Identifying a neuroanatomical signature of schizophrenia, reproducible across sites and stages, using machine learning with structured sparsity

Pierrefeu, Amicie de; Löfstedt, Tommy; Laidi, Charles; Hadj-Selem, Fouad; Bourgin, Julie; Hájek, Tomáš; Španiel, Filip; Kolenič, Marián; Ciuciu, Philippe; Hamdani, Nora; Leboyer, Marion; Fovet, Thomas; Jardri, Renaud; Houenou, Josselin; Duchesnay, Édouard

doi:10.1111/acps.12964

Cited by 22 publications

(28 citation statements)

References 38 publications

(49 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Surprisingly, the trajectory of psychotic symptoms or treatment-related factors had no significant association with SZ-HC pattern expression over the follow-up but were related to alterations in non-disorder-specific brain aging. Our findings of a negative association between SZ-HC pattern expression and poor cognitive performance are in line with the literature on wide-spread cognitive deficits in schizophrenia 33 , 34 and the previous machine learning studies finding similar associations 35 , 36 . The positive relationship between BMI and SZ-HC pattern expression could reflect a high prevalence of obesity in schizophrenia 37 but also disinhibition 38 , a common trait in schizophrenia 39 .…”

Section: Discussionsupporting

confidence: 93%

The progression of disorder-specific brain pattern expression in schizophrenia over 9 years

et al. 2021

View full text Add to dashboard Cite

Age plays a crucial role in the performance of schizophrenia vs. controls (SZ-HC) neuroimaging-based machine learning (ML) models as the accuracy of identifying first-episode psychosis from controls is poor compared to chronic patients. Resolving whether this finding reflects longitudinal progression in a disorder-specific brain pattern or a systematic but non-disorder-specific deviation from a normal brain aging (BA) trajectory in schizophrenia would help the clinical translation of diagnostic ML models. We trained two ML models on structural MRI data: an SZ-HC model based on 70 schizophrenia patients and 74 controls and a BA model (based on 561 healthy individuals, age range = 66 years). We then investigated the two models’ predictions in the naturalistic longitudinal Northern Finland Birth Cohort 1966 (NFBC1966) following 29 schizophrenia and 61 controls for nine years. The SZ-HC model’s schizophrenia-specificity was further assessed by utilizing independent validation (62 schizophrenia, 95 controls) and depression samples (203 depression, 203 controls). We found better performance at the NFBC1966 follow-up (sensitivity = 75.9%, specificity = 83.6%) compared to the baseline (sensitivity = 58.6%, specificity = 86.9%). This finding resulted from progression in disorder-specific pattern expression in schizophrenia and was not explained by concomitant acceleration of brain aging. The disorder-specific pattern’s progression reflected longitudinal changes in cognition, outcomes, and local brain changes, while BA captured treatment-related and global brain alterations. The SZ-HC model was also generalizable to independent schizophrenia validation samples but classified depression as control subjects. Our research underlines the importance of taking account of longitudinal progression in a disorder-specific pattern in schizophrenia when developing ML classifiers for different age groups.

show abstract

Section: Discussionsupporting

confidence: 93%

The progression of disorder-specific brain pattern expression in schizophrenia over 9 years

et al. 2021

View full text Add to dashboard Cite

show abstract

“…After evaluating a number of algorithms, we found that Random Forest-based classification is an effective tool to predict whether participants will experience side effects, with an F1-score of 0.797 and an AUC of 0.804. In the field of translational clinical psychology and psychiatry, machine learning has been widely used for disease diagnosis, differentiation, and outcome prediction (36,37). In our study, we demonstrated that this classifier can accurately differentiate whether patients/clients are likely to experience side effects.…”

Section: Discussionmentioning

confidence: 63%

Influencing Factors and Machine Learning-Based Prediction of Side Effects in Psychotherapy

Yao

Zhao

et al. 2020

Front. Psychiatry

View full text Add to dashboard Cite

Background: Side effects in psychotherapy are a common phenomenon, but due to insufficient understanding of the relevant predictors of side effects in psychotherapy, many psychotherapists or clinicians fail to identify and manage these side effects. The purpose of this study was to predict whether clients or patients would experience side effects in psychotherapy by machine learning and to analyze the related influencing factors.Methods: A self-compiled “Psychotherapy Side Effects Questionnaire (PSEQ)” was delivered online by a WeChat official account. Three hundred and seventy participants were included in the cross-sectional analysis. Psychotherapy outcomes were classified as participants with side effects and without side effects. A number of features were selected to distinguish participants with different psychotherapy outcomes. Six machine learning-based algorithms were then chosen and trained by our dataset to build outcome prediction classifiers.Results: Our study showed that: (1) the most common side effects were negative emotions in psychotherapy, such as anxiety, tension, sadness, and anger, etc. (24.6%, 91/370); (2) the mental state of the psychotherapist, as perceived by the participant during psychotherapy, was the most relevant feature to predict whether clients would experience side effects in psychotherapy; (3) a Random Forest-based machine learning classifier offered the best prediction performance of the psychotherapy outcomes, with an F1-score of 0.797 and an AUC value of 0.804. These numbers indicate a high prediction performance, which allowed our approach to be used in practice.Conclusions: Our Random Forest-based machine learning classifier could accurately predict the possible outcome of a client in psychotherapy. Our study sheds light on the influencing factors of the side effects of psychotherapy and could help psychotherapists better predict the outcomes of psychotherapy.

show abstract

“…The authors used whole-brain grey matter densities from MRI scans with SVM as the classifier and concluded that SVM models trained with less than 130 samples results in an unstable model. The key difference of the study [22] from previous similar studies [23,5], was utilizing a large dataset and using an entirely separate dataset to perform the validation. Additionally, noting that typical schizophrenia medications affect the striatum (part of the brain), they masked it out, ensuring the model doesn't relate medication effects to Schizophrenia detection.…”

Section: Related Workmentioning

confidence: 99%

Diagnosis of Schizophrenia: A comprehensive evaluation

Tanveer,

Jangir,

Ganaie

et al. 2022

Preprint

View full text Add to dashboard Cite

Machine learning models have been successfully employed in the diagnosis of Schizophrenia disease. The impact of classification models and the feature selection techniques on the diagnosis of Schizophrenia have not been evaluated. Here, we sought to access the performance of classification models along with different feature selection approaches on the structural magnetic resonance imaging data. The data consist of 72 subjects with Schizophrenia and 74 healthy control subjects. We evaluated different classification algorithms based on support vector machine (SVM), random forest, kernel ridge regression and randomized neural networks. Moreover, we evaluated T-Test, Receiver Operator Characteristics (ROC), Wilcoxon, entropy, Bhattacharyya, Minimum Redundancy Maximum Relevance (MRMR) and Neighbourhood Component Analysis (NCA) as the feature selection techniques. Based on the evaluation, SVM based models with Gaussian kernel proved better compared to other classification models and Wilcoxon feature selection emerged as the best feature selection approach. Moreover, in terms of data modality the performance on integration of the grey matter and white matter proved better compared to the performance on the grey and white matter individually. Our evaluation showed that classification algorithms along with the feature selection approaches impact the diagnosis of Schizophrenia disease. This indicates that proper selection of the features and the classification models can improve the diagnosis of Schizophrenia.

show abstract

Identifying a neuroanatomical signature of schizophrenia, reproducible across sites and stages, using machine learning with structured sparsity

Cited by 22 publications

References 38 publications

The progression of disorder-specific brain pattern expression in schizophrenia over 9 years

The progression of disorder-specific brain pattern expression in schizophrenia over 9 years

Influencing Factors and Machine Learning-Based Prediction of Side Effects in Psychotherapy

Diagnosis of Schizophrenia: A comprehensive evaluation

Contact Info

Product

Resources

About