Unsupervised Detection of Cell-Assembly Sequences by Similarity-Based Clustering

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Cell assemblies, neurons firing in fixed temporal patterns, are crucial for neural information processing.
• Detecting cell assemblies with temporal structure, especially in large datasets, remains a challenge due to inefficient methods.
• Existing methods often struggle to capture the dynamic, time-dependent nature of neural activity.

Purpose of the Study:

• To develop and validate an efficient method for detecting diverse, time-structured cell assembly activity patterns in noisy neural population data.
• To address the limitations of current methods in handling temporal dynamics within neural assemblies.
• To provide a novel analytical tool for deciphering the neural code during complex cognitive processes.

Main Methods:

• Utilized a computer science technique, "edit similarity," for comparing spike trains to group neuronal activity into assemblies.
• Applied the method to both artificial and experimental neural data from rat hippocampus and prefrontal cortex.
• Validated the detection of cell assemblies across multiple timescales without relying on event-locked averages.

Main Results:

• Successfully detected a variety of cell assembly activity patterns recurring at multiple timescales in noisy neural data.
• Identified simultaneous place-cell sequences on different timescales during navigation and awake replay in the hippocampus.
• Discovered multiple spike sequences in the prefrontal cortex encoding different segments of a goal-directed task.

Conclusions:

• The developed method offers an efficient and novel approach for detecting time-structured cell assemblies in large-scale neural recordings.
• This technique enables the analysis of neural codes during arbitrary behavioral and mental processes without conventional event-driven limitations.
• Provides a powerful new tool for systems neuroscience research, enhancing our ability to understand brain function.