Abstract
Attention deficit hyperactivity disorder (ADHD) is one of the most common neurodevelopmental disorders in childhood. Numerous resting-state functional magnetic resonance imaging (rs-fMRI) studies in ADHD have been performed using traditional low-frequency bands (0.01–0.08 Hz). However, the neural activity patterns of frequency subbands in ADHD still require further investigation. The purpose of this study is to explore the frequency-dependent characteristics and neural activity patterns of ADHD subtypes. We selected the ADHD combined type (ADHD-C, N = 25), ADHD inattentive type (ADHD-I, N = 26) and typically developing (TD, N = 28) children from the ADHD-200 Consortium. Based on the slow-5 band (0.01–0.027 Hz) and slow-4 band (0.027–0.073 Hz), we generated static and dynamic fractional amplitude of low-frequency fluctuation (fALFF) and regional homogeneity (ReHo) maps for each participant. A flexible-factorial analysis of variance model was performed on static and temporal dynamic rs-fMRI measurements within two subbands. Results revealed that the orbital-frontal gyrus, precuneus, superior temporal gyrus and angular gyrus were found to have obvious frequency band and group interaction effects. The intrinsic neural activity differences among three groups were more prominent in the slow-5 frequency band compared to the slow-4 band. In addition, the indices of significant interaction regions showed correlations with the progression of the disease and the features in slow-5 showed an advantageous diagnostic performance compared with those in slow-4. The results suggested the intrinsic neural activities of ADHD subtypes were frequency-dependent. The frequency-specific analysis of static and dynamic brain activity may provide a deeper understanding of neurophysiological dysfunction patterns in ADHD subtypes and provide supplementary information for assessing ADHD subtypes.
1 Introduction
Attention deficit hyperactivity disorder (ADHD) is one of the most common neurodevelopmental disorders in childhood. Patients are characterized as having inattention and/or hyperactivity-impulsivity symptoms, which often exert a negative effect on the quality of their social and academic activities. In China, the total prevalence of ADHD in children and adolescents is 6.26%. In addition, the prevalence rate of ADHD in males is significantly higher than that in females (). ADHD has a high degree of heritability and significant familial aggregation (; ), which may be related to genetic or environmental factors (). Although the symptoms of ADHD partially decrease with age, more than half of cases persist into adulthood ().
The fifth editions of the Diagnostic and Statistical Manual of Mental Disorders (DSM-V) published by the American Psychiatric Association (APA) classified ADHD into different subtypes: inattentive presentation (ADHD-I), hyperactive-impulsive presentation (ADHD-HI) and combined presentation (ADHD-C). The prevalence of ADHD-I is the highest among the subtypes, followed by ADHD-C and ADHD-HI (). In the latent class analysis (LCA), the quantitative analysis of ADHD subtypes in different periods emphasized that the symptoms of ADHD would change with the physical development of patients (). Due to the strong heterogeneity among ADHD individuals (), it is necessary to monitor and investigate characteristic neurophysiological activities of ADHD subtypes.
Functional magnetic resonance imaging (fMRI), including resting-state and task-state imaging, is a well-known technique for measuring brain function and has been widely used to explore the pathophysiological mechanism of ADHD. Structural imaging (; ) and task-state imaging (task-fMRI) (; ; ) were used in early studies of ADHD. Task-fMRI included a variety of tasks in the form of go/no-go, stop-signal or n-back, but multiple meta-analyses showed that the results did not show good consistency (). Resting-state fMRI (rs-fMRI) has gradually garnered attention and it provides a more convenient method compared to task- fMRI (). For example, psychiatric patients, especially preschool children, are often unable to cooperate with the assigned task, while rs-fMRI can be performed when the patient is lying still without performing tasks. Rs-fMRI is designed to examine spontaneous neural fluctuations in the resting state and is a noninvasive tool for obtaining information on brain function.
To characterize the fluctuating patterns of brain activity, various voxel-based rs-fMRI indices have been proposed. ALFF/fALFF (; ) and ReHo () are among the most commonly used functional indices in fMRI studies. ALFF is defined as the average power spectrum of the time series after Fourier transform in a particular low-frequency band (), and fALFF is the ratio of the power spectrum in the low-frequency band to the entire frequency range (). ALFF reflects the intensity of neural activity by the amplitude of the spectrum, while fALFF represents the relative contribution of a particular oscillation to the entire detectable frequency range. ReHo is a method for calculating time series between a particular voxel and its nearest neighboring voxels (). ReHo reflects the degree of concordance in regional neural activity. They have been widely applied to ADHD patients to reveal aberrant neural activities. Previous studies found that, compared to healthy controls, ADHD patients showed decreased ALFF in the left frontal gyrus, while increased ALFF in the right dorsal superior frontal gyrus (). ReHo are observed to show widely-distributed differences in the fronto-cingulo-occipito-cerebellar circuitry (). Additionally, ALFF/fALFF and ReHo measures are data-driven and might not be biased by factors such as seed definition in seed-based correlation analysis and component selection in independent component analysis. In addition, fALFF is reported to show less susceptible to nuisance noise and higher sensitivity and specificity than ALFF (; ). The static measure of fMRI indices assume that the functional activity of brain regions remains stationary throughout the entire resting-state scan. In addition to static rs-fMRI measurements, evidence has shown that spontaneous neuronal activity shows dynamic fluctuations in the resting state (; ). The dynamic measures estimate the temporal fluctuation patterns of interregional neural interactions (; ). A report utilized temporal dynamic rs-fMRI indices in ADHD patients and found a decreased variability in dynamic ALFF in the middle frontal gyrus (). Dynamic functional network analysis revealed that ADHD patients exhibit distinct interactive states in static and dynamic functional connectivity ().
The spontaneous low-frequency oscillations (LFO) of the blood oxygen level-dependent (BOLD) signal measured in rs-fMRI typically are characterized in the range of 0.01–0.08 Hz, which are considered may reflect spontaneous neuronal activity (). Physiological noise such as the respiration rate and heart rate is recorded within the range of 0.1–0.3 Hz and 0.6–1.1 Hz, respectively (). In addition, Buzsáki et al. divided the neural oscillation frequency in the brain into multiple subfrequency bands and found that the oscillation within disparate frequency bands may related to different neural processes (). A study on functional connectivity found that the strengths of functional connections attenuate at different rates as frequency increases across multiple networks (). The LFO can be subdivided into slow-5 (0.01–0.027 Hz) and slow-4 (0.027–0.073 Hz) (). Studies reported that the amplitude of LFO in gray matter is higher than in white matter and the contributions of slow-4 and slow-5 to LFO amplitudes were different in brain regions such as the precuneus, basal ganglia and thalamus in healthy subjects (). Xue et al. replicated the experiment and obtained similar findings (). Furthermore, previous studies have shown that Alzheimer’s disease () and Parkinson’s disease () exhibit discrepant neural activity in slow-5 and slow-4 bands, suggesting frequency-dependent alterations in the amplitude patterns of mental disorders. However, to the best of our knowledge, few studies have focused on the neural activity patterns of ADHD-C and ADHD-I subtypes in specific subfrequency bands. It is currently urgent to clarify the discrepancy of frequency characteristics in ADHD subtypes. Finer frequency-dependent features may provide important implications and help quantify neural functions in the brain. The approach may hold the potential to help explore the mechanisms driving the differences between subtypes of ADHD.
We aimed to identify the representative neural activity patterns and frequency-dependent characteristics of ADHD-C and ADHD-I subtypes. We utilized both static and dynamic measurements of fALFF and ReHo indices to investigate frequency-related characteristics within the slow-5 and slow-4 bands. We expected to improve the understanding of neurophysiological dysfunction patterns in ADHD subtypes.
2 Materials and methods
2.1 Participants
In the present study, data were collected from the ADHD-200 Consortium (). The dataset contains data from eight sites and the consortium is committed to facilitating research on the neural basis of ADHD. Access to these data was approved by the research ethics review committees of the respective institutions, and all participants or legal guardians signed informed consent forms before participating. A previous study showed that the cohorts in ADHD-200 were highly inconsistent, and the author suggested that future studies should be conducted using pooled or single cohort data (). Therefore, we selected data from one site at Peking University, which possessed more subtypes and more complete information. All participants were assessed using the Computerized Diagnostic Interview Schedule IV (C-DIS-IV), a structured clinical interview of computerized version to provide diagnoses of major psychiatric disorders based on the DSM-IV criteria. They were also evaluated using the Schedule of Affective Disorders and Schizophrenia for Children—Present and Lifetime Version (K-SADS-PL). It is a semistructured interview administered by trained assessors, that contains the screening interview and diagnostic supplement to evaluate comorbid psychiatric disorders. The participants underwent MRI scanning using Siemens 3T Trio scanners, and the scan parameters used were 2000 ms for the repetition time (TR) and 30 ms for the echo time (TE). The acquired voxel size was 3 mm and 4.5 mm. The dataset contains three groups, and the scanning parameters of two groups were identical, but one group had slightly different parameters. More detailed scanning parameters can be found in http://fcon_1000.projects.nitrc.org/indi/adhd200/. We excluded female and left-handed individuals to eliminate the effects of sex and handedness factors on the study. We excluded participants with poor-quality images and incomplete clinical information. Then, based on the DSM-IV diagnostic criteria at enrollment, we divided participants into three categories: ADHD-I, ADHD-C and typically developing (TD) groups. We matched the three groups by age. Participants with excessive head movement (translation > 3 mm, rotation > 3°) and mean Jenkinson’s framewise displacement (FD) () greater than 0.2 mm were excluded. Finally, we included 25 ADHD-C, 26 ADHD-I and 28 healthy participants. The flowchart of data exclusion was shown in Figure 1.
FIGURE 1
2.2 Data preprocessing
The Data Processing Assistant for Resting-State fMRI (DPARSF, RRID:SCR_002372) () was used to preprocess resting-state fMRI data based on Statistical Parametric Mapping (SPM, RRID:SCR_007037) and the Resting-State fMRI Data Analysis Toolkit V1.8 (REST 1.8, RRID:SCR_009641) (). (1) The first 10 time points of the imaging data were excluded to avoid signal instability as the participants adapted to the fMRI environment. (2) Time sequence correction was performed to rearrange the images in time and space. (3) Head movement correction was conducted (participants with head translational movement > 3 mm or rotation > 3°, and mean Jenkinson’s FD > 0.2 mm were excluded). (4) Individual structural images were adjusted to the mean functional images for registration. (5) Nuisance covariates, including linear drift, Friston-24 motion parameters, white matter signals and cerebrospinal fluid signals, were regressed out. (6) Functional images were normalized from the original space to the standard space (Montreal Neurological Institute, MNI) with 3 mm isotropic voxels.
The fALFF and ReHo were calculated based on the whole-brain voxel analysis. The calculation method for fALFF was based on the approach proposed by , where it is obtained by the ratio of the power spectrum in the specific frequency band (slow-5 and slow-4 obtained through band-pass filtering in the study) to the entire frequency range (without band-pass filtering). ReHo was measured in both slow-5 and slow-4 bands by calculating Kendall’s coefficient of concordance (KCC), which calculated the similarity of time series between a given voxel and its nearest neighboring voxel (). The criterion for the trueness of adjacent voxels was cornered connections with 26 adjacent voxels. Therefore, the static fALFF and ReHo maps were obtained within the slow-5 and slow-4 bands, respectively. All the fMRI maps were then Z-standardized and smoothed with a 4 mm full width at half maximum Gaussian kernel (FWHM).
Subsequently, as an approach to evaluate the characteristics of the temporal variation in the aforementioned measurements, we used the sliding window method to calculate dynamic fALFF and ReHo indices with Hamming windows. The window length was 48 TR, and the step size was 4 TR. Previous studies showed that different window lengths do not produce significantly different results (; ). Moreover, the step size does not significantly affect the variability of fMRI dynamic characteristics, but the empirical value is equal to one-tenth of the window length (). More importantly, we varied the window length (32 TR/64 TR) to validate the reliability of the results. The results of dynamic indices have been included in the Supplementary materials. The voxel-wised fALFF/ReHo indices were calculated in each window and each voxel was assigned one fMRI value in each window. To measure the variability of the dynamic indices over time, the mean and standard deviation (SD) maps across time windows were computed for each index. Finally, the mean and SD maps were Z-standardized and a 4 mm FWHM was used for smoothing.
2.3 Statistical analysis
Age, IQ and ADHD-related clinical characteristics were assessed by performing normality and homogeneity of variance tests. Then, demographic and clinical characteristics of ADHD-C, ADHD-I and TD groups were compared by analysis of variance (ANOVA). These statistical analyses were conducted using the IBM SPSS Statistics 26.0 software package (RRID: SCR_016479).
We designed a flexible-factorial ANOVA model to perform voxel-based two-way analysis based on fALFF and ReHo maps using the SPM software package. The model included three groups (ADHD-C, ADHD-I and TD) as between-subject factors and frequency bands as (slow-5 and slow-4) as within-subject factors to observe the effects of interactions between subtypes and frequencies. Age and mean FD were added as covariates and the mask was constructed jointly from the gray matter regions of 90% of the participants. All the statistical maps were corrected for multiple comparisons by using Gaussian Random Field (GRF) correction, combining voxel-wised thresholding p < 0.001 and cluster-wised thresholding p < 0.05. Then, we extracted regions showing subtype × frequency interaction effects as regions of interest (ROIs). A post-hoc analysis was performed on ROIs to investigate simple effects of fMRI measures among groups and frequency bands. The post-hoc results were corrected for multiple comparisons by Bonferroni correction with p < 0.05.
We verified the relationship between ROI signals and disease severity of ADHD-C, ADHD-I and TD. Partial correlation analysis was conducted between fMRI indices and clinical scores of three groups with age and FD as the nuisance covariates. Furthermore, we examined the efficiency of ROIs in classifying ADHD-C, ADHD-I and TD groups, respectively. We compared the features of ROIs in the slow-5 and slow-4 bands. The classification efficiency was verified by plotting the receiver operating characteristic (ROC) curve using a logistic regression model. The area under the ROC curve (AUC) and accuracy were calculated.
3 Results
3.1 Demographic and clinical characteristics
Demographic and clinical information was represented in Table 1. The results showed no between-group variance in age and verbal IQ. The between-group comparison of diagnostic index was significant, except for ADHD-C and ADHD-I, which showed no difference in the Inattentive Index. Both ADHD subtypes showed a significant reduction in performance IQ compared with the TD group, but no differences were observed between ADHD-C and ADHD-I groups.
TABLE 1
| ADHD-C | ADHD-I | TD | ANOVAa | Between-group testsb | |||
| C-I | C-TD | I-TD | |||||
| N | 25 | 26 | 28 | – | – | ||
| Gender(male) | 25 | 26 | 28 | – | – | ||
| Age(years) | 11.6 ± 1.7 | 12.5 ± 1.7 | 12.1 ± 1.5 | 0.148 | 0.154 | 0.952 | 0.935 |
| ADHD Index | 56.7 ± 7.5 | 46.0 ± 6.4 | 28.5 ± 5.9 | <0.001 | <0.001 | <0.001 | <0.001 |
| Inattentive Index | 29.6 ± 3.5 | 27.8 ± 3.2 | 15.7 ± 3.5 | <0.001 | 0.239 | <0.001 | <0.001 |
| Hyper/Impulsive Index | 27.2 ± 5.2 | 18.3 ± 4.4 | 12.8 ± 3.5 | <0.001 | <0.001 | <0.001 | <0.001 |
| Verbal IQ | 115.4 ± 17.5 | 110.5 ± 14.7 | 118.0 ± 13.7 | 0.198 | 0.771 | 0.989 | 0.228 |
| Performance IQ | 100.3 ± 12.4 | 95.2 ± 16.6 | 110.6 ± 15.5 | 0.001 | 0.697 | 0.044 | 0.001 |
Demographic and clinical information of participants.
Data are presented as the mean ± standard deviation. C, ADHD-Combined; I, ADHD-Inattentive; TD, typically developing.
ap-values for ANOVA among the three groups;
bp-values for between-group effects after Bonferroni multiple comparison correction.
3.2 Subtype × frequency interaction effect
The two-way ANOVA showed a significant subtype × frequency interaction effect on the static fALFF in the left orbital-frontal gyrus (OFC) (Figure 2A and Table 2) and no significant difference in the static ReHo. In the analysis of dynamic indices, the ANOVA model showed interaction effect on dynamic mean of fALFF value in the left OFC (Figure 2B), dynamic SD of ReHo value in the right superior temporal gyrus (STG), bilateral precuneus (PCUN) and left angular gyrus (ANG) (Figure 2C).
FIGURE 2
TABLE 2
| Functional Index | Region | L/R | Peak MNI coordinates | F | Cluster size (mm2) | Group differencea | |||||
| x | y | z | Slow-5 | Slow-4 | |||||||
| F | Sig. | F | Sig. | ||||||||
| Static fALFF | OFC | L | −24 | 18 | −21 | 11.92 | 189 | 14.47 | <0.001 | 0.37 | 0.691 |
| Dynamic mean fALFF | OFC | L | −24 | 18 | −24 | 12.14 | 189 | 15.40 | <0.001 | 0.06 | 0.942 |
| Dynamic SD ReHo | STG | R | 54 | −42 | 18 | 12.53 | 324 | 13.23 | <0.001 | 1.12 | 0.333 |
| PCUN | L/R | 0 | −63 | 57 | 12.26 | 324 | 4.72 | 0.012 | 7.05 | 0.002 | |
| ANG | L | −39 | −54 | 21 | 11.74 | 270 | 12.27 | <0.0001 | 2.72 | 0.073 | |
Significant results for the subtype and frequency interaction effect and the post-hoc analysis.
SD, standard deviation; OFC, orbital-frontal gyrus; STG, superior temporal gyrus; PCUN, precuneus; ANG, angular gyrus.
aANOVA results of significant interaction effect regions among ADHD-C, ADHD-I and TD groups.
3.3 ROI-based Post-hoc analysis
Regions showing subtype × frequency interaction effects were extracted as ROIs. A post-hoc analysis was performed on ROIs to investigate simple effects. Based on ROI analysis, rs-fMRI indices were compared among ADHD-C, ADHD-I and TD groups in the slow-5 band (Figure 3A) and show-4 band (Figure 3B). ADHD-C and ADHD-I both showed significantly decreased static and dynamic fALFF of OFC in the slow-5 band. Compared to TD, ADHD-C also showed increased dynamic ReHo SD values in the STG, PCUN and ANG, while ADHD-I only showed in the STG. Under the slow-4 band, both ADHD-C and ADHD-I exhibited decreased dynamic ReHo SD values in the PCUN. Furthermore, the ANG of ADHD-C and ADHD-I showed differences in slow-5. Additionally, ROI comparisons between slow-5 and slow-4 bands was shown in Figure 4. Compared with slow-4, ADHD-C showed increased indices of PCUN and ANG in slow-5, while ADHD-I showed increased indices of OFC and STG in slow-5.
FIGURE 3
FIGURE 4
3.4 Correlation and classification model
The results of partial correlation analysis of three groups were shown in Figure 5. In the slow-5 band, all ROIs showed significant correlations with the ADHD index, while OFC also significantly correlated with the performance IQ index. In the slow-4 band, only dynamic SD ReHo of PCUN was negatively correlated with the ADHD index. The results showed that as static and dynamic mean indices decreased and the dynamic variabilities increased, the severity of disease increased. In addition, ROC curves were used to compare the classification efficiency of combined features from five ROIs in the slow-5 and slow-4 bands. The results indicated that features in slow-5 showed higher AUC and accuracy compared to features in slow-4 (Figure 6 and Table 3). Additionally, the features showed high classification efficiency in distinguishing between ADHD-C/ADHD-I and TD groups, while showed low performance in distinguishing between ADHD-C and ADHD-I groups.
FIGURE 5
FIGURE 6
TABLE 3
| Classifier | Features in slow-5 | Features in slow-4 | ||||
| AUC | p-value | Accuracy(%) | AUC | p-value | Accuracy(%) | |
| ADHD-C−ADHD-I | 0.782 | 0.0006 | 76.47 | 0.695 | 0.0170 | 68.63 |
| ADHD-C−TD | 0.956 | <0.0001 | 90.57 | 0.774 | 0.0006 | 77.36 |
| ADHD-I−TD | 0.907 | <0.0001 | 88.89 | 0.729 | 0.0038 | 70.37 |
The results for classifying in ADHD-C, ADHD-I and TD groups with features of ROIs in slow-5 and slow-4.
4 Discussion
To the best of our knowledge, researchers have not previously attempted to focus on comparing the neural activity patterns in specific divided subfrequency bands between different ADHD subtypes. We applied static and dynamic rs-fMRI measurements in subfrequency bands to examine the neural activity of ADHD-C, ADHD-I and TD. The results disclosed that significant subtype and frequency interaction effects were shown in the OFC, STG, PCUN and ANG. The post-hoc analysis showed that the intrinsic brain activity differences among three groups were mainly exhibited in the slow-5 frequency band and the interaction regions were significantly correlated with multiple clinical scores. Additionally, the features of significant interaction regions showed an advantageous diagnostic performance in slow-5 compared with those in slow-4.
Compared to healthy individuals, both ADHD-C and ADHD-I exhibited significantly decreased fALFF in the static and dynamic mean measurements of the OFC. Based on early investigations of the temporal and spatial structure of brain functional connections, that the state of brain activity varies over time (; ; ). In recent years, an increasing number of studies have recognized that the intrinsic brain activity must be considered from the perspective of temporal features. The fALFF is a voxel-wised fMRI measure that provides information about the regional spontaneous amplitude intensity of brain activity within a network (). The inferior frontal gyrus (orbital-frontal gyrus) has been reported to be involved in cognitive, motor control and sustained attention (; ). Furthermore, the orbital-frontal cortex is closely associated with executive functions, mediating top-down cognitive processes which operate in emotional and motivational contexts (; ). This may reflect impaired activity of the motivational executive functions in both ADHD-C and ADHD-I. Research showed that ADHD individuals exhibit reduced activation in the left OFC during interference inhibition tasks (). A previous study also showed that ADHD patients exhibit positive dynamic functional connectivity changes between the inferior frontal gyrus and regions of the cognitive control network following fMRI-neurofeedback training (). The abnormal activity of OFC may serve as the foundation for cognitive control deficits in individuals with ADHD, and the failure in cognitive control plays a crucial role in the main symptoms of ADHD.
In the slow-5 frequency band, ADHD-C showed increased dynamic SD of ReHo in the STG, PCUN, and ANG compared to TD, while ADHD-I only exhibited increased dynamic SD of ReHo in the temporal lobe. ReHo is a voxel-based measurement that assesses the similarity of the time series of a given voxel with those of its nearest neighboring voxels (). We calculated SD of dynamic ReHo to measure the degree of dynamic variability in spontaneous neural activities over time. Cognitive function depends on dynamic interactions among large-scale neural systems (). Most of the abnormal regions in our results overlapped with the default mode network (DMN) (). The DMN is active at rest but its activation decreases continuously during goal-directed behavior, which preserves the privilege of cognitive tasks by reducing unrelated stimulus activation (; ). Some studies assumed that a decrease in the inhibition of the DMN is related to attention deficits (). The increased SD of dynamic ReHo in the DMN reflected a high level of functional disruption, which may be associated with the disrupted balance within the neural network. In addition, ADHD-C exhibited higher dynamic SD of ReHo in the ANG compared to ADHD-C. The ANG acted as a multimodal integration center in reading, comprehension, spatial cognition and attention () and showed reduced functional connectivity related to motor performance in ADHD (). A study examined the performance of ADHD patients in attention tasks revealed that the functional activity of ANG is related to cognitive performance (). Furthermore, research showed that the reduced functional connectivity between the PCUN and other components of the DMN is strongly associated with attentional deficits in ADHD (). In slow-5, ADHD-C showed significant disturbances of functional activities in the PCUN and ANG, while ADHD-I did not. It may suggest that ADHD-C exhibits more severe symptoms of attention deficit than ADHD-I. A previous study also showed that ADHD-I performs better than ADHD-C in the test of attention network tasks (). Additionally, compared to TD, both ADHD-C and ADHD-I showed decreased dynamic ReHo fluctuations in the PCUN in slow-4, which may reflect an excessive stability and ineffective activity. Our results suggested that ANG and PCUN, measured in slow-5, may be the key regions to distinguish ADHD-C from ADHD-I. The results might help explain the behavioral differences between ADHD-C and ADHD-I.
The subtype and frequency interaction effect analysis showed that abnormal fMRI indices of ADHD were not only related to disease factors, but also to specific frequencies. The subtype differences were mainly observed in slow-5 frequency band. Compared to the slow-4 band, functional signals in slow-5 tend to exhibit increased intensity or enhanced fluctuation. The results indicated that different frequency bands are associated with specific pathological states in ADHD-C and ADHD-I. This could suggest a high sensitivity of neural activity patterns in ADHD subtypes to the slow-5 frequency band, which contains more oscillation information in differentiating ADHD subtypes compared to slow-4. Furthermore, our results were consistent with a previous theory that brain regions showed different sensitivities to each frequency band: low-frequency oscillations (slow-5) have higher power, which is conducive to long-distance connections and large-scale neural network construction. Therefore, low-frequency oscillations more easily adjust the large default mode regions (; ; ). High-frequency oscillations (slow-4) have lower power and are mostly related to the spatial structure of small neurons with short connections (; ). Moreover, it has been observed that the PCUN exhibits decreased ReHo in the slow-4 compared to the slow-5 in typically developing individuals. The functional connectivity showed a decreased tendency in the slow-4 band compared to the slow-5 band (). These findings are similar to our results. The findings suggested that it is important to consider selecting sensitive frequency bands when detecting abnormal spontaneous brain activities in ADHD.
According to the correlative analysis among three groups, the spontaneous fluctuation indices of each group could reflect the serious trend in the development of the disease. For example, as the disease progressed, there is a gradual decrease in both static and dynamic mean fALFF of the OFC, and an increased variability in the dynamic ReHo of the PCUN, STG and ANG. Our results suggested that measuring neural fluctuation in specific frequencies may beneficial for understanding the neuropathological basis of ADHD and assist in future monitoring of disease progression. The results from dysfunctional regions applied to the classification showed that the diagnostic efficiency of features constructed in slow-5 was better than that in slow-4. The finding confirmed that the slow-5 frequency band contained more advantageous diagnostic information than slow-4 and it suggested that frequency factors should be considered when evaluating the intrinsic brain activities of ADHD. As more categorical features were derived from the dynamic indices, suggesting the temporal characteristics might be a powerful tool for the detection of pathological changes in different ADHD subtypes. Furthermore, the regions showing interaction effects could be potential neuroimaging markers representing the characteristics of ADHD subtypes.
Our study had some limitations that should be considered. First, our research did not include ADHD-HI patients because no ADHD-HI patients were included in this dataset. Second, the relatively small sample size may have limited the effect size of the study. As such, the replication efforts with larger sample sizes will be necessary to further confirm the robustness and reliability of the findings. Finally, this study focused on the assessment of ADHD in the resting state. More experiments of task state are needed to comprehensively evaluate the differences in brain functional activities of ADHD subtypes in different frequency bands.
5 Conclusion
Our study revealed that the intrinsic neural activities of ADHD subtypes were frequency-dependent and the differences were more evident in slow-5 than in slow-4 band. Furthermore, the abnormal indices were correlated to the trend of disease progression in each group. The study suggested that frequency factors should be considered when evaluating the neural substrates in ADHD subtypes. The frequency-dependent static and dynamic brain activities might provide potential neuroimaging biomarkers of ADHD subtypes and provide Supplementary Information for monitoring ADHD progressions.
Statements
Data availability statement
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: the ADHD-200 Consortium.
Ethics statement
The studies involving humans were approved by the Research Ethics Review Board of Institute of Mental Health, Peking University. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants’ legal guardians/next of kin.
Author contributions
RC: Formal analysis, Investigation, Methodology, Software, Supervision, Validation, Writing−original draft, Writing−review and editing. YJ: Conceptualization, Methodology, Project administration, Supervision, Writing−review and editing. J-SZ: Formal analysis, Software, Supervision, Validation, Writing−review and editing. X-HW: Funding acquisition, Project administration, Resources, Supervision, Validation, Writing−review and editing. M-TZ: Resources, Software, Supervision, Validation, Writing−review and editing.
Funding
The author(s) declare financial support was received for the research, authorship, and/or publication of the article. This work was in part supported by a grant from the National Natural Science Foundation of China (62071158) and the Fundamental Research Funds for the Central Universities of China (No. 2242023K5005).
Acknowledgments
We want to thank the ADHD-200 Consortium for sharing structural and functional MRI datasets.
Conflict of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fnhum.2024.1412572/full#supplementary-material
References
1
AhmadiM.KazemiK.KucK.Cybulska-KlosowiczA.HelfroushM.AarabiA. (2021). Resting state dynamic functional connectivity in children with attention deficit/hyperactivity disorder.J. Neural Eng.18:ac16b3. 10.1088/1741-2552/ac16b3
2
AllenE.DamarajuE.PlisS.ErhardtE.EicheleT.CalhounV. (2014). Tracking whole-brain connectivity dynamics in the resting state.Cereb. Cortex24663–676. 10.1093/cercor/bhs352
3
AnL.CaoQ.SuiM.SunL.ZouQ.ZangY.et al (2013). Local synchronization and amplitude of the fluctuation of spontaneous brain activity in attention-deficit/hyperactivity disorder: A resting-state fMRI study.Neurosci. Bull.29603–613. 10.1007/s12264-013-1353-8
4
BariaA.BalikiM.ParrishT.ApkarianA. (2011). Anatomical and functional assemblies of brain BOLD oscillations.J. Neurosci.317910–7919. 10.1523/jneurosci.1296-11.2011
5
BassettD.SpornsO. (2017). Network neuroscience.Nat. Neurosci.20353–364. 10.1038/nn.4502
6
BucknerR.Andrews-HannaJ.SchacterD. (2008). The brain’s default network: Anatomy, function, and relevance to disease.Ann. N. Y. Acad. Sci.11241–38. 10.1196/annals.1440.011
7
BushG. (2011). Cingulate, frontal, and parietal cortical dysfunction in attention-deficit/hyperactivity disorder.Biol. Psychiatry691160–1167. 10.1016/j.biopsych.2011.01.022
8
BuzsákiG.DraguhnA. (2004). Neuronal oscillations in cortical networks.Science3041926–1929. 10.1126/science.1099745
9
CalhounV.MillerR.PearlsonG.AdalıT. (2014). The chronnectome: Time-varying connectivity networks as the next frontier in fMRI data discovery.Neuron84262–274. 10.1016/j.neuron.2014.10.015
10
CastellanosF.ProalE. (2012). Large-scale brain systems in ADHD: Beyond the prefrontal-striatal model.Trends Cogn. Sci.1617–26. 10.1016/j.tics.2011.11.007
11
CastellanosF.MarguliesD.KellyC.UddinL.GhaffariM.KirschA.et al (2008). Cingulate-precuneus interactions: A new locus of dysfunction in adult attention-deficit/hyperactivity disorder.Biol. Psychiatry63332–337. 10.1016/j.biopsych.2007.06.025
12
Chao-GanY.Yu-FengZ. (2010). DPARSF: A MATLAB toolbox for “Pipeline” data analysis of resting-state fMRI.Front. Syst. Neurosci.4:13. 10.3389/fnsys.2010.00013
13
CoffmanB.TrumboM.ClarkV. (2012). Enhancement of object detection with transcranial direct current stimulation is associated with increased attention.BMC Neurosci.13:108. 10.1186/1471-2202-13-108
14
CordesD.HaughtonV.ArfanakisK.CarewJ.TurskiP.MoritzC.et al (2001). Frequencies contributing to functional connectivity in the cerebral cortex in “resting-state” data.AJNR Am. J. Neuroradiol.221326–1333.
15
CorteseS.AokiY.ItahashiT.CastellanosF.EickhoffS. (2021). Systematic review and meta-analysis: Resting-state functional magnetic resonance imaging studies of attention-deficit/hyperactivity disorder.J. Am. Acad. Child Adolesc. Psychiatry6061–75. 10.1016/j.jaac.2020.08.014
16
CorteseS.KellyC.ChabernaudC.ProalE.Di MartinoA.MilhamM.et al (2012). Toward systems neuroscience of ADHD: A meta-analysis of 55 fMRI studies.Am. J. Psychiatry1691038–1055. 10.1176/appi.ajp.2012.11101521
17
CubilloA.HalariR.GiampietroV.TaylorE.RubiaK. (2011). Fronto-striatal underactivation during interference inhibition and attention allocation in grown up children with attention deficit/hyperactivity disorder and persistent symptoms.Psychiatry Res.19317–27. 10.1016/j.pscychresns.2010.12.014
18
de Oliveira RosaV.Rosa FrancoA.JúniorG.Moreira-MaiaC. R.WagnerF.SimioniA.et al (2020). Effects of computerized cognitive training as add-on treatment to stimulants in ADHD: A pilot fMRI study.Brain Imaging Behav.141933–1944. 10.1007/s11682-019-00137-0
19
DengL.SunJ.ChengL.TongS. (2016). Characterizing dynamic local functional connectivity in the human brain.Sci. Rep.6:26976. 10.1038/srep26976
20
Ellison-WrightI.Ellison-WrightZ.BullmoreE. (2008). Structural brain change in Attention Deficit Hyperactivity Disorder identified by meta-analysis.BMC Psychiatry8:51. 10.1186/1471-244x-8-51
21
FaraoneS.LarssonH. (2019). Genetics of attention deficit hyperactivity disorder.Mol. Psychiatry24562–575. 10.1038/s41380-018-0070-0
22
FaraoneS.BiedermanJ.MickE. (2006). The age-dependent decline of attention deficit hyperactivity disorder: A meta-analysis of follow-up studies.Psychol. Med.36159–165. 10.1017/s003329170500471x
23
FrodlT.SkokauskasN. (2012). Meta-analysis of structural MRI studies in children and adults with attention deficit hyperactivity disorder indicates treatment effects.Acta Psychiatr. Scand.125114–126. 10.1111/j.1600-0447.2011.01786.x
24
HanY.WangJ.ZhaoZ.MinB.LuJ.LiK.et al (2011). Frequency-dependent changes in the amplitude of low-frequency fluctuations in amnestic mild cognitive impairment: A resting-state fMRI study.Neuroimage55287–295. 10.1016/j.neuroimage.2010.11.059
25
HD-200 Consortium (2012). The ADHD-200 consortium: A model to advance the translational potential of neuroimaging in clinical neuroscience.Front. Syst. Neurosci.6:62. 10.3389/fnsys.2012.00062
26
HutchisonR.WomelsdorfT.AllenE.BandettiniP.CalhounV.CorbettaM.et al (2013). Dynamic functional connectivity: Promise, issues, and interpretations.Neuroimage80360–378. 10.1016/j.neuroimage.2013.05.079
27
JenkinsonM.BannisterP.BradyM.SmithS. (2002). Improved optimization for the robust and accurate linear registration and motion correction of brain images.Neuroimage17825–841. 10.1016/s1053-8119(02)91132-8
28
KrasnerA.TurnerJ.FeldmanJ.SilbermanA.FisherP.WorkmanC.et al (2018). ADHD symptoms in a non-referred low birthweight/preterm cohort: Longitudinal profiles, outcomes, and associated features.J. Atten. Disord.22827–838. 10.1177/1087054715617532
29
LaheyB.PelhamW.LoneyJ.LeeS.WillcuttE. (2005). Instability of the DSM-IV subtypes of ADHD from preschool through elementary school.Arch. Gen. Psychiatry62896–902. 10.1001/archpsyc.62.8.896
30
LeiD.DuM.WuM.ChenT.HuangX.DuX.et al (2015). Functional MRI reveals different response inhibition between adults and children with ADHD.Neuropsychology29874–881. 10.1037/neu0000200
31
LiF.HeN.LiY.ChenL.HuangX.LuiS.et al (2014). Intrinsic brain abnormalities in attention deficit hyperactivity disorder: A resting-state functional MR imaging study.Radiology272514–523. 10.1148/radiol.14131622
32
LiaoW.LiJ.JiG.WuG.LongZ.XuQ.et al (2019). Endless fluctuations: Temporal dynamics of the amplitude of low frequency fluctuations.IEEE Trans. Med. Imaging382523–2532. 10.1109/tmi.2019.2904555
33
LouF.TaoJ.ZhouR.ChenS.QianA.YangC.et al (2021). Altered variability and concordance of dynamic resting-state fMRI indices in patients with attention deficit hyperactivity disorder.Front. Neurosci.15:731596. 10.3389/fnins.2021.731596
34
LurieD.KesslerD.BassettD.BetzelR.BreakspearM.KheilholzS.et al (2020). Questions and controversies in the study of time-varying functional connectivity in resting fMRI.Netw. Neurosci.430–69. 10.1162/netn_a_00116
35
MantiniD.PerrucciM.Del GrattaC.RomaniG.CorbettaM. (2007). Electrophysiological signatures of resting state networks in the human brain.Proc. Natl. Acad. Sci. U.S.A.10413170–13175. 10.1073/pnas.0700668104
36
McCarthyH.SkokauskasN.FrodlT. (2014). Identifying a consistent pattern of neural function in attention deficit hyperactivity disorder: A meta-analysis.Psychol. Med.44869–880. 10.1017/s0033291713001037
37
McLeodK.LangevinL.GoodyearB.DeweyD. (2014). Functional connectivity of neural motor networks is disrupted in children with developmental coordination disorder and attention-deficit/hyperactivity disorder.Neuroimage Clin.4566–575. 10.1016/j.nicl.2014.03.010
38
OberlinB.AlfordJ.MarroccoR. (2005). Normal attention orienting but abnormal stimulus alerting and conflict effect in combined subtype of ADHD.Behav. Brain Res.1651–11. 10.1016/j.bbr.2005.06.041
39
PenttonenM.BuzsákiG. (2003). Natural logarithmic relationship between brain oscillators.Thalam. Relat. Syst.2145–152. 10.1016/S1472-9288(03)00007-4
40
PretiM.BoltonT.Van De VilleD. (2017). The dynamic functional connectome: State-of-the-art and perspectives.Neuroimage16041–54. 10.1016/j.neuroimage.2016.12.061
41
RaichleM.MacLeodA.SnyderA.PowersW.GusnardD.ShulmanG. L. (2001). A default mode of brain function.Proc. Natl. Acad. Sci. U.S.A.98676–682. 10.1073/pnas.98.2.676
42
RheeS.WaldmanI.HayD.LevyF. (1999). Sex differences in genetic and environmental influences on DSM-III-R attention-deficit/hyperactivity disorder.J. Abnorm. Psychol.10824–41. 10.1037/0021-843x.108.1.24
43
RubiaK.CriaudM.WulffM.AlegriaA.BrinsonH.BarkerG.et al (2019). Functional connectivity changes associated with fMRI neurofeedback of right inferior frontal cortex in adolescents with ADHD.Neuroimage18843–58. 10.1016/j.neuroimage.2018.11.055
44
SalehinejadM.GhanavatiE.RashidM.NitscheM. (2021). Hot and cold executive functions in the brain: A prefrontal-cingular network.Brain Neurosci. Adv.5:23982128211007769. 10.1177/23982128211007769
45
SameaF.SolukiS.NejatiV.ZareiM.CorteseS.EickhoffS.et al (2019). Brain alterations in children/adolescents with ADHD revisited: A neuroimaging meta-analysis of 96 structural and functional studies.Neurosci. Biobehav. Rev.1001–8. 10.1016/j.neubiorev.2019.02.011
46
SongX.DongZ.LongX.LiS.ZuoX.ZhuC.et al (2011). REST: A toolkit for resting-state functional magnetic resonance imaging data processing.PLoS One6:e25031. 10.1371/journal.pone.0025031
47
TistarelliN.FagnaniC.TroianielloM.StaziM.AdrianiW. (2020). The nature and nurture of ADHD and its comorbidities: A narrative review on twin studies.Neurosci. Biobehav. Rev.10963–77. 10.1016/j.neubiorev.2019.12.017
48
VanniniP.AlmkvistO.FranckA.JonssonT.VolpeU.Kristoffersen WibergM.et al (2004). Task demand modulations of visuospatial processing measured with functional magnetic resonance imaging.Neuroimage2158–68. 10.1016/j.neuroimage.2003.09.033
49
WangT.LiuK.LiZ.XuY.LiuY.ShiW.et al (2017). Prevalence of attention deficit/hyperactivity disorder among children and adolescents in China: A systematic review and meta-analysis.BMC Psychiatry17:32. 10.1186/s12888-016-1187-9
50
WangJ.ZhengL.CaoQ.WangY.SunL.ZangY.et al (2017). Inconsistency in abnormal brain activity across cohorts of ADHD-200 in children with attention deficit hyperactivity disorder.Front. Neurosci.11:320. 10.3389/fnins.2017.00320
51
WangY.ZouQ.AoY.LiuY.OuyangY.WangX.et al (2020). Frequency-dependent circuits anchored in the dorsal and ventral left anterior insula.Sci. Rep.10:16394. 10.1038/s41598-020-73192-z
52
WangZ.LiuY.RuanX.LiY.LiE.ZhangG.et al (2020). Aberrant amplitude of low-frequency fluctuations in different frequency bands in patients with Parkinson’s disease.Front. Aging Neurosci.12:576682. 10.3389/fnagi.2020.576682
53
XueS.LiD.WengX.NorthoffG.LiD. (2014). Different neural manifestations of two slow frequency bands in resting functional magnetic resonance imaging: A systemic survey at regional, interregional, and network levels.Brain Connect.4242–255. 10.1089/brain.2013.0182
54
YanC.CraddockR.ZuoX.ZangY.MilhamM. (2013). Standardizing the intrinsic brain: Towards robust measurement of inter-individual variation in 1000 functional connectomes.Neuroimage80246–262. 10.1016/j.neuroimage.2013.04.081
55
YangL.YanY.LiY.HuX.LuJ.ChanP.et al (2020). Frequency-dependent changes in fractional amplitude of low-frequency oscillations in Alzheimer’s disease: A resting-state fMRI study.Brain Imaging Behav.142187–2201. 10.1007/s11682-019-00169-6
56
YangZ.CraddockR.MarguliesD.YanC.MilhamM. (2014). Common intrinsic connectivity states among posteromedial cortex subdivisions: Insights from analysis of temporal dynamics.Neuroimage93124–137. 10.1016/j.neuroimage.2014.02.014
57
YuR.ChienY.WangH.LiuC.LiuC.HwangT.et al (2014). Frequency-specific alternations in the amplitude of low-frequency fluctuations in schizophrenia.Hum. Brain Mapp.35627–637. 10.1002/hbm.22203
58
ZangY.HeY.ZhuC.CaoQ.SuiM.LiangM.et al (2007). Altered baseline brain activity in children with ADHD revealed by resting-state functional MRI.Brain Dev.2983–91. 10.1016/j.braindev.2006.07.002
59
ZangY.JiangT.LuY.HeY.TianL. (2004). Regional homogeneity approach to fMRI data analysis.Neuroimage22394–400. 10.1016/j.neuroimage.2003.12.030
60
ZelazoP.CarlsonS. (2012). Hot and cool executive function in childhood and adolescence: Development and plasticity.Child Dev. Perspect.6354–360. 10.1111/j.1750-8606.2012.00246.x
61
ZouQ.ZhuC.YangY.ZuoX.LongX.CaoQ.et al (2008). An improved approach to detection of amplitude of low-frequency fluctuation (ALFF) for resting-state fMRI: Fractional ALFF.J. Neurosci. Methods172137–141. 10.1016/j.jneumeth.2008.04.012
62
ZuoX.Di MartinoA.KellyC.ShehzadZ.GeeD.KleinD.et al (2010). The oscillating brain: Complex and reliable.Neuroimage491432–1445. 10.1016/j.neuroimage.2009.09.037
Summary
Keywords
attention deficit hyperactivity disorder, resting-state fMRI, ADHD subtype, frequency band, temporal dynamics
Citation
Chen R, Jiao Y, Zhu J-S, Wang X-H and Zhao M-T (2024) Frequency-specific static and dynamic neural activity indices in children with different attention deficit hyperactivity disorder subtypes: a resting-state fMRI study. Front. Hum. Neurosci. 18:1412572. doi: 10.3389/fnhum.2024.1412572
Received
05 April 2024
Accepted
17 July 2024
Published
09 August 2024
Volume
18 - 2024
Edited by
Weidong Cai, Stanford University, United States
Reviewed by
Weizheng Yan, National Institute on Alcohol Abuse and Alcoholism (NIH), United States
Jin Liu, Stanford University, United States
Updates
Copyright
© 2024 Chen, Jiao, Zhu, Wang and Zhao.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Yun Jiao, yunjiao@seu.edu.cn
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.