Although brain science spans a history of several centuries, it still lacks a systematic and comprehensive theoretical framework to this day. This highly atypical situation has left the field in an immature state. As a result, research findings obtained across different network levels are difficult to connect, reinforce, and integrate with each other, creating a significant barrier to major breakthroughs in cognitive neuroscience and brain-inspired artificial intelligence. The core scientific questions that must be addressed include: (1) How are different brain network levels interconnected and how do they interact? (2) What neural mechanisms underlie the interactions between different network levels? (3) What theories and methods could unify experimental data from brain networks at different levels within a single interpretive framework? (4) What mechanisms allow the brain to operate with low power consumption and high efficiency? To address these questions, this review, building upon a previous article published in Artificial Intelligence Review (Wang et al. in Artif Intell Rev 56:285–350, 2023), proposes a theoretical research framework for brain science centered on the Wang–Zhang (W−Z) neuron energy model and neural energy analysis methods. By leveraging the core principle of energy-information symmetry embedded in the W−Z neuron model, this framework highlights the value and uniqueness of applying the W−Z model, which reflects global functional properties of the brain, to cross-level combinatorial analysis. Traditional reductionist and holistic research approaches in neuroscience have encountered fundamental contradictions when attempting to explain the global information processing and computational mechanisms underlying brain structure and function. In response, we introduce a new methodological perspective termed “synergy integration theory,” guided by the W−Z neuron energy model. This approach allows clear differentiation between the brain’s operational modes and its working principles, thereby enabling large-scale, cross-level self-consistency in information encoding and decoding across hierarchical brain regions and systems. The core contribution of this work lies in the symmetry between neural energy and neural information as captured by the W−Z model, along with its representation of low-power, high-efficiency neural computation. It is expected to establish a scientific foundation for deeper integration between brain science and brain-inspired artificial intelligence in the future.
Wang et al. (Wed,) studied this question.