We introduce LoongX, which effectively integrates multimodal neural signals to guide image editing through novel Cross-Scale State Space (CS3) encoder and Dynamic Gated Fusion (DGF) modules.