Fig 1.
Behavioral setup and adaptive echolocation call changes.
(A) Bats are trained to rest on a platform and track a moving target (tethered mealworm) by echolocation in the dark. The target(s) is moved via a rotary stepper motor attached to monofilament wire that is looped around a set of four pulleys. While the bat tracks the target(s), one microphone records the bat’s vocalizations, and a second microphone records the returning echoes. The 3-D positions of the head and pinna are recorded with a set of four infrared (IR) motion capture cameras. (B) Example sonar oscillograms (top) and target distance versus time plot (bottom) for each target motion condition presented to the bat (top to bottom): one-target-simple motion, two-target simple motion, two-target-pass motion, and one-target complex motion. (C) Top, mean +/- standard error of the mean (s.e.m) change in pulse interval over target distance for different varieties of target motion (colors as in B, significant differences are indicated in S1 Fig). Bottom, mean +/- s.e.m. change in pulse duration over target distance for each target motion condition (colors as in B, significant differences are indicated in S2 Fig). Data for this figure can be found at http://dx.doi.org/10.7281/T1W66HPZ.
Table 1.
Motion parameters and number of trials for each variety of target motion.
Fig 2.
Head waggles and target motion.
(A) Incidence of head waggle for Bat 1 during one-target simple (left) and one-target complex (right) motion. Displayed are the z-positions (height) of right ear (black) and left ear (red), with times of head waggles displayed in green. Notice the increase in head waggles for one-target complex motion. Far right panel displays a cartoon of the head waggle motion. (B) Summary across all bats for normalized head waggles per second for four different types of target motion. There is a significantly higher number of head waggles per second for one-target complex motions than all other types of target motion (permutation test, p < 0.0001). (C) Mean +/- s.e.m. of head waggles per second at the conclusion of a sonar sound group. The zero time point is the offset of the last sonar pulse in a sonar sound group. Plotted over these data are the mean waggles per second outside the context of sonar sound group production (black) and the 95% confidence interval on the mean in red. (D) Left to right, normalized head waggles per second for one-target simple, two-target simple, two-target pass, and one-target complex motions. The top row is the target position throughout trials; the bottom row is the target distance aligned normalized mean +/- s.e.m. of head waggles per second for each variety of target motion. Data for this figure can be found at http://dx.doi.org/10.7281/T1W66HPZ.
Fig 3.
Changes in pinna separation with target distance for each target motion condition (colors as in Fig 1).
(A) Top left, change in inter-pinna separation as a function of target distance for one-target simple motion. Plotted is the mean +/- s.e.m. for all trials. Values are normalized to zero for the starting position at the beginning of each trial. Top right, normalized data for two-target same motions; bottom left, normalized data for two-target pass motions (asterisks indicate when the second target begins to move, and when it overtakes the first target); bottom right, normalized data for one-target complex motions (asterisks indicate times when the target changes motion direction). (B) Normalized inter-pinna separation as a function of target distance for all varieties of target motion. There is a significant correlation between decreasing target distance and increasing pinna separation (Pearson’s correlation, r = 0.45, p < 0.0001). Inset details inter-pinna separation measurement. Data for this figure can be found at http://dx.doi.org/10.7281/T1W66HPZ.
Fig 4.
(A) Top, normalized inter-pinna separations for one-target simple motion (blue) versus two-target simple motions (green) combined across bats. Plotted are the mean +/- s.e.m. of the inter-pinna separation distances as a function of target distance. Bottom, dʹ calculation, or discriminability index, between one-target simple motion and two-target simple motion inter-pinna separation. Red shaded region indicates dʹ values above the 95% confidence interval as determined by a permutation test, indicating time points of significant differences in inter-pinna separation. (B) Top, normalized inter-pinna separations for one-target simple motion (blue) versus two-target pass motions (black) combined across bats. Bottom, dʹ calculation between one-target simple and two-target pass motion conditions. Red shaded region indicates significant differences. (C) Top, normalized inter-pinna separations for two-target simple motion (green) versus two-target pass motions (black) combined across bats. Bottom, dʹ calculation between two-target simple and two-target pass motion conditions. Red shaded region indicates significant differences. Data for this figure can be found at http://dx.doi.org/10.7281/T1W66HPZ.
Fig 5.
Local changes in inter-pinna separation are tied to sonar signals.
(A) Example of changes in inner-pinna distance for a bat tracking a tethered insect in a one-target complex motion trial. On a global scale, the bat changes inter-pinna separation with changing target distance; locally, there is also a small increase in inter-pinna separation near vocal onset time. Red dots indicate time of sonar call emission; green dots indicate time of sonar echo arrival. (B) Top, normalized individual pulse aligned (top panel) and echo aligned (bottom panel) inter-pinna separations. Red indicates the largest separation between the pinna, blue the smallest, in a 60 ms window aligned to the time of pulse emission or echo arrival. Bottom, mean +/- s.e.m. of normalized local peaks in inter-pinna distance aligned to either the sonar pulse onset (red) or echo arrival time (green). (C) All shifts in the peak of inter-pinna separation aligned to pulse onset (red) and echo arrival (green). Asterisks indicate the average shift in the peak displacement of inter-pinna separation around the time of pulse emission/echo arrival. The two distributions are significantly different (permutation test, p < 0.001). (D) Velocity of pinna motion during pulse production (red) and echo arrival (green). Pinna velocity is significantly higher during echo reception than during pulse production (permutation test, p < 0.0001). Data for this figure can be found at http://dx.doi.org/10.7281/T1W66HPZ.