.014	.536	0.11
Left medial orbitofrontal	0.006	.908	0.23	0.029	.324	0.24
Left middle temporal	0.024	.394	0.19	0.010	.708	0.19
Left paracentral	−0.036	.139	0.08	−0.031	.236	0.08
Left parahippocampal	−0.044	.029	0.05	−0.043	.029	0.05
Left pars opercularis	0.001	.956	0.08	0.010	.680	0.08
Left pars orbitalis	−0.003	.923	0.13	0.007	.770	0.13
Left pars triangularis	0.003	.923	0.08	0.021	.388	0.08
Left pericalcarine	−0.004	.923	0.04	−0.008	.744	0.04
Left postcentral	−0.017	.626	0.16	−0.017	.503	0.15
Left posterior cingulate	−0.010	.736	0.10	−0.016	.522	0.10
Left precentral	−0.022	.532	0.19	−0.020	.397	0.18
Left precuneus	−0.010	.809	0.19	0.000	.980	0.19
Left rostral anterior cingulate	−0.032	.218	0.08	−0.001	.984	0.09
Left rostral middle frontal	−0.005	.923	0.17	−0.002	.948	0.17
Left superior frontal	−0.013	.698	0.17	−0.009	.732	0.17
Left superior parietal	−0.011	.736	0.15	0.001	.985	0.15
Left superior temporal	−0.014	.653	0.14	−0.020	.397	0.13
Left supramarginal	−0.019	.626	0.15	−0.020	.397	0.15
Left temporal pole	−0.006	.908	0.10	0.008	.770	0.10
Left transverse temporal	−0.016	.626	0.07	−0.019	.397	0.07
Right banks of superior temporal sulcus	0.008	.833	0.10	0.016	.522	0.10
Right caudal anterior cingulate	−0.046	<.001	0.04	−0.034	.145	0.03
Right caudal middle frontal	−0.019	.626	0.09	−0.016	.522	0.10
Right cuneus	0.017	.626	0.13	0.019	.400	0.13
Right entorhinal	−0.002	.956	0.09	0.000	.995	0.09
Right frontal pole	−0.016	.626	0.12	−0.010	.685	0.12
Right fusiform	0.004	.923	0.20	0.001	.984	0.20
Right inferior parietal	0.005	.908	0.18	0.020	.397	0.18
Right inferior temporal	−0.001	.956	0.21	0.016	.522	0.21
Right insula	−0.002	.956	0.20	0.013	.607	0.20
Right isthmus cingulate	0.012	.728	0.11	0.012	.622	0.15
Right lateral occipital	0.016	.626	0.26	0.024	.364	0.26
Right lateral orbitofrontal	−0.006	.908	0.21	−0.002	.948	0.21
Right lingual	−0.012	.728	0.09	−0.009	.729	0.09
Right medial orbitofrontal	−0.008	.833	0.15	0.015	.526	0.15
Right middle temporal	−0.008	.881	0.22	−0.012	.622	0.21
Right paracentral	−0.040	.087	0.08	−0.024	.364	0.08
Right parahippocampal	−0.032	.199	0.05	−0.037	.104	0.05
Right pars opercularis	0.017	.626	0.08	0.009	.715	0.08
Right pars orbitalis	−0.003	.923	0.13	0.001	.984	0.11
Right pars triangularis	0.011	.736	0.08	0.026	.329	0.08
Right pericalcarine	−0.005	.908	0.05	−0.007	.770	0.04
Right postcentral	−0.018	.626	0.13	−0.012	.607	0.13
Right posterior cingulate	−0.028	.274	0.10	−0.023	.364	0.09
Right precentral	−0.025	.370	0.17	−0.028	.324	0.16
Right precuneus	0.006	.908	0.21	0.016	.522	0.21
Right rostral anterior cingulate	−0.029	.245	0.07	−0.024	.364	0.07
Right rostral middle frontal	0.001	.956	0.16	0.006	.814	0.16
Right superior frontal	−0.016	.626	0.17	−0.007	.770	0.17
Right superior parietal	0.005	.908	0.17	0.020	.397	0.17
Right superior temporal	−0.019	.626	0.12	−0.020	.397	0.12
Right supramarginal	−0.019	.626	0.10	−0.019	.397	0.10
Right temporal pole	−0.003	.923	0.06	0.012	.622	0.06
Right transverse temporal	−0.014	.653	0.09	−0.022	.364	0.09

Note: Analyses controlled for age, sex, race/ethnicity, and MRI scanner model.

TABLE 4.

Associations between inattention and hyperactivity/impulsivity symptoms and gray matter volume of 19 subcortical regions.

Brain region	Inattention			Hyperactivity/impulsivity
	β	p fdr	R 2	β	p fdr	R 2
Left accumbens	−0.030	.245	0.15	0.008	.759	0.15
Left amygdala	0.015	.653	0.26	0.029	.329	0.26
Left caudate	−0.027	.280	0.08	−0.048	<.001	0.07
Left cerebellum	0.015	.653	0.27	0.030	.324	0.27
Left hippocampus	−0.009	.833	0.18	−0.005	.874	0.18
Left pallidum	0.000	.977	0.17	0.029	.324	0.17
Left putamen	−0.018	.626	0.14	0.009	.744	0.14
Left thalamus	−0.003	.923	0.18	−0.007	.781	0.18
Left ventral diencephalon	−0.007	.908	0.20	0.015	.526	0.20
Brain stem	−0.005	.908	0.20	0.005	.868	0.21
Right accumbens	−0.030	.245	0.09	−0.014	.526	0.09
Right amygdala	0.011	.736	0.19	0.020	.397	0.19
Right caudate	−0.034	.160	0.08	−0.044	.029	0.07
Right cerebellum	0.027	.319	0.28	0.040	.104	0.29
Right hippocampus	−0.001	.969	0.17	−0.002	.948	0.17
Right pallidum	0.002	.956	0.12	0.022	.364	0.13
Right putamen	−0.028	.274	0.14	0.003	.935	0.15
Right thalamus	0.018	.626	0.19	0.020	.397	0.20
Right ventral diencephalon	−0.001	.956	0.20	0.026	.329	0.21

Note: Analyses controlled for age, sex, race/ethnicity, and MRI scanner model.

FIGURE 1.

Cortical and subcortical gray matter volume (GMV) regions associated with inattention and hyperactivity/impulsivity factors. Blue indicates negative associations (more psychopathology, smaller GMV). No positive associations were found. L, Left; R, Right.

When covarying for income and medication use, these associations were not only retained but also became stronger for the left and right caudal anterior cingulate as well as the left parahippocampal gyrus (see Table 5 for all sensitivity analyses performed). In addition, we tested the associations between inattention endorsement and regional volume while accounting for total gray matter volume. The associations also increased in strength for all three regions based on the standardized estimates. As a variation of this, we also examined the effects after controlling for total intracranial volume to account for differences in head size and the results also increased in strength, suggesting that when global differences in volume or head size are controlled, there is a strong focal effect for inattention in these regions. Finally, we examined these associations while accounting for in‐scanner motion; results were retained for all associations (Table 5).

3.3. Hyperactivity/impulsivity shows common GMV associations with inattention as well as distinct associations

We then performed the same analyses, but this time, we extracted the variance associated with inattention from the hyperactivity/impulsivity factor to examine the unique associations between hyperactivity/impulsivity and gray matter volumes. The results showed that the hyperactivity/impulsivity factor was inversely associated with left parahippocampal gray matter volume ($\beta$ = −0.043, p _fdr = .029), similar to the results with the inattentive factor (Table 3). However, the hyperactivity/impulsivity factor diverged from inattention by showing no associations with bilateral caudal anterior cingulate (Table 3), suggesting that this region is specific to inattentive traits. Furthermore, the hyperactivity/impulsivity factor also showed unique associations with gray matter volume in the left ($\beta$ = −0.048, p _fdr < .001) and right caudate ($\beta$ = −0.044, p _fdr  = .029). No other cortical or subcortical regions showed significant associations with hyperactivity/impulsivity (Table 4; Figure 1).

When covarying for income and medication status, these aforementioned associations were strengthened for the bilateral caudate and left parahippocampal gyrus (Table 5). In addition, we assessed the associations between hyperactivity/impulsivity endorsement and volume while accounting for total cortical or subcortical gray matter volume as appropriate. Results were retained and became stronger for the left and right caudate when accounting for total subcortical gray matter volume and for the left parahippocampal gyrus while accounting for total cortical gray matter volume. Similar to the inattention results, the effects between hyperactivity/impulsivity and these regions also became stronger when accounting for total intracranial volume. Finally, we examined these associations while accounting for in‐scanner motion; results were retained for all associations (Table 5).

In addition, we examined neural analyses without regressing out the overlapping effects of inattention and hyperactivity/impulsivity to assess the specificity of the significant findings. Across the two dimensions, the bilateral caudal anterior cingulate, bilateral caudate, and the left parahippocampal gyrus were each significant for both factors (p _fdr's < .001).

4. DISCUSSION

In the present study, we examined the neurobiological mechanisms underlying ADHD by analyzing the structural properties associated with two ADHD dimensions—inattention and hyperactivity/impulsivity—in a sample of 10,692 9‐to‐10‐year‐old children from the ABCD Study. We used an approach that accounts for the overlapping variance between these trait clusters and allows us to examine common and dissociable effects. The findings revealed that smaller bilateral caudal anterior cingulate volume was uniquely associated with endorsement of inattentive traits, while smaller bilateral caudate volume was uniquely associated with hyperactive/impulsive traits. Additionally, smaller left parahippocampal volume was associated with both inattentive and hyperactive/impulsive traits, suggesting a common structural feature across ADHD presentations. Sensitivity analyses revealed that these findings retained significance and increased in magnitude when accounting for global indicators of gray matter volume (total cortical volume, total subcortical volume, intracranial volume), suggesting a focal effect of our brain‐behavior findings. Significance was also retained when accounting for the additional covariates of income, medication, and in‐scanner motion. Overall, these findings emphasize both common and specific volume differences associated with inattentive and hyperactive/impulsive traits.

The finding that greater endorsement of inattention is linked to smaller caudal anterior cingulate gray matter volume is in line with prior empirical studies acknowledging the anterior cingulate's potential importance in the pathophysiology of ADHD (Bernanke et al., 2022; Kim et al., 2021; Seidman et al., 2006). Dysfunction or alterations to the anterior cingulate are associated with impaired attention allocation and sustainment, domains important to ADHD's inattentive phenotype (Wu et al., 2017). Moreover, the caudal portion of the anterior cingulate cortex specifically has been associated with activation during cognitive control tasks and tasks with increasing mental demands. For example, past research has shown that activation in the caudal anterior cingulate cortex increases with working memory load (Gray & Braver, 2002). The caudal anterior cingulate cortex is also posited to be involved in performance monitoring, with increases in activation during error detection (Ursu et al., 2009). Deficits in regions associated with error processing and cognitive control may be related to inattention traits; however, while these fMRI studies reveal the functional significance of this region, they do not explain why we would find smaller volumes in the caudal anterior cingulate cortex in inattention traits. While some studies that focus on structural differences in ADHD substantiate the finding of smaller area or volume in the cingulate cortex (Bernanke et al., 2022; Durham et al., 2021; Makris et al., 2010; Seidman et al., 2006); one meta‐analysis found the opposite: larger cingulate volumes associated with ADHD (Nakao et al., 2011). These mixed findings are likely due the common practice of combining presentations in analyses of ADHD traits, which may obscure differences if the effects for inattention and hyperactivity/impulsivity are in opposite directions. Our study expands and improves upon prior work by disentangling the relative contributions of each core feature to allow for greater specificity in the results. Using such an approach, we show that smaller volume in the caudal portion of the cingulate cortex is specific to inattention traits and not hyperactivity/impulsivity traits.

In addition, our analyses showed that greater endorsement of hyperactivity/impulsivity traits was associated with smaller gray matter volume in the bilateral caudate. These subcortical findings align with ENIGMA ADHD Working Group mega‐analysis findings, which similarly report smaller caudate volume associated with ADHD (Hoogman et al., 2017). The caudate is associated with movement planning and control (Grahn et al., 2008). As such, it makes sense that a region associated with motor control would be implicated in hyperactivity/impulsivity traits. Notably, the findings of caudate volume differences in ADHD have not always been consistent in the literature. Reviews and meta‐analyses show that some studies report smaller caudate volumes in those with ADHD compared to controls while many other studies report no differences in the caudate (Seidman et al., 2005; Valera et al., 2007). However, many of these prior studies were limited by small samples sizes. A more recent study with over 900 children with ADHD also did not find significant differences in the caudate in ADHD (Bernanke et al., 2022). This is in contrast to the ENIGMA ADHD Working Group findings mentioned earlier which found smaller caudate volumes in over 1700 individuals with ADHD (Hoogman et al., 2017). We show in the current study that these discrepancies can be clarified when the relative contributions of inattention and hyperactivity/impulsivity are accounted for, which allows us to identify a relationship between smaller caudate volumes and ADHD traits, while showing that this relationship is specific to hyperactivity/impulsivity traits. In support of this, others have also suggested the existence of separate neural phenotypes of ADHD based on inattention and hyperactivity/impulsivity (Carmona et al., 2009). Thus, the results of the current study identified regions with dissociable relationships with inattention and hyperactivity/impulsivity.

In terms of commonalities, the present study found differences in parahippocampal cortex volume as a neurobiological similarity across ADHD's core features. Specifically, we found that smaller left parahippocampal gyrus volume was associated with greater trait endorsement for both inattentive and hyperactive/impulsive traits. The parahippocampal cortex is implicated in contextual associations, which encompass many broad functions including episodic memory, contextual processing, navigation, and scene processing (Aminoff et al., 2013). Previous studies have indicated atypical neural properties in the parahippocampal cortex in relation to ADHD. For example, studies show smaller volume (Carmona et al., 2005; Gehricke et al., 2017), thinner cortices (Hoogman et al., 2019), and atypical function (Braet et al., 2011) in the parahippocampal cortex, with most studies implicating the left parahippocampal gyrus in particular. Atypical functioning in the left parahippocampal gyrus in those with ADHD has been suggested to be related to the regulation of arousal while performing a task, thus, deficits in this region could contribute to failure to engage top‐down processing to improve task performance (Braet et al., 2011). This is supported by a large number of studies documenting impaired task performance in those with ADHD on a wide variety of paradigms (Pievsky & McGrath, 2018), many of which could be affected by both inattentive and hyperactivity/impulsivity traits. Thus, the finding of smaller left parahippocampal gyrus volumes common across inattention and hyperactivity/impulsivity may suggest the existence of a non‐specific risk factor for ADHD traits in general.

The statistical methods leveraged in our analyses represent a notable strength. Traditional diagnosis of ADHD, by nature of its trait threshold, reduces the disorder into binary categories: an individual either has ADHD or does not. Many research studies rely on diagnostic categories to define an ADHD group. However, clinical symptomatology exists on a continuous spectrum as opposed to finite categories. As such, we operationalized ADHD continuously, distinguishing between inattention and hyperactivity/impulsivity factors while retaining the dimensional nature of psychopathology. This allowed us the ability to capture how individual differences in trait presentation are reflected in the brain by showing that not only were inattention and hyperactivity/impulsivity associated with smaller volumes, but that this also maps onto the severity of clinical features. Because the ABCD Study uses a longitudinal design, future research examining trajectories of brain development in relation to inattentive versus hyperactive/impulsive traits will be highly informative for allowing us to understand how ADHD presentations diverge over time.

There are several limitations to consider in interpreting the current study's results. The effect sizes found in the primary analyses are relatively small in magnitude. However, it should be noted that our sensitivity analyses largely strengthened the associations between brain volume and ADHD core features, supporting that our findings are robust. Additionally, prior studies using large samples consistently yield brain‐behavior associations that are reliable, despite being small (Paulus & Thompson, 2019). Additionally, while not a clinical sample, the results of the current study still provide important information about the neurostructural correlates of inattention and hyperactivity/impulsivity and demonstrate that meaningful associations are apparent across the spectrum of ADHD symptoms. Lastly, this study's cross‐sectional design limits our capacity to make inferences about changes over time but serves as a promising starting point for future longitudinal investigations of these associations.

5. CONCLUSION

In sum, prior literature has highlighted the need to determine biologically homogeneous groups within the broader heterogeneous disorder of ADHD (Luo et al., 2019). The results of the current study build upon existing literature to reveal the neurostructural commonalities and differences associated with inattention and hyperactivity/impulsivity traits. The brain regions that emerged as significant in the present findings—the caudal anterior cingulate and the caudate—have clear implications for inattention and hyperactivity/impulsivity, respectively. These results suggest that, when working with patients with ADHD, it is important to differentiate between inattention and hyperactivity/impulsivity, which show dissociable underlying neural mechanisms that may suggest the need for different interventions. Future work should further explore the importance of hyperactivity traits in models of ADHD and whether or not the inattention and hyperactivity/impulsivity qualifiers should be retained in DSM‐5‐TR criteria. Ultimately, findings from the present study will help to refine our understanding of the neural substrates of ADHD presentations, allowing for more accurate diagnosis and targeted treatments in hopes of mitigating the long‐term negative effects of this disorder.

CONFLICT OF INTEREST STATEMENT

The authors declare no competing financial interests.

Reimann, G. E. , Jeong, H. J. , Durham, E. L. , Archer, C. , Moore, T. M. , Berhe, F. , Dupont, R. M. , & Kaczkurkin, A. N. (2024). Gray matter volume associations in youth with ADHD features of inattention and hyperactivity/impulsivity. Human Brain Mapping, 45(5), e26589. 10.1002/hbm.26589

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are openly available in the NIMH Data Archive at https://nda.nih.gov/abcd/.