Pre-registration: Interface Pathology Staging and Pathology-Adaptive Decoding for BCI Long-Term Stability and Energy Efficiency

中文标题:电极-组织界面病理分期标准与病理自适应解码:一项面向脑机接口长期稳定性与能效的预注册研究计划 Abstract:The long-term effectiveness of invasive brain–computer interfaces (BCIs) depends on the coupling between the biological state of the electrode–tissue interface and the computational cost of downstream decoding. Current BCI evaluation frameworks lack a unified interface pathology staging standard, and engineering teams have limited means to incorporate interface pathology as a design constraint for decoding architectures. This pre-registration proposes a testable hypothesis chain: interface pathology scores are associated with signal quality degradation (H1); after controlling for baseline signal quality, higher interface pathology scores are associated with greater computational cost or equivalent energy consumption required to maintain decoding accuracy (H2); and the energy-efficiency advantage of event-driven/neuromorphic decoding over conventional frame-based decoding is more pronounced in samples with higher interface pathology scores (H3, exploratory). The study also proposes a strategic positioning: pathology can both define an “interface pathology staging standard” (definition authority) and provide a constraint variable for “pathology-adaptive decoding architectures” (engineering interface). This study is positioned as an initial feasibility assessment; effect-size estimates will inform the design of larger subsequent studies.