The Showcase is implemented as a series of interactive Jupyter notebooks covering the individual logical steps and can be accessed in a dedicated public EBRAINS collab. 

An EBRAINS collab (as this one dedicated to the SGA3 D1.5 Deliverable) is a virtual environment providing access to a few interlinked services: Wiki, Drive, Bucket and Lab. The Wiki is this web interface that allows to visually navigate through the components of a collab and allows for basic documentation to be presented in the form of webpages. The Drive provides storage for small files. It is the local filesystem of a collab in which the documentation, the notebooks and all the supporting code are placed. The Bucket is a storage service for larger files. In the present collab, the Bucket holds the pre-computed results from the extensive parameter sweeps and model optimizations to allow skipping the computationally demanding steps. The Lab service is an instance of JupyterLab—an interactive computing environment—where the notebooks can be run and worked with. The Lab displays the code and the notebooks that are stored in the Drive.

The Jupyter notebooks in this collab will load all required Python modules including Siibra and The Virtual Brain, and the interfaces for launching the computationally demanding parts in the HPC infrastructure.

The Showcase 1 aimed at investigations related to variability in neuroscience from two perspectives: (a) the interpersonal variability  studied by the virtual ageing study, and (b) the variability across different cortical regions within an individual brain.

(a) Interpersonal variability—virtual ageing

See the details of the first study in the following publication:

M. Lavanga, J. Stumme, B. H. Yalcinkaya, J. Fousek, C. Jockwitz, H. Sheheitli, N. Bittner, M. Hashemi, S. Petkoski, S. Caspers, and V. Jirsa, The Virtual Aging Brain: A Model-Driven Explanation for Cognitive Decline in Older Subjects.

Simulation of resting-state activity

The fist notebook in the inter-individual variability workflow explores the resting-state simulation for a subject of the 1000BRAINS dataset. Functional data are simulated by means of a brain network model implemented in TVB, which is an ensemble of neural mass models linked via the weights of the structural connectivity (SC) matrix. Following topics are covered:

1. The neural-mass model by Montbrió, Pazó and Roxin.
2. Construction of the TVB model for a particular subject.
3. Dynamics of the model, and execution of a parameter study.
4. Summary of the simulation results for the whole cohort.

Link to the notebook:

• virtual_ageing/notebooks/1_BNM_for_resting_state.ipynb

Virtual ageing trajectories

The second steps shows the investigation of virtual ageing trajectory for each subject. In this context, we are going to show:

1. What we mean by virtual ageing and what is the empirical basis to investigate this approach.
2. How we can virtually age a subject using whole-brain modelling.
3. How the increase structure-function relationship relates to virtual ageing.

Link to the notebook:

• virtual_ageing/notebooks/2_virtual_ageing_trajectories.ipynb

Inference with SBI

The last step of the inter-individual variability workflow employs Simulation Based Inference for estimation of the full posterior values of the parameters. Here, a deep neural estimator is trained to provide a relationship between the parameters of a model (black box simulator) and selected descriptive statistics of the observed data. 

Link to the notebook:

• virtual_ageing/notebooks/3_inference_with_SBI.ipynb

(b) Regional variability

Aims at demonstrating the construction of whole-brain network models of the brain's activity, accounting for differences in receptor densities across cortical regions.

Loading the data from EBRAINS

The first step of this workflow consists in loading the data from the Knowledge Graph via the siibra interface, including the regional bias on the model. In this case we require:

1. Structural connectivity matrices,
2. GABAa and AMPA receptor densities for each brain region, and
3. empirical resting-state fMRI data for fitting and validation of the simulations.

The three datasets shall be characterised in the same parcellation. Link to the notebook:

• regional_variability/notebooks/1_load_data.ipynb 

Fitting model parameters for models with regional bias

A series of simulations of the whole-brain network model are launched in the EBRAINS HPC facilities in order to identify the optimal model parameters leading to simulated resting-state brain activity that best resembles the empirically observed activity. In this branch of the showcase, brain regions are simulated using the mean-field AdEx population model; specifically modified to account for the regional densities of GABAa and AMPA neuroreceptors. See the details following document.

Link to the notebook:

• regional_variability/notebooks/2_parameter_swep.ipynb