Measuring abnormal brains: building normative rules in neuroimaging using one-class support vector machines

Iranian Brain Imaging Database: A Neuropsychiatric Database of Healthy Brain

INTRODUCTION

The Iranian Brain Imaging Database (IBID) was initiated in 2017, with 5 major goals: provide researchers easy access to a neuroimaging database, provide normative quantitative measures of the brain for clinical research purposes, study the aging profile of the brain, examine the association of brain structure and function, and join the ENIGMA consortium. Many prestigious databases with similar goals are available. However, they were not done on an Iranian population, and the battery of their tests (e.g. cognitive tests) is selected based on their specific questions and needs.

METHODS

The IBID will include 300 participants (50% female) in the age range of 20 to 70 years old, with an equal number of participants (#60) in each age decade. It comprises a battery of cognitive, lifestyle, medical, and mental health tests, in addition to several Magnetic Resonance Imaging (MRI) protocols. Each participant completes the assessments on two referral days.

RESULTS

The study currently has a cross-sectional design, but longitudinal assessments are considered for the future phases of the study. Here, details of the methodology and the initial results of assessing the first 152 participants of the study are provided.

CONCLUSION

IBID is established to enable research into human brain function, to aid clinicians in disease diagnosis research, and also to unite the Iranian researchers with interests in the brain.

Density-based clustering of static and dynamic functional MRI connectivity features obtained from subjects with cognitive impairment

Various machine-learning classification techniques have been employed previously to classify brain states in healthy and disease populations using functional magnetic resonance imaging (fMRI). These methods generally use supervised classifiers that are sensitive to outliers and require labeling of training data to generate a predictive model. Density-based clustering, which overcomes these issues, is a popular unsupervised learning approach whose utility for high-dimensional neuroimaging data has not been previously evaluated. Its advantages include insensitivity to outliers and ability to work with unlabeled data. Unlike the popular k-means clustering, the number of clusters need not be specified. In this study, we compare the performance of two popular density-based clustering methods, DBSCAN and OPTICS, in accurately identifying individuals with three stages of cognitive impairment, including Alzheimer’s disease. We used static and dynamic functional connectivity features for clustering, which captures the strength and temporal variation of brain connectivity respectively. To assess the robustness of clustering to noise/outliers, we propose a novel method called recursive-clustering using additive-noise (R-CLAN). Results demonstrated that both clustering algorithms were effective, although OPTICS with dynamic connectivity features outperformed in terms of cluster purity (95.46%) and robustness to noise/outliers. This study demonstrates that density-based clustering can accurately and robustly identify diagnostic classes in an unsupervised way using brain connectivity.

Using Machine Learning and Structural Neuroimaging to Detect First Episode Psychosis: Reconsidering the Evidence

Despite the high level of interest in the use of machine learning (ML) and neuroimaging to detect psychosis at the individual level, the reliability of the findings is unclear due to potential methodological issues that may have inflated the existing literature. This study aimed to elucidate the extent to which the application of ML to neuroanatomical data allows detection of first episode psychosis (FEP), while putting in place methodological precautions to avoid overoptimistic results. We tested both traditional ML and an emerging approach known as deep learning (DL) using 3 feature sets of interest: (1) surface-based regional volumes and cortical thickness, (2) voxel-based gray matter volume (GMV) and (3) voxel-based cortical thickness (VBCT). To assess the reliability of the findings, we repeated all analyses in 5 independent datasets, totaling 956 participants (514 FEP and 444 within-site matched controls). The performance was assessed via nested cross-validation (CV) and cross-site CV. Accuracies ranged from 50% to 70% for surfaced-based features; from 50% to 63% for GMV; and from 51% to 68% for VBCT. The best accuracies (70%) were achieved when DL was applied to surface-based features; however, these models generalized poorly to other sites. Findings from this study suggest that, when methodological precautions are adopted to avoid overoptimistic results, detection of individuals in the early stages of psychosis is more challenging than originally thought. In light of this, we argue that the current evidence for the diagnostic value of ML and structural neuroimaging should be reconsidered toward a more cautious interpretation.

Using deep autoencoders to identify abnormal brain structural patterns in neuropsychiatric disorders: A large-scale multi-sample study

Machine learning is becoming an increasingly popular approach for investigating spatially distributed and subtle neuroanatomical alterations in brain-based disorders. However, some machine learning models have been criticized for requiring a large number of cases in each experimental group, and for resembling a "black box" that provides little or no insight into the nature of the data. In this article, we propose an alternative conceptual and practical approach for investigating brain-based disorders which aim to overcome these limitations. We used an artificial neural network known as "deep autoencoder" to create a normative model using structural magnetic resonance imaging data from 1,113 healthy people. We then used this model to estimate total and regional neuroanatomical deviation in individual patients with schizophrenia and autism spectrum disorder using two independent data sets (n = 263). We report that the model was able to generate different values of total neuroanatomical deviation for each disease under investigation relative to their control group (p < .005). Furthermore, the model revealed distinct patterns of neuroanatomical deviations for the two diseases, consistent with the existing neuroimaging literature. We conclude that the deep autoencoder provides a flexible and promising framework for assessing total and regional neuroanatomical deviations in neuropsychiatric populations.

Association between abnormal brain functional connectivity in children and psychopathology: A study based on graph theory and machine learning

OBJECTIVES

One of the major challenges facing psychiatry is how to incorporate biological measures in the classification of mental health disorders. Many of these disorders affect brain development and its connectivity. In this study, we propose a novel method for assessing brain networks based on the combination of a graph theory measure (eigenvector centrality) and a one-class support vector machine (OC-SVM).

METHODS

We applied this approach to resting-state fMRI data from 622 children and adolescents. Eigenvector centrality (EVC) of nodes from positive- and negative-task networks were extracted from each subject and used as input to an OC-SVM to label individual brain networks as typical or atypical. We hypothesised that classification of these subjects regarding the pattern of brain connectivity would predict the level of psychopathology.

RESULTS

Subjects with atypical brain network organisation had higher levels of psychopathology (p < 0.001). There was a greater EVC in the typical group at the bilateral posterior cingulate and bilateral posterior temporal cortices; and significant decreases in EVC at left temporal pole.

CONCLUSIONS

The combination of graph theory methods and an OC-SVM is a promising method to characterise neurodevelopment, and may be useful to understand the deviations leading to mental disorders.

Detection of Lesions Underlying Intractable Epilepsy on T1-Weighted MRI as an Outlier Detection Problem

Pattern recognition methods, such as computer aided diagnosis (CAD) systems, can help clinicians in their diagnosis by marking abnormal regions in an image. We propose a machine learning system based on a one-class support vector machine (OC-SVM) classifier for the detection of abnormalities in magnetic resonance images (MRI) applied to patients with intractable epilepsy. The system learns the features associated with healthy control subjects, allowing a voxelwise assessment of the deviation of a test subject pattern from the learned patterns. While any number of various features can be chosen and learned, here we focus on two texture parameters capturing image patterns associated with epileptogenic lesions on T1-weighted brain MRI e.g. heterotopia and blurred junction between the grey and white matter. The CAD output consists of patient specific 3D maps locating clusters of suspicious voxels ranked by size and degree of deviation from control patterns. System performance was evaluated using realistic simulations of challenging detection tasks as well as clinical data of 77 healthy control subjects and of eleven patients (13 lesions). It was compared to that of a mass univariate statistical parametric mapping (SPM) single subject analysis based on the same set of features. For all simulations, OC-SVM yielded significantly higher values of the area under the ROC curve (AUC) and higher sensitivity at low false positive rate. For the clinical data, both OC-SVM and SPM successfully detected 100% of the lesions in the MRI positive cases (3/13). For the MRI negative cases (10/13), OC-SVM detected 7/10 lesions and SPM analysis detected 5/10 lesions. In all experiments, OC-SVM produced fewer false positive detections than SPM. OC-SVM may be a versatile system for unbiased lesion detection.

One-Class Support Vector Machines Identify the Language and Default Mode Regions As Common Patterns of Structural Alterations in Young Children with Autism Spectrum Disorders

The identification of reliable brain endophenotypes of autism spectrum disorders (ASD) has been hampered to date by the heterogeneity in the neuroanatomical abnormalities detected in this condition. To handle the complexity of neuroimaging data and to convert brain images in informative biomarkers of pathology, multivariate analysis techniques based on Support Vector Machines (SVM) have been widely used in several disease conditions. They are usually trained to distinguish patients from healthy control subjects by making a binary classification. Here, we propose the use of the One-Class Classification (OCC) or Data Description method that, in contrast to two-class classification, is based on a description of one class of objects only. This approach, by defining a multivariate normative rule on one class of subjects, allows recognizing examples from a different category as outliers. We applied the OCC to 314 regional features extracted from brain structural Magnetic Resonance Imaging (MRI) scans of young children with ASD (21 males and 20 females) and control subjects (20 males and 20 females), matched on age [range: 22-72 months of age; mean = 49 months] and non-verbal intelligence quotient (NVIQ) [range: 31-123; mean = 73]. We demonstrated that a common pattern of features characterize the ASD population. The OCC SVM trained on the group of ASD subjects showed the following performances in the ASD vs. controls separation: the area under the receiver operating characteristic curve (AUC) was 0.74 for the male and 0.68 for the female population, respectively. Notably, the ASD vs. controls discrimination results were maximized when evaluated on the subsamples of subjects with NVIQ ≥ 70, leading to AUC = 0.81 for the male and AUC = 0.72 for the female populations, respectively. Language regions and regions from the default mode network-posterior cingulate cortex, pars opercularis and pars triangularis of the inferior frontal gyrus, and transverse temporal gyrus-contributed most to distinguishing individuals with ASD from controls, arguing for the crucial role of these areas in the ASD pathophysiology. The observed brain patterns associate preschoolers with ASD independently of their age, gender and NVIQ and therefore they are expected to constitute part of the ASD brain endophenotype.

The effect of gender on the neuroanatomy of children with autism spectrum disorders: a support vector machine case-control study

Background

Genetic, hormonal, and environmental factors contribute since infancy to sexual dimorphism in regional brain structures of subjects with typical development. However, the neuroanatomical differences between male and female children with autism spectrum disorders (ASD) are an intriguing and still poorly investigated issue. This study aims to evaluate whether the brain of young children with ASD exhibits sex-related structural differences and if a correlation exists between clinical ASD features and neuroanatomical underpinnings.

Methods

A total of 152 structural MRI scans were analysed. Specifically, 76 young children with ASD (38 males and 38 females; 2–7 years of age; mean = 53 months, standard deviation = 17 months) were evaluated employing a support vector machine (SVM)-based analysis of the grey matter (GM). Group comparisons consisted of 76 age-, gender- and non-verbal-intelligence quotient-matched children with typical development or idiopathic developmental delay without autism.

Results

For both genders combined, SVM showed a significantly increased GM volume in young children with ASD with respect to control subjects, predominantly in the bilateral superior frontal gyrus (Brodmann area –BA– 10), bilateral precuneus (BA 31), bilateral superior temporal gyrus (BA 20/22), whereas less GM in patients with ASD was found in right inferior temporal gyrus (BA 37). For the within gender comparisons (i.e., females with ASD vs. controls and males with ASD vs. controls), two overlapping regions in bilateral precuneus (BA 31) and left superior frontal gyrus (BA 9/10) were detected. Sex-by-group analyses revealed in males with ASD compared to matched controls two male-specific regions of increased GM volume (left middle occipital gyrus—BA 19—and right superior temporal gyrus—BA 22). Comparisons in females with and without ASD demonstrated increased GM volumes predominantly in the bilateral frontal regions. Additional regions of significantly increased GM volume in the right anterior cingulate cortex (BA 32) and right cerebellum were typical only of females with ASD.

Conclusions

Despite the specific behavioural correlates of sex-dimorphism in ASD, brain morphology as yet remains unclear and requires future dedicated investigations. This study provides evidence of structural brain gender differences in young children with ASD that possibly contribute to the different phenotypic disease manifestations in males and females.

Electronic supplementary material

The online version of this article (doi:10.1186/s13229-015-0067-3) contains supplementary material, which is available to authorized users.

PrediQt-Cx: post treatment health related quality of life prediction model for cervical cancer patients

Background

Cervical cancer is the third largest cause of cancer mortality in India. The objectives of the study were to compare the pre and the post treatment quality of life in cervical cancer patients and to develop a prediction model to provide an insight into the possibilities in the treatment modules.

Methodology/Principal Findings

A total of 198 patients were assessed with two structured questionnaires of Health Related Quality of Life (The European Organisation for Research and Treatment of Cancer, EORTC QLQ C-30 and CX-24). The baseline observations were recorded when the patients first reported (T1) and second evaluation was done at 6 months post treatment (T2). The mean age of detection was 50.9 years with the literacy level being non-educated or less than high school. Majority of them were married/cohabiting 179 (90.4%). On histopathological examination (HPE) squamous cell carcinoma was found to be the most common cell type carcinoma 147 (74.2%) followed by Adenocarcinoma 31 (15.7%). Radical hysterectomy was the most common treatment modality 76 (38.4%), followed by Wertheims Hysterectomy 46 (23.2%) and Radiochemotherapy 59 (29.8%). The mean score of global health of cervical cancer patients post treatment was 77.90, which was significantly higher than the pre - treatment score (54.32). Mean “symptoms score” post treatment was 21.69 with an aggravation of 7.32 compared to pre treatment scores. Patients experienced substantial decrease in sexual activity post treatment.

Conclusions/Significance

The prediction model(PrediQt-Cx), based on Support Vector Machine(SVM) for predicting post treatment HRQoL in cervical cancer patients was developed and internally cross validated. After external validation PrediQt-Cx can be easily employed to support decision making by clinicians and patients from north India region, through openly made available for access at http://prediqt.org.