Neuroscience informatics最新文献

In this modern era, brain tumour is one of the dreadful diseases that occur due to the growth of abnormal cells or by the accumulation of dead cells in the brain. If these abnormalities are not detected in the early stages, they lead to severe conditions and may cause death to the patients. With the advancement of medical imaging, Magnetic Resonance Images (MRI) are developed to analyze the patients manually. However, this manual screening is prone to errors. To overcome this, a novel depth search-based network termed light weight channel attention and residual network (LWCAR-Net) is proposed by integrating with a novel depth search block (DSB) and a CAR module. The depth search block extracts the pertinent features by performing a series of convolution operations enabling the network to restore low-level information at every stage. On other hand, CAR module in decoding path refines the feature maps to increase the representation and generalization abilities of the network. This allows the network to locate the brain tumor pixels from MRI images more precisely. The performance of the depth search based LWCAR-Net is estimated by testing on different globally available datasets like BraTs 2020 and Kaggle LGG dataset. This method achieved a sensitivity of 95%, specificity of 99%, the accuracy of 99.97%, and dice coefficient of 95% respectively. Furthermore, the proposed model outperformed the existing state-of-the-art models like U-Net++, SegNet, etc by achieving an AUC of 98% in segmenting the brain tumour cells.

In recent years, multiple data-driven fiber orientation distribution function (fODF) estimation algorithms and automatic tractography pipelines have been proposed to address the limitations of traditional methods. However, these approaches lack precision and generalizability. To tackle these shortcomings, we introduce CTtrack, a CNN+Transformer-based pipeline to estimate fODFs and perform tractography. In this approach, a convolutional neural network (CNN) module is employed to project the resampled diffusion-weighted magnetic resonance imaging (DW-MRI) data to a lower dimension. Then, a transformer model estimates the fiber orientation distribution functions using the projected data within a local block around each voxel. The proposed model represents the extracted fODFs by spherical harmonics coefficients. The predicted fiber ODFs can be used for both deterministic and probabilistic tractography. Our pipeline was tested in terms of the precision and robustness in estimating fODFs and performing tractography using both simulated and real diffusion data. The Tractometer tool was employed to compare our method with the classical and data-driven tractography approaches. The qualitative and quantitative assessments illustrate the competitive performance of our framework compared to other available algorithms.

Background

Alzheimer's disease (AD) is the most common cause of dementia. Effective connectivity (EC) methods signify the direction of brain interactions. The identified inter-system mappings can be helpful in characterizing the pathophysiology of the disease.

Methods and Results

We conducted a systematic review of the alterations in EC findings in individuals with AD or Mild Cognitive Impairment (MCI) from PubMed, Scopus, and Google Scholar from fMRI studies. We extracted EC alterations and altered network findings related to specific cognitive impairments. Additionally, we brought a narrative synthesis on the clinical-pathologic relevance of the utilized computational methods. Thirty-nine studies retrieved from the full-text screening. A general pattern of disconnection in several hub centers and changes in inter-network interactions was identified.

Conclusion

In summary, this study demonstrated the beneficial role of EC analyses and network measures in understanding the pathophysiology of AD. Future studies are needed to bring out methodologically consistent data for more structured meta-analytic views.

Alzheimer's disease (AD) as an advanced brain disorder may cause damage to the memory and tissue loss in the brain. Since AD is a mostly costly disease, various deep learning-based models have been presented to achieve a high accuracy classifier to Alzheimer's diagnosis. While a model with high discriminative features is required to obtain a high performance classifier, the lack of image samples in the datasets causes over-fitting and deteriorates the performance of deep learning models. To overcome this problem, few-shot learning such as deep metric learning was introduced. In this paper we employ a novel deep triplet network as a metric learning approach to brain MRI analysis and Alzheimer's detection. The proposed deep triplet network uses a conditional loss function to overcome the lack of limited samples and improve the accuracy of the model. The basic network of this model inspired by VGG16 and experiments are conducted on open access series of imaging studies (OASIS). The experiments show that the proposed model outperforms the state-of-the-art models in term of accuracy.

Background: Artificial intelligence (AI) is one of the active research fields to develop systems that mimic human intelligence and is helpful in many fields, particularly in medicine. (“Role of Artificial Intelligence Techniques ... - PubMed”) Physiotherapy is mainly involving in curing bone-related pain and injuries. The recent emergence of artificially intelligent machines has seen human cognitive capacity enhanced by computational agents that can recognize previously hidden patterns within massive data sets. (“(PDF) Artificial intelligence in clinical practice ...”) In this context, artificial intelligence in pediatric physiotherapy could be one of the most important modalities in delivering better medical and healthcare services to needy people. It is an attempt to identify the types, as well as to assess the effectiveness of interventions provided by artificial intelligence on pediatric physical therapy optimization-related outcomes.

Methods: Data acquisition was carried out by systematic searches from various academic and research databases i.e., google scholar, PubMed, and IEEE from March 2011 to March 2021. Besides, numerous trial registries and grey literature resources were also explored. A total of 187 titles/abstracts were screened, and forty-eight full-text articles were assessed for eligibility.

Conclusions: This research describes some of the possible influences of artificial intelligence technologies on pediatric physiotherapy practice, and the subsequent ways in which physiotherapy education will need to change to graduate professionals who are fit for practice in the 21^st century health system for promoting safe and effective use of artificial intelligence and the delivery of Pediatric Physical Therapy care to people.

Purpose

Cadaveric and electronic dissections are well-established procedures to examine the brain. They are typically variable, content-specific, and not determined in any stereotactic space. We propose a novel approach to use systematically designed stereotactic multi-sequences of neuroimages of the dissected brain with non-dissected 3D structures/systems of interest. Our purpose is three-fold to propose a method for systematic brain dissecting, create a gallery of systematically dissected brain images located in a stereotactic space, and integrate this gallery with the NOWinBRAIN 3D neuroimage repository for public use.

Basic procedures

Systematic brain sectioning consists in the generation of a sequence of dissected image sequences and providing an image naming syntax. Brain dissections are defined by four parameters, dissection direction, dissection location, view of presentation, and appearance (parcellation and labeling).

Main findings

The created dissection gallery contains brain dissections with non-dissected cerebral ventricles, deep gray nuclei, white matter tracts, intracranial arteries, deep cerebral veins, and cranial nerve nuclei. It has 1,942 images organized in 6 albums and 32 sub-albums.

Principal conclusion

Systematic and stereotactic virtual brain dissections cum labeling facilitates exploration of location, course, continuity, extent, and cerebral context of structures and systems which are otherwise fully or partly obscured by the parenchyma. Because of its advantages, user simplicity, and free availability, the dissection gallery with NOWinBRAIN of overall 7,761 images is vital in medicine and beyond for medical students, residents, educators, medical professionals, neuroscientists, medical illustrators, patients, and brain enthusiasts for brain studying, teaching, testing, exploring, referencing, and communicating. This is the first work introducing stereotaxy to brain sectioning.

Background and objective

The hemodynamic model is a fundamental approach for successfully monitoring and possibly forecasting brain activities in the biomedical engineering field. The hemodynamic model describes the inner scenario of a blood flowing voxel in a human brain and it is most popular hypothesis on the brain related research activities. The hemodynamic model has nonlinearities in nature. The solution of such type hemodynamic model is researchable work.

Method

There are many model solving algorithms by using fMRI images; recently, Haifeng Wu presented Confounds Square-root Cubature Kalman Filtering and Confounds Square-root Cubature Smoothing (CSCKF-CSCKS) is the latest approach for solving hemodynamic models. The relative accuracy of this model was shown 84%. In this article, in order to achieve better accuracy, the data analysis and model algorithms are presented differently and find new result that was not mentioned earlier.

Result

The data analysis of this experiment shows that if the maximum number of iterations increases three times, the overall accuracy for solving the hemodynamic model raises by 5.76% under the exact type of fMRI measurements used in both cases. We also represent a formula for calculating a relative error to evaluate the performance of these estimations.

Conclusion

A recommendation is made for solving the hemodynamic model algorithm by using fMRI images to get better performance for estimating the model's biophysical parameters and hidden states. As a result, we will find out more accurate scenario of a specific region of human brain by using fMRI images of that region.

Human activity recognition (HAR) is an active field of research for the classification of human movements and applications in a wide variety of areas such as medical diagnosis, health care systems, elderly care, rehabilitation, surveillance in a smart home, and so on. HAR data are collected from wearable devices which include different types of sensors and/or with the smartphone sensor's aid. In recent years, deep learning algorithms have been showed a significant robustness for classifying human activities on HAR data. In the architecture of such deep learning networks, there are several hyperparameters to control the model efficiency which are mainly set by experiment. In this paper, firstly, we introduced one dimensional Convolutional neural network (CNN) as a model among supervised deep learning for an online HAR data classification. In order to automatically choose the optimum hyperparameters of the CNN model, seven approaches based on metaheuristic algorithms were investigated. The optimization algorithms were evaluated on the HAR dataset from the UCI Machine Learning repository. Furthermore, the performance of the proposed method was compared with several state-of-the-art evolutionary algorithms and other deep learning models. The experimental results showed the robustness of using metaheuristic algorithms to optimize the hyperparameters in CNN.

This article has been withdrawn at the request of the author(s) and/or editor. The Publisher apologizes for any inconvenience this may cause.

The full Elsevier Policy on Article Withdrawal can be found at https://www.elsevier.com/about/our-business/policies/article-withdrawal.

Building a highly efficient machine learning model requires sufficient data to allow robust feature extraction capable of recognizing patterns in each class; thus, the model can distinguish among different classes. It is important to extract effective features from the available amount of data without the need for more real data or improve them using an augmentation technique. The matter gets more complicated if the data is of the image type. In this paper, a new approach for feature extraction called Feature Extraction Based on Region of Mines (FE_mines) is presented that includes three versions to deal with different medical images; this approach obtains multiple formulas for each image using the signal and image processing, then data distribution skew is used to calculate three statistical measurements that include the hidden features, which leads to increased discrimination among classes to build powerful models with better performance and high efficiency. Three experiments were conducted using three types of medical image datasets, namely: Diabetic Retinopathy (Color Fundus photography); Brain Tumor (MRI); and COVID-19 chest (X-ray). The results proved that the FE_mines approach achieved higher accuracy ranges (1 to 13)% within the three experiments than the two traditional methods (RGB and ASPS approaches). In addition, an augmentation technique to increase the size of the dataset is not required which has negative effects on performance. Furthermore, the approach simultaneously included three preprocessing techniques: feature selection, reduction, and extraction.