Narrow microtunnel technology for the isolation and precise identification of axonal communication among distinct hippocampal subregion networks

Communication between different sub regions of the hippocampus is fundamental tolearning and memory. However accurate knowledge about information transferbetween sub regions from access to the activity in individual axons is lacking. MEMSdevices with microtunnels connecting two sub networks have begun to approach thisproblem but the commonly used 10 μm wide tunnels frequently measure signals frommultiple axons. To reduce this complexity, we compared polydimethylsiloxane (PDMS)microtunnel devices each with a separate tunnel width of 2.5, 5 or 10 μm bridging twowells aligned over a multi electrode array (MEA). Primary rat neurons were grown inthe chambers with neurons from the dentate gyrus on one side and hippocampal CA3on the other. After 2-3 weeks of culture, spontaneous activity in the axons inside thetunnels was recorded. We report electrophysiological, exploratory data analysis forfeature clustering and visual evidence to support the expectation that 2.5 μm widetunnels have fewer axons per tunnel and therefore more clearly delineated signals than10 or 5 μm wide tunnels. Several measures indicated that fewer axons per electrodeenabled more accurate detection of spikes. A clustering analysis comparing thevariations of spike height and width for different tunnel widths revealed tighter clustersrepresenting unique spikes with less height and width variation when measured innarrow tunnels. Wider tunnels tended toward more diffuse clusters from a continuumof spike heights and widths. Standard deviations for multiple cluster measures, suchas Average Dissimilarity, Silhouette Value (S) and Separation Factor (averagedissimilarity/S value), support a conclusion that 2.5 μm wide tunnels containing feweraxons enable more precise determination of individual action potential peaks, theirpropagation direction, timing, and information transfer between sub networks.