We need a student to finish up experiments

Example of ON/OFF traces.  And then the wiggling was measured in both conditions (with and without light). It seems like they move more when they are in the light side. I also did measure the derivative with a tau=2 (in case the sampling is over the temporal frequency of fly behavior)

Comparing scores in both setups show some lines that match their scores and other lines that have opposite scores. This filters the ones that are context dependent to the ones that show a context independent reinforcement valence.

It seems that there is something more there. There were 5 batches of 3 flies per side reinforced. I put the intensity higher than ever before. I will try to get the maximum intensity for the next experiment and see what it looks like.

I have decreased the light intensity to around half of what it previously was. The result do not show any clear phenotype. The ‘logical PI’ looks kind of inverted because I actually have plotted it so that positive means light, negative means no light. A total of 30 flies were tested: some of them with right, other with left reinforcement

Example of a trace within the different trainning/test segments

Inter sampling interval just to check

10 experiment segments with alternating no reinforcement/reinforcement

The same as above but for the 3 platforms that ran in the same batch

Trace dynamics when the light is ON (blue) and OFF (red). Working on nicer visualization in general

So this first picture shows graphically how I get the valences contributions for each of the dopaminergic clusters. On the Y-axis you see the lines I used for the modelling and on the x-axis the clusters. This is the expression pattern for all the drivers (split G4 and the dirtier G4s). I also made this expression pattern binary, to avoid the errors I could add by trying to estimate the expression intensity from the literature.

Here below are the results I obtained for one of the metrics. I wont explain to much here because the main result I see is that the results change drastically upon changes in the model. This tells me that there is something wrong there. Since making the expression table binary or weighted, or using a subset of the G4s used should not give me so random values for the dopaminergic clusters.

With this, I am quite stuck and do not know what to do next. Results seem not to show that much. Considering planning another experiment while there is time or continuing analyizing. Comments please!