Temporally correlated fluctuations drive epileptiform dynamics

Area of Science:

• Computational neuroscience
• Neural dynamics modeling
• Epilepsy research

Background:

• Macroscopic brain network models often assume Gaussian white noise for afferent input.
• Brain activity typically exhibits 1/f^b power spectra, deviating from white noise.
• Understanding deviations from white noise is critical for modeling brain dynamics, especially in epilepsy.

Purpose of the Study:

• Investigate the impact of temporally correlated noise on neural mass models.
• Determine the role of noise characteristics in eliciting aberrant rhythms in the epileptic brain.
• Explore how specific noise temporal correlations influence healthy versus epileptiform brain dynamics.

Main Methods:

• Studied a neural mass model driven by stochastic, temporally correlated input (Ornstein-Uhlenbeck process).
• Characterized model dynamics under varying temporal correlations and noise amplitudes.
• Analyzed model response to sinusoidal driving to understand underlying mechanisms.

Main Results:

• Identified specific temporal correlations that promote epileptiform dynamics.
• Found that these correlations lead to maximal noise power in the delta and theta frequency bands.
• Observed that delta and theta band power increases precede seizures in epilepsy models.

Conclusions:

• Temporally correlated noise, particularly in delta and theta bands, can facilitate epileptiform activity.
• The bifurcation structure of neural mass models explains the generation of epileptiform dynamics by specific noise types.
• Large-scale brain models should incorporate diverse noise types to accurately represent brain function and dysfunction.