​MULTI-CHANNEL LOCALIZATION THROUGH AN EPSP-BASED ALGORITHM

Investigation of brain dynamics and neural networks can be achieved through high-throughput multichannel electrophysiological recordings with silicon probes. A key challenge of this technique is the determination of the precise localization of the recording sites within the brain, an aspect which is fundamental for accurate data analysis and interpretation. Analysis of electrophysiological signatures of local field potentials can support localization coupled with postmortem histological analysis, thereby providing correlated information about electrode localization. However, analysis of electrophysiologically evoked field excitatory post-synaptic potentials can substantially improve the accuracy of real-time multichannel localization. This is because evoked potentials possess stereotypical characteristics, as well as depth profiles that are defined by the spread of current sources and sinks along activated neurons. Here, we developed a multichannel localization strategy based on these features. Our approach combines template matching to standard potential waveforms, as well as clustering of the waveforms in a low-dimensional space. We tested this approach by the scrutiny of potentials evoked by Schaffer collaterals stimulation in the dorsal hippocampus and the retrosplenial cortex of freely moving rats. Our results show that a robust and stable channel localization was achieved by the template matching strategy, and a consistent clustering of the cortical potentials. This was further validated by histological probe reconstruction. Overall, this approach enhances the ability to study intra and trans-structural dynamics with layer-wise anatomical specificity.
Funded by a scholarship from the German Academic Exchange Service (www.DAAD.de) to TT and a Deutsche Forschungsgemeinschaft grant to DMV (SPP2411, project number: 520284247).