Event Abstract Back to Event Radar chart to visualize connectivity of neuronal network on planner type MEA Gu-Haeng Lee1 and Yoonkey Nam1* 1 College of Engineering, KAIST, Department of Bio and Brain Engineering, Republic of Korea 1. Motivation Many researchers use connectivity analysis in neuronal network research. In particular, when studying plasticity, it is necessary to analyze the connectivity of consecutive data. In this case, many visualization methods are used. Since MEAs can simultaneously measure signals from multiple neurons, they are suitable for neuronal network studies as compared to studies using patch clamps. Methods for studying connectivity using MEA data are cross-correlation analysis of two specific channels, using cross-correlation matrix and dendrogram, and drawing connectivity map. However, existing connectivity maps are difficult to identify overall connectivity due to too many connection lines if all neurons in the network are strongly connected. [1] Therefore, in this study, we studied how to see the connection trend more easily. 2. Material and Methods For the measurement, a 60-channel MEA from MCS with a 30 um diameter titanium oxide electrodes and a 200 um space was used. E18 SD rat hippocampal neurons were randomly cultured on MEA. Spike data were measured from 8 DIV to 30 DIV for one hour daily. All analysis was done using Matlab. After cross-correlation analysis, the connection direction was identified through peak timing and the connection strength was defined using normalized peak amplitude. The connection score of one channel had a direction and size of 360 degrees. The size of the score was determined by the connection strength and distance between channels. The direction and range of the score were determined according to the direction and distance between the channels. To see the direction and distance of the connection at a glance, if the two channels were close to each other, the score had a smaller value in a wider range, and in a far distance, the score had a narrower range in a larger value. The obtained connection score was drawn as a radar chart at the position corresponding to each channel. In order to show the distance of connectivity, the distribution of scores was shown as uniform, triangle, and Gaussian distribution. 3. Result & Discussion Even in the case of data with a maximum of 47 active channels, the connectivity of all channels can be seen without overlapping. At the initial stage of culture, even though there was an active channel, the spike number was not enough to confirm the connectivity. After two weeks of culture, connectivity began to appear. As the time passes, the connection strength increases. After 3 weeks, the active channel was reduced, but connectivity was still present. Even if spike recording data has thousands of channels, it will be able to see the connectivity at a glance by normalizing the score. Through the shape of the radar chart, we can see that the direction and size of connectivity change in a short time. It can be considered that the visualization method using the radar chart can be used for the study of plasticity. 4. Conclusion Rader chart was used to visualize the connectivity of neuronal networks cultured in MEA. Even fully developed neuronal networks with very complex connectivity were able to see connectivity at a glance using this method. Acknowledgements This work was supported by National Research Foundation grant(NRF-2015K1A3A1A14021018) funded by Korean government.
Read full abstract