Dynamic Neuroplastic Networks for Financial Decision-Making: A Self-Adaptive Approach for Mitigating Catastrophic Forgetting in Continual Learning.

Continual learning remains a critical challenge in financial decision-making due to catastrophic forgetting where models struggle to retain past knowledge while adapting to evolving market conditions. This study proposes a Dynamic Neuroplastic Network (DNN) that integrates self-adaptive learning mechanisms inspired by neuroplasticity principles to mitigate catastrophic forgetting. The model was trained on a comprehensive financial dataset spanning 2015 to 2025, covering stock market indices, forex exchange rates, cryptocurrency data, and macroeconomic indicators. Experimental results demonstrate that the proposed model achieved a steady accuracy improvement from 48.3% to 60.9%, with a corresponding reduction in loss values from 0.7086 to 0.6427 over 50 epochs. Comparative analysis with traditional continual learning methods, including Elastic Weight Consolidation (EWC) and Synaptic Intelligence (SI), revealed that the DNN exhibited superior knowledge retention and adaptability across bullish, bearish, and high-volatility market conditions. Statistical evaluation confirmed the effectiveness of neuroplasticity-inspired mechanisms, with a mean accuracy of 57.76% (±2.30%) and a minimum loss of 0.0024, highlighting the model's capacity for long-term financial forecasting. The findings underscore the practical applicability of neuroplastic networks in financial modeling, where continual adaptation to market fluctuations is crucial. The study provides a strong foundation for future research in hybrid continual learning frameworks integrating reinforcement learning and advanced neural architectures for robust financial decision-making.