Critical dynamics on a large human Open Connectome network.

Area of Science:

• Computational neuroscience
• Network science
• Complex systems

Background:

• Human brain networks exhibit complex dynamics.
• Threshold models are used to simulate neuronal activity.
• Previous models often lack critical dynamics or rely on self-organized criticality assumptions.

Purpose of the Study:

• Investigate threshold models on a large-scale human brain network.
• Explore the emergence of power-law scaling and critical dynamics.
• Examine the impact of network heterogeneity and connection properties.

Main Methods:

• Extended numerical simulations using variable threshold models.
• Analysis of a human brain network (N=836,733 nodes) from the Open Connectome Project.
• Inclusion of link directness and inhibitory connections in simulations.

Main Results:

• Variable threshold models exhibit extended power-law scaling regions.
• Griffiths effects become relevant due to network heterogeneity.
• Nonuniversal power-law avalanche size and time distributions were found.
• Exponents agree with human brain electrode experiment data.

Conclusions:

• Variable threshold models can reproduce critical brain dynamics.
• Network heterogeneity is crucial for observing Griffiths effects and power-law scaling.
• Critical dynamics emerge in an extended parameter space, not requiring self-organized criticality.