First published online September 15, 2004
Journal of Experimental Biology 207, 3693-3706 (2004)
Published by The Company of Biologists 2004
doi: 10.1242/jeb.01201
Spatio-temporal patterns of antennal movements in the searching cockroach
Jiro Okada* and
Yoshihiro Toh
Department of Biology, Graduate School of Sciences, Kyushu
University, Fukuoka 812-8581, Japan
View larger version (31K):
[in a new window]
|
Fig. 1. Experimental apparatus for recording antennal movements. An animal was
tethered on a free-moving Styrofoam ball to allow its unrestrained locomotion
(Okada and Toh, 2000 ). Three
video cameras were set just in front (1), top (2) and side (3) of the animal.
Video images were processed by a 4-split multi-viewer, videotaped, and
analyzed using a PC. Photograph on the right is an example of synthesized
video frame after the processing, showing frontal (1), top (2) and side (3)
views.
|
|
View larger version (31K):
[in a new window]
|
Fig. 2. Method for measuring positions of the antenna, head and foreleg. (A) Top
view (i) depicts horizontal angular positions of the right antenna and the
foreleg, and yaw position of the head. Lateral view (ii) is a diagram for the
antennal vertical and head pitch positions. (B) Relationships of parameters
for the antennal (i) and head (ii) positions. Note that three-dimensional
coordinates for antennal (i) and head (ii) positions are different from each
other. The rostrocaudal y axis for the former is identical with the
head axis, and the latter (y' axis) with the body axis. See
Materials and methods for details of these parameters. O, origin for both
antennal and head positions; O', origin for foreleg position; P,
plotting point registering antennal position; P', plotting point
registering foreleg position.
|
|
View larger version (52K):
[in a new window]
|
Fig. 3. Antennal and head movements during pausing (A) and walking (B) in a
searching cockroach. These two examples were obtained from the same animal. In
pausing, large deflections were frequently observed in traces of the antennal
horizontal components and the yaw and pitch components of the head (see shaded
area). Antennal vertical arcs were lowered entirely in the walking state.
Deflections in the head roll were relatively small in both behavioral states.
Vertical calibration bars show angular positions for antennae and head. L,
left; R, right.
|
|
View larger version (12K):
[in a new window]
|
Fig. 4. Angular ranges and central positions for the horizontal and vertical
components in both antennae and head. Bars and symbols indicate the mean
angular ranges and central positions, respectively (N=10). Working
ranges for antennal horizontal and head yaw components decreased,
respectively, by about 25–30° and 15° during walking, but were
not largely changed for antennal vertical and head pitch component s. Vertical
central positions for both antennae were lowered by 15–25° during
walking. R, right; L, left.
|
|
View larger version (38K):
[in a new window]
|
Fig. 5. Power-spectra for antennal and leg movements in a searching cockroach. (A)
Spectra for horizontal (H) and vertical (V) deflections in pausing. Their
major peaks were distributed mainly at <3 Hz (H) and <4 Hz (V),
respectively. (B) Power-spectra for antennal (H and V) and foreleg (L)
movements in walking. Lower frequency components at <2 Hz were largely
suppressed for horizontal deflections (H). Higher frequency components around
4 Hz became remarkable for vertical deflections (V) instead. Note that no
conspicuous peak is observable in the H and V spectra at the frequency
corresponding to major peaks in the L spectrum (2–3.5 Hz). Inset in each
spectrum is part of the time course before FFT. These spectra were obtained
from data for a 31 s period each.
|
|
View larger version (59K):
[in a new window]
|
Fig. 6. Trajectories of both antennae and head. Data in cases 1–3 were
obtained from three different animals. The sampling period for each graph was
15 s. The right and left antennal trajectories are indicated, respectively, by
red and blue traces, and the head by green traces. (1) The random pattern was
most commonly observed in both pausing (A) and walking (B). (2) The loop-like
pattern was apparent only during walking. Insets in the right panel show
typical examples of the three basic patterns forming the loop-like trajectory.
These were derived from a trajectory of the left antenna. L, loop; A, arch; V,
vertical line. (3) The `consistent' loop-like pattern independent of the
searching mode.
|
|
View larger version (30K):
[in a new window]
|
Fig. 7. Kinematics of the scape–pedicel (S–P) joint in two behavioral
states. By immobilizing the head–scape (H–S) joints (see inset at
bottom), movement of the S–P joint could be analyzed independently. (A)
Trajectory patterns formed simple vertical lines in both pausing and walking.
Vertical positions were lowered entirely by 20–30° during pausing.
(B) Power-spectra for vertical components (V) in the two states. These spectra
were obtained from the time-series whose trajectories are shown in A. Discrete
peaks appear at 1.5–3.5 Hz during pausing, but distributed over a wider
range up to 4.5 Hz in walking. (C) Spatial profiles of the S–P joints in
pausing and walking. Bars and symbols indicate averages of the angular ranges
and the central positions, respectively (N=4). Both antennae
significantly lowered by about 25° during walking. Inset shows a rotation
axis of the free S–P joint in this condition. Sampling periods were 30
s(trajectories) and 34 s (power-spectra). H, head; S, scape; P, pedicel; F,
flagellum.
|
|
View larger version (29K):
[in a new window]
|
Fig. 8. Kinematics of the head–scape (H–S) joint in two behavioral
states. The S–P joint is immobilized in this condition (see inset at
bottom). (A) Trajectories showed two-dimensional expansion in both states.
During walking, both antennae concentrated at lower positions. (B)
Power-spectra of horizontal (H) and vertical (V) components in the two states.
Peaks at <2 Hz decreased in walking for both components. These spectra were
obtained from the time-series whose trajectories are shown in A. (C) Mean
angular ranges (bars) and central positions (symbols) of both antennae during
pausing and walking (N=4). Vertical central positions for both
antennae were significantly lowered as observed in the S–P joint free
condition (see Fig. 7).
Horizontal working ranges were slightly but significantly narrowed during
walking. Inset indicates two rotation axes of the free H–S joint.
Sampling periods and abbreviations, see
Fig. 7.
|
|
View larger version (37K):
[in a new window]
|
Fig. 9. Cross-correlations between right (R) and left (L) antennal movements during
pausing. (A) Trajectories of both antennae show typical random patterns
(sampling period 18 s). (B) Cross-correlation for horizontal components. The
waves indicate the time courses of the right and left horizontal deflections
for a period of 42 s. Note that polarities for the horizontal position are
opposite between the right and left antennae (refer to calibration bars). RH,
horizontal deflection of the right antenna; LH, horizontal deflection of the
left antenna. Correlation coefficients (r) in the correlation map
were calculated by reference to the left antenna. Significant positive
correlations (r>0.25) are indicated by yellow to red, and the
negative ones (r<–0.25) by blue to black. Significant
negative correlations were observed as broad blue or black spots. Backgrounds
filled by gray color represent the areas where r values were not
statistically significant (|r|<0.25). (C)
Cross-correlation for vertical components. Significant correlations were few
on the map. RV, vertical deflection of the right antenna; LV, vertical
deflection of the left antenna.
|
|
View larger version (40K):
[in a new window]
|
Fig. 10. Cross-correlations between right and left antennal movements in walking.
The data were from the same animal as in
Fig. 9. (A) Trajectories of
both antennae were still random, and their scanning areas were smaller than
those seen during pausing (cf. Fig.
9A). (B) Cross-correlation for horizontal components. Significant
negative correlations existed close to the zero time lag as relatively compact
spreads along the ordinate (compare with
Fig. 9B). Inset shows expansion
of the boxed area (1). Though individual small waves did not exactly
synchronize with each other, coordination is recognizable in both larger and
slower components. (C) Cross-correlation for vertical components. Significant
positive correlation spots were continuously observed over the map at zero
time lag as sharp peaks. Expansion of the boxed area (2) shows synchronization
of the pair of waves at 2–3 Hz. Note the time scale difference from
Inset 1.
|
|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati 
Twitter What's this?
© The Company of Biologists Ltd 2004
