Gridsearch speedups TODO list

[ollieparish]

• [x] (1) Combine morph maps and inv-sols into numpy array thing (ch_names stored separtely)
  • [MVP DONE] #276
  • check if sparse [They are]
• [ ] (2) Functions to 2s format / EMEG to 1s + zeros format
  • _This means getstds (normalisation to turn dot-product into valid correlation) happens on func instead
  • Also means the corrs will be slightly different? (as a result of the cutting up of the file)
• [ ] (3) FFT and element-wise multiply in sensor-space, then map up
  • Need to map up separately (just summary stats for the new normalisation std)
  • _Func getstds indep. of map up (now done on func) - The EMEG summary stat norms
  • derangements probably stay in EMEG space now that they're in 1s format (?)
• [ ] (4) Store EMEG data in FFT-ed format (along with norm vals maybe) <- quite a big speed-up probably
• [ ] (5)Think about the pval calculation - maybe think about this later?

[neukym]

For

(5)Think about the pval calculation - maybe think about this later?

we should perhaps think about moving to pure probabilities at this stage. Kaibo and I have been working on this and I would like to share with you and Tianyi at some point.