Summary: weekly mtg 20170613 (Byron, Steve, Emily, me) - mobeets/nullSpaceControl GitHub Wiki

Discussion:

• first sentences should be the main point.
• OFC paragraph example
  • Almost all OFC models proposed to date have been framed in terms of muscle activity. How could we apply these models to neural activity? Certain implementations of OFC cost functions specify minimal intervention, but we find that there's yoking etc. Others suggest null space activity might be uncontrolled. Our results are not inconsistent with this, as we find that the entirety of the support of the null space ends up being used, etc.
• Emily:
  • beginning summary could be stronger?
  • end kinda trails off?
• Steve: maybe as spine for last section, reiterate that previous studies of neural activity focus on signal dependence, tuning curves, potent dimensions. however, the principles that dictate how activity is organized in the null space can also be informative about brain function. now bring up kaufman: activity in the null space in the null space can permit motor preparation without overt movements; it can permit dynamic processing while still allowing for stable perception. null space allows for multiplexing, e.g.
  • raposo, churchland
  • mante, newsome
  • d'avella bizzi pnas

Habitual -> Persistent Strategy