Classifying Alzheimer's Disease Using a Finite Basis Physics Neural Network DOI Open Access
Logeshwari Dhavamani, Sagar Joshi, Pavan Kumar Varma Kothapalli

et al.

Microscopy Research and Technique, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 20, 2024

The disease amyloid plaques, neurofibrillary tangles, synaptic dysfunction, and neuronal death gradually accumulate throughout Alzheimer's (AD), resulting in cognitive decline functional disability. challenges of dataset quality, interpretability, ethical integration, population variety, picture standardization must be addressed using deep learning for the magnetic resonance imaging (MRI) classification AD order to guarantee a trustworthy practical therapeutic application. In this manuscript Classifying finite basis physics neural network (CAD-FBPINN) is proposed. Initially, images are collected from Neuroimaging Initiative (ADNI) dataset. fed Pre-processing segment. During preprocessing phase reverse lognormal Kalman filter (RLKF) used enhance input images. Then preprocessed given feature extraction process. Feature done by Newton-time-extracting wavelet transform (NTEWT), which extract statistical features such as mean, kurtosis, skewness. Finally extracted FBPINNs early mild impairment (EMCI), AD, (MCI), late (LMCI), normal control (NC), subjective memory complaints (SMCs). General, FBPINN does not express adapting optimization strategies determine optimal factors ensure correct classification. Hence, sea-horse algorithm (SHOA) optimize FBPINN, accurately classifies AD. proposed technique implemented python efficacy CAD-FBPINN assessed with support numerous performances like accuracy, precision, Recall, F1-score, specificity negative predictive value (NPV) analyzed. Proposed method attain 30.53%, 23.34%, 32.64% higher accuracy; 20.53%, 25.34%, 29.64% precision; NP values analyzed existing Stages through Brain Modifications Optimized optimizer. Then, effectiveness compared other methods that currently use, diagnosis convolution (DC-AD-AlexNet), Predicting 4 years before incident (PDP-ADI-GCNN), Using DC-AD-AlexNet algorithm, diagnose classify

Language: Английский

A regularized volumetric ConvNet based Alzheimer detection using T1-weighted MRI images DOI Creative Commons
Nitika Goenka, Akhilesh Sharma, Shamik Tiwari

et al.

Cogent Engineering, Journal Year: 2024, Volume and Issue: 11(1)

Published: Feb. 11, 2024

Alzheimer’s disease is a gradual neurodegenerative condition affecting the brain, causing decline in cognitive function by progressively damaging nerve cells over time. While cure for remains elusive, detection of (AD) through brain biomarkers crucial to impede its advancement. High-resolution structural MRI scans, particularly T1-weighted images, are commonly used detection. These images provide detailed information about brain’s structure, allowing researchers and clinicians identify abnormalities. Our study employs deep learning methodology using binary classification task—distinguishing between AD normal/healthy control (NC). The volumetric convolutional neural network model deployed on pre-processed validated MIRIAD datasets, achieving an impressive accuracy 97%, surpassing other models. Addressing challenge limited datasets models, we incorporated various augmentation techniques such as rotation rescaling, resulting outstanding effective discerning normal controls.

Language: Английский

Citations

8
Deep learning-based classification of dementia using image representation of subcortical signals DOI Creative Commons

Shivani Ranjan,

Ayush Tripathi,

Harshal Shende

et al.

BMC Medical Informatics and Decision Making, Journal Year: 2025, Volume and Issue: 25(1)

Published: March 6, 2025

Dementia is a neurological syndrome marked by cognitive decline. Alzheimer's disease (AD) and frontotemporal dementia (FTD) are the common forms of dementia, each with distinct progression patterns. Early accurate diagnosis cases (AD FTD) crucial for effective medical care, as both conditions have similar early-symptoms. EEG, non-invasive tool recording brain activity, has shown potential in distinguishing AD from FTD mild impairment (MCI). This study aims to develop deep learning-based classification system analyzing EEG derived scout time-series signals regions, specifically hippocampus, amygdala, thalamus. Scout time series extracted via standardized low-resolution electromagnetic tomography (sLORETA) technique utilized. The converted image representations using continuous wavelet transform (CWT) fed input learning models. Two high-density datasets utilized validate efficacy proposed method: online BrainLat dataset (128 channels, comprising 16 AD, 13 FTD, 19 healthy controls (HC)) in-house IITD-AIIA (64 including subjects 10 9 MCI, 8 HC). Different strategies classifier combinations been mapping classes data sets. best results were achieved product probabilities classifiers left right subcortical regions conjunction DenseNet model architecture. It yield accuracies 94.17 % 77.72 on datasets, respectively. highlight that representation-based approach differentiate various stages dementia. pave way more early diagnosis, which treatment management debilitating conditions.

Language: Английский

Citations

0
A feature-aware multimodal framework with auto-fusion for Alzheimer’s disease diagnosis DOI
Meiwei Zhang, Qiushi Cui, Yang Lü

et al.

Computers in Biology and Medicine, Journal Year: 2024, Volume and Issue: 178, P. 108740 - 108740

Published: June 19, 2024

Language: Английский

Citations

3
Detection of Alzheimer’s disease using Otsu thresholding with tunicate swarm algorithm and deep belief network DOI Creative Commons

Praveena Ganesan,

G. P. Ramesh,

Przemysław Falkowski‐Gilski

et al.

Frontiers in Physiology, Journal Year: 2024, Volume and Issue: 15

Published: July 9, 2024

Introduction: Alzheimer’s Disease (AD) is a degenerative brain disorder characterized by cognitive and memory dysfunctions. The early detection of AD necessary to reduce the mortality rate through slowing down its progression. prevention emerging research topic for many researchers. structural Magnetic Resonance Imaging (sMRI) an extensively used imaging technique in AD, because it efficiently reflects variations. Methods: Machine learning deep models are widely applied on sMRI images accelerate diagnosis process assist clinicians timely treatment. In this article, effective automated framework implemented AD. At first, Region Interest (RoI) segmented from acquired employing Otsu thresholding method with Tunicate Swarm Algorithm (TSA). TSA finds optimal segmentation threshold value method. Then, vectors extracted RoI applying Local Binary Pattern (LBP) Directional variance (LDPv) descriptors. last, passed Deep Belief Networks (DBN) image classification. Results Discussion: proposed achieves supreme classification accuracy 99.80% 99.92% Neuroimaging Initiative (ADNI) Australian Imaging, Biomarker Lifestyle flagship work ageing (AIBL) datasets, which higher than conventional models.

Language: Английский

Citations

2
A multimodal learning machine framework for Alzheimer’s disease diagnosis based on neuropsychological and neuroimaging data DOI
Meiwei Zhang, Qiushi Cui, Yang Lü

et al.

Computers & Industrial Engineering, Journal Year: 2024, Volume and Issue: unknown, P. 110625 - 110625

Published: Oct. 1, 2024

Language: Английский

Citations

1
A Data-Driven Boosting Cognitive Domain-Based Multimodal Framework for Alzheimer's Disease Diagnosis DOI
Meiwei Zhang, Qiushi Cui, Yang Lü

et al.

Published: Jan. 1, 2024

Language: Английский

Citations

0
ViTBayesianNet: An adaptive deep bayesian network-aided alzheimer disease detection framework with vision transformer-based residual densenet for feature extraction using MRI images DOI
Rajasekar Mohan, Rajesh Arunachalam, Neha Verma

et al.

Network Computation in Neural Systems, Journal Year: 2024, Volume and Issue: unknown, P. 1 - 41

Published: Dec. 11, 2024

One of the most familiar types disease is Alzheimer's (AD) and it mainly impacts people over age limit 60. AD causes irreversible brain damage in humans. It difficult to recognize various stages AD, hence advanced deep learning methods are suggested for recognizing its initial stages. In this experiment, an effective model-based detection approach introduced provide treatment patient. Initially, essential MRI collected from benchmark resources. After that, gathered MRIs provided as input feature extraction phase. Also, important features image extracted by Vision Transformer-based Residual DenseNet (ViT-ResDenseNet). Later, retrieved applied stage. phase, detected using Adaptive Deep Bayesian Network (Ada-DBN). Additionally, attributes Ada-DBN optimized with help Enhanced Golf Optimization Algorithm (EGOA). So, implemented model accomplishes relatively higher reliability than existing techniques. The numerical results framework obtained accuracy value 96.35 which greater 91.08, 91.95, 93.95 attained EfficientNet-B2, TF- CNN, ViT-GRU, respectively.

Language: Английский

Citations

0
Classifying Alzheimer's Disease Using a Finite Basis Physics Neural Network DOI Open Access
Logeshwari Dhavamani, Sagar Joshi, Pavan Kumar Varma Kothapalli

et al.

Microscopy Research and Technique, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 20, 2024

The disease amyloid plaques, neurofibrillary tangles, synaptic dysfunction, and neuronal death gradually accumulate throughout Alzheimer's (AD), resulting in cognitive decline functional disability. challenges of dataset quality, interpretability, ethical integration, population variety, picture standardization must be addressed using deep learning for the magnetic resonance imaging (MRI) classification AD order to guarantee a trustworthy practical therapeutic application. In this manuscript Classifying finite basis physics neural network (CAD-FBPINN) is proposed. Initially, images are collected from Neuroimaging Initiative (ADNI) dataset. fed Pre-processing segment. During preprocessing phase reverse lognormal Kalman filter (RLKF) used enhance input images. Then preprocessed given feature extraction process. Feature done by Newton-time-extracting wavelet transform (NTEWT), which extract statistical features such as mean, kurtosis, skewness. Finally extracted FBPINNs early mild impairment (EMCI), AD, (MCI), late (LMCI), normal control (NC), subjective memory complaints (SMCs). General, FBPINN does not express adapting optimization strategies determine optimal factors ensure correct classification. Hence, sea-horse algorithm (SHOA) optimize FBPINN, accurately classifies AD. proposed technique implemented python efficacy CAD-FBPINN assessed with support numerous performances like accuracy, precision, Recall, F1-score, specificity negative predictive value (NPV) analyzed. Proposed method attain 30.53%, 23.34%, 32.64% higher accuracy; 20.53%, 25.34%, 29.64% precision; NP values analyzed existing Stages through Brain Modifications Optimized optimizer. Then, effectiveness compared other methods that currently use, diagnosis convolution (DC-AD-AlexNet), Predicting 4 years before incident (PDP-ADI-GCNN), Using DC-AD-AlexNet algorithm, diagnose classify

Language: Английский

Citations

0