for example, in the famous within- and off-manifold BCI learning task, we show that we can understand the learning speeds and recovery of individual decoders, including variability of within-manifold decoders, with "dynamical constraints" - low/high controllability and dynamical (mis)alignment.

We brought a control theoretic perspective to understand what kinds of readouts/behaviors are easy to learn and generate. We can differentiate desired neural _trajectories_ based on alignment with underlying flowfields and controllability, even when all trajectories lie in the same subspace.

Comp #Neuroskyence hivemind - I've noticed multiple eqs for RNN dynamics that vary in interpretation and implications due to placement of nonlinearity. Similar mixture in theory papers... @loradrian.bsky.social @jbarbosa.org @bio-emergent.bsky.social @matthijspals.bsky.social @kenmiller.bsky.social

Not biased at all... but I think this is one of the most exciting places to be a systems neuroscientist and develop an interdisciplinary skill set.

Link: www.ucl.ac.uk/gatsby/study...

Since this is a fresh start, I replace Eliot's urban discontent with some more energetic, optimistic, even whimsical takes on being human from Ted Kooser, Louïse Gluck, James Tate, Wisława Szymorska and Czesław Miłosz.