Discrete velocity decoding worked very well for a monkey using an intracortical #BCI as well. (37% target hits with the Wiener filter vs 61% with DDS).

The discrete direction decoder (DDS, purple) makes staircase-looking cursor paths, but people do really well with it in our iBCI model, hitting way more targets than standard decoders.

tl;dr on the #BCI model we used (the jaBCI): scholarblogs.emory.edu/danziger/202...

Many science SM posts are about new grants and papers, which makes it hard to get a realistic sense picture of what its like applying for grants.

To help, especially for #newPI, I am sharing a complete record of grant submissions.

Full details on the blog: scholarblogs.emory.edu/danziger/202...

The secret sauce for this approach is to use sequential (Newton) optimization that at each step 1st solves for the system dynamics then 2nd updating the ANN parameters using the step-1 results.

It works great in many simulated ODE systems.

Say we want to simulate a neuron and have eqns. for Na+ activation, Na+ inactivation, and membrane voltage. We know it has important K+ dynamics, but don't know the eqn. or have measurements of it.

We can embed an ANN into the system and fully recover the missing K+ dynamics. 🥳