A Parkinson's disease-related nuclei segmentation network based on CNN-Transformer interleaved encoder with feature fusion.

Brain Magnetic Resonance Imaging (MRI) image segmentation is one of the critical technologies of clinical medicine, and is the basis of three-dimensional reconstruction and downstream analysis between normal tissues and diseased tissues. However, there are various limitations in brain MRI images, such as gray irregularities, noise, and low contrast, reducing the accuracy of the brain MRI images segmentation. In this paper, we propose two optimization solutions for the fuzzy clustering algorithm based on local Gaussian probability fuzzy C-means (LGP-FCM) model and anisotropic weight fuzzy C-means (AW-FCM) model and apply it in brain MRI image segmentation. An FCM clustering algorithm is proposed based on AW-FCM. By introducing the new neighborhood weight calculation method, each point has the weight of anisotropy, effectively overcomes the influence of noise on the image segmentation. In addition, the LGP model is introduced in the objective function of fuzzy clustering, and a fuzzy clustering segmentation algorithm based on LGP-FCM is proposed. A clustering segmentation algorithm of adaptive scale fuzzy LGP model is proposed. The neighborhood scale corresponding to each pixel in the image is automatically estimated, which improves the robustness of the model and achieves the purpose of precise segmentation. Extensive experimental results demonstrate that the proposed LGP-FCM algorithm outperforms comparison algorithms in terms of sensitivity, specificity and accuracy. LGP-FCM can effectively segment the target regions from brain MRI images.

In order to improve the accuracy of brain magnetic resonance imaging (MRI) classification and to reduce classification time, this paper proposes a brain MRI image classification algorithm combined with transfer learning and support vector machine (SVM). Firstly, the three deep convolutional neural networks pre-trained on the ImageNet database, AlexNet, VGG16 and GoogleNet, are used as feature extractors to extract the features of brain MRI images in the Harvard Medical School Website database. The feature extraction process does not require to fine-tune the transferred networks. Then, the features extracted in each convolutional neural network are combined to form a feature vector of each brain MRI image and it is input to the SVM for classification. Finally the SVM classified the brain MRI images into healthy, Alzheimer's disease, and stroke. The experimental results show that the classification accuracy can achieve 100%, and the classification time is only 26 seconds. Compared with the brain MRI image classification algorithm proposed by the literature, the accuracy of the proposed method is increased by 8.67%, 1.09%, 0.55%, respectively. The proposed method can provide effective help for the diagnosis of brain diseases.

For the problem of low accuracy using K-means clustering algorithm to segment noisy brain magnetic resonance imaging (MRI) images, this paper proposed a strategy to improve segmentation accuracy. Firstly, the strategy uses wavelet transform to brain MRI image denoising, secondly, brain MRI image is segmented by k-means clustering algorithm. Experimental results show that the proposed strategy can effectively improve the segmentation accuracy of the noisy MRI brain image and is universal to some extent.

Parkinson's disease (PD) and Parkinson plus syndromes (PPS) are neurodegenerative movement disorders caused by loss of dopamine in the basal ganglia. The diagnosis of both PD and PPS is complex as it is made solely on the basis of clinical features, with no established imaging modality to aid in the diagnosis. Technetium-99m-labeled tropane derivative (Tc-TRODAT-1) binds to the dopamine transporters present in the presynaptic membrane of the dopaminergic nerve terminal. The aim of this prospective study was to investigate the potential usefulness of Tc-TRODAT-1 imaging in the diagnosis of PD and PPS. Fifty-eight patients with a clinical diagnosis of idiopathic PD or PPS were recruited. The severity of the disease was assessed using the Hoehn and Yahr scale. Patients in stage I and II were considered as cases of Early PD. Twenty-five apparently healthy volunteers served as controls. Brain single-photon emission computed tomography/computed tomography in all the participants was performed 3-4 h after an injection of Tc-TRODAT-1. Specific uptake ratios (SURs) of striatum were calculated for both the left and right striatum, and the values were compared between PD, PPS, and healthy volunteers. A significant lower uptake of tracer activity was found in either of the striatum in PD and PPS cases compared with the control group, which showed a symmetrical comma-shaped striatal uptake. This was also reflected in the SUR values, which were significantly higher in the control group in comparison with the PD and PPS patients (P<0.001). A significant difference was also found in the SUR values between the cases of early PD and control group (P<0.001).No significant difference was noted among the SUR values in different Hoehn and Yahr stages. For clinical practice, both the visual analysis and the quantitative parameters of Tc-TRODAT-1 single-photon emission computed tomography/computed tomography showed usefulness in distinguishing cases of PD and PPS from the healthy individuals.

Substantia nigra hyperechogenicity (SN+) detected by transcranial ultrasound (TUS) is useful for Parkinson's disease (PD) diagnosis. Approximately 15% false negative results of unknown significance are reported. However, most TUS studies are transversal, and diagnosis of PD may change during follow-up. Analysis of our prospective registry of TUS in clinical practice, selecting patients with sufficient bone window, to whom TUS was performed because of suspected PD, and a minimum of 3-year follow-up. Subjects were classified regarding SN echogenicity (SN+/SN-). 172 patients (122 SN+, 50 SN-), mean age 71years (25-90), were included. At the end of follow-up, PD diagnosis was retained by 91% SN+ vs. 54% SN- subjects (p < 0.0001), while final diagnosis of atypical parkinsonism (3%SN+ vs. 16%SN-, p:0.0059) was more frequent in SN-. Dopaminergic therapy response was associated with SN+ (88% SN+ vs. 50% SN-, p < 0.0001), as were abnormal DaTSCANs (90%SN+ vs. 56%SN-, p 0.0027). SN echogenicity had 80% sensitivity and 68% specificity for PD diagnosis, while SPECT had 91% and 73%, respectively. SN+ was the only baseline predictor of keeping PD diagnosis at the end of follow-up, with an odds ratio of 12 (95% CI 3-42) (p < 0.001). In our sample of patients with suspected PD, SN hyperechogenicity predicted PD diagnosis in the long term with a high odds ratio. Conversely, a baseline normal SN echogenicity was associated with a poorer response to PD therapy and change to a different diagnosis from PD. Normal SN appears to be a caveat for clinicians to check for atypical parkinsonism features during follow-up.

Magnetic resonance imaging (MRI) is the imaging technique used to diagnosing brain tumor disease. Early diagnosis of brain tumors is an essential task in medical work to find out whether the tumor can potentially become cancerous. Deep learning is a handy and efficient method for image classification. Deep learning has been widely applied in various fields including medical imaging, because its application does not require the reliability of an expert in the related field, but requires the amount of data and diverse data to produce good classification results. Convolutional Neural Network (CNN) is the deep learning technique to perform image classification. In this paper, we compared two model CNN find the best model CNN to classify tumours in Brain MRI Image and at the end, we have trained CNN and obtained a prediction accuracy of up to 93%.

Millions of people worldwide suffer from Parkinson’s disease (PD), a neurodegenerative disorder marked by motor symptoms such as tremors, bradykinesia, and stiffness. Accurate early diagnosis is crucial for effective management and treatment. This article presents a novel review of Machine Learning (ML) and Deep Learning (DL) techniques for PD detection and progression monitoring, offering new perspectives by integrating diverse data sources. We examine the public datasets recently used in studies, including audio recordings, gait analysis, and medical imaging. We discuss the preprocessing methods applied, the state-of-the-art models utilized, and their performance. Our evaluation included different algorithms such as support vector machines (SVM), random forests (RF), convolutional neural networks (CNN). These algorithms have shown promising results in PD diagnosis with accuracy rates exceeding 99% in some studies combining data sources. Our analysis particularly showcases the effectiveness of audio analysis in early symptom detection and gait analysis, including the Unified Parkinson’s Disease Rating Scale (UPDRS), in monitoring disease progression. Medical imaging, enhanced by DL techniques, has improved the identification of PD. The application of ML and DL in PD research offers significant potential for improving diagnostic accuracy. However, challenges like the need for large and diverse datasets, data privacy concerns, and data quality in healthcare remain. Additionally, developing explainable AI is crucial to ensure that clinicians can trust and understand ML and DL models. Our review highlights these key challenges that must be addressed to enhance the robustness and applicability of AI models in PD diagnosis, setting the groundwork for future research to overcome these obstacles.

Diagnosing Brain Tumor with the aid of Magnetic Resonance Imaging (MRI) has gained enormous prominence over the years, primarily in the field of medical science. Detection and/or partitioning of brain tumors solely with the aid of MR imaging is achieved at the cost of immense time and effort and demands a lot of expertise from engaged personnel. This substantiates the necessity of fabricating an autonomous model brain tumor diagnosis. Our work involves implementing a deep convolutional neural network (DCNN) for diagnosing brain tumors from MR images. The dataset used in this paper consists of 253 brain MR images where 155 images are reported to have tumors. Our model can single out the MR images with tumors with an overall accuracy of 96%. The model outperformed the existing conventional methods for the diagnosis of brain tumor in the test dataset (Precision = 0.93, Sensitivity = 1.00, and F1-score = 0.97). Moreover, the proposed model's average precision-recall score is 0.93, Cohen's Kappa 0.91, and AUC 0.95. Therefore, the proposed model can help clinical experts verify whether the patient has a brain tumor and, consequently, accelerate the treatment procedure.

Background Cerebral small vessel disease (CSVD) is associated with an increased risk of stroke, dementia, and death. The risk of CSVD in people with HIV (PWH) is twice the general population but the CSVD risk in PWH with substance abuse is poorly defined. With national surveys showing approximately two-thirds of PWH admitting to using an illicit drug in their lifetime, the evaluation of CSVD risk in this population is needful.Table 1:Demographic and Clinical information of CSVD in PWH with substance use disorder (N=23) Methods We retrospectively analyzed brain magnetic resonance imaging (MRI) findings of PWH attending our ambulatory HIV needle exchange clinic. Brain MRIs done in the previous 5 years before study analysis were included. CSVD was determined by Cerebral small vessel disease score (0 to 4), assigning 1 point each for the following: (i) presence of lacunar infarct, (ii) microbleeds, (iii) prominent perivascular spaces, and (iv) white matter hyperintensities. The findings suggestive of CSVD such as age-related white matter changes and chronic microvascular ischemia were also included in the CSVD assessment. Substance use disorder (uncontrolled use of a substance despite harmful consequences) was determined to be present if this was included on the patient diagnosis list.Table 2:Cerebral Small Vessel disease Score in PWH with substance use disorder. 5 participants had a CSVD score of 0 but "age related white matter disease" on Brain MRI Results 51 patients (PWH with substance use disorder) were included in the analysis. 23 participants out of this group had brain MRI imaging out of which 12 (52.2%) had CSVD. The mean age was 50 years. The demographic and clinical data of the cohort are shown in Table 1. The CSVD score ranged from 0 to 3 with an average score of 1. Table 2 illustrates the CSVD score of the cohort. 5 participants had a CSVD score of 0 but a brain MRI report of “age-related white matter changes”. The Pearson Correlation between age and CSVD score was 0.057 with a p-value of 0.86. Conclusion 52% of PWH who use substance have CSVD on MRI brain imaging. This could translate to an increased risk of stroke, cognitive impairment, and death if not recognized early and aggressively addressed. CSVD score provides objective quantification of CSVD but age-related white matter changes, which are not captured on this score may be equally important especially as PWH continues to age. A systematic approach for screening for CSVD in PWH with substance use disorder is warranted to minimize the unfavorable outcomes associated with this cohort. Disclosures All Authors: No reported disclosures

Parkinson's disease (PD) is a common neurodegenerative disorder that affects millions of people worldwide. Parkinson's disease is driven by low amounts of the neurotransmitter serotonin and it is developed by the subtle degradation of dopamine-producing cells within the brain. Still, there is no real treatment for Parkinson's disease, although the symptoms can be constrained as allowing it to be delayed. Research shows that "diagnosis" is vital for symptom management. However, PD diagnosis is complex, particularly in the early stages of the illness. This study mentions variety of approaches for detecting Parkinson's disease (PD), like deep learning (DL), transfer learning, convolutional neural networks (CNN), machine learning (ML), and the Internet of things (IOT), but it also displays the results of multiple studies conducted by other scientists using different techniques.

Alzheimer&amp;rsquo;s disease (AD) is a progressive neurodegenerative disorder that severely impacts memory and cognitive functions. Early diagnosis remains crucial for timely intervention and care. This research aims to explore the use of artificial intelligence, specifically deep learning, for the early prediction and classification of AD using structural magnetic resonance imaging (MRI) images. A dataset comprising approximately 44,000 brain MRI images with four diagnostic classes (mild, moderate, severe, and very severe dementia) was used to train and evaluate multiple convolutional neural network (CNN) architectures. Three deep learning models were developed and tested: a custom CNN built from scratch, a spatial-channel convolutional attention network (SCCAN), and a pre-trained Visual Geometry Group VGG16 model using transfer learning. The methodology included extensive preprocessing, data augmentation, normalization, and a train&amp;ndash;validation&amp;ndash;test split to ensure robust performance. Evaluation metrics such as accuracy, precision, recall, F1-score, and confusion matrices were used to assess classification efficacy. Among the models tested, the Visual Geometry Group VGG16 model achieved the highest classification accuracy, closely followed by the SCCAN, while the custom CNN demonstrated competitive performance with fewer layers. Grad-CAM visualizations were integrated to provide insight into model decision-making, enhancing interpretability. The results confirm the effectiveness of deep learning in classifying early AD stages with high accuracy and support its integration into clinical diagnostic tools. However, the study also identifies limitations, including dataset diversity, class imbalance, and generalizability across diverse populations. Future research should consider using larger, multi-center datasets (including PET and EEG modalities). This project demonstrates that deep learning can offer reliable, scalable, and interpretable solutions for the early detection of AD, potentially transforming the diagnostic pathway and enabling earlier therapeutic interventions.

The classification of brain tumor is a very important section of medical diagnosis because correct and early diagnosis can greatly enhance the outcome of the patient. To aid in automated and dependable diagnostic methods, this paper examines the efficacy of already trained deep learning model to classify brain magnetic resonance imaging (MRI) scans into four categories, namely, pituitary tumor, meningioma, glioma, and no tumor. This was done using a publicly available set of 7,023 brain MRI images. DenseNet121, ResNet50, Xception and MobileNet four advanced convolutional neural network (CNN) models were also fine-tuned with transfer learning, as well as image preprocessing and data augmentation methods. Transfer learning was used to facilitate effective model adaptation with a low computation complexity and improved classification. According to the experimental findings, DenseNet121 performed better as it obtained the highest classification accuracy of 98.47% and the highest F1 score of 98.47%. The assessed models presented a high level of generalization and steady performance under a variety of evaluation criteria, which also indicates their possible application in the clinical decision support systems. These encouraging findings, however, need additional performance to improve recall on individual tumor classes, as well as the interpretability of deep-learning predictions in clinical settings. On the whole, the results highlight the great potential of the deep learning and transfer learning methods in medical image analysis, which is an opportunity to establish a more reliable, scalable and efficient solution to diagnosis. The next phase of work will involve diversifying the datasets, enhancing the model explainability, and assessing the performance of AI-oriented healthcare systems in practice to make them more widely adopted.

Olfactory Function, Eating Ability, and Visceral Obesity Associated with MMSE Three Years after Parkinson's Disease Diagnosis.

Since there is no diagnostic biological marker for Parkinson disease (PD), the diagnosis is based on the results of clinical assessment. The accuracy of diagnosis improves with time and repeated assessments. Studies that require only inclusion of early cases of PD present a diagnostic challenge. Previous studies concluded that initial diagnoses of PD made by general neurologists were incorrect in 24% to 35% of the cases when patients were examined at autopsy. Experts in movement disorders are expected to have greater accuracy of initial diagnosis of PD. To determine the evolution of clinical diagnosis in patients with early PD made initially by experts in PD. Eight hundred patients with mild parkinsonian symptoms (Hoehn and Yahr stage 1 or 2) who received a diagnosis of PD less than 5 years before the beginning of the study were included in the original Deprenyl and Tocopherol Antioxidative Therapy for Parkinson's Disease study. These patients were followed up prospectively with repeated clinical assessments. The following clinical criteria were used to reassess the initial diagnosis: investigator's confidence in the diagnosis of PD, presence of atypical clinical features, findings of imaging studies, response to levodopa, and results of autopsy examinations. The mean +/- SD duration of illness in the 800 cases at enrollment was 2.2+/-1.3 years, and the mean +/- SD Hoehn and Yahr stage was 1.6+/-0.5. The mean +/- SD follow-up was 6.0+/-1.4 years (range, 0.2-7.6 years). In 5 cases, PD was not confirmed at autopsy, and in 15 patients, the results of imaging studies indicated the presence of other pathological conditions. Of the 550 cases treated with levodopa, 49 (8.9%) had little or no improvement; 6 of these cases overlap with either autopsy or imaging study exclusion criteria. Two other cases had at least 4 of the 6 atypical clinical features arguing against the diagnosis of PD. Thus, of the 800 patients, 65 (8.1%) did not have PD according to the study criteria. Compared with those patients with the final diagnosis of PD, in the diagnoses of 60 patients without autopsy, the duration of symptoms (mean +/- SD, 7.2+/-2.0 years vs. 8.3+/-1.9 years; P<.001) and the duration of follow-up (5.3+/-1.6 years vs. 6.1+/-1.3 years; P<.001) were shorter. We found that 65 (8.1%) of patients initially diagnosed as having PD were later found to have an alternate diagnosis based on multifactorial clinical diagnostic criteria. This alternate diagnosis indicated that experts in PD changed their diagnoses infrequently during the 7.6-year follow-up.

Parkinson's disease (PD) is associated with a higher risk of hip fracture. However, the risk of hip fracture at the prodromal stage of PD is unknown. Our study aimed to investigate the incidence of hip fractures in persons with and without PD in Finland from 5years before to 10years after the diagnosis of PD. We included persons diagnosed with PD between 2000 and 2009 in Finland and 1:2 matched comparison cohort. Hazard ratio [HR] was computed to assess overall risk of incident hip fracture. The annual incidence rate per 1000 person-year [IR/1000PY] and incidence rate ratios [IRR] were calculated for each year of follow-up from 5years before to 10years after PD diagnosis. During the follow-up among 33,153 eligible persons, 13.4% persons with and 5.3% persons without PD had an incident hip fracture (HR 1.82, 95% confidence interval [CI] 1.47-2.26). Persons with PD had higher annual incidence rates of hip fracture starting from 3years before diagnosis (with PD: IR/1000PY 2.83, 95% CI 1.93-4.02; without PD: IR/1000PY 1.64, 95% CI 1.15-2.28). The higher annual incidence rates continued until 10years after diagnosis (with PD: IR/1000PY 15.7, 95% CI 11.7-20.5; without PD: IR/1000PY 4.53, 95% CI 3.08-6.44). Prevalence of fall-risk-increasing medications but not anti-osteoporosis medications was higher at hip fractures among persons with PD than without PD. As the risk of hip fracture is elevated already years before the diagnosis of PD, it is important to manage fall risks concomitantly while applying the diagnostic procedures for PD.