Accuracy of diagnostic classification algorithms using cognitive-, electrophysiological-, and neuroanatomical data in antipsychotic-naïve schizophrenia patients.
Adult
Algorithms
Antipsychotic Agents
Brain
/ diagnostic imaging
Cognition Disorders
/ diagnosis
Diffusion Tensor Imaging
Evoked Potentials
Female
Humans
Logistic Models
Machine Learning
Magnetic Resonance Imaging
Male
Multivariate Analysis
Neuropsychological Tests
Psychiatric Status Rating Scales
Schizophrenia
/ classification
Schizophrenic Psychology
Young Adult
Antipsychotic-naïve first-episode schizophrenia
cognition
diffusion tensor imaging
electrophysiology
machine learning
structural magnetic resonance imaging
Journal
Psychological medicine
ISSN: 1469-8978
Titre abrégé: Psychol Med
Pays: England
ID NLM: 1254142
Informations de publication
Date de publication:
12 2019
12 2019
Historique:
pubmed:
19
12
2018
medline:
4
9
2020
entrez:
19
12
2018
Statut:
ppublish
Résumé
A wealth of clinical studies have identified objective biomarkers, which separate schizophrenia patients from healthy controls on a group level, but current diagnostic systems solely include clinical symptoms. In this study, we investigate if machine learning algorithms on multimodal data can serve as a framework for clinical translation. Forty-six antipsychotic-naïve, first-episode schizophrenia patients and 58 controls underwent neurocognitive tests, electrophysiology, and magnetic resonance imaging (MRI). Patients underwent clinical assessments before and after 6 weeks of antipsychotic monotherapy with amisulpride. Nine configurations of different supervised machine learning algorithms were applied to first estimate the unimodal diagnostic accuracy, and next to estimate the multimodal diagnostic accuracy. Finally, we explored the predictability of symptom remission. Cognitive data significantly classified patients from controls (accuracies = 60-69%; p values = 0.0001-0.009). Accuracies of electrophysiology, structural MRI, and diffusion tensor imaging did not exceed chance level. Multimodal analyses with cognition plus any combination of one or more of the remaining three modalities did not outperform cognition alone. None of the modalities predicted symptom remission. In this multivariate and multimodal study in antipsychotic-naïve patients, only cognition significantly discriminated patients from controls, and no modality appeared to predict short-term symptom remission. Overall, these findings add to the increasing call for cognition to be included in the definition of schizophrenia. To bring about the full potential of machine learning algorithms in first-episode, antipsychotic-naïve schizophrenia patients, careful a priori variable selection based on independent data as well as inclusion of other modalities may be required.
Sections du résumé
BACKGROUND
A wealth of clinical studies have identified objective biomarkers, which separate schizophrenia patients from healthy controls on a group level, but current diagnostic systems solely include clinical symptoms. In this study, we investigate if machine learning algorithms on multimodal data can serve as a framework for clinical translation.
METHODS
Forty-six antipsychotic-naïve, first-episode schizophrenia patients and 58 controls underwent neurocognitive tests, electrophysiology, and magnetic resonance imaging (MRI). Patients underwent clinical assessments before and after 6 weeks of antipsychotic monotherapy with amisulpride. Nine configurations of different supervised machine learning algorithms were applied to first estimate the unimodal diagnostic accuracy, and next to estimate the multimodal diagnostic accuracy. Finally, we explored the predictability of symptom remission.
RESULTS
Cognitive data significantly classified patients from controls (accuracies = 60-69%; p values = 0.0001-0.009). Accuracies of electrophysiology, structural MRI, and diffusion tensor imaging did not exceed chance level. Multimodal analyses with cognition plus any combination of one or more of the remaining three modalities did not outperform cognition alone. None of the modalities predicted symptom remission.
CONCLUSIONS
In this multivariate and multimodal study in antipsychotic-naïve patients, only cognition significantly discriminated patients from controls, and no modality appeared to predict short-term symptom remission. Overall, these findings add to the increasing call for cognition to be included in the definition of schizophrenia. To bring about the full potential of machine learning algorithms in first-episode, antipsychotic-naïve schizophrenia patients, careful a priori variable selection based on independent data as well as inclusion of other modalities may be required.
Identifiants
pubmed: 30560750
pii: S0033291718003781
doi: 10.1017/S0033291718003781
pmc: PMC6877469
doi:
Substances chimiques
Antipsychotic Agents
0
Banques de données
ClinicalTrials.gov
['NCT01154829']
Types de publication
Controlled Clinical Trial
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
2754-2763Références
Neuroimage. 2017 Jan 15;145(Pt B):166-179
pubmed: 27989847
Biol Psychiatry. 1982 Jun;17(6):639-54
pubmed: 7104417
Nature. 2016 Oct 05;538(7623):20-23
pubmed: 27708329
Neuroinformatics. 2011 Dec;9(4):321-33
pubmed: 21246418
Schizophr Res. 2018 Feb;192:479-480
pubmed: 28576547
Psychol Med. 2013 Dec;43(12):2547-62
pubmed: 23507081
Schizophr Res. 2015 Jan;161(1):19-28
pubmed: 24893909
Am J Psychiatry. 2010 Jul;167(7):748-51
pubmed: 20595427
Schizophr Bull. 2018 Oct 17;44(6):1332-1340
pubmed: 29373756
Arch Gen Psychiatry. 2012 Dec;69(12):1195-204
pubmed: 22868877
Neuropsychopharmacology. 2015 Jun;40(7):1742-51
pubmed: 25601228
Schizophr Res. 2004 Jun 1;68(2-3):283-97
pubmed: 15099610
Schizophr Bull. 2014 Jul;40(4):878-85
pubmed: 23934819
Transl Psychiatry. 2017 Apr 11;7(4):e1087
pubmed: 28398342
World J Biol Psychiatry. 2015;16(5):280-90
pubmed: 26213111
Neuropsychology. 2009 May;23(3):315-36
pubmed: 19413446
Biol Psychiatry. 1991 Nov 15;30(10):1059-62
pubmed: 1756198
IEEE Trans Biomed Eng. 2017 Feb;64(2):395-407
pubmed: 28113193
Psychol Med. 2019 Apr;49(5):754-763
pubmed: 29734953
Schizophr Bull. 2013 Sep;39(5):1129-38
pubmed: 23042112
Dementia. 1994 Sep-Oct;5(5):266-81
pubmed: 7951684
Schizophr Bull. 1987;13(2):261-76
pubmed: 3616518
Schizophr Bull. 2015 Sep;41(5):1143-52
pubmed: 25698711
Neuroimage. 2006 Jul 15;31(4):1487-505
pubmed: 16624579
Neuroimage. 2012 Aug 15;62(2):782-90
pubmed: 21979382
Psychiatry (Edgmont). 2007 Jul;4(7):28-37
pubmed: 20526405
Mol Psychiatry. 2012 Dec;17(12):1174-9
pubmed: 22869033
JAMA Psychiatry. 2013 Oct;70(10):1107-12
pubmed: 23925787
Behav Neurol. 2016;2016:7849526
pubmed: 27843197
Biol Psychiatry. 2012 May 15;71(10):898-905
pubmed: 22418013
Neurosci Biobehav Rev. 2012 Apr;36(4):1342-56
pubmed: 22244985
Am J Psychiatry. 2005 Mar;162(3):441-9
pubmed: 15741458
Comput Math Methods Med. 2013;2013:867924
pubmed: 24489603
Acta Psychiatr Scand. 2012 Jul;126(1):59-71
pubmed: 22384856
Arch Gen Psychiatry. 1990 Feb;47(2):181-8
pubmed: 2405807
Schizophr Res. 2017 Dec 2;:null
pubmed: 29208422
Int J Neuropsychopharmacol. 2015 Oct 09;19(3):pyv109
pubmed: 26453696
J Psychiatry Neurosci. 2016 Mar;41(2):133-41
pubmed: 26599135
PLoS One. 2016 Oct 10;11(10):e0164464
pubmed: 27723782
Curr Opin Psychiatry. 2013 Mar;26(2):172-87
pubmed: 23324948
Schizophr Bull. 2007 Jan;33(1):49-68
pubmed: 17101692
Schizophr Res. 2007 Mar;91(1-3):87-96
pubmed: 17306506
Neuropsychopharmacology. 2014 Dec;39(13):3000-8
pubmed: 24954063
Cortex. 1978 Jun;14(2):234-44
pubmed: 679704
Dialogues Clin Neurosci. 2014 Dec;16(4):491-503
pubmed: 25733954
World Psychiatry. 2016 Feb;15(1):26-31
pubmed: 26833601
Schizophr Res. 2017 Dec 20;:null
pubmed: 29274736
Arch Gen Psychiatry. 1990 Jun;47(6):589-93
pubmed: 2190539
Schizophr Res. 2016 Aug;175(1-3):72-78
pubmed: 27117677
Am J Psychiatry. 2016 Mar 1;173(3):232-43
pubmed: 26621570
Neurosci Biobehav Rev. 2017 Sep;80:586-604
pubmed: 28757454
Psychol Med. 2018 Jun;48(8):1325-1340
pubmed: 29094675
Biol Psychiatry. 2001 Jan 1;49(1):71-7
pubmed: 11163782
Schizophr Res. 2017 Mar;181:6-12
pubmed: 27613509