Cultured Nerve Cell Networks as Biosensor using Artificial Neural Nets
Martin Bogdan, Adelmo Luis Cechin, Sorin Breit, Wolfgang Rosenstiel · 1997
Mammalian neural networks grown on microelectrode arrays represent transducers that respond to changes in their chemical environment. Their response profile to a great variety of neuroactive compounds is often substance- and concentration-specific and might be used as a biosensor for the detection of a variety of chemical substances. Employing appropriate data processing and analysis, these biological systems may potentially be used for certain sensory tasks as cellular biosensors. In this paper, the concentration dependent response of the network for strychnine and N-methyl-D-aspartic acid (NMDA) are explored, using artificial neural nets (ANN). The recorded data has been investigated using Backpropagation nets as well as Kohonen’s self-organizing map (SOM). Using the information containing within the trained Backpropagation net, FAGNIS (Fuzzy Automatically Generated Neural Inferred System) was applied to obtain fuzzy rules of the system.