Simulating Large-scale Models of Brain Neuronal Circuits using Google Cloud Platform.

Area of Science:

• Computational Neuroscience
• Neuroscience
• High-Performance Computing

Background:

• Biophysically detailed modeling is crucial for integrating experimental data and simulating brain circuits with high precision.
• Large-scale neural simulations require significant computational resources.

Purpose of the Study:

• To describe the setup and utilization of Google Compute Platform (GCP) with Slurm for large-scale biophysically detailed brain circuit simulations.
• To present best practices and solutions for challenges encountered during high-performance computing for neuroscience.

Main Methods:

• Developed a detailed model of motor cortex circuits, including over 10,000 neurons and 30 million synaptic connections.
• Employed parameter exploration using grid search and evolutionary algorithms for model optimization.
• Utilized Google Compute Platform (GCP) with the Slurm workload manager for high-throughput simulations.

Main Results:

• Successfully configured and executed tens of thousands of large-scale neural simulations on GCP.
• Identified and addressed practical challenges associated with cloud-based high-performance computing for neuroscience.
• Obtained preliminary results from the simulated motor cortex circuits.

Conclusions:

• Google Compute Platform (GCP) with Slurm provides a viable and powerful environment for large-scale, biophysically detailed neural simulations.
• The described methodology and solutions facilitate efficient computational neuroscience research.
• This work enables advanced exploration of brain circuit dynamics through scalable simulation.