NEURONAL NETWORKS FOR BIOCHEMICAL SENSING

Embryonic mouse spinal nerve cells grown in cell culture develop networks displaying complex spontaneous activity and histiotypic pharmacological responses that remain stable for over three months in vitro. However, as living systems growing without the homeostatic control mechanisms of the central nervous system, they show a remarkable sensitivity to minute changes in their environment and, therefore, display some of the properties of sensory tissue. This paper explores the chemical sensitivity expressed in terms of changes in the spontaneous spike (i.e., action potential) activity patterns. We suggest the possibility of transforming non-sensory networks into 'sensomimes' that can reliably recognize specific minute chemical or physical changes. As presently envisioned, such systems would consist of genetically altered or enzymatically modified biological neuronal networks maintained in cell culture, and would utilize gross changes in the spontaneous spike activity of the network as the sensory output. With this approach, co-culturing with sensory epithelium would not be necessary, knowledge of complex sensory coding would not be required, the neural tissues which are most robust under culture conditions could be utilized, and novel physical and chemical sensitivities could be specifically engineered into the networks.