Comparing the performances of PDF and PCA on Parkinson's disease classification using structural MRI images

Özkan, Çiğdem; Yılmaz, Arif; Beheshti, Iman; Demirel, Hasan

doi:10.1109/siu.2018.8404697

Cited by 14 publications

(8 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, they used SVM to combine complex network features with clinical scores typical of the PD prodromal phase and provide a diagnostic index. Cigdem et al 30 feature selection method based on PCA and probability distribution function, using VBM analysis, to perform feature selection on the volume of the region of interest in MRI images, and perform feature selection through SVMs classification. Rana et al 14 trained and tested axial T2 MRI images via a transfer‐learned Alexnet network.…”

Section: Resultsmentioning

confidence: 99%

Diagnosis of Parkinson's disease based on feature fusion on T2 MRI images

Cui

Lou

et al. 2022

Int J of Intelligent Sys

View full text Add to dashboard Cite

Deep‐learning methods (especially convolutional neural networks) using magnetic resonance imaging (MRI) data have been successfully applied to computer‐aided diagnosis of Parkinson's Disease (PD). Early detection and prior care may help patients improve their quality of life, although this neurodegenerative disease has no known cure. In this study, we propose a FResnet18 model to classify MRI images of PD and Health Control (HC) by fusing image texture features with deep features. First, Local Binary Pattern and Gray‐Level Co‐occurrence Matrix are used to extract the handcrafted features. Second, the modified ResNet18 network is used to extract deep features. Finally, the fused features are classified by Support Vector Machine. The classification accuracy rate for MRI images reaches 98.66%, and the findings demonstrate that the model can successfully differentiate between PD and HC. The suggested FResnet18 provides greater performance compared with existing approaches, and it is shown through extensive experimental findings on the Parkinson's Disease Progression Markers Initiative data set that feature fusion may improve classification performance.

show abstract

Section: Resultsmentioning

confidence: 99%

Diagnosis of Parkinson's disease based on feature fusion on T2 MRI images

Cui

Lou

et al. 2022

Int J of Intelligent Sys

View full text Add to dashboard Cite

show abstract

“…Based on this, machine learning models have been developed that aim to make use of such high-dimensional data for individual-level classification ( Maceachern and Forkert, 2021 ). Within this context, a variety of machine learning models for PD classification have been developed based on structural neuroimaging with accuracy levels ranging from 71.5% to 100% ( Adeli et al, 2017 , Adeli et al, 2016 , Amoroso et al, 2018 , Chakraborty et al, 2020 , Cigdem et al, 2018 , Esmaeilzadeh et al, 2018 , Solana-Lavalle and Rosas-Romero, 2021 ). Most of these works followed a classical machine learning setup, typically including image preprocessing, feature extraction, and training and testing of a conventional machine learning model ( e.g., support vector machine) using a cross-validation scheme.…”

Section: Introductionmentioning

confidence: 99%

Explainable classification of Parkinson’s disease using deep learning trained on a large multi-center database of T1-weighted MRI datasets

Camacho

Wilms

Mouchès

et al. 2023

NeuroImage: Clinical

View full text Add to dashboard Cite

“…An increasing number of feature selection methods are being applied in the field of neuroscience. Chu et al (2011) Cigdem et al (2018) classified PD and controls using a probability distribution function based on feature selection methods to build separate decision models for GM and WM, which achieved good classification accuracy of 75.00 and 72.50%, respectively. However, the histogram technology used in that study leads to the loss of information.…”

Section: Introductionmentioning

confidence: 99%

“…and rank them in terms of their ability to detect group-level differences; (2) wrapper methods, which are based on the cost function, and sort all features based on their degree of correlation (Kohavi and John, 1997 ; ChenZhiHong et al, 2020 ); and (3) embedded methods (Wang et al, 2015 ), which select relevant features by imposing certain “penalties” to obtain a subset of relevant features. Filtering methods have the benefit of low computational cost, while wrapper methods are superior to filtering methods in performance due to their discriminative ability (Lee and Verleysen, 2007 ; Chu et al, 2011 ; Adeli et al, 2016 ; Cigdem et al, 2018 ). Considering the interaction among features, embedded methods have shown excellent performance in pattern classification research (Wang et al, 2015 ).…”

Section: Introductionmentioning

confidence: 99%

“…Chu et al ( 2011 ) used a variety of different feature extraction techniques to construct a classification model, and experiments have revealed that feature selection methods can enhance the accuracy of classification. Cigdem et al ( 2018 ) classified PD and controls using a probability distribution function based on feature selection methods to build separate decision models for GM and WM, which achieved good classification accuracy of 75.00 and 72.50%, respectively. However, the histogram technology used in that study leads to the loss of information.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Stability Evaluation of Brain Changes in Parkinson's Disease Based on Machine Learning

Song

Zhang

Jiang

et al. 2021

Front. Comput. Neurosci.

View full text Add to dashboard Cite

Structural MRI (sMRI) has been widely used to examine the cerebral changes that occur in Parkinson's disease (PD). However, previous studies have aimed for brain changes at the group level rather than at the individual level. Additionally, previous studies have been inconsistent regarding the changes they identified. It is difficult to identify which brain regions are the true biomarkers of PD. To overcome these two issues, we employed four different feature selection methods [ReliefF, graph-theory, recursive feature elimination (RFE), and stability selection] to obtain a minimal set of relevant features and nonredundant features from gray matter (GM) and white matter (WM). Then, a support vector machine (SVM) was utilized to learn decision models from selected features. Based on machine learning technique, this study has not only extended group level statistical analysis with identifying group difference to individual level with predicting patients with PD from healthy controls (HCs), but also identified most informative brain regions with feature selection methods. Furthermore, we conducted horizontal and vertical analyses to investigate the stability of the identified brain regions. On the one hand, we compared the brain changes found by different feature selection methods and considered these brain regions found by feature selection methods commonly as the potential biomarkers related to PD. On the other hand, we compared these brain changes with previous findings reported by conventional statistical analysis to evaluate their stability. Our experiments have demonstrated that the proposed machine learning techniques achieve satisfactory and robust classification performance. The highest classification performance was 92.24% (specificity), 92.42% (sensitivity), 89.58% (accuracy), and 89.77% (AUC) for GM and 71.93% (specificity), 74.87% (sensitivity), 71.18% (accuracy), and 71.82% (AUC) for WM. Moreover, most brain regions identified by machine learning were consistent with previous findings, which means that these brain regions are related to the pathological brain changes characteristic of PD and can be regarded as potential biomarkers of PD. Besides, we also found the brain abnormality of superior frontal gyrus (dorsolateral, SFGdor) and lingual gyrus (LING), which have been confirmed in other studies of PD. This further demonstrates that machine learning models are beneficial for clinicians as a decision support system in diagnosing PD.

show abstract

Comparing the performances of PDF and PCA on Parkinson's disease classification using structural MRI images

Cited by 14 publications

References 7 publications

Diagnosis of Parkinson's disease based on feature fusion on T2 MRI images

Diagnosis of Parkinson's disease based on feature fusion on T2 MRI images

Explainable classification of Parkinson’s disease using deep learning trained on a large multi-center database of T1-weighted MRI datasets

Stability Evaluation of Brain Changes in Parkinson's Disease Based on Machine Learning

Contact Info

Product

Resources

About