Accuracy of diagnostic classification algorithms using cognitive-, electrophysiological-, and neuroanatomical data in antipsychotic-naïve schizophrenia patients

Ebdrup, Bjørn H; Axelsen, Martin Christian; Bak, Nikolaj; Fagerlund, Birgitte; Oranje, Bob; Raghava, Jayachandra Mitta; Nielsen, Mette Ødegaard; Rostrup, Egill; Hansen, Lars Kai; Glenthøj, Birte

doi:10.1017/s0033291718003781

Cited by 19 publications

(30 citation statements)

References 58 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some other studies did not find an impairment in sensory gating in FEP (150)(151)(152)(153), FES (154,155), and HR (150,151,154,156,157) subjects. Furthermore, a study involving the innovative implementation of machine learning (ML) to distinguish FES from HCs with P50-related measures (amplitude and ratio), in addition to other neuroimaging and clinical evaluations, highlighted that this EEG-index did not contribute significantly to the discrimination performed by the mathematical model (158).…”

Section: P50mentioning

confidence: 99%

“…In a recent longitudinal study (255), machine learning was used to discriminate two subgroups of FEP subjects, according to changes in dMMN amplitude, revealing that subjects with improvement of dMMN had better clinical, cognitive, and functional follow-up outcomes than those with worsening of dMMN (255). However, another study did not find a significant contribution of electrophysiological indices, such as P50 and MMN, for the discrimination between FES and HCs (158).…”

Section: Current and Future Perspective On The Employment Of Eeg-indices In Clinical Settingsmentioning

confidence: 99%

See 1 more Smart Citation

EEG-Based Measures in At-Risk Mental State and Early Stages of Schizophrenia: A Systematic Review

et al. 2021

View full text Add to dashboard Cite

Introduction: Electrophysiological (EEG) abnormalities in subjects with schizophrenia have been largely reported. In the last decades, research has shifted to the identification of electrophysiological alterations in the prodromal and early phases of the disorder, focusing on the prediction of clinical and functional outcome. The identification of neuronal aberrations in subjects with a first episode of psychosis (FEP) and in those at ultra high-risk (UHR) or clinical high-risk (CHR) to develop a psychosis is crucial to implement adequate interventions, reduce the rate of transition to psychosis, as well as the risk of irreversible functioning impairment. The aim of the review is to provide an up-to-date synthesis of the electrophysiological findings in the at-risk mental state and early stages of schizophrenia.Methods: A systematic review of English articles using Pubmed, Scopus, and PsychINFO was undertaken in July 2020. Additional studies were identified by hand-search. Electrophysiological studies that included at least one group of FEP or subjects at risk to develop psychosis, compared to healthy controls (HCs), were considered. The heterogeneity of the studies prevented a quantitative synthesis.Results: Out of 319 records screened, 133 studies were included in a final qualitative synthesis. Included studies were mainly carried out using frequency analysis, microstates and event-related potentials. The most common findings included an increase in delta and gamma power, an impairment in sensory gating assessed through P50 and N100 and a reduction of Mismatch Negativity and P300 amplitude in at-risk mental state and early stages of schizophrenia. Progressive changes in some of these electrophysiological measures were associated with transition to psychosis and disease course. Heterogeneous data have been reported for indices evaluating synchrony, connectivity, and evoked-responses in different frequency bands.Conclusions: Multiple EEG-indices were altered during at-risk mental state and early stages of schizophrenia, supporting the hypothesis that cerebral network dysfunctions appear already before the onset of the disorder. Some of these alterations demonstrated association with transition to psychosis or poor functional outcome. However, heterogeneity in subjects' inclusion criteria, clinical measures and electrophysiological methods prevents drawing solid conclusions. Large prospective studies are needed to consolidate findings concerning electrophysiological markers of clinical and functional outcome.

show abstract

Section: P50mentioning

confidence: 99%

Section: Current and Future Perspective On The Employment Of Eeg-indices In Clinical Settingsmentioning

confidence: 99%

EEG-Based Measures in At-Risk Mental State and Early Stages of Schizophrenia: A Systematic Review

et al. 2021

View full text Add to dashboard Cite

show abstract

“…refs. 16,19,20 ), and a complete list of publications is provided at www.cinsr.dk. Patients in cohort A had been randomized to treatment with either risperidone or zuclopenthixol for 3 months.…”

Section: Participants and Interventionsmentioning

confidence: 99%

“…In recent years, advances have been made towards combining data from multiple modalities in order to improve prediction. Clinical studies applying multimodal approaches are scarce, but may improve classification of schizophrenia patients from healthy controls compared to unimodal approaches 15 , although findings are equivocal 16 . We recently reported that the treatment response of psychopathologically indistinguishable patient subgroups was significantly predicted by an ML model based on cognitive and electrophysiological data 17 .…”

Section: Introductionmentioning

confidence: 99%

A machine-learning framework for robust and reliable prediction of short- and long-term treatment response in initially antipsychotic-naïve schizophrenia patients based on multimodal neuropsychiatric data

et al. 2020

Self Cite

View full text Add to dashboard Cite

The reproducibility of machine-learning analyses in computational psychiatry is a growing concern. In a multimodal neuropsychiatric dataset of antipsychotic-naïve, first-episode schizophrenia patients, we discuss a workflow aimed at reducing bias and overfitting by invoking simulated data in the design process and analysis in two independent machine-learning approaches, one based on a single algorithm and the other incorporating an ensemble of algorithms. We aimed to (1) classify patients from controls to establish the framework, (2) predict short-and long-term treatment response, and (3) validate the methodological framework. We included 138 antipsychotic-naïve, firstepisode schizophrenia patients with data on psychopathology, cognition, electrophysiology, and structural magnetic resonance imaging (MRI). Perinatal data and long-term outcome measures were obtained from Danish registers. Shortterm treatment response was defined as change in Positive And Negative Syndrome Score (PANSS) after the initial antipsychotic treatment period. Baseline diagnostic classification algorithms also included data from 151 matched controls. Both approaches significantly classified patients from healthy controls with a balanced accuracy of 63.8% and 64.2%, respectively. Post-hoc analyses showed that the classification primarily was driven by the cognitive data. Neither approach predicted short-nor long-term treatment response. Validation of the framework showed that choice of algorithm and parameter settings in the real data was successfully guided by results from the simulated data. In conclusion, this novel approach holds promise as an important step to minimize bias and obtain reliable results with modest sample sizes when independent replication samples are not available.

show abstract

“…To date, however, there are few studies of multimodal MRI integration entirely based on samples of first-episode patients. Two examples are the study by Peruzzo et al (2015), which used small sample sizes or the more recent study by Ebdrup et al (2018), which found no predictive power in the MRI datasets.…”

Section: Discussionmentioning

confidence: 99%

Multimodal Integration of Brain Images for MRI-Based Diagnosis in Schizophrenia

et al. 2019

View full text Add to dashboard Cite

Magnetic resonance imaging (MRI) has been proposed as a source of information for automatic prediction of individual diagnosis in schizophrenia. Optimal integration of data from different MRI modalities is an active area of research aimed at increasing diagnostic accuracy. Based on a sample of 96 patients with schizophrenia and a matched sample of 115 healthy controls that had undergone a single multimodal MRI session, we generated individual brain maps of gray matter vbm, 1back, and 2back levels of activation (nback fMRI), maps of amplitude of low-frequency fluctuations (resting-state fMRI), and maps of weighted global brain connectivity (resting-state fMRI). Four unimodal classifiers (Ridge, Lasso, Random Forests, and Gradient boosting) were applied to these maps to evaluate their classification accuracies. Based on the assignments made by the algorithms on test individuals, we quantified the amount of predictive information shared between maps (what we call redundancy analysis). Finally, we explored the added accuracy provided by a set of multimodal strategies that included post-classification integration based on probabilities, two-step sequential integration, and voxel-level multimodal integration through one-dimensional-convolutional neural networks (1D-CNNs). All four unimodal classifiers showed the highest test accuracies with the 2back maps (80% on average) achieving a maximum of 84% with the Lasso. Redundancy levels between brain maps were generally low (overall mean redundancy score of 0.14 in a 0–1 range), indicating that each brain map contained differential predictive information. The highest multimodal accuracy was delivered by the two-step Ridge classifier (87%) followed by the Ridge maximum and mean probability classifiers (both with 85% accuracy) and by the 1D-CNN, which achieved the same accuracy as the best unimodal classifier (84%). From these results, we conclude that from all MRI modalities evaluated task-based fMRI may be the best unimodal diagnostic option in schizophrenia. Low redundancy values point to ample potential for accuracy improvements through multimodal integration, with the two-step Ridge emerging as a suitable strategy.

show abstract

Accuracy of diagnostic classification algorithms using cognitive-, electrophysiological-, and neuroanatomical data in antipsychotic-naïve schizophrenia patients

Cited by 19 publications

References 58 publications

EEG-Based Measures in At-Risk Mental State and Early Stages of Schizophrenia: A Systematic Review

EEG-Based Measures in At-Risk Mental State and Early Stages of Schizophrenia: A Systematic Review

A machine-learning framework for robust and reliable prediction of short- and long-term treatment response in initially antipsychotic-naïve schizophrenia patients based on multimodal neuropsychiatric data

Multimodal Integration of Brain Images for MRI-Based Diagnosis in Schizophrenia

Contact Info

Product

Resources

About