arXiv:2511.00042v1 Announce Type: new Abstract: Purpose: Optimization challenges in science, engineering, and real-world applications often involve complex, high-dimensional, and multimodal search spaces. Traditional optimization methods frequently struggle with local optima entrapment, slow convergence, and inefficiency in large-scale environments. This study aims to address these limitations by proposing a novel optimization algorithm inspired by neural mechanisms. Design/methodology/approach: The paper introduces Neuron Synapse Optimization (NSO), a new metaheuristic algorithm inspired by neural interactions. NSO features key innovations such as fitness-based synaptic weight updates to improve search influence, adaptive pruning to minimize computational overhead, and dual guidance from global and local best solutions to balance exploration and exploitation. The algorithm was benchmarked against popular metaheuristics and the recently published Hippopotamus Optimization Algorithm (HOA) using the CEC 2014 test suite, encompassing unimodal, multimodal, and composition function landscapes. Findings: Benchmark results reveal that NSO consistently outperforms HOA and other major algorithms in terms of convergence speed, robustness, and scalability. NSO demonstrates superior adaptability and efficiency, particularly in complex, high-dimensional search spaces. Originality: NSO introduces a unique blend of neural-inspired mechanisms with dynamic resource allocation, setting it apart from existing algorithms. Its innovative design enhances search performance while reducing computational cost. With promising applications in technology, healthcare, data science, and engineering, NSO paves the way for future research into dynamic and multi-objective optimization, machine learning hyperparameter tuning, and real-world engineering design problems.