Scoring sequences of hippocampal activity using hidden Markov models.

Area of Science:

• Computational Neuroscience
• Systems Neuroscience
• Neuroscience

Background:

• Analyzing neural activity sequences is crucial for understanding brain function.
• Latent ensemble states, or virtual place fields, represent neural representations of environments.
• Existing methods may not fully capture the sequential and contextual information in neural data.

Purpose of the Study:

• To introduce a novel sequence score for quantifying neural activity consistency with latent states.
• To demonstrate the utility of hidden Markov models (HMMs) for analyzing neural sequences.
• To decouple sequential and contextual information within neural activity patterns.

Main Methods:

• Utilizing hidden Markov models (HMMs) to model neural activity sequences.
• Developing a two-component sequence score based on joint probability of observations and states.
• Implementing a method to decouple sequential and contextual information.

Main Results:

• The proposed sequence score effectively measures consistency between neural activity and latent state trajectories.
• The HMM-based approach successfully models and analyzes neural activity sequences.
• The score demonstrated the ability to differentiate true hippocampal neural sequences from shuffled data.

Conclusions:

• The novel sequence score provides a robust tool for analyzing neural activity in relation to spatial environments.
• HMMs offer a powerful framework for dissecting complex neural sequence data.
• This approach enhances the ability to interpret neural representations of experience.