Place cells and grid cells are important neurons involved in spatial navigation in the mammalian brain. Grid cells are believed to play an important role in forming a cognitive map of the environment. Experimental observations in recent years showed that the grid pattern is not invariant but is influenced by the shape of the spatial environment. However, the cause of this deformation remains elusive. Here, we focused on the functional interactions between place cells and grid cells, utilizing the information of location relationships between the firing fields of place cells to optimize the previous grid cell feedforward generation model and expand its application to more complex environmental scenarios. Not only was the regular equilateral triangle periodic firing field structure of the grid cells reproduced, but the expected results were consistent with the experiment for the environment with various complex boundary shapes and environmental deformation. Even in the field of three-dimensional spatial grid patterns, forward-looking predictions have been made. This provides a possible model explanation for how the coupling of grid cells and place cells adapt to the diversity of the external environment to deepen our understanding of the neural basis for constructing cognitive maps.

位置细胞和网格细胞是哺乳动物大脑中参与空间导航的重要神经元。网格细胞被认为在形成环境认知地图方面发挥着重要作用。近年来的实验观察表明，网格模式不是一成不变的，而是受空间环境形状的影响。然而，这种变形的原因仍然难以捉摸。在这里，我们专注于位置单元和网格单元之间的功能交互，利用位置单元的发射场之间的位置关系信息来优化先前的网格单元前馈生成模型，并将其应用扩展到更复杂的环境场景。不仅再现了格子单元规则的等边三角形周期性发射场结构，但对于各种复杂边界形状和环境变形的环境，预期结果与实验一致。甚至在三维空间网格格局领域，也做出了前瞻性的预测。这为网格细胞和位置细胞的耦合如何适应外部环境的多样性提供了可能的模型解释，以加深我们对构建认知地图的神经基础的理解。