Figure 4. Odors independently modulate multiple behavioral parameters in an odor dependent manner.

(A) 17 parameters which are all significantly modulated by ACV0 at p <0.05 (each parameter is described in detail in Materials and methods). Bars on top indicate the variables that are significantly different from the solvent control in a rank sum test after Bonferroni correction for multiple comparisons (p <0.003). Dashed line marks the expected value when there is no odor modulation. (B) Principal component analysis on the 17-dimensional odor space shows that ACV0 and BUN3 both activate multiple independent motor programs. (C) Left: The first canonical variates for the response of individual flies to ACV0 (magenta) and BUN3 (green). The responses due to BUN3 and ACV0 are clearly segregated along the first canonical variate. Each circle is a single fly. Right: The distribution of distances between medians in 50,000 trials in which odor labels were randomized. Less than 0.003 (0.3%) of the trials had medians greater than the original distribution (gray shaded area).

DOI: http://dx.doi.org/10.7554/eLife.11092.011

Figure 4—figure supplement 1. Evidence for independence of different motor parameters.

If attraction is a singular motor program, then we would expect other motor programs to scale with the level of attraction. We examine the linear correlations between these parameters for a fly’s response to ACV0. (A) On a fly-by-fly basis, time to return is uncorrelated with time spent inside implying that these two programs are independent. Line represent the best linear fit to data. (B) Same data as in A but normalized to show fractional change in the two parameter. Line represent the best linear fit to data. (C) Cross-covariance between the parameters in our 17-parameter description of behavior show that most parameters are not significantly correlated (left). Only 15 out 136 pairwise comparisons are significantly correlated (right). 1. attraction index, 2. time inside/transit, 3. time to return, 4. radial density, 5. speed in, 6. speed out, 7. speed crossing in, 8. speed crossing out, 9. run duration, 10. stop duration, 11. run probability in, 12. run probability out, 13. angular speed in, 14. angular speed out, 15. smooth turns in, 16. smooth turns out, 17. sharp turn at odor bounary.

DOI: http://dx.doi.org/10.7554/eLife.11092.012