Partial amplitude synchronization detection in brain signals using Bayesian Gaussian mixture models.

Area of Science:

• Neuroscience
• Signal Processing
• Computational Biology

Background:

• Understanding brain function relies on analyzing complex neural signals.
• Identifying synchronized neural activity is crucial for mapping brain networks.
• Existing methods often require baseline data or are limited in scope.

Purpose of the Study:

• To develop a novel method for detecting instantaneous synchronization in multivariate time series.
• To identify synchronized neural activity without relying on baseline information.
• To assess the stability and relevance of detected synchronized subsets.

Main Methods:

• A Bayesian Gaussian mixture model was applied to time-frequency maps of multivariate signals.
• A stability measure was employed to evaluate the relevance of identified synchronized subsets.
• The method was tested on Local Field Potentials (LFPs) from a visuo-motor experiment in monkeys.

Main Results:

• The new method successfully detected subsets of instantaneously synchronized time series.
• No baseline information was required for the synchronization detection.
• A subset of synchronized time series was identified within the visual cortex of monkeys during the experiment.

Conclusions:

• The developed Bayesian approach offers a robust method for instantaneous synchronization detection.
• This technique can reveal functional neural connectivity without preprocessing steps like baseline subtraction.
• The findings highlight the utility of the method in uncovering task-specific neural synchrony in the visual cortex.