Brain activity | Jesús Cabrera-Álvarez

Understanding the Principles of Brain Simulation

Brain activity emerges from the complex interplay of neural populations and their structural connections. Computational models allow us to simulate these dynamics, providing insights into how different regions interact and how large-scale brain function arises from local neural mechanisms. This research line explores how to construct and refine such models to better capture real brain activity.

Neural Mass Models: Simulating Local Population Dynamics

Neural mass models serve as a mathematical framework for approximating the collective activity of large groups of neurons. By adjusting their parameters, we can reproduce different oscillatory behaviors observed in electrophysiological recordings. Exploring these models helps us understand how local circuit dynamics contribute to overall brain function and dysfunction.

Structural Constraints: Extracting Connectivity from dMRI

Long-range communication between neural populations is constrained by the brain’s structural connectivity. Diffusion-weighted MRI (dMRI) provides a means to reconstruct white matter pathways and quantify connectivity strength between regions. By integrating these structural features into computational models, we can simulate biologically plausible network interactions and study their role in shaping brain activity.

Brain Network Models: Merging Dynamics and Structure

Combining neural mass models with structural connectivity data leads to brain network models—simulations that capture how distributed regions interact dynamically. These models provide a powerful tool to investigate large-scale functional connectivity, spectral properties, and topographical organization, helping bridge the gap between structural and functional neuroimaging findings.

Parameter Exploration: Matching Models to Real Brain Activity

Brain network models must be carefully parameterized to reproduce empirical observations. Key aspects include:

Spectral properties: Matching frequency-specific activity patterns seen in MEG/EEG data.
Topographical organization: Ensuring spatial correspondence between simulated and real activation maps.
Functional connectivity: Reproducing the network interactions observed in resting-state and task-based neuroimaging. By systematically tuning model parameters, we aim to generate simulations that align closely with empirical brain dynamics.

Related citations (Cabrera-Álvarez et al., 2023)

Future Directions: Expanding the Scope of Brain Simulations

As computational methods advance, several promising avenues for research emerge:

Surface-Based Simulations: Moving beyond regional models to simulate activity directly on the cortical surface, capturing finer spatial details of brain function.
Minimal Parcellations: Exploring ultra-fine-grained models where parcels are as small as individual mesh triangles, allowing for more precise representations of neural dynamics.

This research line contributes to the broader goal of developing biologically plausible and computationally efficient models that enhance our understanding of brain function and its disruptions in neurological conditions.

References

2023

PLoS Comput. Biol.
Modeling the role of the thalamus in resting-state functional connectivity: Nature or structure

Jesús Cabrera-Álvarez, Nina Doorn, Fernando Maestú, and 1 more author

PLOS Computational Biology, Aug 2023

Abs DOI Bib

The thalamus is a central brain structure that serves as a relay station for sensory inputs from the periphery to the cortex and regulates cortical arousal. Traditionally, it has been regarded as a passive relay that transmits information between brain regions. However, recent studies have suggested that the thalamus may also play a role in shaping functional connectivity (FC) in a task-based context. Based on this idea, we hypothesized that due to its centrality in the network and its involvement in cortical activation, the thalamus may also contribute to resting-state FC, a key neurological biomarker widely used to characterize brain function in health and disease. To investigate this hypothesis, we constructed ten in-silico brain network models based on neuroimaging data (MEG, MRI, and dwMRI), and simulated them including and excluding the thalamus, and raising the noise into thalamus to represent the afferences related to the reticular activating system (RAS) and the relay of peripheral sensory inputs. We simulated brain activity and compared the resulting FC to their empirical MEG counterparts to evaluate model’s performance. Results showed that a parceled version of the thalamus with higher noise, able to drive damped cortical oscillators, enhanced the match to empirical FC. However, with an already active self-oscillatory cortex, no impact on the dynamics was observed when introducing the thalamus. We also demonstrated that the enhanced performance was not related to the structural connectivity of the thalamus, but to its higher noisy inputs. Additionally, we highlighted the relevance of a balanced signal-to-noise ratio in thalamus to allow it to propagate its own dynamics. In conclusion, our study sheds light on the role of the thalamus in shaping brain dynamics and FC in resting-state and allowed us to discuss the general role of criticality in the brain at the mesoscale level.
@article{Cabrera-Alvarez2023-Thalamus, title = {Modeling the role of the thalamus in resting-state functional connectivity: Nature or structure}, volume = {19}, issn = {1553-7358}, url = {http://dx.doi.org/10.1371/journal.pcbi.1011007}, doi = {10.1371/journal.pcbi.1011007}, number = {8}, journal = {PLOS Computational Biology}, publisher = {Public Library of Science (PLoS)}, author = {Cabrera-Álvarez, Jesús and Doorn, Nina and Maestú, Fernando and Susi, Gianluca}, editor = {Cagnan, Hayriye}, year = {2023}, month = aug, pages = {e1011007}, dimensions = {true}, selected = true, }