[Paper Review] Situation-based memory in spiking neuron-astrocyte network

Susanna Gordleeva, Yuliya Tsybina+5 more|arXiv (Cornell University)|Feb 15, 2022

Advanced Memory and Neural Computing4 citations

TL;DR

This paper proposes a spiking neuron-astrocyte network model that enables situation-based memory—rapid, context-specific recall of stimuli patterns within localized environments. Using a two-net neuromorphic architecture with astrocyte-mediated modulation, the model demonstrates enhanced retrieval quality over standard spiking neural networks, highlighting astrocytes' structural necessity for effective situation-driven learning and memory.

ABSTRACT

Mammalian brains operate in a very special surrounding: to survive they have to react quickly and effectively to the pool of stimuli patterns previously recognized as danger. Many learning tasks often encountered by living organisms involve a specific set-up centered around a relatively small set of patterns presented in a particular environment. For example, at a party, people recognize friends immediately, without deep analysis, just by seeing a fragment of their clothes. This set-up with reduced "ontology" is referred to as a "situation". Situations are usually local in space and time. In this work, we propose that neuron-astrocyte networks provide a network topology that is effectively adapted to accommodate situation-based memory. In order to illustrate this, we numerically simulate and analyze a well-established model of a neuron-astrocyte network, which is subjected to stimuli conforming to the situation-driven environment. Three pools of stimuli patterns are considered: external patterns, patterns from the situation associative pool regularly presented to the network and learned by the network, and patterns already learned and remembered by astrocytes. Patterns from the external world are added to and removed from the associative pool. Then we show that astrocytes are structurally necessary for an effective function in such a learning and testing set-up. To demonstrate this we present a novel neuromorphic model for short-term memory implemented by a two-net spiking neural-astrocytic network. Our results show that such a system tested on synthesized data with selective astrocyte-induced modulation of neuronal activity provides an enhancement of retrieval quality in comparison to standard spiking neural networks trained via Hebbian plasticity only. We argue that the proposed set-up may offer a new way to analyze, model, and understand neuromorphic artificial intelligence systems.

Motivation & Objective

• To investigate how neuron-astrocyte networks support rapid, context-specific memory recall in dynamic, localized environments.
• To address the challenge of efficient memory storage and retrieval in environments with limited, recurring stimulus patterns (i.e., 'situations').
• To demonstrate the necessity of astrocytes in enhancing memory performance in situation-driven learning tasks.
• To develop and validate a neuromorphic model that leverages astrocyte-induced modulation for improved short-term memory function.
• To provide a biologically plausible framework for modeling and designing neuromorphic artificial intelligence systems with context-sensitive memory.

Proposed method

• Numerical simulation of a well-established spiking neuron-astrocyte network model under situation-driven stimulus conditions.
• Implementation of three stimulus pools: external stimuli, situation-specific associative patterns, and already-learned astrocyte-remembered patterns.
• Use of Hebbian plasticity for neuronal learning, combined with astrocyte-mediated modulation of neuronal activity to enhance memory retrieval.
• Design of a two-net spiking neural-astrocytic network architecture to simulate short-term memory function.
• Application of selective astrocyte-induced modulation to dynamically regulate synaptic efficacy and improve pattern recognition fidelity.
• Evaluation of retrieval quality using synthesized data under controlled stimulus presentation and removal protocols.

Experimental results

Research questions

• RQ1 How do neuron-astrocyte networks support rapid, context-specific memory recall in localized, time-limited environments?
• RQ2 What role do astrocytes play in enhancing memory retrieval quality in situation-based learning scenarios?
• RQ3 Can a neuromorphic spiking network with astrocyte modulation outperform standard spiking neural networks trained via Hebbian plasticity alone?
• RQ4 How does the dynamic addition and removal of external stimuli into the associative pool affect memory performance?
• RQ5 To what extent is astrocyte presence structurally necessary for effective situation-based memory function in the proposed model?

Key findings

• The inclusion of astrocyte-mediated modulation significantly enhanced retrieval quality compared to standard spiking neural networks using only Hebbian plasticity.
• Astrocytes were found to be structurally necessary for optimal performance in situation-based memory tasks, as removing their influence degraded memory fidelity.
• The two-net neuromorphic model successfully implemented short-term memory function with high accuracy in retrieving situation-specific patterns.
• The system demonstrated robust performance when external stimuli were dynamically added to and removed from the associative pool, reflecting real-world environmental changes.
• The model achieved improved pattern recognition and memory recall under selective astrocyte-induced modulation, confirming its functional advantage in context-sensitive learning.
• The results support the hypothesis that neuron-astrocyte networks are inherently adapted to handle situation-based memory due to their dynamic, modulatory architecture.