The attachment of cells onto solid supports is fundamental in the development of advanced biosensors or biochips. In this work, we characterize cortical neuron adhesion, growth, and distribution of an adhesive layer, depending on the molecular structure and composition . Neuronal networks are successfully grown on amino-terminated alkanethiol self-assembled monolayer (SAM) on a gold substrate without adhesion protein interfaces. Neuron adhesion efficiency was studied for amino-terminated, carboxy-terminated, and 1:1 mixed alkanethiol SAMs deposited on gold substrates. Atomic force microscopy and X-ray photoelectron spectroscopy were used to measure the roughness of gold substrate and thickness of SAM monolayers. Conformational ordering and ionic content of SAMs were characterized by vibrational sum frequency generation (VSFG) spectroscopy. Only pure amino-terminated SAMs provide efficient neuronal cell attachment. Ordering of the terminal amino groups does not affect efficiency of neuron adhesion. VSFG analysis shows that ordering of the terminal groups improves with decreasing surface roughness; however the number of gauche defects in alkane chains is independent of surface roughness. We monitor partial dissociation of carboxy groups in mixed SAMs that implies formation of NH3+ neighbors and appearance of catanionic structure. Such catanionic environment proved inefficient for neuron adhesion. Surface roughness of metal within the 0.7-2 nm range has little effect on the efficiency of neuron adhesion. This approach can be used to create new methods that help map structure-property relationships of biohybrid systems.