Related Experiment Video
Updated: Jun 23, 2026
16:01
Multi-electrode Array Recordings of Neuronal Avalanches in Organotypic Cultures
Published on: August 1, 2011
Subsampling effects in neuronal avalanche distributions recorded in vivo
Viola Priesemann1, Matthias H J Munk, Michael Wibral
1Department of Neurophysiology, Max Planck Institute for Brain Research, Deutschordenstrasse 46, D-60528 Frankfurt am Main, Germany. v.priesemann@gmx.de
BMC Neuroscience
|April 30, 2009
Summary
Self-organized criticality (SOC) may govern brain activity, but subsampling neuronal data can obscure power-law behavior. Researchers found that subsampling affects avalanche size distributions and branching parameters, impacting model selection for brain function.
Area of Science:
- Neuroscience
- Complex Systems
- Computational Biology
Background:
- Many natural systems exhibit complex behavior with intermittent large cascades (avalanches) and periods of low activity.
- Self-organized criticality (SOC) describes systems where avalanche size distributions follow a power law and the branching parameter (sigma) is unity.
- SOC has been hypothesized to govern in vivo neuronal dynamics due to similarities in variability, stability, and memory, but testing is hindered by data subsampling.
Purpose of the Study:
- To investigate the impact of spatial or temporal subsampling on the avalanche size distribution (f(s)) and branching parameter (sigma) in self-organized critical (SOC) systems.
- To compare the characteristics of subsampled SOC models with multielectrode local field potential (LFP) activity recorded from macaque monkeys during a memory task.
Main Methods:
- Three different SOC models were subjected to imposed subsampling.
- The avalanche size distribution (f(s)) and branching parameter (sigma) were analyzed for the subsampled models.
- These metrics were compared to f(s) and sigma calculated from LFP data recorded in monkeys performing a short-term memory task.
Main Results:
- Neither the LFP data nor the subsampled SOC models exhibited a power-law distribution for f(s).
- Both f(s) and sigma were highly sensitive to the subsampling geometry and the specific dynamics of the SOC model.
- Only the Abelian Sandpile Model, when subsampled, showed f(s) and sigma values comparable to those derived from LFP recordings.
Conclusions:
- Subsampling can mask the characteristic power-law behavior of SOC systems, potentially leading to misclassification as sub- or supercritical.
- The study highlights the importance of system-specific scaling of f(s) and sigma under subsampling conditions.
- This approach can aid in selecting more physiologically relevant computational models of brain function by comparing their performance against empirical data under realistic subsampling constraints.
