Closed-Loop Deep-Brain Stimulation for Controlling Synchronization of Spiking Neurons.
The overall architecture of this work.
git submodule update --init --recursive
pip install -r requirements.txt
pip install ./dbsenv
python src/run_env_cls.py
tensorboard --logdir ./ppo_tensorboard
# open http://localhost:6006/ to visualize training results in TensorBoard.- ftsts-ol - implementation based on original open-loop regime.
- dbsenv - custom OpenAI Gym environment for simulating closed-loop DBS.
- ftsts - original work on open-loop stimulation regime.
- ftsts (post-ictal) - extension of original work to target post-ictal desynchronization.
- ftsts (pi) - extension of original work to close the loop using proportional-integral (PI) control. (contact for material)
- rl driven cl-dbs - a study on using reinforcement learning agents for adaptive parameter control in a closed loop DBS environment based on the basal ganglia-thamalic (BGT) model.