Constraining inference of across-region interactions using neural activity perturbations

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Functional interactions between brain regions are crucial for understanding brain function.
• Multi-region models are commonly used to infer these interactions from neural recordings in behaving animals.
• Current methods face challenges due to the inherent variability in neural activity.

Purpose of the Study:

• To investigate the limitations of current methods for inferring brain region interactions.
• To explore novel approaches for improving the accuracy of these inferences.
• To determine the impact of region-wide activity perturbations on model-based interaction estimates.

Main Methods:

• Analysis of neural activity data from behaving animals.
• Development and application of multi-region models.
• Simulations incorporating region-wide activity perturbations.
• Evaluation of model fit constraints and estimation accuracy.

Main Results:

• Inference of across-region interactions is hindered by the wide breadth of model fits to naturally occurring neural activity.
• Models incorporating region-wide activity perturbations yield well-constrained estimates of across-region interactions.
• Simulations indicate that these perturbed-data-based estimates can be accurate.

Conclusions:

• The variability in natural neural activity poses a significant challenge for accurately inferring functional brain connectivity.
• Introducing controlled region-wide perturbations during neural recording can significantly enhance the precision of multi-region models.
• This approach offers a promising strategy for more reliable estimation of functional interactions between brain regions.