From estimating activation locality to predicting disorder: a review of pattern recognition for neuroimaging-based psychiatric diagnostics
Open accessAbstract
Psychiatric disorders are increasingly being recognised as having a biological basis, but their diagnosis is made exclusively behaviourally. A promising approach for 'biomarker' discovery has been based on pattern recognition methods applied to neuroimaging data, which could yield clinical utility in future. In this review we survey the literature on pattern recognition for making diagnostic predictions in psychiatric disorders, and evaluate progress made in translating such findings towards clinical application. We evaluate studies on many criteria, including data modalities used, the types of features extracted and algorithm applied. We identify problems common to many studies, such as a relatively small sample size and a primary focus on estimating generalisability within a single study. Furthermore, we highlight challenges that are not widely acknowledged in the field including the importance of accommodating disease prevalence, the necessity of more extensive validation using large carefully acquired samples, the need for methodological innovations to improve accuracy and to discriminate between multiple disorders simultaneously. Finally, we identify specific clinical contexts in which pattern recognition can add value in the short to medium term.
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.
Integrity signals
Registry
Cites
Statistics
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.
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?
Multimodal multivariate approaches including several imaging modalities and non-imaging data are likely to achieve superior classification accuracy than univariate approaches.
“Multimodal multivariate approaches including several imaging modalities, including functional and structural imaging data as well as non-imaging data such as cognitive and genetic measures are likely to achieve superior classification accuracy than univariate approaches (Orrù et al., 2012; Wolfers et al., 2015).”
From Cognitive Neuroscience of Attention Deficit Hyperactivity Disorder (ADHD) and Its Clinical Translation (2018) · cites this paper as Wolfers et al., 2015
Section 4.3 recommends combining different data types, including genetics, and discusses multimodal techniques, but says their potential benefits depend on the information in the data; it does not predict superior classification accuracy. Section 2.6 suggests multivariate estimates may be more sensitive to spatially distributed effects than univariate approaches, which is not the same as demonstrating or predicting higher classification accuracy. Page numbers were not available in our copy. (LLM, conf=0.83)
Experimental — accuracy not yet measuredMultivariate pattern analyses have been shown to provide sensitive and specific diagnostic indicators for individual patients with other pathologies, in particular neurological disorders such as Alzheimer's disease but also autism and depression.
“These methods have been shown to provide sensitive and specific diagnostic indicators for individual patients with other pathologies in particular for neurological disorders such as Alzheimer’s disease but also for autism and depression (for review, see Orrù et al., 2012; Wolfers et al., 2015).”
From Cognitive Neuroscience of Attention Deficit Hyperactivity Disorder (ADHD) and Its Clinical Translation (2018) · cites this paper as Wolfers et al., 2015
Sections 3.2 and 3.5 report promising pattern-recognition classification accuracies for depression and autism. The copy does not establish comparable diagnostic performance for Alzheimer’s disease; the Alzheimer’s classification studies found in the search appear in the References. Section 4.1 also cautions that considerable work remains before these methods are applicable to individual patients in clinical practice. (LLM, conf=0.86)
Experimental — accuracy not yet measuredClassification results were very low with accuracies not exceeding 61% on the test sample, and clinical data were more predictive of ADHD classification than imaging data.
“The classification results were very low with accuracies not exceeding 61% on the test sample and clinical data were more predictive of ADHD classification than imaging data (Wolfers et al., 2015).”
From Cognitive Neuroscience of Attention Deficit Hyperactivity Disorder (ADHD) and Its Clinical Translation (2018) · cites this paper as Wolfers et al., 2015
Section 3.4 (page unavailable) reports that classifiers in the ADHD-200 competition did not exceed 61% accuracy on a held-out test sample. It says the classifiers used inputs ranging from demographic tables to MRI, but does not say clinical data predicted ADHD better than imaging data. The 61% figure applies to ADHD-200, not to every ADHD classification study reviewed. (LLM, conf=0.89)
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
The citation lists haven’t been fetched yet
References 176 · Cited by 290
Research works printed but not linked: 4
- Abdulkadir, A., Peter, J., Brox, T., Ronneberger, O., Klöppel, S., 2014. Voxel-based multi-class classification of AD, MCI, and elderly controls: Blind evaluation on an independent test set. In: Proc MICCAI Workshop Challenge on Computer-Aided Diagnosis of Dementia Based on Structural MRI Data, pp. 8-15.
- Craddock, R.C., Holtzheimer, P.E., Hu, X., Mayberg, H.S., 2013. Disease state prediction form resting state functional connectivity. Magn. Reson. Med. 62, 1619-1628, http://dx.doi.org/10.1002/mrm.22159.Disease. · DOI
- Frangou, S., Schwarz, E., Meyer-Lindenberg, A., Imaging Genetics for Mental Disorders (IMAGEMEND). World Psychiatry, in press.
- Zhu, D., Shen, D., Jiang, X., Liu, T., 2014. Connectomics signature for characterization of mild cognitive impairment and schizophrenia. In: Proc IEEE IntSymp Biomed Imaging, pp. 325-328, http://dx.doi.org/10.1109/ISBI.2014. 6867874.CONNECTOMICS. · DOI
- Authors
- Wolfers, T., Buitelaar, J. K., Beckmann, C. F., Franke, B., & Marquand, A. F.
- Affiliations
- Radboud University Nijmegen (NL)
- Corresponding author
- Wolfers, Thomas
- Venue
- Neuroscience & Biobehavioral Reviews
- Publisher
- Elsevier BV
- Published
- 2015-10
- Volume
- 57
- Pages
- 328-349
- ISSN
- 0149-7634 (ISSN-L)
- OpenAlex
- W1097741600
- Cited by
- 290
- References
- 176
- Citations by year
- 2026: 6 · 2025: 23 · 2024: 26 · 2023: 19 · 2022: 19 · 2021: 20 · 2020: 27 · 2019: 41 · 2018: 34 · 2017: 32 · 2016: 41 · 2015: 2
- FWCI
- 3.61
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- Neuroscience & Biobehavioral Reviews (publishedVersion, cc-by-nc-nd)
- PubMed (publishedVersion)
- Radboud Repository (Radboud University) (submittedVersion)
- Radboud Repository (Radboud University) (submittedVersion, other-oa)
- Funders
- Nederlandse Organisatie voor Wetenschappelijk Onderzoek (433-09-242, 016-130-669, 056-13-015)
- Seventh Framework Programme (602450, 602805, 603016, 278948)
- Topics
- Life Sciences › Neuroscience › Cognitive Neuroscience › Functional Brain Connectivity Studies
- Health Sciences › Medicine › Psychiatry and Mental health › Dementia and Cognitive Impairment Research
- Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
- Keywords
- pattern recognition, neuroimaging, psychiatric diagnosis, biomarkers, machine learning, clinical translation, disease prevalence, generalizability, sample size, feature extraction
- MeSH terms
- Mental Disorders, Humans, Pattern Recognition, Automated, Neuroimaging
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
