Brody-Lab / jbreda_animal_training

Repository for the ingestion, cleaning and visualization of behavioral experiments
0 stars 0 forks source link

R032 #37

Open jess-breda opened 6 months ago

jess-breda commented 6 months ago

anti guide is still water water fyi

jess-breda commented 6 months ago

hit switch threshold at 0.65

jess-breda commented 6 months ago

hit switch threshold 0.65 -> 0.75 to maintain pro performance. note, this might be a little high need to keep an eye on.

jess-breda commented 6 months ago

alpha minus at 0.75 (default)

jess-breda commented 6 months ago

dropping alpha minus 0.75 -> 0.68 to encourage growth.

jess-breda commented 6 months ago

violating a lot- dropping post .4 -> 0.25

jess-breda commented 6 months ago

seems to be a bug where the give rate is low, but non-give stim performance is only for the pro pairs.

Image

jess-breda commented 5 months ago

fully off give delay and performance is looking good. moving into stage 16 (no give) since the animal isn't using it. Making block sizes random and smaller (mean 10, std 2). hit switch threshold is at 0.73 (was at 0.70) and switch sound duration went from 5 to 2 seconds

Image

jess-breda commented 5 months ago

changes above did not save properly- now officially into stage 16, re-weaned itself off

Image

jess-breda commented 5 months ago

putting into sampled block size with mean of 10. should be moving into blocked -> random soon

jess-breda commented 5 months ago

turning off switch sound- let's see how performance looks

jess-breda commented 5 months ago

fully off light guide and switch sound- putting into random

Image

jess-breda commented 4 months ago

doing well, dropping max blocks a bit from 7 -> 5 and pro prob 0.6 -> 0.5

Image