NeuMoSync: End-to-End Neuromodulatory Control for Plasticity and Adaptability in Continual Learning

2026-08-04

AI Summary

This work addresses the challenges of catastrophic forgetting and limited knowledge transfer in continual learning with deep neural networks, which often suffer from degraded plasticity over time. Inspired by global neuromodulatory mechanisms in the brain, the authors propose NeuMoSync, a novel architecture that introduces neuron-level dynamic modulation into end-to-end deep networks for the first time. NeuMoSync employs a learnable historical context vector and a high-level abstract modulation module to adaptively regulate neuronal activation and synaptic plasticity based on both input stimuli and internal network states. Evaluated across multiple continual learning benchmarks, the method significantly outperforms existing approaches in preserving plasticity and enhancing both forward and backward knowledge transfer. Ablation studies confirm the contribution of each component and reveal interpretable cross-task modulation coordination patterns.

Abstract

Continual learning (CL) requires models to learn tasks sequentially, yet deep neural networks often suffer from plasticity loss and poor knowledge transfer, which can impede their long-term adaptability. Drawing high-level inspiration from global neuromodulatory mechanisms in the brain, we introduce Neuromodulation and Synchronization (NeuMoSync), a novel architecture that integrates dynamic, neuron-specific modulation into deep neural networks to enhance their adaptability and plasticity. NeuMoSync extends standard neural network architectures with learnable feature vectors for each neuron that track network-wide historical context and with a module operating at a higher level of abstraction. This module synthesizes neuron-specific signals, conditioned on both current inputs and the network's evolving state, to adaptively regulate activation dynamics and synaptic plasticity. Evaluated on diverse CL benchmarks, including memorization (Random Label CIFAR-10 and Random Label MNIST), concept drift (Shuffle CIFAR-10 and Shuffle Mini-ImageNet), class-incremental learning (Class Split ImageNet and Class Split CIFAR-100), and domain-incremental learning (Permuted MNIST), NeuMoSync demonstrates strong performance in retaining plasticity and achieves improvements in both forward and backward adaptation compared with existing methods. Ablation studies validate the necessity of each component, while analysis of the learned modulatory signals reveals interpretable coordination patterns across tasks. Our work underscores the potential of integrating global coordination mechanisms into deep learning systems to advance robust, adaptive continual learning. The code is publicly available at https://github.com/RoozbehRazavi/NeuMoSync.

Problem

• continual learning
• plasticity loss
• knowledge transfer
• adaptability
• neuromodulation

Innovation

• neuromodulation
• continual learning
• synaptic plasticity
• adaptive control
• dynamic modulation