Unitary events in multiple single-neuron spiking activity: II. Nonstationary data

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Detecting functional groups (cell assemblies) in neuronal activity is crucial for understanding brain function.
• Existing methods like unitary event analysis assume stationary neuronal firing rates, which is often not true for real neural data.

Purpose of the Study:

• To develop a novel method for detecting cell assemblies that accounts for non-stationary neuronal firing rates.
• To introduce the unitary events by moving window analysis (UEMWA) for analyzing time-varying neuronal synchrony.

Main Methods:

• Unitary events by moving window analysis (UEMWA) was developed to analyze neuronal spiking activity.
• The method performs unitary event analysis in overlapping time segments, assuming stationarity within each window.
• Simulated spike trains and recordings from awake, behaving monkeys were used to validate the method.

Main Results:

• UEMWA successfully normalizes for changes in neuronal firing rates.
• The method can detect coincident spiking activity, indicating cell assembly membership, even in non-stationary data.
• Analysis of recordings from monkey frontal and motor cortices demonstrated the method's potential.

Conclusions:

• UEMWA is a robust statistical technique for identifying dynamic cell assemblies in neuronal recordings.
• The method's ability to handle non-stationarities allows for the temporal localization of coordinated neuronal activity.
• This approach advances the study of functional neural networks in behaving animals.