Theory of the asynchronous state of structured rotator networks and its application to recurrent networks of excitatory and inhibitory units
Recurrently coupled oscillators that are sufficiently heterogeneous and/or randomly coupled can show an asynchronous activity in which there are no significant correlations among the units of the network. The asynchronous state can nevertheless exhibit a rich temporal correlation statistics, that is generally difficult to capture theoretically. For randomly coupled rotator networks, it is possible to derive differential equations that determine the autocorrelation functions of the network noise and of the single elements in the network. So far, the theory has been restricted to statistically homogeneous networks, making it difficult to apply this framework to real-world networks, which are structured with respect to the properties of the single units and their connectivity. A particularly striking case are neural networks for which one has to distinguish between excitatory and inhibitory neurons, which drive their target neurons towards or away from firing threshold. To take into account network structures like that, here we extend the theory for rotator networks to the case of multiple populations. Specifically, we derive a system of differential equations that govern the self-consistent autocorrelation functions of the network fluctuations in the respective populations. We then apply this general theory to the special but important case of recurrent networks of excitatory and inhibitory units in the balanced case and compare our theory to numerical simulations. We inspect the effect of the network structure on the noise statistics by comparing our results to the case of an equivalent homogeneous network devoid of internal structure. Our results show that structured connectivity and heterogeneity of the oscillator type can both enhance or reduce the overall strength of the generated network noise and shape its temporal correlations.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
DenoiseRotator: Enhance Pruning Robustness for LLMs via Importance Concentration
Pruning is a widely used technique to compress large language models (LLMs) by removing unimportant weights, but it often suffers from significant performance degradation - especially under semi-structured sparsity const…
A self-consistent analytical theory for rotator networks under stochastic forcing: effects of intrinsic noise and common input
Despite the incredible complexity of our brains' neural networks, theoretical descriptions of neural dynamics have led to profound insights into possible network states and dynamics. It remains challenging to develop the…
New tissue engineered scaffolds for rotator cuff tendon-bone interface regeneration
Healing of Tendon-bone interface(TBI) injuries is slow and is often repaired with scar tissue formation that compromises normal function. Despite the increasing maturity of surgical techniques, re-tearing of the rotator …
Spectral Theory for Edge Pruning in Asynchronous Recurrent Graph Neural Networks
Graph Neural Networks (GNNs) have emerged as a powerful tool for learning on graph-structured data, finding applications in numerous domains including social network analysis and molecular biology. Within this broad cate…
A Hybrid Approach for Aspect-Based Sentiment Analysis Using a Lexicalized Domain Ontology and Attentional Neural Models
This work focuses on sentence-level aspect-based sentiment analysis for restaurant reviews. A two-stage sentiment analysis algorithm is proposed. In this method, first a lexicalized domain ontology is used to predict …
Aspect-Based Sentiment AnalysisAspect-Based Sentiment Analysis (ABSA)SentenceSentiment Analysis