Use of Overlapping Group LASSO Sparse Deep Belief Network to Discriminate Parkinson's Disease and Normal Control.
ABSTRACT: As a medical imaging technology which can show the metabolism of the brain, 18F-fluorodeoxyglucose (FDG)-positron emission tomography (PET) is of great value for the diagnosis of Parkinson's Disease (PD). With the development of pattern recognition technology, analysis of brain images using deep learning are becoming more and more popular. However, existing computer-aided-diagnosis technologies often over fit and have poor generalizability. Therefore, we aimed to improve a framework based on Group Lasso Sparse Deep Belief Network (GLS-DBN) for discriminating PD and normal control (NC) subjects based on FDG-PET imaging. In this study, 225 NC and 125 PD cohorts from Huashan and Wuxi 904 hospitals were selected. They were divided into the training & validation dataset and 2 test datasets. First, in the training & validation set, subjects were randomly partitioned 80:20, with multiple training iterations for the deep learning model. Next, Locally Linear Embedding was used as a dimension reduction algorithm. Then, GLS-DBN was used for feature learning and classification. Different sparse DBN models were used to compare datasets to evaluate the effectiveness of our framework. Accuracy, sensitivity, and specificity were examined to validate the results. Output variables of the network were also correlated with longitudinal changes of rating scales about movement disorders (UPDRS, H&Y). As a result, accuracy of prediction (90% in Test 1, 86% in Test 2) for classification of PD and NC patients outperformed conventional approaches. Output scores of the network were strongly correlated with UPDRS and H&Y (R = 0.705, p < 0.001; R = 0.697, p < 0.001 in Test 1; R = 0.592, p = 0.0018, R = 0.528, p = 0.0067 in Test 2). These results show the GLS-DBN is feasible method for early diagnosis of PD.
Project description:Sequence classification plays an important role in metagenomics studies. We assess the deep neural network approach for fungal sequence classification as it has emerged as a successful paradigm for big data classification and clustering. Two deep learning-based classifiers, a convolutional neural network (CNN) and a deep belief network (DBN) were trained using our recently released barcode datasets. Experimental results show that CNN outperformed the traditional BLAST classification and the most accurate machine learning based Ribosomal Database Project (RDP) classifier on datasets that had many of the labels present in the training datasets. When classifying an independent dataset namely the "Top 50 Most Wanted Fungi", CNN and DBN assigned less sequences than BLAST. However, they could assign much more sequences than the RDP classifier. In terms of efficiency, it took the machine learning classifiers up to two seconds to classify a test dataset while it was 53 s for BLAST. The result of the current study will enable us to speed up the taxonomic assignments for the fungal barcode sequences generated at our institute as ~?70% of them still need to be validated for public release. In addition, it will help to quickly provide a taxonomic profile for metagenomics samples.
Project description:The development of therapeutic interventions for Parkinson disease (PD) is challenged by disease complexity and subjectivity of symptom evaluation. A Parkinson's Disease Related Pattern (PDRP) of glucose metabolism via fluorodeoxyglucose positron emission tomography (FDG-PET) has been reported to correlate with motor symptom scores and may aid the detection of disease-modifying therapeutic effects.We sought to independently evaluate the potential utility of the PDRP as a biomarker for clinical trials of early-stage PD.Two machine learning approaches (Scaled Subprofile Model (SSM) and NPAIRS with Canonical Variates Analysis) were performed on FDG-PET scans from 17 healthy controls (HC) and 23 PD patients. The approaches were compared regarding discrimination of HC from PD and relationship to motor symptoms.Both classifiers discriminated HC from PD (p?<?0.01, p?<?0.03), and classifier scores for age- and gender- matched HC and PD correlated with Hoehn & Yahr stage (R2?=?0.24, p?<?0.015) and UPDRS (R2?=?0.23, p?<?0.018). Metabolic patterns were highly similar, with hypometabolism in parieto-occipital and prefrontal regions and hypermetabolism in cerebellum, pons, thalamus, paracentral gyrus, and lentiform nucleus relative to whole brain, consistent with the PDRP. An additional classifier was developed using only PD subjects, resulting in scores that correlated with UPDRS (R2?=?0.25, p?<?0.02) and Hoehn & Yahr stage (R2?=?0.16, p?<?0.06).Two independent analyses performed in a cohort of mild PD patients replicated key features of the PDRP, confirming that FDG-PET and multivariate classification can provide an objective, sensitive biomarker of disease stage with the potential to detect treatment effects on PD progression.
Project description:BACKGROUND:With the advance of subthalamic nucleus (STN) deep brain stimulation (DBS) in the treatment of Parkinson's disease (PD), it is desired to identify objective criteria for the monitoring of the therapy outcome. This paper explores the feasibility of metabolic network derived from positron emission tomography (PET) with 18F-fluorodeoxyglucose in monitoring the STN DBS treatment for PD. METHODS:Age-matched 33 PD patients, 33 healthy controls (HCs), 9 PD patients with bilateral DBS surgery and 9 controls underwent 18F-FDG PET scans. The DBS patients were followed longitudinally to investigate the alternations of the PD-related metabolic covariance pattern (PDRP) expressions. RESULTS:The PDRP expression was abnormally elevated in PD patients compared with HCs (P?<?0.001). For DBS patients, a significant decrease in the Unified Parkinson's Disease Rating Scale (UPDRS, P?=?0.001) and PDRP expression (P?=?0.004) was observed 3 months after STN DBS treatment, while a rollback was observed in both UPDRS and PDRP expressions (both P?<?0.01) 12 months after treatment. The changes in PDRP expression mediated by STN DBS were generally in line with UPDRS improvement. The graphical network analysis shows increased connections at 3 months and a return at 12 months confirmed by small-worldness coefficient. CONCLUSIONS:The preliminary results demonstrate the potential of metabolic network expression as complimentary objective biomarker for the assessment and monitoring of STN DBS treatment in PD patients. Clinical Trial Registration ChiCTR-DOC-16008645. http://www.chictr.org.cn/showproj.aspx?proj=13865 .
Project description:Parkinson's disease (PD) is a degenerative disorder that affects the central nervous system. PD-related alterations in structural and functional neuroimaging have not been fully explored. This study explored multi-modal PD neuroimaging and its application for predicting clinical scores on the Movement Disorder Society-sponsored Unified Parkinson's Disease Rating Scale (MDS-UPDRS). Multi-modal imaging that combined 123I-Ioflupane single-photon emission computed tomography (SPECT) and diffusion tensor imaging (DTI) were adopted to incorporate complementary brain imaging information. SPECT and DTI images of normal controls (NC; n?=?45) and PD patients (n?=?45) were obtained from a database. The specific binding ratio (SBR) was calculated from SPECT. Tractography was performed using DTI. Group-wise differences between NC and PD patients were quantified using SBR of SPECT and structural connectivity of DTI for regions of interest (ROIs) related to PD. MDS-UPDRS scores were predicted using multi-modal imaging features in a partial least-squares regression framework. Three regions and four connections within the cortico-basal ganglia thalamocortical circuit were identified using SBR and DTI, respectively. Predicted MDS-UPDRS scores using identified regions and connections and actual MDS-UPDRS scores showed a meaningful correlation (r?=?0.6854, p?<?0.001). Our study provided insight on regions and connections that are instrumental in PD.
Project description:Background:Quality of life (QoL) was worse in Parkinson's disease patients with mild cognitive impairment (PD-MCI) or dementia (PDD) than PD patients with normal cognition (PD-NC). The aim of this study was to investigate and compare the potential heterogeneous determinants of QoL in PD patients with different cognitive statuses. Methods:We recruited 600 PD patients, including 185 PD-NC patients, 336 PD-MCI patients and 79 PDD patients, in this cross-sectional study. All patients completed the QoL assessment by the 39-item Parkinson's Disease Questionnaire (PDQ-39), as well as clinical evaluations and neuropsychological tests. The determinants of the QoL were analyzed by multiple stepwise regression analysis. Results:QoL was more impaired across the three groups (PD-NC < PD-MCI < PDD). The Unified Parkinson's Disease Rating Scale part III (UPDRS-III) score, Geriatric Depression Rating Scale (GDS) score and daily levodopa equivalent dose (LED) were independent variables of PDQ-39 in PD-NC patients. The GDS score, disease duration, UPDRS-III score, Epworth Sleepiness Score (ESS) and sex were independent variables of PDQ-39 in PD-MCI patients. The GDS score and disease duration were independent variables of PDQ-39 in PDD patients. Conclusion:The determinants of QoL in PD-NC, PD-MCI and PDD patients were heterogeneous. Motor function was considered to be the most crucial determinant for QoL in PD-NC, while depression was indicated to be the most vital determinant for PD-MCI and PDD. For QoL improvement, clinicians might need to focus more on motor function in PD-NC patients and on depression in PD-MCI and PDD patients.
Project description:Background:There is a critical need to develop valid, non-invasive biomarkers for Parkinsonian syndromes. The current 17-site, international study assesses whether non-invasive diffusion MRI (dMRI) can distinguish between Parkinsonian syndromes. Methods:We used dMRI from 1002 subjects, along with the Movement Disorders Society Unified Parkinson's Disease Rating Scale part III (MDS-UPDRS III), to develop and validate disease-specific machine learning comparisons using 60 template regions and tracts of interest in Montreal Neurological Institute (MNI) space between Parkinson's disease (PD) and Atypical Parkinsonism (multiple system atrophy - MSA, progressive supranuclear palsy - PSP), as well as between MSA and PSP. For each comparison, models were developed on a training/validation cohort and evaluated in a test cohort by quantifying the area under the curve (AUC) of receiving operating characteristic (ROC) curves. Findings:In the test cohort for both disease-specific comparisons, AUCs were high in the dMRI + MDS-UPDRS (PD vs. Atypical Parkinsonism: 0·962; MSA vs. PSP: 0·897) and dMRI Only (PD vs. Atypical Parkinsonism: 0·955; MSA vs. PSP: 0·926) models, whereas the MDS-UPDRS III Only models had significantly lower AUCs (PD vs. Atypical Parkinsonism: 0·775; MSA vs. PSP: 0·582). Interpretations:This study provides an objective, validated, and generalizable imaging approach to distinguish different forms of Parkinsonian syndromes using multi-site dMRI cohorts. The dMRI method does not involve radioactive tracers, is completely automated, and can be collected in less than 12 minutes across 3T scanners worldwide. The use of this test could thus positively impact the clinical care of patients with Parkinson's disease and Parkinsonism as well as reduce the number of misdiagnosed cases in clinical trials.
Project description:Studying a common architecture reflecting both brain's structural and functional organizations across individuals and populations in a hierarchical way has been of significant interest in the brain mapping field. Recently, deep learning models exhibited ability in extracting meaningful hierarchical structures from brain imaging data, e.g., fMRI and DTI. However, deep learning models have been rarely used to explore the relation between brain structure and function yet. In this paper, we proposed a novel multimodal deep believe network (DBN) model to discover and quantitatively represent the hierarchical organizations of common and consistent brain networks from both fMRI and DTI data. A prominent characteristic of DBN is that it is capable of extracting meaningful features from complex neuroimaging data with a hierarchical manner. With our proposed DBN model, three hierarchical layers with hundreds of common and consistent brain networks across individual brains are successfully constructed through learning a large dimension of representative features from fMRI/DTI data.
Project description:BACKGROUND:Attenuation correction (AC) of PET data is usually performed using a second imaging for the generation of attenuation maps. In certain situations however-when CT- or MR-derived attenuation maps are corrupted or CT acquisition solely for the purpose of AC shall be avoided-it would be of value to have the possibility of obtaining attenuation maps only based on PET information. The purpose of this study was to thus develop, implement, and evaluate a deep learning-based method for whole body [18F]FDG-PET AC which is independent of other imaging modalities for acquiring the attenuation map. METHODS:The proposed method is investigated on whole body [18F]FDG-PET data using a Generative Adversarial Networks (GAN) deep learning framework. It is trained to generate pseudo CT images (CTGAN) based on paired training data of non-attenuation corrected PET data (PETNAC) and corresponding CT data. Generated pseudo CTs are then used for subsequent PET AC. One hundred data sets of whole body PETNAC and corresponding CT were used for training. Twenty-five PET/CT examinations were used as test data sets (not included in training). On these test data sets, AC of PET was performed using the acquired CT as well as CTGAN resulting in the corresponding PET data sets PETAC and PETGAN. CTGAN and PETGAN were evaluated qualitatively by visual inspection and by visual analysis of color-coded difference maps. Quantitative analysis was performed by comparison of organ and lesion SUVs between PETAC and PETGAN. RESULTS:Qualitative analysis revealed no major SUV deviations on PETGAN for most anatomic regions; visually detectable deviations were mainly observed along the diaphragm and the lung border. Quantitative analysis revealed mean percent deviations of SUVs on PETGAN of - 0.8 ± 8.6% over all organs (range [- 30.7%, + 27.1%]). Mean lesion SUVs showed a mean deviation of 0.9 ± 9.2% (range [- 19.6%, + 29.2%]). CONCLUSION:Independent AC of whole body [18F]FDG-PET is feasible using the proposed deep learning approach yielding satisfactory PET quantification accuracy. Further clinical validation is necessary prior to implementation in clinical routine applications.
Project description:Neuroimaging-based models contribute to increasing our understanding of schizophrenia pathophysiology and can reveal the underlying characteristics of this and other clinical conditions. However, the considerable variability in reported neuroimaging results mirrors the heterogeneity of the disorder. Machine learning methods capable of representing invariant features could circumvent this problem. In this structural MRI study, we trained a deep learning model known as deep belief network (DBN) to extract features from brain morphometry data and investigated its performance in discriminating between healthy controls (N?=?83) and patients with schizophrenia (N?=?143). We further analysed performance in classifying patients with a first-episode psychosis (N?=?32). The DBN highlighted differences between classes, especially in the frontal, temporal, parietal, and insular cortices, and in some subcortical regions, including the corpus callosum, putamen, and cerebellum. The DBN was slightly more accurate as a classifier (accuracy?=?73.6%) than the support vector machine (accuracy?=?68.1%). Finally, the error rate of the DBN in classifying first-episode patients was 56.3%, indicating that the representations learned from patients with schizophrenia and healthy controls were not suitable to define these patients. Our data suggest that deep learning could improve our understanding of psychiatric disorders such as schizophrenia by improving neuromorphometric analyses.
Project description:fMRI data decomposition techniques have advanced significantly from shallow models such as Independent Component Analysis (ICA) and Sparse Coding and Dictionary Learning (SCDL) to deep learning models such Deep Belief Networks (DBN) and Convolutional Autoencoder (DCAE). However, interpretations of those decomposed networks are still open questions due to the lack of functional brain atlases, no correspondence across decomposed or reconstructed networks across different subjects, and significant individual variabilities. Recent studies showed that deep learning, especially deep convolutional neural networks (CNN), has extraordinary ability of accommodating spatial object patterns, e.g., our recent works using 3D CNN for fMRI-derived network classifications achieved high accuracy with a remarkable tolerance for mistakenly labelled training brain networks. However, the training data preparation is one of the biggest obstacles in these supervised deep learning models for functional brain network map recognitions, since manual labelling requires tedious and time-consuming labours which will sometimes even introduce label mistakes. Especially for mapping functional networks in large scale datasets such as hundreds of thousands of brain networks used in this paper, the manual labelling method will become almost infeasible. In response, in this work, we tackled both the network recognition and training data labelling tasks by proposing a new iteratively optimized deep learning CNN (IO-CNN) framework with an automatic weak label initialization, which enables the functional brain networks recognition task to a fully automatic large-scale classification procedure. Our extensive experiments based on ABIDE-II 1099 brains' fMRI data showed the great promise of our IO-CNN framework.