Limits...
Discriminating external and internal causes for heading changes in freely flying Drosophila.

Censi A, Straw AD, Sayaman RW, Murray RM, Dickinson MH - PLoS Comput. Biol. (2013)

Bottom Line: The remaining turning decisions, not explained by this feature of the visual input, may be attributed to a combination of deterministic processes based on unobservable internal states and purely stochastic behavior.We cannot distinguish these contributions using external observations alone, but we are able to provide a quantitative bound of their relative importance with respect to stimulus-triggered decisions.We discuss how this technique could be generalized for use in other systems and employed as a tool for classifying effects into sensory, decision, and motor categories when used to analyze data from genetic behavioral screens.

View Article: PubMed Central - PubMed

Affiliation: Control & Dynamical Systems, California Institute of Technology, Pasadena, California, United States of America.

ABSTRACT
As animals move through the world in search of resources, they change course in reaction to both external sensory cues and internally-generated programs. Elucidating the functional logic of complex search algorithms is challenging because the observable actions of the animal cannot be unambiguously assigned to externally- or internally-triggered events. We present a technique that addresses this challenge by assessing quantitatively the contribution of external stimuli and internal processes. We apply this technique to the analysis of rapid turns ("saccades") of freely flying Drosophila melanogaster. We show that a single scalar feature computed from the visual stimulus experienced by the animal is sufficient to explain a majority (93%) of the turning decisions. We automatically estimate this scalar value from the observable trajectory, without any assumption regarding the sensory processing. A posteriori, we show that the estimated feature field is consistent with previous results measured in other experimental conditions. The remaining turning decisions, not explained by this feature of the visual input, may be attributed to a combination of deterministic processes based on unobservable internal states and purely stochastic behavior. We cannot distinguish these contributions using external observations alone, but we are able to provide a quantitative bound of their relative importance with respect to stimulus-triggered decisions. Our results suggest that comparatively few saccades in free-flying conditions are a result of an intrinsic spontaneous process, contrary to previous suggestions. We discuss how this technique could be generalized for use in other systems and employed as a tool for classifying effects into sensory, decision, and motor categories when used to analyze data from genetic behavioral screens.

Show MeSH

Related in: MedlinePlus

The estimated decision feature . Panel A shows the estimated one dimensional feature . This is the best one dimensional spatial feature that explains the left and right saccade rates. It is a dimensionless quantity, which we normalize in the interval . Panel B-i shows, for each cell , the rates  as a function of the estimated feature ; Panel B-ii shows the same data, but with error bars corresponding to 95% confidence intervals (the bars are not symmetric because the posterior distribution of the estimated rates is not Gaussian; see Supplemental Materials for details). The single feature  is sufficient to predict the rate in  of the environment, in the sense that 93% of the rates can be considered (with the error bars) as lying on the same curve; these curves are the functions  and  discussed previously that allow predicting the rates from the feature. The remaining  of data that this model cannot fit correspond to configurations with the fly pointing directly against the wall at a small distance (<0.3 m).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3585425&req=5

pcbi-1002891-g005: The estimated decision feature . Panel A shows the estimated one dimensional feature . This is the best one dimensional spatial feature that explains the left and right saccade rates. It is a dimensionless quantity, which we normalize in the interval . Panel B-i shows, for each cell , the rates as a function of the estimated feature ; Panel B-ii shows the same data, but with error bars corresponding to 95% confidence intervals (the bars are not symmetric because the posterior distribution of the estimated rates is not Gaussian; see Supplemental Materials for details). The single feature is sufficient to predict the rate in of the environment, in the sense that 93% of the rates can be considered (with the error bars) as lying on the same curve; these curves are the functions and discussed previously that allow predicting the rates from the feature. The remaining of data that this model cannot fit correspond to configurations with the fly pointing directly against the wall at a small distance (<0.3 m).

Mentions: Figure 4A shows the estimated saccade generation function across the reduced configuration space. These rates are obtained by first computing the observed generation rates by averaging the number of saccades (Figure 1E) by the time spent in each cell (Figure 1D). Then the rates are obtained from by correcting for an estimated inhibition interval s. Panels B and C show the data separately for left and right saccades ( and ). The most evident phenomenon is that the fly tends to turn left when the wall is on the right (and vice versa), however, there are many saccades of the opposite direction initiated, even when the turning would orient the fly towards the wall rather than away from it. This is the phenomenon that we want the feature to explain: we want to find the best spatial scalar value such that both and can be written as a function of . Figure 5Ai-ii shows the estimated feature as a function of the reduced configuration c. This is the unidimensional feature that best explains both the left and saccade rates. The estimated feature using the alternative saccade detector is qualitatively similar (Figure S1). We now have the spatial feature as well as the rates as a function of the reduced configuration and can now plot as a function of (using as an implicit variable). This is shown in Figure 5B, which shows, for each cell , the value of as a function of . Figure 5Bi shows the data as a scatter plot, while Figure 5Bii shows the error bars on the estimated rates at the 95% significance level.


Discriminating external and internal causes for heading changes in freely flying Drosophila.

Censi A, Straw AD, Sayaman RW, Murray RM, Dickinson MH - PLoS Comput. Biol. (2013)

The estimated decision feature . Panel A shows the estimated one dimensional feature . This is the best one dimensional spatial feature that explains the left and right saccade rates. It is a dimensionless quantity, which we normalize in the interval . Panel B-i shows, for each cell , the rates  as a function of the estimated feature ; Panel B-ii shows the same data, but with error bars corresponding to 95% confidence intervals (the bars are not symmetric because the posterior distribution of the estimated rates is not Gaussian; see Supplemental Materials for details). The single feature  is sufficient to predict the rate in  of the environment, in the sense that 93% of the rates can be considered (with the error bars) as lying on the same curve; these curves are the functions  and  discussed previously that allow predicting the rates from the feature. The remaining  of data that this model cannot fit correspond to configurations with the fly pointing directly against the wall at a small distance (<0.3 m).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3585425&req=5

pcbi-1002891-g005: The estimated decision feature . Panel A shows the estimated one dimensional feature . This is the best one dimensional spatial feature that explains the left and right saccade rates. It is a dimensionless quantity, which we normalize in the interval . Panel B-i shows, for each cell , the rates as a function of the estimated feature ; Panel B-ii shows the same data, but with error bars corresponding to 95% confidence intervals (the bars are not symmetric because the posterior distribution of the estimated rates is not Gaussian; see Supplemental Materials for details). The single feature is sufficient to predict the rate in of the environment, in the sense that 93% of the rates can be considered (with the error bars) as lying on the same curve; these curves are the functions and discussed previously that allow predicting the rates from the feature. The remaining of data that this model cannot fit correspond to configurations with the fly pointing directly against the wall at a small distance (<0.3 m).
Mentions: Figure 4A shows the estimated saccade generation function across the reduced configuration space. These rates are obtained by first computing the observed generation rates by averaging the number of saccades (Figure 1E) by the time spent in each cell (Figure 1D). Then the rates are obtained from by correcting for an estimated inhibition interval s. Panels B and C show the data separately for left and right saccades ( and ). The most evident phenomenon is that the fly tends to turn left when the wall is on the right (and vice versa), however, there are many saccades of the opposite direction initiated, even when the turning would orient the fly towards the wall rather than away from it. This is the phenomenon that we want the feature to explain: we want to find the best spatial scalar value such that both and can be written as a function of . Figure 5Ai-ii shows the estimated feature as a function of the reduced configuration c. This is the unidimensional feature that best explains both the left and saccade rates. The estimated feature using the alternative saccade detector is qualitatively similar (Figure S1). We now have the spatial feature as well as the rates as a function of the reduced configuration and can now plot as a function of (using as an implicit variable). This is shown in Figure 5B, which shows, for each cell , the value of as a function of . Figure 5Bi shows the data as a scatter plot, while Figure 5Bii shows the error bars on the estimated rates at the 95% significance level.

Bottom Line: The remaining turning decisions, not explained by this feature of the visual input, may be attributed to a combination of deterministic processes based on unobservable internal states and purely stochastic behavior.We cannot distinguish these contributions using external observations alone, but we are able to provide a quantitative bound of their relative importance with respect to stimulus-triggered decisions.We discuss how this technique could be generalized for use in other systems and employed as a tool for classifying effects into sensory, decision, and motor categories when used to analyze data from genetic behavioral screens.

View Article: PubMed Central - PubMed

Affiliation: Control & Dynamical Systems, California Institute of Technology, Pasadena, California, United States of America.

ABSTRACT
As animals move through the world in search of resources, they change course in reaction to both external sensory cues and internally-generated programs. Elucidating the functional logic of complex search algorithms is challenging because the observable actions of the animal cannot be unambiguously assigned to externally- or internally-triggered events. We present a technique that addresses this challenge by assessing quantitatively the contribution of external stimuli and internal processes. We apply this technique to the analysis of rapid turns ("saccades") of freely flying Drosophila melanogaster. We show that a single scalar feature computed from the visual stimulus experienced by the animal is sufficient to explain a majority (93%) of the turning decisions. We automatically estimate this scalar value from the observable trajectory, without any assumption regarding the sensory processing. A posteriori, we show that the estimated feature field is consistent with previous results measured in other experimental conditions. The remaining turning decisions, not explained by this feature of the visual input, may be attributed to a combination of deterministic processes based on unobservable internal states and purely stochastic behavior. We cannot distinguish these contributions using external observations alone, but we are able to provide a quantitative bound of their relative importance with respect to stimulus-triggered decisions. Our results suggest that comparatively few saccades in free-flying conditions are a result of an intrinsic spontaneous process, contrary to previous suggestions. We discuss how this technique could be generalized for use in other systems and employed as a tool for classifying effects into sensory, decision, and motor categories when used to analyze data from genetic behavioral screens.

Show MeSH
Related in: MedlinePlus