Interfacing insect neuronal networks with microelectronic devices

Information processing systems in animals have evolved to be extraordinarily efficient. The present work deals with the challenge of designing neuronal networks in vitro in order to mimic these sensory systems. The cercal sensory system of crickets is one of the most sensitive sensory systems known. Neurons from the terminal abdominal ganglion (TAG) are responsible for the first information processing of the cercal sensory system that yields the escape reaction. Using in vitro systems, information processing steps within the neuronal networks can be studied without disturbing side input, and network formation can be monitored over developmental stages. Knowledge gained from such studies offers a thorough understanding of biological processes during network formation and function, and this knowledge can be implemented in engineering of prostheses and biosensors. Construction of functional in vitro neuronal networks requires a reliable and universal cell patterning method in order to enable the reconstruction of an in vivo-like network with appropriate connection patterns as well as interfacing with a sensitive recording system. Microelectronic devices provide a non-destructive way of studying networks of neurons over the long-term with the additional advantage of simultaneously recording several single neurons within the network. However, interfacing neurons with microelectronic devices is challenging, as the device must provide a growth-permissive surface and the possibility to guide neuronal growth. This work presents an approach to control insect neuroarchitecture in vitro on both glass surfaces as well as silicon devices. Using isocynanate functionalized star-shaped PEG which changes reactivity over time, a patterning method can be developed such that covalently bound cell adhesive proteins are surrounded by cell and protein repulsive background. A geometrical patterning protocol using microcontact printing was initially used to control network formation of TAG neurons. The initial patterning method was modified further to offer the possibility to control neuronal growth on to extracellular recording devices. In the further developed method, star PEG is applied uniformly on the device surface and is subsequently removed by means of oxygen plasma, which can be combined with backfill of an adhesive protein to the inverse pattern. The second approach is especially interesting due to the reduction in neuron-electrode distance by removing the PEG layer. Detailed investigations of the pre-patterned neurons proved that patterning supports a more in vivo-like morphology in the cell culture system and the physiology is maintained. The investigation of basic electrophysiological characteristics of single neurons showed that neurons from the TAG are more excitable than in vivo neurons, which is typical for insect neurons after axotomy. However, compared to neurons grown on un-patterned, uniformly coated substrates with the same adhesive protein, the patterned neurons show the same distribution of electrophysiological response types and basic electrophysiological characteristics. Additionally, paired patch clamp experiments providing information about synaptic connectivity showed that functionality is restored within reconstructed, patterned networks. In all, this work takes a step forward in the development of cell-based microsystems and biosensors, which seek to mimic natural information processing systems. The results presented here indicate that geometric patterning of insect neurons can simplify identifying network connectivity and modulate morphological characteristics of neurons to a more in vivo-like condition without modifying the electrophysiological characteristics of the neurons