Alzheimer's Disease Detection Using CNN Model

Alzheimer's disease is a degenerative disease associated with memory loss, communication difficulties, mental health, thinking skills, and other psychological disorders that affect a person's daily activities. Alzheimer's disease is a disease that causes disability for people aged 70 years and over and is the seventh highest contributor to death in the world. However, until now there has not been found an effective treatment to cure Alzheimer's disease. Thus, early detection of Alzheimer's disease is very important so that sufferers of Alzheimer's disease can immediately receive intensive medical care so as to reduce the death rate from Alzheimer's disease. One method that can be used to detect Alzheimer's disease is by utilizing a machine learning algorithm model. The machine learning model in this study was carried out using the Decision Tree C4.5 algorithm classification method based on Binary Particle Swarm Optimization (BPSO). The C4.5 Decision Tree algorithm is used to classify Alzheimer's disease, while the BPSO algorithm is used to perform feature selection. By performing feature selection with the BPSO algorithm, the results show that the BPSO algorithm can improve accuracy and can increase the performance of the C4.5 algorithm in the Alzheimer's disease classification process. The results of the accuracy of the C4.5 algorithm using the BPSO feature selection are greater, namely 98.2% compared to the C4.5 algorithm without BPSO feature selection, which is only 96.4%.

Alzheimer’s disease (AD) is a progressive neurological disorder characterised by cognitive decline and memory loss. Early diagnosis is essential for effective treatment, although the complexity of the initial symptoms sometimes delays it. This review addresses the development of a deep learning-based model to aid in the early diagnosis of Alzheimer’s disease using neuroimaging data. Using MRI and PET scans from public datasets such as the Alzheimer’s Disease Neuroimaging Initiative, the proposed version makes use of convolutional neural networks (CNNs) to extract feature extraction and types by looking at brain structure in and on models associated with early Alzheimer’s ailment at the version that tries to decide the values ​​Performance is measured using key parameters along with sensitivity, specificity, accuracy, and area under the curve. The purpose is to develop a predictive tool which could help medical doctors diagnose Alzheimer’s disorder in advance and, in all likelihood, enhance the affected person’s effects via timely intervention.

As the global population ages, the prevalence of dementia caused by Alzheimer's disease is expected to increase, posing a significant healthcare challenge. Dementia is a leading cause of disability and dependency among older people, and it also has broad physical, psychological, social, and economic impacts on individuals, their caregivers, families, and communities. Despite the existence of treatments to manage dementia symptoms, a lack of awareness and understanding of the condition often leads to stigmatization and difficulties in accessing diagnosis and care. Neuroimaging is a viable method for reaching this objective because early Alzheimer's disease detection is essential for successful intervention. Although dementia cannot be cured, controlling symptoms can enhance the quality of life for those who have the condition. The objective is to employ structural magnetic resonance imaging (sMRI) information to examine deep learning algorithms for the identification of significant biomarkers related to Alzheimer's disease. The suggested method employs Convolutional Neural Networks (CNNs) to classify brain images into one of categories of Alzheimer disease as Non Demented, Very Mild Demented, Mild Demented and Moderate Demented.

This study develops an automatic algorithm for detecting Alzheimer's disease (AD) using magnetic resonance imaging (MRI) through deep learning and feature selection techniques. It utilizes a dataset of 6400 MRI images from Kaggle, categorized into four classes. Initially, the study employs pretrained CNN architectures—DenseNet-201, MobileNet-v2, ResNet-18, ResNet-50, ResNet-101, and ShuffleNet—for classification using five fold cross-validation, with DenseNet-201 achieving the highest accuracy of 82.11%. Due to the dataset's size and imbalance, as well as the long training times, the study aims to create a more efficient algorithm. The CNNs are used as deep feature extractors from AD images, and the extracted features are reduced using a new fine-tuning neighborhood component analysis (FTNCA) algorithm, which minimizes loss and determines the optimal tolerance value. The essential features are then classified using various machine learning algorithms, including artificial neural network (ANN), K-nearest neighbor (KNN), Naïve Bayes, and support vector machine (SVM). Experimental results reveal that reducing the feature set from 2048 to 344 allows the ResNet-50-FTNCA-KNN model to achieve 100% accuracy, significantly enhancing AD detection. This approach will aid in the early diagnosis and treatment of AD patients.

Early detection of Alzheimer's disease (AD) ensures that affected individuals and their caregivers can make appropriate plans for health care needs, yet many ethnic minorities delay seeking care for AD until the disease has progressed. This study examined attitudes toward care-seeking for AD among Korean Americans (KAs) and identified factors affecting their attitudes. A cross-sectional survey was used to collect data from 234 KA adults. We conducted hierarchical regression analyses to examine the effects of sociocultural background (age, gender, education, cultural orientation), AD knowledge and exposure to AD, and beliefs about AD (stigma of pity, shame, and public avoidance) and AD care (subjective norm) on KAs' attitudes toward seeking AD care from primary care physicians (PCPs) and AD specialists. We also tested whether knowledge of AD moderated the impact of beliefs about AD and AD care on KAs' attitudes toward seeking AD care. For both PCPs and AD specialists, the subjective norm had the strongest effect on KAs' attitudes toward care seeking (β = 0.557 for PCPs, β = 0.360 for specialists). Effects of stigma beliefs disappeared in the presence of the subjective norm. AD knowledge moderated the impact of the subjective norm on the attitudes toward both PCPs (β = -1.653) and specialists (β = -1.742). The significance of the subjective norm in KAs' attitudes toward AD care-seeking underscores the importance of public education, and our study suggests that increasing AD knowledge could facilitate a change in public attitudes toward seeking AD care.

The early detection of Alzheimer's disease (AD) in cognitively healthy individuals remains a major preclinical challenge. EEG is a promising tool that has shown effectiveness in detecting AD risk. Task-related EEG has been rarely used in Alzheimer's disease research, as most studies have focused on resting-state EEG. An interpretable deep learning framework-Interpretable Convolutional Neural Network (InterpretableCNN)-was utilized to identify AD-related EEG features. EEG data were recorded during three cognitive task conditions, and samples were labeled based on APOE genotype and polygenic risk scores. A 100-fold leave-p%-subjects-out cross-validation (LPSO-CV) was used to evaluate model performance and generalizability. The model achieved an ROC AUC of 60.84% across the tasks and subjects, with a Kappa value of 0.22, indicating fair agreement. Interpretation revealed a consistent focus on theta and alpha activity in the parietal and temporal regions-areas commonly associated with AD pathology. Task-related EEG combined with interpretable deep learning can reveal early AD risk signatures in healthy individuals. InterpretableCNN enhances transparency in feature identification, offering a valuable tool for preclinical screening.

The existing literature on the magnetic resonance imaging of murine models of Alzheimer's disease is reviewed. Particular attention is paid to the possibilities for the early detection of the disease. To this effect, not only are relaxometric and volumetric approaches discussed, but also mathematical models for plaque distribution and aggregation. Image analysis plays a prominent role in this line of research, as stochastic image models and texture analysis have shown some success in the classification of subjects affected by Alzheimer's disease. It is concluded that relaxometric approaches seem to be a promising candidate for the task at hand, especially when combined with sophisticated image analysis, and when data from more than one time-point is available. There have been few longitudinal studies of mice models so far, so this direction of research warrants future efforts.

Image recognition and neuroimaging are increasingly being used to understand the progression of Alzheimer's disease (AD). However, image data from single-photon emission computed tomography (SPECT) are limited. Medical image analysis requires large, labeled training datasets. Therefore, studies have focused on overcoming this problem. In this study, the detection performance of five convolutional neural network (CNN) models (MobileNet V2 and NASNetMobile (lightweight models); VGG16, Inception V3, and ResNet (heavier weight models)) on medical images was compared to establish a classification model for epidemiological research. Brain scan image data were collected from 99 subjects, and 4711 images were used. Demographic data were compared using the chi-squared test and one-way analysis of variance with Bonferroni's post hoc test. Accuracy and loss functions were used to evaluate the performance of CNN models. The cognitive abilities screening instrument and mini mental state exam scores of subjects with a clinical dementia rating (CDR) of 2 were considerably lower than those of subjects with a CDR of 1 or 0.5. This study analyzed the classification performance of various CNN models for medical images and proved the effectiveness of transfer learning in identifying the mild cognitive impairment, mild AD, and moderate AD scoring based on SPECT images.

Alzheimer's disease (AD) is characterized by amyloid-beta plaques, tau tangles, and neuroinflammation. C-X3-C motif chemokine ligand 1 (CX3CL1, also known as fractalkine), a neuroimmune chemokine implicated in AD pathogenesis, shows inconsistent alterations in plasma/serum across studies. Specifically examining age-dependency and diagnostic utility, we investigated plasma CX3CL1 levels across the cognitive continuum (cognitively normal [CN], amnestic mild cognitive impairment [aMCI], AD) in a Chinese cohort. A total of 443 participants, including 130 patients with AD, 72 patients with aMCI, and 99 age-and sex-matched CN controls, as well as a cohort of 142 CN subjects of different ages, were enrolled from Chongqing General Hospital. Plasma CX3CL1 levels were determined using Enzyme-Linked Immunosorbent Assay (ELISA). Apolipoprotein E genotypes (APOE) were performed. The correlations between Plasma CX3CL1 levels and cognition test scores or age were analyzed. The optimal diagnostic sensitivity and specificity were determined using receiver operating characteristic curve analysis. Plasma CX3CL1 levels significantly increased with age in CN individuals. No significant sex difference was found. Plasma CX3CL1 levels did not differ significantly between APOE ε4 carriers and non-carriers. Stepwise elevation across continuum: CX3CL1 levels showed a significant stepwise increase: CN controls (1.73 ± 0.51 ng/mL) < aMCI (2.40 ± 1.06 ng/mL) < AD (4.15 ± 1.24 ng/mL) (p < 0.001 between all groups). This pattern persisted in both male and female subgroups, between the AD group and the aMCI group, between the AD group and the CN control group (p < 0.001), between the aMCI group and the CN control group, and between the male and female subgroups (p < 0.05). CX3CL1 levels negatively correlated with Mini-Mental State Examination (MMSE) scores and positively correlated with age. Plasma CX3CL1 levels exhibit a significant age-dependent increase in cognitively normal individuals, peak in midlife (40-49 years), and demonstrate a stepwise elevation across the AD continuum (CN → aMCI → AD). Strong inverse correlations with cognitive scores in disease groups and high diagnostic accuracy for AD, particularly against CN, support its role as a biomarker reflecting both physiological aging and AD-related pathological decline. Its regulation appears independent of APOE ε4 status. The midlife peak suggests potential relevance for preclinical processes, warranting further investigation of CX3CL1 as a biomarker and therapeutic target.

Early detection of Alzheimer's disease (AD) is critical for preventing disease progression. Blood platelets have emerged as a useful peripheral source for AD diagnosis. However, the identification of proteomics-based platelet biomarkers of mild cognitive impairment (MCI) and AD in relation to amyloid β (Aβ) deposition remains largely unexplored. In this study, we compared four groups from 18 participants: subjective memory impairment (SMI, n = 4) as cognitive normal controls, MCI without Aβ deposition (MCI-A(+), n = 5), MCI with Aβ deposition (MCI-A(-), n = 5), and AD (n = 4). We conducted in-depth platelet protein profiling using high-throughput LC-MS/MS with tandem mass tag labeling. Among the total 4,524 proteins detected, we identified both unique and overlapping differentially expressed proteins in MCI-A(+), MCI-A(-), and AD compared with SMI. Hierarchical clustering analysis revealed seven distinct patterns of proteomic alterations across groups. Functional network and gene ontology enrichment analyses indicated that each cluster was associated with specific processes, including platelet activation, AD, and apoptotic signaling pathways. Notably, upregulated proteins in MCI-A(-) and AD were linked to endomembrane system organization. Furthermore, we quantified the relative abundance of multiple protein candidates that were significantly altered in MCI-A(-) and AD compared with SMI and MCI-A(+). Our findings highlight several platelet proteins-ATP6V0C, AP4B1, RAB2B, PSMD9, FKBP1B, and mTOR-as potential molecular targets for predicting AD at the stage of MCI with Aβ deposition, providing new insights into amyloid-related neurodegeneration.

The mango leaf powdery mildew disease poses a serious threat to mango production society globally by significantly lowering yield and quality. For timely intervention and efficient management, early disease detection and classification are important. In this research and education area, a novel hybrid approach utilizes Convolutional Neural Networks (CNN) and Support Vector Machines to identify the mango leaf powdery mildew disease based on four severity levels (SVM). Three phases make up the proposed approach: data structure, CNN-selected attributes, and SVM classification. We collect and preprocess images of mango leaves during the data organization step, and in the CNN - attributes selection phase, we apply a CNN model for feature extraction and selection. For the mango leaf powdery mildew dataset, we improve the CNN model to find the most relevant features for the classification task. The SVM - classification step includes training an SVM model on the obtained features and refining the hyperparameters via k-fold cross-validation. The proposed CNN and SVM hybrid multi-classification model for mango leaf powdery mildew disease achieved an overall accuracy of 89.29%. A dataset of 2559 images with 4 severity levels was utilized. The model works well overall, as a macro-average F1-score of 90.10, the weighted average F1-score's minimal value of 53.85%. The model is less successful in predicting instances for classes with smaller support proportions, as shown by the micro-average F1-score, which is 89.29% and is lower overall than the macro-average F1-score.

This study aims to detect direct cost savings with early detection of Alzheimer's disease (AD) and related dementias. We hypothesize that earlier diagnosis of AD can reduce direct healthcare costs and unnecessary utilization. A retrospective case-control study is conducted using Optum Market Clarity Claims/EHR linked dataset on Medicare Advantage population from Jan 1 2011 to Dec 31 2020. Based on whether the patient was diagnosed with prior Mild Cognitive Impairment (MCI) before progressing to an AD diagnosis per ICD-9/10 cm diagnosis codes, AD patients are categorized into early vs. late diagnosis (for those who were identified in the MCI stage vs not). Both groups are compared on annual all-cause healthcare costs and frequency of encounters longitudinally spanning across 6 years pre-index period and 3 years post-index with AD diagnosis date as the index date. To understand the impact of early diagnosis, we estimated total cost, medical service cost and drug cost post-index within the first 3 years by risk adjustment on baseline demographics, comorbidities, procedures, lab tests, and visits prior to AD diagnosis using propensity score weighting. Early-diagnosed patients (N = 421) have a lower utilization rate of inpatient skilled nursing facility, nursing home and hospitalization than late-diagnosed patients (N = 2817) after being diagnosed with AD. We also observed average medical service cost savings in all 3 years post-index, ranging from $320 to $2556 per patient per year. Per cost allocation analysis, medical service cost savings are observed in inpatient, emergency room, inpatient SNF, urgent care, nursing home/assisted living, and observation unit. Risk adjusted cost savings with statistical significance are observed in year 3 post-index. The mean cost difference between late and early diagnoses: total cost ($15472.11), medical service ($15013.7) and medication cost ($1601.61). The study highlights opportunities for earlier diagnosis and suggests long-term healthcare savings for AD patients who have been diagnosed at MCI stage compared to patients who were directly diagnosed with AD. Further research is needed to drill down into the driver of cost savings and whether it is attributable to disease management of MCI or other complications.

Among the many deep learning methods, the convolutional neural network (CNN) model has an excellent performance in image recognition. Research on identifying and classifying image datasets using CNN is ongoing. Animal species recognition and classification with CNN is expected to be helpful for various applications. However, sophisticated feature recognition is essential to classify quasi-species with similar features, such as the quasi-species of parrots that have a high color similarity. The purpose of this study is to develop a vision-based mobile application to classify endangered parrot species using an advanced CNN model based on transfer learning (some parrots have quite similar colors and shapes). We acquired the images in two ways: collecting them directly from the Seoul Grand Park Zoo and crawling them using the Google search. Subsequently, we have built advanced CNN models with transfer learning and trained them using the data. Next, we converted one of the fully trained models into a file for execution on mobile devices and created the Android package files. The accuracy was measured for each of the eight CNN models. The overall accuracy for the camera of the mobile device was 94.125%. For certain species, the accuracy of recognition was 100%, with the required time of only 455 ms. Our approach helps to recognize the species in real time using the camera of the mobile device. Applications will be helpful for the prevention of smuggling of endangered species in the customs clearance area.

Automated study of brain sub-anatomic region like Corpus Callosum (CC) is challenging due to its complex topology and varying shape. The development of reliable Computer Aided Diagnosis (CAD) systems would help in the early detection of Alzheimer's Disease (AD) and to perform drug trails to palliate the effect of AD. In this work, an attempt has been made to analyse the shape changes of CC using shape based Laplace Beltrami (LB) eigen value features and machine learning techniques. CC from the normal and AD T1-weighted magnetic resonance images are segmented using Reaction Diffusion (RD) level set method and the obtained results are validated against the Ground Truth (GT) images. Ten LB eigen values are extracted from the segmented CC images. LB eigen values are positive sequence of infinite series that describe the intrinsic geometry of objects. These values capture the shape information of CC by solving the eigen value problem of LB operator on the triangular meshes. The significant features are selected based on Information Gain (IG) ranking and subjected to classification using K-Nearest Neighbour (KNN), Support Vector Machine (SVM) and Naïve Bayes (NB). The performance of LB eigen values in the AD diagnosis is evaluated using classifiers’ accuracy, specificity and sensitivity measures.Results show that, RD level set is able to segment CC in normal and AD images with high percentage of similarity with GT. The extracted LB eigen values are found to show high difference in the mean values between normal and AD subjects with high statistical significance. The LB eigen modes λ2, λ7 and λ8 are identified as prominent features by IG based ranking. KNN is able to give maximum classification accuracy of 93.37% compared to linear SVM and NB classifiers. This value is observed to be high than the results obtained using geometric features. The proposed CAD system focuses solely on the geometric variations of CC extracted using LB eigen value spectrum. The extraction of eigen modes in the LB spectrum is easy to compute, does not involve too many parameters and less time consuming. Thus this CAD study seems to be clinically significant in the shape investigation of brain structures for AD diagnosis.

AI-based tool for early detection of Alzheimer's disease