Extended Data Fig. 4: Reanalysis of the postural sway in humans, Homo sapiens¹⁷, and odor plume tracking response of American cockroach, Periplaneta americana¹⁸ show evidence in support of sensory salience dependent mode-switching strategy.

(a,b) Representative temporal traces of mediolateral movement of center of mass, ML CoM (a) and the histograms of ML CoM velocities (b) for different experimental conditions-both vision and touch (top), only touch (middle) and only vision (bottom). The magenta dashed and the blue solid curves in (b) correspond to a normal and GMM fit with three components, respectively. The dataset analyzed here, comprised of 7 subjects (N), with 2-3 replicate trials per experimental condition, was collected at 50 Hz. (c,d) Comparison of the RMS velocities (c) and switching frequency (d) for different experimental conditions. Different shades of gray denotes different human subjects. The p-values were computed using the sign test. (e,f) Representative temporal traces of the lateral head movement (e) and the histograms of the lateral velocities (f) for different antennae length as indicated. The colors of fitted curves are same as in (a,b). The dataset analyzed here, was collected at 30 Hz but later was subsampled at 15 Hz by the original study authors. The kurtosis (κ) values and the total number of trajectories (N, single trajectory per subject) analyzed are indicated next to the respective panels in (f). (g,h) Comparison of the RMS lateral velocities (g) and residence time at slow mode (h) for different experimental conditions. All box and whisker plots include the median line, the box denotes the interquartile range (IQR), whiskers denote the rest of the data distribution and outliers are denoted by points greater than ± 1.5 × IQR. Sample sizes (n) are shown in each boxplot. The one-sided p-values were computed using the Mann-Whitney-Wilcoxon test.