BCI-based intervention re-normalizes brain functional network topology in children with ADHD

Submission No:


Submission Type:

Abstract Submission 


Xing Qian1, Francisco Castellanos2, Lucina Uddin3, Beatrice Rui-Yi Loo4, Hui Li Koh5, Daniel Fung6, Michael Chee4, Tih-Shih Lee5, Choon Guan Lim7, Juan Zhou8


1Duke-NUS medical school, National university of Singapore, Singapore, Singapore, 2NYU Child Study Center, NYU Langone Medical Center, New York City, United States, 3Department of Psychology, University of Miami, Miami, United States, 4Duke-NUS Medical School, Singapore, Singapore, 5Duke-National University of Singapore Graduate Medical School, Singapore, Singapore, 6Department of Child and Adolescent Psychiatry, Institute of Mental Health, Singapore, Singapore, Singapore, 7Department of Child and Adolescent Psychiatry, Institute of Mental Health, Singapore, Singapore, 8Duke-National University of Singapore Medical School, Singapore, Singapore


Attention deficit/hyperactivity disorder (ADHD) is one of the most commonly diagnosed neuropsychiatric disorders of childhood and can be difficult to treat. Studies have found that children with ADHD show abnormal small-world architecture of brain functional network characterized by higher local efficiency and lower global efficiency, suggesting a developmental lag of brain maturation [2]. Recently, a brain-computer-interface (BCI) based attention training game system has shown perspectives for treating children with significant inattentive symptoms [1]. However, how the brain network changes relate to the behavior improvement following this treatment in ADHD children remain unknown. To cover this gap, we aimed to examine the topological alterations of large-scale brain functional networks induced by the BCI-based attention intervention and its clinical relevance in ADHD children by using resting state functional magnetic resonance imaging (rsfMRI) method.


ADHD patients were randomly divided into two groups: 8-week BCI intervention group (ADHD-I) and non-intervention group (ADHD-NI). The BCI intervention method was described in our previous work [1]. MRI data and clinical assessment (Child Behavior Checklist and ADHD Rating Scales) at baseline and follow-up were obtained for all subjects. After quality control, 18 subjects from ADHD-I group (mean (SD) age: 9.00 (1.50) years) and 11 subjects from ADHD-NI group (mean (SD) age: 9.45 (1.29) years) have good neuroimaging data at both time points.
RsfMRI and structural MRI data were preprocessed using a standard pipeline [3]. The two groups were matched in demographic, motion and number of frames. Region-of-interest (ROI) time series were extracted using a 144 ROI rsfMRI-based brain functional parcellation scheme [4]. A correlation matrix of functional connectivity (FC) between all pairs of ROIs was generated [3]. Graph theoretical metrics including degree centrality, clustering coefficient and closeness were derived from individual FC matrices. We examined the group-time interaction effect on the intra- and inter-network FC strength and global and nodal graph metrics using two-way repeated ANOVA (alpha level of 0.05). We also sought to test whether changes in FC measures related to clinical improvement over time.


Compared to the ADHD-NI group, the ADHD-I group showed greater reduction of inattention symptoms (inattention scores of ADHD Rating Scales) after 8-week BCI-based training (p = 0.038). Following the BCI intervention, the ADHD-I group had greater FC reductions in the salience/ventral attention network (SN) compared to the ADHD-NI group. For inter-network FC, the ADHD-I group showed greater FC reductions between the SN and dorsal attention (DAN), somatomotor (SMN) and subcortical network, as well as between the executive control (ECN) and SMN (fig. 1). More importantly, such decrease in intra-SN FC and inter-network FC between SN and DAN were associated with behavior improvement of internalizing problems across all ADHD patients.
Moreover, the graph theoretical analysis revealed that following BCI intervention, the nodal degree centrality and clustering coefficient were reduced and the nodal closeness was increased in the SN, ECN, default mode, DAN, SMN and visual networks (fig. 2). Similarly, these nodal topological changes (especially in the SN) related to improvement of inattention and internalizing problems in all ADHD patients.
Supporting Image: fig1.jpg
   ·figure 1
Supporting Image: fig2-2.jpg
   ·figure 2


The BCI-based attention training could help renormalize salience processing and accelerate brain maturation in ADHD children by enhancing network functional specialization and segregation, supporting behavior improvement. Our findings shed light on the neural mechanism underlying the effective BCI-based attention training for ADHD childhood. Future long-term longitudinal neuroimaging studies are needed to further develop the BCI-based intervention approach for personalized therapy and treatment monitoring.

Disorders of the Nervous System:

Other Psychiatric Disorders 2

Imaging Methods:


Modeling and Analysis Methods:

fMRI Connectivity and Network Modeling

Motor Behavior:

Brain Machine Interface

Perception and Attention:

Attention: Visual


Attention Deficit Disorder
Other - functional connectivity, brain-computer-interface

1|2Indicates the priority used for review