Deriving reproducible biomarkers from multi-site resting-state data: An Autism-based example.

Published on Feb 1, 2017in NeuroImage5.812
· DOI :10.1016/j.neuroimage.2016.10.045
Alexandre Abraham7
Estimated H-index: 7
(IRIA: French Institute for Research in Computer Science and Automation),
Michael P. Milham84
Estimated H-index: 84
(MIND Institute)
+ 4 AuthorsGaël Varoquaux34
Estimated H-index: 34
(IRIA: French Institute for Research in Computer Science and Automation)
Abstract Resting-state functional Magnetic Resonance Imaging (R-fMRI) holds the promise to reveal functional biomarkers of neuropsychiatric disorders. However, extracting such biomarkers is challenging for complex multi-faceted neuropathologies, such as autism spectrum disorders. Large multi-site datasets increase sample sizes to compensate for this complexity, at the cost of uncontrolled heterogeneity. This heterogeneity raises new challenges, akin to those face in realistic diagnostic applications. Here, we demonstrate the feasibility of inter-site classification of neuropsychiatric status, with an application to the Autism Brain Imaging Data Exchange (ABIDE) database, a large (N=871) multi-site autism dataset. For this purpose, we investigate pipelines that extract the most predictive biomarkers from the data. These R-fMRI pipelines build participant-specific connectomes from functionally-defined brain areas. Connectomes are then compared across participants to learn patterns of connectivity that differentiate typical controls from individuals with autism. We predict this neuropsychiatric status for participants from the same acquisition sites or different, unseen, ones. Good choices of methods for the various steps of the pipeline lead to 67% prediction accuracy on the full ABIDE data, which is significantly better than previously reported results. We perform extensive validation on multiple subsets of the data defined by different inclusion criteria. These enables detailed analysis of the factors contributing to successful connectome-based prediction. First, prediction accuracy improves as we include more subjects, up to the maximum amount of subjects available. Second, the definition of functional brain areas is of paramount importance for biomarker discovery: brain areas extracted from large R-fMRI datasets outperform reference atlases in the classification tasks.
Figures & Tables
  • References (76)
  • Citations (138)
📖 Papers frequently viewed together
751 Citations
8,955 Citations
115 Citations
78% of Scinapse members use related papers. After signing in, all features are FREE.
#1Shlomi HaarH-Index: 7
#2Sigal Berman (BGU: Ben-Gurion University of the Negev)H-Index: 15
Last. Ilan DinsteinH-Index: 20
view all 4 authors...
Substantial controversy exists regarding the presence and significance of anatomical abnormalities in autism spectrum disorders (ASD). The release of the Autism Brain Imaging Data Exchange (∼1000 participants, age 6–65 years) offers an unprecedented opportunity to conduct large-scale comparisons of anatomical MRI scans across groups and to resolve many of the outstanding questions. Comprehensive univariate analyses using volumetric, thickness, and surface area measures of over 180 anatomically d...
123 CitationsSource
#1Eva LothH-Index: 30
#2Will SpoorenH-Index: 27
Last. Declan G M Murphy D G MH-Index: 76
view all 38 authors...
60 CitationsSource
#1Simon B. Eickhoff (HHU: University of Düsseldorf)H-Index: 86
#2Bertrand ThirionH-Index: 37
Last. Danilo BzdokH-Index: 24
view all 4 authors...
Regional specialization and functional integration are often viewed as two fundamental principles of human brain organization. They are closely intertwined because each functionally specialized brain region is probably characterized by a distinct set of long-range connections. This notion has prompted the quickly developing family of connectivity-based parcellation (CBP) methods in neuroi-maging research. CBP assumes that there is a latent structure of parcels in a region of interest (ROI). Firs...
128 CitationsSource
#1William R. Shirer (Stanford University)H-Index: 11
#2Heidi Jiang (Stanford University)H-Index: 4
Last. Michael D. Greicius (Stanford University)H-Index: 42
view all 5 authors...
Abstract Resting-state functional magnetic resonance imaging (rs-fMRI) has become an increasingly important tool in mapping the functional networks of the brain. This tool has been used to examine network changes induced by cognitive and emotional states, neurological traits, and neuropsychiatric disorders. However, noise that remains in the rs-fMRI data after preprocessing has limited the reliability of individual-subject results, wherein scanner artifacts, subject movements, and other noise so...
68 CitationsSource
#1NeuroDataH-Index: 1
#2Craddock CH-Index: 1
44 CitationsSource
#1Mark Plitt (NIH: National Institutes of Health)H-Index: 8
#2Kelly Anne Barnes (NIH: National Institutes of Health)H-Index: 15
Last. Alex Martin (NIH: National Institutes of Health)H-Index: 111
view all 3 authors...
Abstract Objectives Autism spectrum disorders (ASD) are diagnosed based on early-manifesting clinical symptoms, including markedly impaired social communication. We assessed the viability of resting-state functional MRI (rs-fMRI) connectivity measures as diagnostic biomarkers for ASD and investigated which connectivity features are predictive of a diagnosis. Methods Rs-fMRI scans from 59 high functioning males with ASD and 59 age- and IQ-matched typically developing (TD) males were used to build...
87 CitationsSource
#1Pallavi RaneH-Index: 5
#2David M. CochranH-Index: 7
Last. Jean A. FrazierH-Index: 50
view all 6 authors...
Learning ObjectiveAfter participating in this activity, learners should be better able to:Assess the resting state and diffusion tensor imaging connectivity literature regarding subjects with autism spectrum disorder.AbstractAutism spectrum disorder (ASD) affects 1 in 50 children between the ages of
69 CitationsSource
#1Colleen P. Chen (SDSU: San Diego State University)H-Index: 4
#2Christopher L. Keown (SDSU: San Diego State University)H-Index: 15
Last. Ralph-Axel Müller (SDSU: San Diego State University)H-Index: 46
view all 7 authors...
Despite consensus on the neurological nature of autism spectrum disorders (ASD), brain biomarkers remain unknown and diagnosis continues to be based on behavioral criteria. Growing evidence suggests that brain abnormalities in ASD occur at the level of interconnected networks; however, previous attempts using functional connectivity data for diagnostic classification have reached only moderate accuracy. We selected 252 low-motion resting-state functional MRI (rs-fMRI) scans from the Autism Brain...
51 CitationsSource
#1Leanna M. Hernandez (UCLA: University of California, Los Angeles)H-Index: 13
#2Jeffrey D. Rudie (UCLA: University of California, Los Angeles)H-Index: 17
Last. Mirella Dapretto (UCLA: University of California, Los Angeles)H-Index: 46
view all 5 authors...
Neuroimaging investigations of autism spectrum disorders (ASDs) have advanced our understanding of atypical brain function and structure, and have recently converged on a model of altered network-level connectivity. Traditional task-based functional magnetic resonance imaging (MRI) and volume-based structural MRI studies have identified widespread atypicalities in brain regions involved in social behavior and other core ASD-related behavioral deficits. More recent advances in MR-neuroimaging met...
40 CitationsSource
#1Siddharth Ray (OHSU: Oregon Health & Science University)H-Index: 4
#2Meghan Miller (UC Davis: University of California, Davis)H-Index: 15
Last. Damien A. Fair (OHSU: Oregon Health & Science University)H-Index: 46
view all 12 authors...
Attention-deficit/hyperactive disorder (ADHD) and autism spectrum disorders (ASD) are two of the most common and vexing neurodevelopmental disorders among children. Although the two disorders share many behavioral and neuropsychological characteristics, most MRI studies examine only one of the disorders at a time. Using graph theory combined with structural and functional connectivity, we examined the large-scale network organization among three groups of children: a group with ADHD (8–12 years,...
77 CitationsSource
Cited By138
#1Xiaoxiao Li (Princeton University)H-Index: 1
#2James Scott Duncan (Yale University)
Understanding how certain brain regions relate to a specific neurological disorder or cognitive stimuli has been an important area of neuroimaging research. We propose BrainGNN, a graph neural network (GNN) framework to analyze functional magnetic resonance images (fMRI) and discover neurological biomarkers. In contrast to feedforward neural networks (FNN) and convolutional neural networks (CNN) in traditional functional connectivity-based fMRI analysis methods, we construct weighted graphs from...
#1Long Lu (WHU: Wuhan University)
#2Ye Chen (University of Cincinnati Academic Health Center)
Last. Tanya E. Froehlich (University of Cincinnati Academic Health Center)H-Index: 16
view all 8 authors...
BACKGROUND: Autism spectrum disorder (ASD) is a complex neurodevelopmental disorder with unknown etiology. Early diagnosis and intervention are the keys to improving outcomes for patients with ASD. Structural MRI (sMRI) has been widely used in clinic to facilitate the diagnosis of brain diseases such as brain tumors. However, sMRI is less frequently investigated in neurological and psychiatric disorders such as ASD due to subtle, if any, anatomical changes of the brain. In recent years, more and...
#1Barnaly Rashid (Harvard University)H-Index: 6
#2Vince D. Calhoun (Georgia Institute of Technology)H-Index: 93
Neuroimaging-based approaches have been extensively applied to study mental illness in recent years and have deepened our understanding of both cognitively healthy and disordered brain structure and function. Recent advancements in machine learning techniques have shown promising outcomes for individualized prediction and characterization of patients with psychiatric disorders. Studies have utilized features from a variety of neuroimaging modalities, including structural, functional, and diffusi...
#1Dan Jin (CAS: Chinese Academy of Sciences)H-Index: 1
#2Pan Wang (TJU: Tianjin University)
Last. Hongxiang Yao (Chinese PLA General Hospital)H-Index: 9
view all 22 authors...
Alzheimer's disease (AD) is associated with disruptions in brain activity and networks. However, there is substantial inconsistency among studies that have investigated functional brain alterations in AD; such contradictions have hindered efforts to elucidate the core disease mechanisms. In this study, we aim to comprehensively characterize AD-associated functional brain alterations using one of the world's largest resting-state functional MRI (fMRI) biobank for the disorder. The biobank include...
#1Katherine E. Lawrence (UCLA: University of California, Los Angeles)H-Index: 2
#2Leanna M. Hernandez (UCLA: University of California, Los Angeles)H-Index: 13
Last. Raphael Bernier (UW: University of Washington)H-Index: 44
view all 18 authors...
Autism spectrum disorder (ASD) is associated with the altered functional connectivity of 3 neurocognitive networks that are hypothesized to be central to the symptomatology of ASD: the salience network (SN), default mode network (DMN), and central executive network (CEN). Due to the considerably higher prevalence of ASD in males, however, previous studies examining these networks in ASD have used primarily male samples. It is thus unknown how these networks may be differentially impacted among f...
#1André Santos (UC: University of Coimbra)
#2Francisco Caramelo (UC: University of Coimbra)H-Index: 10
Last. Miguel Castelo-Branco (UC: University of Coimbra)H-Index: 29
view all 4 authors...
The neural basis of behavioural changes in Autism Spectrum Disorders (ASD) remains a controversial issue. One factor contributing to this challenge is the phenotypic heterogeneity observed in ASD, which suggests that several different system disruptions may contribute to diverse patterns of impairment between and within study samples. Here, we took a retrospective approach, using SFARI data to study ASD by focusing on participants with genetic imbalances targeting the dopaminergic system. Using ...
#1Seok-Jun Hong (MIND Institute)H-Index: 15
#2Joshua T. Vogelstein (Johns Hopkins University)H-Index: 26
Last. Adriana Di Martino (MIND Institute)H-Index: 44
view all 7 authors...
Abstract There is a general consensus that substantial heterogeneity underlies the neurobiology of autism spectrum disorder (ASD). As such, it has become increasingly clear that a dissection of variation at the molecular, cellular, and brain-network domains is a prerequisite for identifying biomarkers. Neuroimaging has been widely used to characterize atypical brain patterns in ASD, though findings have varied across studies. This is, at least in part, due to a failure to account for neurobiolog...
#1Xin-Bo Wang (Liaoning Normal University)H-Index: 1
#2Qing Wang (Liaoning Normal University)
Last. Dong-Qiang Liu (Liaoning Normal University)H-Index: 12
view all 6 authors...
It has been reported that resting state fluctuation amplitude (RSFA) exhibits extremely large inter-site variability, which limits its application in multisite studies. Although global normalization (GN) based approaches are efficient in reducing the site effects, they may cause spurious results. In this study, our purpose was to find alternative strategies to minimize the substantial site effects for RSFA, without the risk of introducing artificial findings. We firstly modified the ALFF algorit...
#1Kun Zhao (CAS: Chinese Academy of Sciences)H-Index: 27
#2Yanhui Ding (SDNU: Shandong Normal University)H-Index: 1
Last. Chengyuan Song (SDU: Shandong University)
view all 24 authors...
Abstract Hippocampal morphological change is one of the main hallmarks of Alzheimer's disease (AD). However, whether hippocampal radiomic features are robust as predictors of progression from mild cognitive impairment (MCI) to AD dementia and whether these features provide any neurobiological foundation remains unclear. The primary aim of this study was to verify whether hippocampal radiomic features can serve as robust MRI markers for AD. Multivariate classifier-based SVM analysis provided indi...
1 CitationsSource