guest ::
| Home > Publications database > SWAN: A tool to track single units across consecutive electrophysiological recordings |
| Home > Publications database > SWAN: A tool to track single units across consecutive electrophysiological recordings |
| Poster (After Call) | FZJ-2019-02141 |
; ; ; ; ; ;
2019
Please use a persistent id in citations: http://hdl.handle.net/2128/22257
Abstract: Electrophysiological experiments often involve the measurement of extracellular analog voltage signals from brain tissue using implanted microelectrodes. The spiking activity of neurons in the direct vicinity of an electrode is captured as short-lasting voltage deflections. Some of the deflections are action potentials (spikes) of neurons close to the electrode tip [1]. A crucial first step in the analysis of such data is the extraction of spikes and their assortment into clusters corresponding to contributions from different putative single neurons, or in short ‘units’ (spike sorting) [1, 2]. The sorting compares the shape of waveforms and the spiking characteristics. With chronically implanted electrodes, spiking activity is recorded for several months over multiple sessions. It is not clear if each electrode detects identical units over the entire course of an experiment. However, this knowledge would help to monitor long-term changes of neural activity during training for a task. In practice, some single units disappear, (re-)appear or progressively change their spike shape, likely due to small movements of the electrode and/or tissue growth. Thus, it becomes a challenge to keep track of identical units across consecutive chronic recordings [3, 4]. In the absence of such a tracking of neurons across sessions, detected units in one session are assumed to be independent of those in other sessions of the experiment. They may thus be considered more than once in analyses of several sessions, and bias statistics across sessions.Here we present the Sequential Waveform Analyzer (SWAN) - an open-source tool developed to track individual units across sessions, but also to identify units that are different. It provides a graphical user interface (GUI) to visualize and relate spike-sorted data across multiple sessions. The configurable user interface is divided into several windows (see Figure). In each window, a certain set of features (e.g., mean waveforms, inter-spike interval histograms, principal component analysis of mean waveforms, firing rate profiles) are compared between different units and across multiple sessions. Each set of similar units across sessions is then assigned one global unit ID, represented by one common color across all windows. Thus, we visualize the tracking of a certain unit across consecutive sessions. The assignment of units to global unit IDs is performed by published [3,4] and newly developed automatic algorithms, and can be easily edited by the experimenter in the GUI. We demonstrate the capabilities of SWAN and practically illustrate its application on large-scale recordings from macaque monkey motor cortex [5].Acknowledgements: EU Horizon 2020 Framework Programme for Research and Innovation under Specific Grant Agreements No. 720270 and No. 785907 (Human Brain Projects SGA1 and SGA2); DFG Grants DE 2175/2-1 and GR 1753/4-2 Priority Program (SPP 1665); International Associated Laboratory (LIA) “Vision for Action” between CNRS and Aix-Marseille Univ, Marseille, France, and Research Centre Juelich, Germany. Christoph Gollan created the initial version of SWAN.References:1. Einevoll, G. T. et al. Cur Opinion Neurobiol 22, 11 (2012).2. Buzsáki, G. Nat Neurosci 7, 446 (2004).3. Dickey, A. S. et al. J Neurophys 102, 1331 (2012).4. Fraser, G. W. et al. J Neurophys 107, 1970 (2012).5. Brochier, T. et al. Sci Data 5, 180055 (2018).
|
The record appears in these collections: |
PDF
PDF (PDFA)