Breaking the ellipsoid body code
on Monday, October 20th, 2025 12:14 | by Björn Brembs
Now that I’ve reached a sample size of near 30 animals in each group, it was time to break the code to see which group was wich. To my surprise, the worst performing group in term,s of learning was the control group:

It is unusual for between-groups comparisons to become statistically significant at the 0.005 level with a sample size of less than 50, let alone 30, but these DopR2 knock-downs in ellipsoid body ring neurons are just a class of their own:

The DopR2 receptor knock-down not only performed exceedingly well in these leaning experiments, it was also the cross with the strongest optomotor response:

And also that difference is statistically significant:

The TNT-E flies had a slight positive preference, worse avoidance (esp. in the first training period) and flew really bad. One could say they did have some learning deficits, but it doesn’t look as if learning was completely abolished (which is roughly what Andreas had found before):

Andi’s results:

Category: operant self-learning
Leave a Reply