The role of ocelli in cockroach optomotor performance

Abstract

Insect ocelli are relatively simple eyes that have been assigned various functions not related to pictorial vision. In some species they function as sensors of ambient light intensity, from which information is relayed to various parts of the nervous system, e.g., for the control of circadian rhythms. In this work we have investigated the possibility that the ocellar light stimulation changes the properties of the optomotor performance of the cockroach Periplaneta americana. We used a virtual reality environment where a panoramic moving image is presented to the cockroach while its movements are recorded with a trackball. Previously we have shown that the optomotor reaction of the cockroach persists down to the intensity of moonless night sky, equivalent to less than 0.1 photons/s being absorbed by each compound eye photoreceptor. By occluding the compound eyes, the ocelli, or both, we show that the ocellar stimulation can change the intensity dependence of the optomotor reaction, indicating involvement of the ocellar visual system in the information processing of movement. We also measured the cuticular transmission, which, although relatively large, is unlikely to contribute profoundly to ocellar function, but may be significant in determining the mean activity level of completely blinded cockroaches.

Keywords: Behaviour; Cuticular transmission; Ocelli; Optomotor reaction; Virtual reality.

Fig. 1

Experimental setup for cuticle measurements (a) and its schematic principle (b). MI measuring illuminator, FW filter wheel, SH shutter, GF grey filter (optional), XY-VA X-Y variable circular zero aperture, IR-VC infrared sensitised video camera (connected to a monitor; not shown), BS beam splitter, BI background illuminator, FP filter pack, CON condenser, XYS xy-stage, S specimen, OBJ objective, C-OF collimator-optic fibre system, M microscope, SP spectrometer, PC computer with the SpectraSuite software

Fig. 2

Optomotor response strengths ± SD to different temporal frequencies of the stimulus with 60° angular period. Data from the “unmanipulated” group are shown in the left column and data from the “ocelli covered” group on the right. Light intensities are shown in the upper right corner of each row. Solid bars represent the strength of the response during the stationary controls and hatched bars during rotating stimuli. Response strengths range between 1 for the strongest positive and − 1 for the strongest negative (anti-directional) response. The expected control level is zero. One, two, and three asterisks indicate significant differences between the control and rotation distributions at confidence levels of, respectively, 0.05, 0.01, and 0.001 (paired sample Wilcoxon signed-rank test). a–d Attenuation of the response strengths, and the narrowing of the frequency band that is able to elicit the optomotor response, with falling light levels. e–h When the ocelli are covered, the response strength attenuates, and the frequency band narrows, at higher light intensity levels than in unmanipulated cockroaches. Sample sizes were a N = 20 animals, n = 40 measurements; b N = 24, n = 78; c, d N = 23, n = 66 and e–h N = 20, n = 40. Data in a–d are from the data set in Honkanen et al. (2014)

Fig. 3

Response strengths of cockroaches whose compound eyes or all eyes have been covered at 500 lx. See Fig. 2 for the symbol keys. a No significant differences between control and rotation values were found. b Significant difference (p = 0.00256, paired sample Wilcoxon signed-rank test) between control and rotation values was found at 2.4 Hz, where the control value happens to be on the positive and the rotation value on the negative side. The responses in a and b are almost identical. Sample sizes were a N = 20 animals, n = 40 measurements; b N = 19, n = 38

Fig. 4

Total distances and average velocities walked by the cockroach during the 30-s stimulus rotation. a Boxplots and data point distributions of the total angular distances covered by the cockroaches during stimulus rotation. “Unmanipulated” is the condition where none of the eyes are covered. “Ocelli”, “compound eyes”, and “all eyes” denote which eyes are covered in the three treatment groups. The box plot shows the first and third quartiles and the median of the data; the square inside the box plot marks the mean; the “whiskers” of the plot are the 95th and 5th percentile; the maximum and minimum values of the data are marked with “X”. Pairwise differences between treatments were tested with two-sample Kolmogorov–Smirnov test. All other groups differ from each other significantly (p < 0.001) except “compound eyes” and “all eyes” (N.S.; p = 0.55). Data from experiments with 0.4–10 Hz stimuli are combined to produce the plots. Sample sizes are n = 462 (Unmanipulated); n = 200 (Ocelli and compound eyes); n = 190 (All eyes). b Average angular walking velocities ± SD of cockroaches with different eye manipulations (insert) across different light intensities. Each data set contains data from all nine temporal frequencies of the stimulus. The mean velocities of all the groups rise with rising light intensity. In pairwise comparisons, all data points of “unmanipulated” group differ from the other three groups (two-tailed two-sample Kolmogorov–Smirnov test p < 0.001) and all data points of “ocelli covered” group differ from the “compound eyes covered” group (p < 0.01). The averages of “compound eyes covered” and “all eyes covered” groups are not significantly different from each other at 500 lx (p = 0.09) and are closely identical but significantly different (p < 0.05) at all other light intensities. Sample sizes per data point n = 216–920 (unmanipulated); n = 360 (Ocelli and compound eyes covered), n = 90–342 (all eyes covered)

Fig. 5

Absorbances of cuticle and antennal structures in the cockroach head capsule near the visual sense organs. a Schematic view of the cockroach head capsule. b Sample pieces used for the spectral absorbance measurements were chosen: 1: between the ocelli, 2: underside of the ocellus and antennal joint, 3: between the ocellus and antenna, and 4: the antenna base. Scale bar, 500 µm. c Averaged absorbance spectra ± SD from the positions numbered in b of the cockroach head. Sample sizes were: N = 3 animals, n = 5 measurements (sample 1); N = 3, n = 10 (2); N = 3, n = 9 (3); N = 3, n = 11 (4)