Interfacing neurons with silicon can provide insights into neuronal coding and a basis for neuroprostheses. We aim to construct networks on parylene‐patterned SiO 2 , previously used to successfully pattern primary murine hippocampal cells. A key question is whether neurons will pattern in isolation , or whether they require a supporting cellular template. Both undifferentiated (UD) and differentiated (DF) Lund Human Mesencephalic (LUHMES) neurons fail to adhere, nor show morphological signs of differentiation, when cultured alone on‐chip (UD LUHMES parylene adhesion index (PAI) 0.02±0.05, silicon repulsion index (SRI) 1.0±0.01; DF LUHMES PAI 0.0±0.0,SRI 1±0.001). We therefore sought a different cell type to pre‐pattern on parylene. HEK 293 cells patterned accurately (PAI 0.46±0.2,SRI 0.98±0.01) as did a human glioma‐derived cell line (PAI 0.96±0.01,SRI 0.95±0.01). LUHMES, plated secondarily, adhered specifically to either cellular template, extending neurites to form a network. The next challenge is to control network topography. By altering the configuration of an array of parylene nodes, we have generated orthogonal networks. Hypothesizing that parylene can haptotactically guide neurites, different nodular features (teardrop shapes or short tracks) were created. Neurite directionality was not influenced by parylene shape; neurites instead connected with the nearest adjacent cell cluster. Grant Funding Source : Wellcome Trust