کیوان معقولی

Parkinson’s disease (PD) is a neurological disorder caused by the dysfunction of the basal nuclei, especially the substantia nigra, and a decrease in dopamine levels in the neural circuits of this part. In this paper, to investigate the effects of optogenetic stimulation on parkinsonian nervous systems, a complete model of basal ganglia (BG, including subthalamic nucleus [STN], globus pallidus interna [GPi], globus pallidus externa [GPe], and thalamus [TH] neurons) was first selected to simulate the neurons of the basal nuclei and was developed for PD using three- and four-state optogenetic stimulation.

For this purpose, channelrhodopsin-2 (ChETA), ChRwt, and halorhodopsin (NpHR) opsins were evaluated in three-state and four-state stimulation modes, and different stimulation conditions according to different parameters in the BG model were taken into consideration.

To verify the developed model, the obtained results were compared with the results of experimental studies. The amount of compliance for different stimulation conditions and various genes underwent investigation. The value of the error index was calculated to evaluate the performance of the BG model for each gene in three- and four-state stimulation conditions with different values of basic parameters, and the stimulation conditions that created the error index equal to zero were introduced as optimal conditions. Based on the results, the frequencies of 20 Hz and 200 Hz in the four-state ChRwt model and the frequency of 80 Hz in the three-state ChETA model have been suggested as optimal genes, frequencies, and models.

The response of GPe neurons was consistent with the experimental results, and the response of other neurons was also similar to the response of GPe neurons. In optimal conditions, STN and GPe neurons could provide excitatory input and appropriate inhibitory input to GPi, respectively, and GPi neurons could provide appropriate inhibitory input to TH, and as a result, its function improved and the pathological effects of PD disappeared.

In this research, it has been attempted to improve the function of the basal nuclei of the brain by using optogenetic stimulation and the elimination of the pathological symptoms of PD.

With the increase in the number of infections and deaths of covid-19, the need is felt to use artificial intelligence and machine learning tools to diagnose the corona virus quickly and on time. In this study, using chest x-ray images and VGG-16 deep neural network, an automatic system was designed and implemented to detect cases of covid-19.

In this research, chest x-ray images obtained from the Kaggle database were used. The design of this study included: data sampling, training division, creating directories, transferring images to their own folders, classifying images, creating the proposed VGG-16 model in Python programming language and Keras and Tensorflow libraries, evaluation of the proposed model and finally the creation of the confusion matrix and their analysis and interpretation.

The accuracy and Precision of the proposed model for the positive Covid-19 class were 91 and 93%, respectively. Also, the recall rate and F1-Score for cases with covid-19 were 94% and 88% respectively.

Due to the high accuracy and precision of the proposed model, it can be used to diagnose covid-19 and separate COVID-19 cases from healthy cases and also be used as an auxiliary tool to help doctors diagnose this disease.

Chronic hepatitis B (CHB) leads to liver fibrosis, its failure, and death in the long term. The stage of fibrosis in CHB patients can also be detected based on the biochemical markers. The aim of this study was to predict the state of liver fibrosis in CHB patients and determine the possibility of patients shifting from a given state to another one.

This study is a cross-sectional study conducted in 2021. Age, blood platelet count, AST, and ALT enzymes were used as the input variable to create predictive models. Predictive models were Decision Tree (DT), Naïve Bayes, Support Vector Machine (SVM), and Neural Network (NN). The probability of a patient shifting from a given stage of fibrosis to another was calculated using the transition matrix. The 10-fold cross-validation was used to ensure the generalization of predictive models.

The DT had the best precision, recall, and accuracy (100%) among developed algorithms to predict the stage of fibrosis in CHB patients. The NN was the second most efficient algorithm. Its accuracy and mean square error was 99.35±0.60 and 0.058±0.025, respectively. Besides, SVM had the lowest recall, precision, and accuracy values. Based on the transition matrix results, there is a very low probability that the patients with non-significant fibrosis state shifted to the cirrhosis state.

Computational approaches like machine learning algorithms are the non-invasive way to predict the fibrosis state in CHB patients efficiently.

Today, retinal optical coherence tomography (OCT) plays an important role in ophthalmology and automated analysis of OCT images is very important, The need for labeled data and manual segmentation of these data by the ophthalmologist and limited labeled data are challenges in the field of medical imaging and a large number of labeled data is needed to train deep learning algorithms as well as to evaluate the performance of image processing algorithms.

The data of this project include OCT dataset in the presence of retinal abnormality in age-related macular degeneration disease, obtained from AREDS2 Ancillary device at Duke University. Seventy B scans that are randomly selected from the above set are used to train the algorithm. The active shape model is used to generate synthetic boundaries, and new images are produced similar to the shapes in the training set.

Considering the anatomical features of the existing images such as the number and thickness of the layers as well as the brightness related to them, data with high similarity are produced in the presence of anomalies. Synthetic data were reviewed by an ophthalmologist to evaluate the results.

The proposed model is designed to fill the gap in the production of OCT Synthetic data in the presence of anomalies, using important features such as the boundary of the main layers of the retina and disease-damaged tissue, and can be used as a data set for training deep learning-based algorithms as well as performance of segmentation algorithms.

 Keratoconus is a common complication among corneal defects. As a result of expeditious and extensive progress of medical science in recent decades, corneal ring implantation has turned into a successful surgical procedure to correct the vision of Keratoconus patients; however, selecting the right patient is essential in the success of the operation. The prediction of corneal condition or, more precisely, the prediction of corneal topographic indices after implanting the ring has been taken into consideration in the present study.

Neural network framework is one of the optimal methods for the modeling and prediction. In this study, corneal topographic indices of patients have been predicted 6 and 12 months after the ring implantation for the first time using the multilayer feed forward neural network. The study has focused on predicting corneal topographic indices that are applicable to Keratoconus diagnosis and progression using MATLAB software. Therefore, the statistical data of a number of patients, including the effective indices of the cornea topography before and after implanting the ring, were collected.

 Using the collected data, neural network models have been developed and the corneal topographic indices have been predicted 6 and 12 months after the implantation of keraring and myoring rings. The mean error of 7.22% is achieved for the four trained neural network models. 

 Choosing the right surgical candidate is one of the primary concerns of ophthalmologists. The results indicate the great capability of neural networks in assisting ophthalmologists to select right surgical candidates through predicting corneal topographic indices after the ring implantation.

To diagnose keratoconus from healthy eyes, as well as suspected keratoconus.

Certain parameters were extracted from Casia, Corvis, and Pentacam HR devices for 3 groups of healthy, with keratoconus, and suspected keratoconus. This study was performed on 340 eyes with keratoconus, 310 normal eyes, and 350 suspected keratoconus. The processing method involved the fusion of features derived from 3 devices using optimal and superior feature selection method and Genetic Algorithm (GA). Common methods of z-score normalization, minimum-maximum, Softmax, and maximum value normalization were used to normalize the parameters. The classifiers used were also Multiclass Support Vector Machine (MSVM), Multilayer Perceptron (MLP), and K-Nearest Neighbor (KNN). In the GA method, the best case was shown in MSVM classification and minimum-maximum normalization with mean accuracy and a standard deviation of 87.5±2.5%.

The results showed the separation of three classes of keratoconus, keratoconus suspected and healthy with two methods of GA and MSVM with mean accuracy and standard deviation of 2.5 ± 87.5%.

This result demonstrates the superiority of the principal component analysis method for the fusion of parameters derived from the three Pentacam HR, Casia, and Corvis ST devices and the robustness of the MSVM in separating these three classes.

Keratoconus is a common disease characterized by progressive corneal slimming and steepening. The disease progression is generally accompanied by the significant decline in the vision, aggravation of irregular corneal astigmatism and the resultant decrease in the patientchr('39')s quality of life. One of the successful treatments for Keratoconus is the corneal ring implantation. The prediction of post-surgical visual characteristics has been considered in this study to assist the ophthalmologist in appropriately choosing surgery candidates.

In this study, the set of data collected from numerous tests of visual characteristics performed before and after implanting the keraring and myoring rings has been utilized. By using MATLAB software, the visual characteristics of keratoconus patients after implant ing the rings have been estimated through correct training of the proposed neural networks. The characteristics include: uncorrected visual acuity (UCVA), sphere (SPH), astigmatism (Ast), orientation of astigmatism (Axis), and best corrected visual acuity (BCVA).

In this research, for the first time, the visual characteristics of keratoconus patients six and twelve months after implanting the corneal ring have been predicted with the mean error of 9.51% manipulating a novel neural-network-based method.

The results indicate the precision and accuracy of the proposed method in predicting the visual characteristics of keratoconus patients after implanting the corneal rings. The ophthal mologist could precisely choose the right candidate for surgery amongst his patients based on the estimated characteristics.