Skip to content

Instantly share code, notes, and snippets.

@elijahc
Last active July 12, 2017 18:59
Show Gist options
  • Select an option

  • Save elijahc/24f4d21f04c31b6b99063176b6004bff to your computer and use it in GitHub Desktop.

Select an option

Save elijahc/24f4d21f04c31b6b99063176b6004bff to your computer and use it in GitHub Desktop.
Updates - FORCE learning and timeconstant

In walking through the math to derive the equation they use in their code to maintain/model the X(t) neuron state vector I realized I was not using an appropriate time step size. After correcting the step size the model appears to make better predictions as quantified by mean Pearson Correlation Coefficient

N = 1000

214 Cells in Experiment 500964514

Using 50 cell ids: [ 7 11 26 28 40 41 45 47 49 50 54 63 68 69 70 76 77 82 83 87 93 95 100 101 104 105 106 107 115 117 122 125 131 148 151 156 158 163 167 169 174 175 179 182 183 191 198 206 207 212]

alpha = 1

pGG = .1

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment