Theoretical Analysis of Coupled Modified Hindmarsh-rose Model Under Transcranial Magnetic-acoustic Electrical Stimulation

Transcranial magnetic-acoustic electrical stimulation (TMAES) is a new technology with ultrasonic waves and a static magnetic field to generate an electric current in nerve tissues to modulate neuronal firing activities. The existing neuron models only simulate a single neuron, and there are few studies on coupled neurons models about TMAES. Most of the neurons in the cerebral cortex are not isolated but are coupled to each other. It is necessary to study the information transmission of coupled neurons. The types of neuron coupled synapses include electrical synapse and chemical synapse. A neuron model without considering chemical synapses is not comprehensive. Here, we modified the Hindmarsh-Rose (HR) model to simulate the smallest nervous system—two neurons coupled electrical synapses and chemical synapses under TMAES. And the environmental variables describing the synaptic coupling between two neurons and the nonlinearity of the nervous system are also taken into account. The firing behavior of the nervous system can be modulated by changing the intensity or the modulation frequency. The results show that within a certain range of parameters, the discharge frequency of coupled neurons could be increased by altering the modulation frequency, and intensity of stimulation, modulating the excitability of neurons, reducing the response time of chemical postsynaptic neurons, and accelerating the information transferring. Moreover, the discharge frequency of neurons was selective to stimulus parameters. These results demonstrate the possible theoretical regulatory mechanism of the neurons' firing frequency characteristics by TMAES. The study establishes the foundation for large-scale neural network modeling and can be taken as the theoretical basis for TMAES experimental and clinical application.

NEURAL NETWORK CALCULATION OF CARBON DIOXIDE CONCENTRATIONS

Снижение углеродного следа в настоящее время является одной из приоритетных задач мировой экономики. Для достижения этой цели необходимо с одной стороны снижать выбросы парниковых газов, с другой стороны развивать методы мониторинга парниковых газов в атмосферном воздухе для обеспечения контроля эффективности принимаемых решений.Учитывая сложность процессов рассеивания газов в атмосферном воздухе, значительными преимуществами в вопросах определения концентраций атмосферных примесей обладают нейросетевые методы моделирования. В данной статье представлен метод расчета концентраций углекислого газа в атмосферном воздухе с помощью спроектированной и обученной каскадной нейросетевой модели, позволяющей при расчете концентраций учитывать сложное влияние метеорологических факторов и локальных условий рассеивания. Первым уровнем модели является расчет концентрации оксида углерода по известным параметрам источников выбросов этого вещества с использованием регламентированной методики расчета рассеивания примесей в атмосфере в Унифицированной программе расчета рассеивания «Эколог-Город». Вторым уровнем является нейронная сеть, которая корректирует рассчитанную на первом шаге концентрацию по заданным метеорологическим параметрам для увеличения точности моделирования. Третьим уровнем является нейронная сеть, позволяющая по полученной на предыдущем шаге концентрации оксида углерода, а также измеренным значениям коэффициента химической трансформации и концентрации атмосферного озона производить расчет концентрации углекислого газа.Полученная каскадная модель апробирована на территории г. Нижнекамск. Достигнутая точность расчета концентрации углекислого составила более 95%. Таким образом, представленная технология позволяет расширить возможности локальной системы мониторинга в условиях недостаточного количества измерений диоксида углерода. Reducing the carbon footprint is currently one of the priorities for the world economy. To do this, it is necessary to reduce greenhouse gas emissions, as well as to develop methods for monitoring greenhouse gases in the atmospheric air to ensure control over the effectiveness of decisions taken.Considering the complexity of the processes of dispersion of gases in the atmospheric air, neural network modeling methods have significant advantages in determining the concentrations of atmospheric impurities. This article presents a method for calculating the concentration of carbon dioxide in the atmospheric air using a designed and trained cascade neural network model, which makes it possible to take into account the complex influence of meteorological factors and local dispersion conditions when calculating concentrations. The first level of the model is the calculation of the concentration of carbon monoxide according to the known parameters of the emission sources of this substance using the regulated method for calculating the dispersion of impurities in the atmosphere in the Unified program for calculating dispersion "Ecolog-City". The second level is a neural network, which corrects the concentration calculated at the first step according to the specified meteorological parameters to increase the modeling accuracy. The third level is a neural network that allows calculating the concentration of carbon dioxide based on the concentration of carbon monoxide obtained at the previous step, as well as the measured values of the coefficient of chemical transformation and concentration of atmospheric ozone.The resulting cascade model was tested on the territory of Nizhnekamsk. The achieved accuracy of calculating the concentration of carbon dioxide was more than 95%. Thus, the presented technology makes it possible to expand the capabilities of the local monitoring system in conditions of an insufficient number of measurements of carbon dioxide.

USING NEURAL NETWORK MODELING TO PREDICT THE COURSE OF ACUTE PANCREATITIS

The database of studies of 82 patients with acute pancreatitis are presented. Using neural network analysis, the most indicative parameters for predicting acute pancreatitis were revealed: indexes of Kalf-Kalif intoxication modified by Kostyuchenko and Khomich, Reis, Garkavi, the ratio of leukocytes to ESR, leukocyte index, general intoxication index; sonographic parameters – the size of the head of the pancreas, the diameter of the splenic vein, the presence of free fluid in the abdominal cavity; biochemical parameters – blood amylase concentration, urine diastase. When conducting clustering in a multidimensional feature space, a Kohonen neural network was created. All analyzed objects were effectively divided into 3 clusters. The most severe and prognostically unfavorable is cluster 1, which included data from 30 patients, with the maximum mortality rate and maximum hospital stay.