Summary: weekly mtg 20160921 (Byron, Steve, Matt, me) - mobeets/nullSpaceControl GitHub Wiki

Discussed

to discuss in paper:

• clarify that first two factors aren't the mapping
• relate to nature paper
• has monkey fully internalized the mapping? short-term vs. long-term

look at data for artifacts

• those sessions with large cov error

SSS

• check IME vs without
• try trace instead of det of cov
• compare hyps
• look at variance in row spaces, SSS
• scale by total covariance in space per session?
  • e.g. (tr(cov(Y*SSS))/2) / (tr(cov(Y))/10)

To discuss

• main results figures
  • showing observed activity tuning?
  • simple hyps: showing distributions
  • 2-mappings hyps: showing distributions
  • comparing scores of hab and cld?
  • comparing scores of all hyps?
• paper outline