These are the results of Saurabh and Avani together. The positive control, the last one seems to be good (for the barplot and the boxplot). The only objection is that from the metadata it seems like they did all the positive controls on the same day, the 6th of June, which is not  a good scientific practice. The third plot is the amount of experiments of each line.

These are the experiments from Amanda, the second plot is the amount of experiments per line. Since the positive control does not show the aversive phenotype I am afraid I have to throw her data away :(

I can´t see if effects reproduce in different hands because I do not have the data from Amanda, so I will just pool all the data together.

Since we do not reproduce each others results, together with the previous post with the bootstrapping I can confirm that these neurons do not have an effect in reinforcement (in general). But we will focus on TH-D’

Residuals:
Min 1Q Median 3Q Max
-0.26964 -0.15074 0.07699 0.10043 0.24611

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.05663 0.12282 0.461 0.661
new_Christian$mean 1.07180 0.56922 1.883 0.109

Residual standard error: 0.1943 on 6 degrees of freedom
Multiple R-squared: 0.3714, Adjusted R-squared: 0.2667
F-statistic: 3.545 on 1 and 6 DF, p-value: 0.1087

Residuals:
Min 1Q Median 3Q Max
-0.43451 -0.10793 0.02633 0.14283 0.28667

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.05975 0.03595 -1.662 0.1101
new_Gaia$mean 0.33806 0.19627 1.722 0.0984 .
—
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 0.1787 on 23 degrees of freedom
Multiple R-squared: 0.1142, Adjusted R-squared: 0.07574
F-statistic: 2.967 on 1 and 23 DF, p-value: 0.09842

To see if there is an effect of the activation of these neurons in general, I thought of bootstrapping all the flies containing NorpA without G4s to see if the statistics are similar to that of my screen.

here the final screen

Here a total of 37 Tmaze experiments containing NorpA with UAS-GtACRs, UAS-Chrimson or UAS-ChR2XXL but without G4. Experiments are from Naman, Gaia and me, here the pooled effect:

Here the barplot of the result of  12 samples (with repetitions) for 20 times. Considering 37 experiments are closer to the real true distribution of NorpA flies, we sample from them to observe the probability of obtaining false positives, as well as the distributions.

Here the boxplotFrom this results I would deduce that most of these neurons have no effect on their reinforcement. This idea came because I saw that the phenotype scores of th-G4+th-G80>Chrimson was always close to zero in all of the behavioral setups. I thought that this could show that the dopaminergic neurons have an effect that is context dependent and that is why PIs might be more extreme than with the negative control (th-G4+th-G80>Chrimson).

This is a quick edit to see how it would look with 32 lines (the same number as for the real screen)

Tested Flies with the Joystick

Reinforcement

Positive Control: Gr28bd-G4, TrpA1-G4

Parameters: Light: intensity (500 Lux side, 1000 Lux bottom); frequency = 20Hz; Delay = 1 ms; Duration = 9.9 ms; volts = 6.4

Red lines: completed

mb025b: not selected against tubby

Mean trace of the positive control in the Joystick to get to see what are the overall dynamics and maybe to get an idea what might be the best score to pick.Here the standard deviation of the flies along the time axis. This is just to see if all the flies have more similar phenotypes with each other or not at each time.

This is to see if the flies have less degrees of freedom at any segment by measuring the standard deviation at each segment. There does not seem to be any effect. Although this might be mixed with the wiggle scores. I think measuring entropy is a better measure.

All the same plots as above but for TH-D’, the interesting line from the screen.

Standard deviation across flies

Standard deviation across segments

This are the performance indices for the different models performed to estimate the valence of the dopaminergic clusters. AIC: Akaike Information Criteria; BIC: Bayesian information Criteria; LogLikelihood: log Likelihood estimation

lm: linear model

+ int: taking double interaccions into consideration

b lm: bayesian linear model with bayesglm function

b lm MCMC: bayesian linear model with MCMCglm function

nlm: nonlinear model with lm function with splines fitted

b nlm: splines fitted to each cluster and MCMCglm function

GAM: general additive model with gam function

Adding double interactions seems to produce better models, nonlinearities also make models better and frequentist also. To me it seems like this  data might be noise and therefore adding interactions, nonlinearities and frequentist methods is just fitting the noise better (overfitting) and that is why I get better scores with them. In addition, care needs to be taken since I use different functions that calculate the model performance scores differently (although the formulas are theoretically the same for all!)

Below is given the plot of effect-sizes of reinforcement of 30 genotypes. On the y-axis are the PI values for learning effect sizes.  These scores are calculated by taking the average of PI values of training periods and then subtracting pretest PI values from it.

Reinforcement scores = mean of training score – pretest PI score

Yellow 1 (Positive Control): Gr28bd-G4, TrpA1-G4

Parameters: Light: intensity (500 Lux side, 1000 Lux bottom); frequency = 20Hz; Delay = 1 ms; Duration = 9.9 ms; volts = 6.4

This is the correlation from the T-maze experiments from Gaia and Naman. Neither ranked nor regular correlation show any significant effect. This means that these effects seem to be random, at least for most of them, is this an overfitting result?

I would say blue 1 is a line that was negative for all the tests I have so far seen. So this might be an interesting line. What to do next?

I would unblind the blue1, which is TH-D’. It was shown to be required for classical conditionning in shock and temperature learning (Galili et al. 2014). Another interesting observation is that th-g4+th-g80 seems to have like zero PI scores in all of the experiments (Naman and Gaia in the Tmaze, Joystick and Y-mazes). So could it be that all of these neurons have indeed a meaning, but is depending every time in the context?? Maybe Vanessa Ruta´s work might be interesting for that.