CoChat
Sign Up
Cited by 199Functional Brain Connectivity StudiesAttention Deficit Hyperactivity DisorderAutism Spectrum Disorder Research

The DOI matches its Crossref record; 41 of the 41 references it cites were found in a registry or an index.

A general prediction model for the detection of ADHD and Autism using structural and functional MRI

Open access
Sen, Bhaskar, Borle, Neil C., Greiner, Russell, Brown, Matthew R. G.·2018·PLOS ONE
DOI 10.1371/journal.pone.0194856PMC PMC5903601
Open this paper in CoChat — free accountOpen PDF

Abstract

This work presents a novel method for learning a model that can diagnose Attention Deficit Hyperactivity Disorder (ADHD), as well as Autism, using structural texture and functional connectivity features obtained from 3-dimensional structural magnetic resonance imaging (MRI) and 4-dimensional resting-state functional magnetic resonance imaging (fMRI) scans of subjects. We explore a series of three learners: (1) The LeFMS learner first extracts features from the structural MRI images using the texture-based filters produced by a sparse autoencoder. These filters are then convolved with the original MRI image using an unsupervised convolutional network. The resulting features are used as input to a linear support vector machine (SVM) classifier. (2) The LeFMF learner produces a diagnostic model by first computing spatial non-stationary independent components of the fMRI scans, which it uses to decompose each subject's fMRI scan into the time courses of these common spatial components. These features can then be used with a learner by themselves or in combination with other features to produce the model. Regardless of which approach is used, the final set of features are input to a linear support vector machine (SVM) classifier. (3) Finally, the overall LeFMSF learner uses the combined features obtained from the two feature extraction processes in (1) and (2) above as input to an SVM classifier, achieving an accuracy of 0.673 on the ADHD-200 holdout data and 0.643 on the ABIDE holdout data. Both of these results, obtained with the same LeFMSF framework, are the best known, over all hold-out accuracies on these datasets when only using imaging data-exceeding previously-published results by 0.012 for ADHD and 0.042 for Autism. Our results show that combining multi-modal features can yield good classification accuracy for diagnosis of ADHD and Autism, which is an important step towards computer-aided diagnosis of these psychiatric diseases and perhaps others as well.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1371/journal.pone.0194856

Cites

41 of 41 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig 1 Overall pipeline, including feature extraction steps, for each of our methods: LeFMSF uses all of these steps (as it uses both MRI and fMRI data); LeFMS does not use the red “fMRI feature extraction” box; and LeFMF does not use the grey “MRI feature extraction” box. In each case, in addition to the learner that utilizes the combined features, our system also learns the most effective features from the MRI data (grey box) and from the fMRI data (red box); those features are used to pre-process the test data (see boxes of the same color).
Fig 1 Overall pipeline, including feature extraction steps, for each of our methods: LeFMSF uses all of these steps (as it uses both MRI and fMRI data); LeFMS does not use the red “fMRI feature extraction” box; and LeFMF does not use the grey “MRI feature extraction” box. In each case, in addition to the learner that utilizes the combined features, our system also learns the most effective features from the MRI data (grey box) and from the fMRI data (red box); those features are used to pre-process the test data (see boxes of the same color).CC BY · PubMed Central
Fig 2 Process of extracting MRI features to be provided to a learning algorithm. This expands the “MRI feature extraction” box from Fig 1.
Fig 2 Process of extracting MRI features to be provided to a learning algorithm. This expands the “MRI feature extraction” box from Fig 1.CC BY · PubMed Central
Fig 3 Autoencoder that tries to reconstruct the 5 × 5 × 5 patches (represented as a 125-D vector). For each of the k = 3 filters, there is a link (“Encoding Weights” θ1) from each input node to each hidden node, and then from each hidden node to each Output node (“Decoding Weights”, θ2); we show only a subset for clarity.
Fig 3 Autoencoder that tries to reconstruct the 5 × 5 × 5 patches (represented as a 125-D vector). For each of the k = 3 filters, there is a link (“Encoding Weights” θ1) from each input node to each hidden node, and then from each hidden node to each Output node (“Decoding Weights”, θ2); we show only a subset for clarity.CC BY · PubMed Central
Fig 4 (a) Median axial slice from MRI scan and (b) the result after convolution with a filter.
Fig 4 (a) Median axial slice from MRI scan and (b) the result after convolution with a filter.CC BY · PubMed Central
Fig 5 Detailed process of how a classifier is produced and used with fMRI scans. This corresponds to the “Red Box” (i.e., the fMRI processing component) of Fig 1.
Fig 5 Detailed process of how a classifier is produced and used with fMRI scans. This corresponds to the “Red Box” (i.e., the fMRI processing component) of Fig 1.CC BY · PubMed Central
Fig 6 Detailed version of the fMRI data to the spatial maps process that elaborates the boxes “Dimensionality Reduction with 2-step PCA” and “Obtain Spatial maps Using PCA/k-PCA/ICA/NSD” in Fig 5. The first three of four steps show how 2-step PCA is applied to the data, where the concatenation step is temporal concatenation (Fig 7). The first column shows that the scan from subject i is a T × V matrix I[T×V]ⁱ. The next column, after “Subject Level PCA”, shows the ith subject is described as Y[Tred×V]ⁱ, with dimension Tred × V. The subsequent “Concatentation” step produces a group matrix with dimension nTred × V. The “Group Level PCA” step reduces this to a K × V matrix, which the “Source Separation” step factorizes to produce the mixing matrix size K × C. This final step could be the identify transformation, kernel-PCA, standard ICA or non-linear ICA; each of these might lead to different values of C.
Fig 6 Detailed version of the fMRI data to the spatial maps process that elaborates the boxes “Dimensionality Reduction with 2-step PCA” and “Obtain Spatial maps Using PCA/k-PCA/ICA/NSD” in Fig 5. The first three of four steps show how 2-step PCA is applied to the data, where the concatenation step is temporal concatenation (Fig 7). The first column shows that the scan from subject i is a T × V matrix I[T×V]ⁱ. The next column, after “Subject Level PCA”, shows the ith subject is described as Y[Tred×V]ⁱ, with dimension Tred × V. The subsequent “Concatentation” step produces a group matrix with dimension nTred × V. The “Group Level PCA” step reduces this to a K × V matrix, which the “Source Separation” step factorizes to produce the mixing matrix size K × C. This final step could be the identify transformation, kernel-PCA, standard ICA or non-linear ICA; each of these might lead to different values of C.CC BY · PubMed Central
Fig 7 Temporal vs. spatial concatenation. Each ‘fMRI data i” box (for i = 1..n) corresponds to one of the n subjects. This shows two ways we can concatenate these boxes from subjects in Fig 6 to produce X, Here, we use temporal concatenation.
Fig 7 Temporal vs. spatial concatenation. Each ‘fMRI data i” box (for i = 1..n) corresponds to one of the n subjects. This shows two ways we can concatenate these boxes from subjects in Fig 6 to produce X, Here, we use temporal concatenation.CC BY · PubMed Central
Fig 8 Temporal concatenation with ICA elaborated from Fig 7. This corresponds to the ICA decomposition X ≈ AS, except here, for illustrative purposes, we have replaced X with Y (the concatenated fMRI data from Fig 6 before the Group-level PCA).
Fig 8 Temporal concatenation with ICA elaborated from Fig 7. This corresponds to the ICA decomposition X ≈ AS, except here, for illustrative purposes, we have replaced X with Y (the concatenated fMRI data from Fig 6 before the Group-level PCA).CC BY · PubMed Central
Fig 9 5-fold cross validation results of classification.
Fig 9 5-fold cross validation results of classification.CC BY · PubMed Central
Fig 10 Spatial maps (aka components) 1–25 for ADHD-200 using NSD. Each 3×3 box shows 9 axial slices of one component. The colorbar is same as Fig 4.
Fig 10 Spatial maps (aka components) 1–25 for ADHD-200 using NSD. Each 3×3 box shows 9 axial slices of one component. The colorbar is same as Fig 4.CC BY · PubMed Central
Fig 11 Spatial maps or components 26–45 for ADHD-200 using NSD. Each component is shown in a box and 9 axial slices are shown. The colorbar is same as Fig 4.
Fig 11 Spatial maps or components 26–45 for ADHD-200 using NSD. Each component is shown in a box and 9 axial slices are shown. The colorbar is same as Fig 4.CC BY · PubMed Central
Fig 12 Spatial maps or components 1-25 for ABIDE using NSD. Each component is shown in a box and 9 axial slices are shown. The colorbar is same as Fig 4.
Fig 12 Spatial maps or components 1-25 for ABIDE using NSD. Each component is shown in a box and 9 axial slices are shown. The colorbar is same as Fig 4.CC BY · PubMed Central
Fig 13 Spatial maps or components 26-45 for ABIDE using NSD. Each component is shown in a box and 9 axial slices are shown. The colorbar is same as Fig 4.
Fig 13 Spatial maps or components 26-45 for ABIDE using NSD. Each component is shown in a box and 9 axial slices are shown. The colorbar is same as Fig 4.CC BY · PubMed Central
Fig 14 Permutation tests.
Fig 14 Permutation tests.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 41 references

Research works printed but not linked: 2

  1. Ghiassian S, Greiner R, Jin P, Brown M. Learning to classify psychiatric disorders based on fMR images: Autism vs healthy and ADHD vs healthy. In: Proceedings of 3rd NIPS Workshop on Machine Learning and Interpretation in NeuroImaging; 2013.
  2. Souloumiac A. Blind source detection and separation using second order non-stationarity. In: Acoustics, Speech, and Signal Processing, 1995. ICASSP-95., 1995 International Conference on. vol. 3. IEEE; 1995. p. 1912–1915. · DOI

Cited by

Sorted by citations

See all 199

Related

Metadata

Authors
Sen, B., Borle, N. C., Greiner, R., & Brown, M. R. G.
Affiliations
University of Alberta (CA)
Corresponding author
Brown, Matthew R. G.
ORCID
Venue
PLOS ONE
Publisher
Public Library of Science (PLoS)
Published
2018-04-17
Volume
13
Issue
4
Pages
e0194856
ISSN
1932-6203 (ISSN-L)
DOI
10.1371/journal.pone.0194856
PMC
PMC5903601
OpenAlex
W2801745895
Cited by
198
References
41
Citations by year
2026: 9 · 2025: 20 · 2024: 28 · 2023: 25 · 2022: 31 · 2021: 31 · 2020: 36 · 2019: 13 · 2018: 5
FWCI
12.79
Type
journal article
Language
en
Open access
Licence
Locations
  • PLoS ONE (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
Funders
  • Canadian Institutes of Health Research (CSE125741)
  • Natural Sciences and Engineering Research Council of Canada (NSERC) (203245)
  • Alberta Machine Intelligence Institute (AMII)
Topics
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Functional Brain Connectivity Studies
  • Health Sciences › Medicine › Psychiatry and Mental health › Attention Deficit Hyperactivity Disorder
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Autism Spectrum Disorder Research
Keywords
ADHD diagnosis, autism diagnosis, structural MRI, resting-state fMRI, functional connectivity, texture features, sparse autoencoders, independent component analysis, support vector machine, computer-aided diagnosis, multimodal fusion, mental illness
MeSH terms
Adolescent, Adult, Attention Deficit Disorder with Hyperactivity, Autistic Disorder, Brain, Child, Decision Support Techniques, Diagnosis, Computer-Assisted, Female, Humans, Image Processing, Computer-Assisted, Magnetic Resonance Imaging, Male, Prognosis, Young Adult, Datasets as Topic

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review