Computational NeuroscienceJoint analysis of extracellular spike waveforms and neuronal network bursts

- Spike and burst statistics give limited information on changes in networks.
- Here, spike sorting combined with burst detection.
- Spike waveform type participation in bursts revealed.
- Spike type compositions of bursts change under network modifications.
- New kind of information obtained on the changes in bursting networks.

BackgroundNeuronal networks are routinely assessed based on extracellular electrophysiological microelectrode array (MEA) measurements by spike sorting, and spike and burst statistics. We propose to jointly analyze sorted spikes and detected bursts, and hypothesize that the obtained spike type compositions of the bursts can provide new information on the functional networks.New methodSpikes are detected and sorted to obtain spike types and bursts are detected. In the proposed joint analysis, each burst spike is associated with a spike type, and the spike type compositions of the bursts are assessed.ResultsThe proposed method was tested with simulations and MEA measurements of in vitro human stem cell derived neuronal networks under different pharmacological treatments. The results show that the treatments altered the spike type compositions of the bursts. For example, 6-cyano-7-nitroquinoxaline-2,3-dione almost completely abolished two types of spikes which had composed the bursts in the baseline, while bursts of spikes of two other types appeared more frequently. This phenomenon was not observable by spike sorting or burst analysis alone, but was revealed by the proposed joint analysis.Comparison with existing methodsThe existing methods do not provide the information obtainable with the proposed method: for the first time, the spike type compositions of bursts are analyzed.ConclusionsWe showed that the proposed method provides useful and novel information, including the possible changes in the spike type compositions of the bursts due to external factors. Our method can be employed on any data exhibiting sortable action potential waveforms and detectable bursts.