List view

• Circuit model with model inputs and outputs

  A generic `Circuit` class which accepts (different) models as nodes and connectivity matrices as edges. `Circuit` class should be able to integrate all models by one `dt`, gather all `outputs` and couple all nodes. Coupling is the supplied as `inputs` to all models. - Models need to be able to receive `inputs`. - Models need to be able to integrate by single `dt` *efficiently*. This requires a good solution to quickly reinitialize or continue an ongoing integration.

  No due date
  0% complete0 open0 closed

• Model-agnostic (chunkwise) integration and parallel BOLD simulation

  Chunkwise integration has to be implemented by the user. This could be solved more generally by the simulation framework itself. This would also make it possible to couple all models to a BOLD simulation. Todos: - chunkwise time integration - chunkwise integration should support integration of models by a single `dt` - universal BOLD simulation support - BOLD simulation on a separate thread / process? - `scipy.integrate.ode` and `scipy.integrate.odeint` integration support

  No due date

  •1/1 issues closed
  100% complete0 open1 closed

• Integration of more models

  More builtin models, including - [x] Wilson-Cowan model - [x] Fitz-Hugh Nagumo oscillator - ...

  No due date

  •1/1 issues closed
  100% complete0 open1 closed

• Whole-brain optimization.

  Working whole-brain optimization with one or more models. Parameter exploration should be possible with a whole-brain network and optimization using the evolutionary algorithm.

  No due date

  •2/2 issues closed
  100% complete0 open2 closed

• Introduce general architecture of the models

  Decide on, implement and test base classes for nodes (`NeuralMass`); circuits (representing whole-brain networks); integrators (only `numba` based? more?). Rewrite existing models (Hopf and ALN to the new setting using base classes)

  No due date

  •5/5 issues closed
  100% complete0 open5 closed