Network-Wide Adaptive Burst Detection Depicts Neuronal Activity with Improved Accuracy

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Neuronal networks exhibit complex spiking and bursting dynamics.
• Previous adaptive burst analysis methods focused on local network behavior.
• Analyzing different network parts with varying rules can hinder comprehensive assessment.

Purpose of the Study:

• To investigate the impact of multi-channel and multi-timepoint analysis on adaptive burst detection.
• To develop a network-wide adaptive burst detection method for unified assessment of neuronal activity.
• To explore if network-wide analysis provides novel insights into overall network function.

Main Methods:

• Modification of the inter-spike interval (ISI) histogram based cumulative moving average (CMA) algorithm for simultaneous multi-spike train analysis.
• Application of original and network-wide CMA algorithms on artificial spike trains and in vitro rat cortical networks (microelectrode array recordings).
• Introduction of a new burst synchrony measure and application of k-means clustering for network classification based on bursting statistics.

Main Results:

• The network-wide CMA algorithm successfully unifies burst definition across the entire network.
• The method demonstrates adaptability to neuronal networks with varying dynamics.
• Improved assessment and classification of neuronal activity, including pharmaceutical effects and developmental network comparisons.

Conclusions:

• Network-wide adaptive burst detection offers a unified approach to analyzing complex neuronal network activity.
• The proposed method enhances the assessment and classification of neuronal network dynamics, facilitating comparisons between networks with different firing patterns.
• This technique is particularly valuable for studying developing neuronal networks and the impact of various interventions.