Information processing, memories, and synchronization in chaotic neural network with the time delay

Relevance. Modern artificial neural network models require significant energy and other resources for training and operation. Training generative models involves vast amounts of data. At the same time, these models face challenges related to the trustworthiness of the information they generate. An alternative to current paradigms of building and training neural networks is the development of energy-based models, which could potentially overcome these shortcomings and bring information processing closer to biologically and physically grounded processes. However, existing energy-based models differ little from classical models in terms of their limitations and drawbacks. Therefore, developing new approaches to modeling energy-based information processing in neural networks is highly relevant. The object of research is the process of information processing in artificial neural networks. The subject of the research is the mathematical models for the construction and training of artificial neural networks. The purpose of this paper is to develop and experimentally validate a theoretical framework that postulates the energetic nature of information and its role in the self-organization and evolution of complex information systems. Research Results. A fundamental theory is proposed, describing information as a structure of perceived external energy parameters that govern the processes of forming the internal energetic structure of a system—its model of the external world. This theory encompasses concepts of energy landscapes, principles of energy-based structural and parametric reduction, and a critical analysis of existing computational paradigms. Experimental studies on the construction and training of the developed energy-based model confirm its high generalization ability in one-pass training without using the backpropagation algorithm on ultra-small training datasets.

Chemical reactions provide new insights concerning information processing in neural networks. Batch reactions are considered in first place and it is presented a model in which computational instructions will be conceptualised as being sent in parallel to multiple operational structures in sites where they are transformed in a way that generates a variety of computational viewpoints associated to a set of data.Anastomotic computations are considered as pre-processing operations implemented by convolutional weighing functions. Quantum valence theory is used to construct computational models for cognitive processes. The model of isolobal fragments is interpreted in a way that gives rise to classes of equivalence that represent concepts in extenso.The conscious quality of cognition is derived from the emergence of new macroscopic qualities from the relationships between valence bonds. Binding of representations is considered as dependent on dendro-dendritic relationships homeomorphic with orbital valence frontiers.

Dissociated neuronal cultures provide a powerful, simplified model for investigating self-organized prediction and information processing in neural networks. This review synthesizes and critically examines research demonstrating their fundamental computational abilities, including predictive coding, adaptive learning, goal-directed behavior, and deviance detection. A unique contribution of this work is the integration of findings on network self-organization, such as the development of critical dynamics optimized for information processing, with emergent predictive capabilities, the mechanisms of learning and memory, and the relevance of the free energy principle within these systems. Building on this, we discuss how insights from these cultures inform the design of neuromorphic and reservoir computing architectures, aiming to enhance energy efficiency and adaptive functionality in artificial intelligence. Finally, this review outlines promising future directions, including advancements in three-dimensional cultures, multi-compartment models, and brain organoids, to deepen our understanding of hierarchical predictive processes in both biological and artificial systems, thereby paving the way for novel, biologically inspired computing solutions.

As a background to a comprehensive discussion on recent work in neural modelling, the thesis provides a brief biological background description of some of the components of Natural Neural Systems. The review of the work of earlier neural modellers concentrates on the work of the last fifteen years, with the exception of a few earlier, classic papers. The research work performed for this thesis employs an Experimental System evolved to test the ability of a digital, discrete-time model of a neural net to process patterns of signals, provided as input. The aim of the Experiments was to find the type of networks that can perform useful pulse processing functions. The Experimental work is divided into two sections, the first analysing non-adaptive nets and the second analysing networks which use Hebb-type algorithms to alter the strength of interconnections between cells. The first section describes and displays activity of many neural nets. Different inputs are applied to the net and the effects noted. The study is extended by employing Spectral Analysis techniques. The effect of many parameters on frequency of firing of the net are examined including, for example, the decay rates used in the cells and the frequency of the input signal. The second section simulates adaptive nets and examined the relationship between input signals and the final activity of the adapted net. It also employs Spectral Analysis and a specially defined form of display, the Cell Firing Histogram which provides information on how the circuits are being altered by the algorithm. A simple mechanism for recognising signal patterns is proposed that employs several of the properties discovered using the Cell Firing Histogram.

This paper reviews some central notions of the theoretical biophysics of neural networks, viz., information coding through coherent firing of the neurons and spatio-temporal spike patterns. After an introduction to the neural coding problem we first turn to oscillator models and analyze their dynamics in terms of a Lyapunov function. The rest of the paper is devoted to spiking neurons, a pulse code. We review the current neuron models, introduce a new and more flexible one, the spike response model (SRM), and verify that it offers a realistic description of neuronal behavior. The corresponding spike statistics is considered as well. For a network of SRM neurons we present an analytic solution of its dynamics, analyze the possible asymptotic states, and check their stability. Special attention is given to coherent oscillations. Finally we show that Hebbian learning also works for low activity spatio-temporal spike patterns. The models which we study always describe globally connected networks and, thus, have a high degree of feedback. We only touch upon functional feedback, that is, feedback between areas that have different tasks. Information processing in conjunction with functional feedback is treated explicitly in a companion paper [94].

One of the central questions in neuroscience is how particular tasks, or computations, are implemented by neural networks to generate behavior. The prevailing view has been that information processing in neural networks results primarily from the properties of synapses and the connectivity of neurons within the network, with the intrinsic excitability of single neurons playing a lesser role. As a consequence, the contribution of single neurons to computation in the brain has long been underestimated. Here we review recent work showing that neuronal dendrites exhibit a range of linear and nonlinear mechanisms that allow them to implement elementary computations. We discuss why these dendritic properties may be essential for the computations performed by the neuron and the network and provide theoretical and experimental examples to support this view.

In this paper, anti-synchronization problem of two identical chaotic neural networks with time-varying delays is proposed. By using time-delay feedback control technique, mean value theorem and the Leibniz-Newton formula, and by constructing appropriately Lyapunov-Krasovskii functional, sufficient condition is proposed to guarantee the asymptotically anti-synchronization of two identical chaotic neural networks. This condition, which is expressed in terms of linear matrix inequality, rely on the connection matrix in the drive and response networks as well as the suitable designed feedback gains in the response network. Finally, the anti-synchronization of two chaotic cellular neural network and Hopfield neural network with time-varying delays are considered to illustrate the effectiveness of the proposed control scheme, in which, when compared with the nonlinear feedback control method, the proposed method shows superior performance. Key words: Lyapunov-Krasovskii functional, chaotic neural networks, anti-synchronization, time-varying delay, linear matrix inequality.

Stochastic resonance (SR) has been shown to enhance the signal-to-noise ratio and detection of low level signals in neurons. It is not yet clear how this effect of SR plays an important role in the information processing of neural networks. The objective of this article is to test the hypothesis that information transmission can be enhanced with SR when sub-threshold signals are applied to distal positions of the dendrites of hippocampal CA1 neuron models. In the computer simulation, random sub-threshold signals were presented repeatedly to a distal position of the main apical branch, while the homogeneous Poisson shot noise was applied as a background noise to the mid-point of a basal dendrite in the CA1 neuron model consisting of the soma with one sodium, one calcium, and five potassium channels. From spike firing times recorded at the soma, the mutual information and information rate of the spike trains were estimated. The simulation results obtained showed a typical resonance curve of SR, and that as the activity (intensity) of sub-threshold signals increased, the maximum value of the information rate tended to increased and eventually SR disappeared. It is concluded that SR can play a key role in enhancing the information transmission of sub-threshold stimuli applied to distal positions on the dendritic trees.

This paper presents a novel neuron model for information processing in neural networks of the spike-neuron class. The model called hybrid biological neuron, HBN is trained to be an autonomous oscillator, thus is able to model cardiac action potentials and frequency of cardiac rhythmicity. The HBN is a new model that allows pulse trains to be interpreted on one hand as a logic function, and on the other hand as a continuous time system in which the pulse shape represents a second order modulation of the information not encoded in the patterns only. The HBN belongs to a spiking neuron class but models chemical synapses and considers the receptive field caused by the direct influence of other dendrites, specifically the ones that do not pass through the pre-synaptic filters.

With the macroeconomy entering a new normal, many new problems are exposed in all walks of life, and the risk of default in the financial sector is also being exposed at an accelerated pace. In the context of big data, internet finance, as an important part of the financial market, also faces many risks in the process of its rapid development. Reasonable, scientific, and effective prediction and prevention of financial default risk have become a key link in the process of risk management practice in the financial industry. Based on the powerful prediction function of the neural network, this paper combined neural network and chaos theory to construct a chaotic RBF neural network. It was applied to financial default risk prediction, which made the prediction accuracy and efficiency higher. The chaotic neural network solves the shortcomings of unstable prediction in the basic neural network and can comprehensively and accurately predict the financial default risk, so as to take measures to prevent risks. The experimental results of this paper show that the accuracy rate of the chaotic RBF neural network reaches 95%, while the accuracy rates of the BP neural network and the RBF neural network are 67% and 78%, respectively. Although the prediction accuracy of these two methods is also high, it is still not as high as the chaotic RBF neural network. Therefore, it is very meaningful to choose the chaotic RBF neural network to predict financial default risk in this paper.

The chaotic neural network is studied under the threshold activated coupling, which provides a controlled output patterns. In general the chaotic neural network constructed with chaotic neurons exhibits very rich dynamic behaviors with a nonperiodic associative memory. In the chaotic neural network, it is difficult to distinguish the stored patterns from others, because of the chaotic states of output of the network. In order to apply the nonperiodic associative memory into information search and pattern recognition, etc, it is necessary to control chaos in this chaotic neural network. Chaos in the chaotic neural network is controlled with threshold activated coupling method and the network converges on one of its stored patterns or their reverses which has the smallest Hamming distance with the initial state of the network or provides a controlled neural network with spatio-temporal patterns depending upon the nature of control.

BackgroundThe aim of our study is to analyze gender‐specific patterns of Parkinson’s disease dementia (PDD) incidence. In addition, we are investigating the extent to which gender differences in survival after initial Parkinson’s disease (PD) diagnosis influence differences in PDD risk among PD patients.MethodWe used a random sample of German longitudinal health claims data of persons at ages 65 years or older (2006‐2017). We identified incident PD cases who were followed‐up for a PDD diagnosis or death. We computed age‐ and gender‐specific PDD incidence rates, death rates, and cumulative PDD incidence functions with death as competing risk. In addition, we performed Cox regression models and Fine and Gray competing‐risk regression models, considering death as a competing event, to calculate PDD hazard ratios (HR), adjusted for important risk factors such as age at PD onset, PD severity in terms of modified Hoehn&Yahr (HY) scale, comorbidities, and medications.ResultOf 2,195 incident PD cases, 602 people died and 750 people received a dementia diagnosis by the end of 2017. 44% of PD men and 45% of PD women had received a dementia diagnosis, 45% of PD men and 32% of PD women had died, and only 11% of PD men but 23% of PD women had survived without a PDD diagnosis. PDD risk among PD patients differs by gender when controlled for age at onset, severity of PD, comorbidities and medications, with men having a higher PDD risk than women (HR = 1.18, p = 0.029). When accounting for death, men and women do not differ regarding their PDD risk (HR = 1.02, p = 0.770). Gender‐specific analyses showed significant age and severity effects in women (Age: HR = 1.05, p<0.001, HY 0‐2.5: reference, HY 3‐5: HR = 1.46, p = 0.011), but not in men.ConclusionWe found that older age at first PD diagnosis and higher disease severity increase PDD risk, but this association is attenuated for PD men when we consider death as a competing risk. This implies that the most frail PD men die rapidly before receiving a dementia diagnosis, whereas women with PD survive at higher rates than men, regardless of their age at onset and disease severity.

The Sixth International Symposium on Neural Networks (ISNN 2009)

The cognitive impairment seen in Parkinson's disease (PD) results in patient disability and reduced quality of life. However, using cognitive screening scales specific to PD in routine clinical practice is difficult because of limited time, resources, and skills. We studied the ability of routine cognitive tests to differentiate between Parkinson's disease dementia (PDD) and PD and among the neuropsychological profiles of elderly individuals with PD, PDD, Alzheimer's disease (AD), and normal cognition. This cross-sectional study involved 124 subjects. Subjects were 35 cognitively normal elderly and 37 elderly individuals with PD, 22 with PDD, and 30 with AD. All subjects were diagnosed by a specialist using standard criteria. Clinically relevant data and scores from the Montreal Cognitive Assessment and the Thai Mental State Examination were collected. Cognitive test scores were compared among groups. Receiver operating characteristic curves were constructed for a range of cut-off points to explore the sensitivity and specificity of the screening tools to detect PDD. There were 74 female subjects (59.7%), and the average age of all subjects was 75.6 years. The median score on the modified Hoehn and Yahr scale was 2.5 in subjects with PD and 4 in those with PDD (P < 0.001). The cut-offs for differentiating PDD from PD were 25 on the Thai Mental State Examination and 14 on the Montreal Cognitive Assessment. The sensitivity of the Thai Mental State Examination was 78.4%, and the specificity was 66.7% (area under the curve: 0.828). The sensitivity of the Montreal Cognitive Assessment was 81.1%, and the specificity was 75% (area under the curve: 0.876). There was a significant difference in the memory and language subdomains between AD and PDD (P < 0.05). The cut-offs used to differentiate PDD from PD were not the same as routine cut-offs in distinguishing AD from normal elderly. The cognitive profile deficit in PDD differed from that in AD. Interpretations of positive screenings test should take this finding into consideration.

Background: The group of diseases including dementia with Lewy bodies (DLB), Parkinson’s disease (PD), and Parkinson’s disease dementia (PDD) are related to synucleinopathies. In addition to Alzheimer’s disease (AD), these are common neurodegenerative diseases that share clinical features leading to difficulties in diagnosis. Hence, molecular imaging showing correlation with pathogenesis is recognized as a critical tool for improving the efficiency of differential diagnosis. Objective: To determine characteristic patterns of dopamine transporter (DAT) density using ⁹⁹ᵐTc-TRODAT-1 (TRODAT-1), glucose metabolism using ¹⁸F-fludeoxyglucose (FDG) and amyloid deposition using ¹¹C-Pittsburgh compound B (PiB) among these diseases. Materials and Methods: Twenty-five individuals participated in the present study and included four healthy controls (HC), eight DLB, two PD, six PDD, and five AD. All subjects underwent FDG and PiB positron emission tomography (PET) and TRODAT-1 single-photon emission computed tomography (SPECT), with visual and semi-quantitative analysis of imaging. Results: All subjects for synucleinopathy groups (DLB, PD, and PDD) had positive TRODAT-1 SPECT scintigraphy, while the AD and HC groups showed negative results. The positive PiB PET results were 100% in the AD, 33.33% with two of six cases in the PDD group, 37.5% with three of eight cases in the DLB group, and 0% in both the HC and PD groups. The hypometabolism patterns in PDD and AD groups were noted at parietotemporal region and parietotemporal, posterior cingulate, and precuneus regions, respectively. Meanwhile, the DLB hypometabolism pattern was found at parietotemporal, posterior cingulate, precuneus regions, and occipital lobe regions. None of the subjects in the HC or PD groups showed hypometabolism. Conclusion: Positive TRODAT-1 SPECT provided an informative pattern for differential diagnosis among alpha-synuclein-related diseases, AD, and HC. Remarkably, hypometabolism in the occipital region can be used to differentiate DLB from AD, PDD, and PD. The amyloid deposition pattern in DLB and PDD groups was not significantly different. Keywords: alpha-synucleinopathies; PET; SPECT; ⁹⁹ᵐTc-TRODAT-1; Alzheimer’s disease