Computational protocol: Oscillatory brain activity in spontaneous and induced sleep stages in flies

Similar protocols

Protocol publication

[…] All statistical analyses for data gathered from the overnight and exposed-brain recording setup were performed using Prism 7 for Windows (GraphPad). A subset of behavioral and LFP power data set did not pass the Shapiro–Wilk normality test (p < 0.05). Depending on the outcome of the Shapiro–Wilk normality test, a Wilcoxon signed rank test or a t test was used to test for significant effects between two matched conditions. The appropriate tests used are mentioned in the figure legends. Friedman test with Dunn’s post hoc multiple comparisons test were used to compare three or more matched conditions, and Kruskal–Wallis test with Dunn’s post hoc for unmatched data. All the data presented in figures are as means ± S.E.M. for bar and line graphs while box and whiskers plot presents median and 10–90 percentiles as whiskers. All tests for significance were two-tailed and confidence levels set at α = 0.05.For the multichannel statistical analysis, the following R packages were used: ARTool, , car, dplyr, influence.ME, lattice, lme4, magrittr, MASS, Matrix, nortest, phia, and plyr. The data.frame was organized by splitting the data set into 104 y and C5 groups to be analyzed separately. In the case of the frequency cluster analysis, the data were further divided into individual frequency bands. The data for the 2–40 Hz band were not normally distributed (Lilliefors (Kolomogorv–Smirnov) Test p < 0.001). Therefore, a non-parametric test was used for the log transformed data, which allowed the test of multiple factors and their interactions called the Aligned-Rank ANOVA from the R ARTool package. The Aligned-Rank ANOVA allows multi-factor or mixed model regression to be performed on a non-parametric dataset or one that violates the normal assumptions of parametric models. For the 2–40 Hz and frequency cluster analysis, Aligned-Ranks were constructed using the art function from ARTool. The ARTool package makes use of the lmer function for testing mixed models from the lme4 package and thus uses its syntax.To perform contrasts on significant higher-order interactions, the testInteractions function from the phia package was used to test post hoc contrasts between categorical variables, employing a scheme called Helmert coding. Unlike other types of factor level coding, Helmert contrasts allows flexibility in the equivalence assigned to factor levels. In this instance, it allows the mean across both optic lobes to be compared to the center for the Region factor (e.g., −1/2 for each optic lobe and 1 for the center, summing to zero). The contrasts also compared the TRP-lines to GAL4 or UAS controls (TRP = 1, GAL4 = −1), Heat On to Baseline (Baseline = −1, Heat On = 1, Heat Off = 0) or Heat Off to Baseline (Baseline = −1, Heat On = 0, Heat Off = 1), unless otherwise specified. The testInteractions function takes the model output provided by ARTool. The Aligned-Rank ANOVA has two diagnostic tests associated with it which tests whether the aligned-rank transformation was performed successfully. For the first test, the columns of aligned-rank responses should all sum to zero. All analyses performed passed this test. The second test checks whether a full-factorial ANOVA on ranked (but not aligned) responses has all main effects stripped out as indicated by an F value of 0 (Pr = 1). […]

Pipeline specifications

Software tools dplyr, lme4
Application Mathematical modeling
Organisms Drosophila melanogaster