43 results about "Neuron structure" patented technology

The activity of a pharmaceutical preparation or medicament on a neuronal structure is analyzed by subjecting a neuronal structure to the influence of a pharmaceutical preparation. Signals describing neuronal activities in the neuronal structure under the influence of the pharmaceutical preparation are detected and statistically evaluated to determine indicators for the neuronal structure under the influence of the pharmaceutical preparation. The indicators describe the activity of the pharmaceutical preparation.

The invention discloses a method for classifying multi-variable time sequences based on a deep long/short-term memory neural network. The method includes the steps of by selecting long/short-term memory neuron structures as hidden layer neuron structures of a recurrent neural network, overlapping long/short-term memory neurons, and designing a deep layer recurrent neural network classification framework, thereby achieving the aim of improving the classification accuracy of multi-variable time sequence data. It is found through experimental comparison that accuracy is higher compared with an existing classification model and universality is achieved on time sequence data set classification tasks in multiple fields.

A process of correlating cognitive response to a stimulus is provided by collecting biometric data with a camera having a microprocessor for modifying an image by localizing eyes within a face in the image, collecting neural activity data from a subject while exposed to a stimulus and mapping neural activation in response to the stimulus onto neuronal structure of a subject brain. A unique three-dimensional cognitive engram representative of the neural activity caused by the stimulus is then constructed. The cognitive engram is then compiled in a database containing a plurality of engrams, each of which is associated with a different stimulus for interpreting cognitive response. Then, each unique three-dimensional cognitive engram stored in the database is correlated with the biometric data representative of a stimulus.

The invention discloses a multilayer perceptron artificial nerve network based on a residual error network; the multilayer perceptron artificial nerve network based on the residual error network comprises a plurality of network module structures; a full connection mode is employed to replace the convolution in the residual error nerve network; Neuron units in the network module structures can obtain the output of the complete residual error module through the output of each hidden layer, wherein the output of each hidden layer refers to si=ReLU[BN(neti)]; the complete residual error module output refers to oi=ReLU[BN(neti+1)+neti]. The multilayer perceptron artificial nerve network is based on the residual error network, small in computational complexity, high in accuracy, and can be applied to more fields besides the image field.

This disclosure relates to modulation of the interactions between proNTs and p75^NTR/SorCS2 expressed on neuronal cells. Inhibition of such interactions is useful for reducing unwanted synaptic elimination, neurite pruning and/or other neuronal structural collapses, and for treating early stage and mild neurological disorders including mild cognitive impairment.

The invention discloses a rapid neuron staining method based on a Golgi silver staining method. The rapid neuron staining method based on the Golgi silver staining method is characterized in that according to mass volume ratio concentration, the aqueous solution with 5 % of potassium dichromate, the aqueous solution with 5 % of mercuric chloride and the aqueous solution with 5 % of potassium chromate are prepared into a Golgi silver staining solution in the volume ratio of 1:1:1, a brain tissue is soaked in the silver staining solution at 37 DEGCfor 36-48 hours, the soaked brain tissue is sliced on a vibrating slicer, tissue sections are arranged on a glass slide coated with gelatin, the tissue sections stay overnight, then conventional ammonia developing is conducted, gradient alcohol dehydration is conducted, the transparency process is conducted by xylene, finally, the tissue sections are sealed by netrual gum, and the tissue sections are observed through a microscope. According to the rapid neuron staining method based on the Golgi silver staining method, staining can be conducted on each encephalic region of the brain, the staining result is clear, details are obvious, and research staff can observe the neuron structure of the brain tissues conveniently, the convenient and fast means is provided for case analysis of nervous system lesions, and the rapid neuron staining method based on the Golgi silver staining method has significant meanings for the field of foundational research of neurology.

The invention provides a user behavior feature extraction method, device, equipment and system. The method comprises the steps of collecting behavior data of a target user; inputting the behavior datainto a preset self-encoding network model for learning, wherein the preset self-encoding network model is obtained by training based on a preset neuron structure,the preset neuron structure comprisesa time interval sequence and an attention mechanism, the time interval sequence is used for reflecting the influence of an operation time interval on neuron memory updating, and the attention mechanism is used for reflecting the difference between a historical behavior and a current behavior; and taking a result output by a cognitive network in the preset self-encoding network model as behavior characteristic information of the target user. By utilizing the embodiment of the invention, the user behavior characteristics can be extracted more accurately, and the anti-fraud detection difficultyis reduced.

A synthetic neuronal structure makes use of a semiconductor-metal phase transition material having material regions separated by discontinuities. The discontinuities represent interfaces such that different phases in two adjacent regions result in a metal-semiconductor interface. The interface supports a charge accumulation and a discharge of accumulated charge when an activation energy provided, for example, by electrical current, localized heating or optical energy, reaches a threshold necessary for breakdown of a potential barrier presented by the interface, and thus mimics a leaky integrate-and-fire neuron. With many such interfaces distributed through the structure, the local inputs to a neuron become a weighted sum of energy from neighboring neurons. Thus, different combinations of signals at one or more inputs connected to the structure will favor different neural pathways through the structure, thereby resulting in a neural network.

The invention relates to an MXene/nitrogen-doped carbon foam composite material with a 3D porous neuron-like structure and a preparation method thereof. The composite material takes porous nitrogen-doped carbon foam as a skeleton, pores of the porous nitrogen-doped carbon foam are of a neuron-like structure, and MXene nanosheets are uniformly deposited on the surface of the skeleton and dispersedamong the pores of the porous nitrogen-doped carbon foam. According to the MXene/nitrogen-doped carbon foam composite material provided by the invention, stacking and aggregation of the MXene nanosheets are effectively prevented, the composite material as an electrode material can provide more contact sites, the permeation of electrolyte is promoted, an effective channel is provided for the transmission of ions in electrodes, the ion transfer is accelerated, the capacitance and the rate performance can be significantly improved, and the assembled all-solid-state supercapacitor device has highcompressibility and excellent cycle stability, and has a good application prospect.

The invention relates to a displacement test method which is suitable for static and dynamic displacement measurement of a structure of scale in a long line range or a space range. The neuron model displacement or deformation monitoring method of an electric transducer embedded with a microcomputer is characterized in that a multi-point neuron structure model, the transducer as a neuron body and a supplement rigid rod as a neuron myelin are connected to form a neuron basic unit assembly; a number of neuron assemblies are connected in series to form a nerve chain; and a start node is used as a benchmark to acquire the relative displacement of each unit to determine the attached structure deformation. Compared with the prior art, the structure deformation and dynamic characteristic measuring method provided by the invention can overcome the limits of traditional mechanical temporary test and optical test method environment impacts and high cost price, and is especially suitable for large range or long scale and multi-point long-term monitoring, such as longitudinal and transverse deformation measurement and real-time monitoring of a building, a bridge, a tunnel, a dam, a slope and other structures.

The invention discloses a system and a method of multiple-target control of an intelligent power grid. Load current harmonics and idle components are directly separated from the system, the fluctuation situations of voltage are monitored at the same time, a multi-objective decision theory is adopted, targets are regulated according to different power qualities, and the work modes of APF, SVG, DVR and other devices are coordinated, so that the aim of improving power grid quality to the maximum extent is reached. The system of the multiple-target control of the intelligent power grid is mainly formed by an idle detecting unit based on self-adaptive linear neuron structural harmonics and a controller, wherein the idle detecting unit based on the self-adaptive linear neuron structural harmonics is mainly formed by a sampling unit, a current control circuit and a control unit. The sampling unit is connected on a power grid power supply with load, and the sampling unit and the control unit are connected through the current control circuit.

The invention relates to the technical field of agricultural machinery, and specifically relates to an intelligent agricultural machinery fertilization method and device based on the Internet of things. According to the invention, an agricultural information server terminal sends water and fertilizer ratio information and motor speed recommendation information to a main control module through a wireless signal module according to a geographical location information matching result, a water and fertilizer ratio module builds a neuron structure model, the main control module controls the water and fertilizer ratio module, a speed control module and a liquid pump on the basis of the neuron model and controls an agricultural machinery fertilization device to perform a fertilization operation by combining an environment compensation module, an Internet of things monitoring terminal realizes remote operation monitoring through a GPS data terminal switch and a man-machine interaction unit, the water and fertilizer ratio module is controlled to realize different ratios, the motor is controlled to realize different rotating speeds and the liquid pump is controlled to realize different operating conditions according to the recommended operation information, uncertain systems can be learnt and adapted by using the neural network, a method basis is provided for variable rate fertilization, and the scientificity and the reliability of the operation are improved.

The invention discloses an artificial sensory neuron structure based on a multi-side grid synaptic device and a preparation method thereof. The artificial sensory neuron structure comprises at least two piezoelectric nano-generators used for external force induction and a synaptic device used for processing at least two voltage input signals, wherein the synaptic device is an electric double-layer transistor; the double-electrode-layer transistor takes an electrolyte material as a gate medium, takes an oxide semiconductor as a channel layer, and is provided with at least two plane side gates; and each piezoelectric nano generator is electrically connected to one side grid of the electric double-layer transistor. According to the artificial sensory neuron structure, the planar multi-side grid structure and the double-electric-layer coupling characteristic of the double-electric-layer transistor are fully exerted, and two or more sensors can be connected to different side grids of one transistor at the same time, so that induction and processing of various external signals are realized; the limitation that one synaptic device of a traditional artificial sensory neuron structure processes one sensing device signal is broken through.

An exemplary system, method, and computer-accessible medium for generating an image(s) of an anatomical structure(s) in a biological sample(s) can include receiving first wide field microscopy imaging information for the biological sample, generating second imaging information by applying a gradient-based distance transform to the first imaging information, and generating the image(s) based on the second imaging information. The second imaging information can be generated by applying an anisotropic diffusion procedure to the first imaging information. The second imaging information can be generated by applying a curvilinear filter and a Hessian-based enhancement filter after the application of the gradient-based distance transform. The second information can be generated by applying (i) a tube enhancement procedure or (ii) a plate enhancement procedure after the application of the gradient-based distance transform.

A method for solving a multi-target shortest path in a time-varying environment based on a neural network comprises the following steps: monitoring and collecting data of an urban traffic network in real time by using map software, collecting a cost function of each edge in the network, and constructing a traffic network model; in combination with the topological structure of the network, designing a brand-new neuron structure and constructing a neural network model. Topological information of a traffic network is loaded into a neural network model, a neuron wave generator and a neuron filter are used for sending automatic waves to subsequent neuron, other neuron in the network is activated, and each neuron ignores response to part of waves through the filter. According to the method, non-optimal sub-paths are pruned by designing the neural network, so that the calculation complexity is greatly reduced, and the problem of accurately solving the multi-target shortest path in the network optimization field in a time-varying environment is solved. In addition, the problem solving speed is increased through parallel computing of the neural network.

The invention discloses a multiple cerebral infarction modeling device and method prepared from soft autologous blood microembolus, and relates to the field of animal modeling. The method comprises the following steps: injecting autologous blood soft microembolus coated with carrageenan into the brain of a rat through an external carotid artery, establishing a multiple cerebral infarction animal model, detecting cerebral infarction disease related experimental markers, observing pathological changes of brain tissues, neuronal structure changes and generation of spatial behavior dysmnesia, and establishing a multiple cerebral infarction animal model. Comprising the steps of red tetrazole staining, hematoxylin-eosin staining, immunohistochemistry, water maze and Bass maze. The rat model proves that the rat has spatial behavior memory disorder, neurons and brain tissues have pathological mechanism changes, and an animal model with similar pathogenesis processes is provided for researching clinical multiple cerebral infarction. The establishment of the animal model has an important value for deepening the treatment research of multiple cerebral infarction diseases.

The embodiment of the invention provides a method, a device and equipment for determining public opinion categories. The method, the device and the equipment for determining the public opinion categories can be applied to the technical field of big data. The method comprises the following steps: acquiring specified public opinion abstract information; and processing the specified public opinion summary information by utilizing a public opinion classification model to obtain a public opinion category corresponding to the specified public opinion summary information, wherein the public opinion classification model is obtained by training a preset long and short term memory model based on public opinion abstract information and a corresponding public opinion category, and the neuron structure of the preset long and short-term memory model comprises an external connection gate, and the external connection gate is used for enhancing the connection between input data and output data. The public opinion category can be efficiently and accurately determined by utilizing the embodiment of the specification.

The invention discloses a self-organizing T-S fuzzy neural network PID (Proportion Integration Differentiation) controller for controlling a grid-connected inverter. A concept of intensity activation is added to a controller algorithm on the basis of a T-S fuzzy neural network. The essence of the method is that a fuzzy rule layer activation intensity method is adopted, a neuron structure is adaptively adjusted according to an actual environment, a proper control structure is constructed, and the performance and scale of the whole network can be better determined and improved due to the online learning capability and the powerful fault-tolerant capability of the method, so that the control precision is improved. And meanwhile, a gradient descent method is adopted to adjust each parameter of the controller in real time, so that the control system has better anti-interference capability and dynamic performance.

The invention discloses a constrained time-varying shortest path solving method based on a time wave delay neural network, and the method comprises the steps of solving a constrained time-varying shortest path problem(CTSPP for short) through the construction of a time wave delay neural network (TWDNN for short). The designed TWDNN is a novel neural network based on time wave delay neurons, the difference between the TWDNN and a traditional neural network is that the TWDNN does not need any training, and the design of the TWDNN comprises neuron structure design and neural network structure design. Each neuron on the time wave delay neural network is composed of an input part, a wave receiver part, a neuron state part, a wave generator part, a wave transmitter part and an output part. The wave is a medium for exchanging information among the neurons, the propagation delay of the wave among the neurons is equivalent to the transmission delay of a data packet on a network, and the output of the shortest path is based on the time wave received by the target neuron. Experimental results show that the method provided by the invention is superior to the existing method.