Optokinetic Responses, Visual Adaptation and Multisensory Control of Eye Movements in the Spiny Lobster, Palinurus Vulgaris

Compensatory eye movements and equilibrium reactions in decapod crustaceans are performed with reference to three spatial parameters: gravity, substrate orientation and visual field (Schöne, 1980; Neil, 1982). These responses contribute in different ways and to different extents to the overall reaction in the species so far investigated (Procambarus clarkii:Wiersma & Oberjat, 1968; Hisada, Sugawara & Higuch 1969; Fay, 1973; Olivo & Jazak, 1980; Panulirus interruptus: York, Yanagisawa & Wiersma, 1972; Carcinus maenas:Horridge, 1966; Wiersma & Fiore, 1971), but little is known about the interrelationships of the responses. This paper forms one of a series to investigate the mechanism of multisensory interaction in the spiny lobster, Palinurus vulgaris.

The contribution of mechanosensory inputs from statocysts and from leg proprioceptors to compensatory eye movements in Palinurus has been studied in some detail, with particular reference to responses in the roll plane (Schöne, Neil, Stein & Carlstead, 1976; Schöne, Neil & Scapini, 1978; Schöne & Neil, 1977; Neil & Schöne, 1979; Birkner, 1978; Scapini, Neil & Schöne, 1978; Neil, Schöne & Scapini, 1979). Studies in other decapods, particularly on the optokinetic system, suggest a major role for vision in the control of eyestalk position (Horridge & Sandeman, 1964; Wiersma & Oberjat, 1968; York et al. 1972). The response characteristics of the visual system in Palinurus are described in this paper, and we also consider how visual inputs interact with proprioceptive inputs from the legs. An analysis of the combined control system, which involves in addition the response to gravity, is presented elsewhere (Schöne, Neil, Scapini & Dreissman, 1983).

Spiny lobsters {Palinurus vulgaris Latreille = Palinurus elephas Fabricius) from the Mediterranean were kept under semi-natural conditions in tanks of recirculated sea water. General procedures for preparing and mounting the animals are described elsewhere (Schöne et al. 1976). Optokinetic stimulation was provided by a halfcylinder of translucent Plexiglass, radius 14·5 cm and length 21 cm, upon which 11 black stripes 2 cm wide were mounted 2 cm apart. This drum was mounted over the animal with the axis of the frame coincident with that of the platform, and the stripes parallel to the animal’s longitudinal axis (Fig. 1). With the lobster in position, the angle subtended at the eye by a stripe period was approximately 30°. Additional panels carrying a pattern of radiating stripes were inserted in front of the animal. The drum was illuminated either by two 60 W (J s⁻¹) lamps at 5-10cm distance (Fig. 1), or by one such light source 20 cm above. The irradiance measured inside the drum at the level of the eye was 1200 lx with the former arrangement and 50 lx with the latter. For experiments with varying irradiance, neutral density filters (Kodak Wratten) were placed in the light path from the single overhead source. In some experiments, a translucent drum without stripes was used, in conjunction with translucent panels in front of the animal. Both drums produced a bright upper field of illumination. The platform itself, being made of translucent material, significantly reduced the visibility of stationary edges in the animal’s ventral field of view.

Movements of the drum assembly and the platform were produced by eccentric drives from d.c. motors which could be driven over a wide range of speeds by the appropriate choice of gearbox and voltage. These stimuli were monitored using precision potentiometers. Movements of the left eye were measured using a miniature angle transducing device (Marrelli & Hsiao, 1976). The transducer wire could be mounted on a narrow plastic ring which fitted firmly around the base of the outer segment without restricting vision or movement. Alternatively, caps of silicone rubber could be used either to restrict pattern vision (translucent) or to exclude all light (black). Leg forces were recorded by connecting the animal’s recording frame to a strain gauge device (Grass FT.03C). All stimulus and response monitors were fed to a paper chart recorder for later analysis. In experiments with sinusoidal drum movement, measurements of response amplitude and velocity were obtained from the original records by averaging values from 6-10 cycles. Waveforms were digitized manually for Fourier analysis by computer (Mills, 1981).

When the striped half-cylinder was rotated around the long axis of the body, both eyes followed the movement of the stripes without nystagmus until they came to rest with maintained displacements. If the drum was oscillated through an arc of less than 20°, eye movements were also continuous, and the amplitude of their excursions provided a suitable measure of tracking performance. This may not represent the maximum performance of the optokinetic system, because of the limited extent of the striped pattern in the visual field.

The strength of coupling between the two eyes in optokinetic tracking was measured in this way. When the measured eye alone viewed the striped drum the response was reduced by 30—40% compared with its value with both eyes viewing (Fig. 2A, B). When the other eye provided the only visual input, the movement of the capped and measured eye was reduced to approximately 50% (Fig. 2C). These results indicate that a contralateral optokinetic drive exists, but it is weaker than the ipsilateral effect.

Responses of the eyestalk to trapezoid movements of the striped drum were investigated with both eyes seeing. For stimulus arcs up to 20°, eye movement closely followed drum movement with an amplitude gain approaching 1 (Fig. 3A). The eye then maintained its positional relationship to the stripes as long as the drum remained displaced. This property distinguishes these visually-evoked responses from those elicited by leg proprioceptor stimulation, which display a distinct phaso-tonic form (Fig. 3B, C; cf. Neil & Schöne, 1979).

Using sinusoidal drum movements through a 20° arc, the optokinetic response in the roll plane was tested over four decades of frequency. Above 0·05 Hz the eye response was both small and strongly habituating, declining sharply after several Cycles (Fig. 4A, B), until eye movement in many cases ceased completely (see Fig. 11B2). This habituation was not seen at lower frequencies, where the eye continued to follow the stripes with a constant tracking performance for as long as the stimulus was presented (Fig. 4C, D). The optokinetic system also displayed less predictable changes in responsiveness. Novel mechanical stimuli often induced transient increases in tracking gain (Fig. 4E) which were accompanied by active movements of the legs. Similar effects occurred apparently spontaneously during periods of increased arousal (Fig. 4F).

Amplitude gain increased steadily with decreasing frequency until it reached a value close to unity at 0·01 Hz (Fig. 5). This level of performance was maintained down to the lowest frequency tested, 0·0002 Hz, at which the maximum velocity of stripe movement was 0·012°s⁻¹.

The amplitude of the optokinetic response remained almost constant as the light intensity was reduced (with neutral density filters) by three log units from 50 lx, until it failed completely at levels below 0·05 lx (Fig. 6). However, if the animal was kept at this light level while the drum continued to oscillate, the response showed a consistent recovery over several minutes to levels which equalled, or sometimes exceeded, those found at higher light intensities (Fig. 6A-C). This adaptation process to low luminance could also be demonstrated in the experimental condition in which stationary stripes antagonized the eye response, mediated by leg proprioceptors, to a moving platform. Eye movement was completely suppressed in the presence of a stationary visual field. This effect persisted with decreasing illuminance, until at 0·05 lx the eye response to leg input appeared with an amplitude approaching that found in the dark (Fig. 6E). After several minutes at this light level, however, eye movement was again abolished

To show that irradiance was a stimulus separable from the optokinetic stimulus of the high-contrast stripes, the striped drum was replaced by a translucent drum which transmitted light, but not pattern, to the upper half of the visual field. Translucent panels in front, and a translucent platform below, minimized the effect of stationary edges in the animal’s field of view. Under these conditions movements of the drum itself elicit no optokinetic response (Schöne et al. 1983). The effect of the unpattemed visual stimulus on the eye response was tested in combination experiments with the platform stimulus. The amplitude of eye movement was much reduced, compared with the response in darkness (Fig. 7C, F).

To test for the possibility that stationary edges in the visual field were responsible for this effect, the following experiment was performed. The eyes were covered with translucent caps of silicone rubber, which transmitted light but not pattern. Their effectiveness was checked by measuring the eye response to the moving striped drum. With the caps on, response amplitude was reduced by more than 80%, compared with that obtained with both eyes seeing (Fig. 7A, B). With the combination of translucent drum and platform, however, the inhibitory effect of the visual stimulus was still significant (Fig. 7D, F), despite a reduced amount of light reaching the eyes through the translucent caps. Furthermore, this inhibition was reduced when the drum was Removed, revealing the point light source above, but reducing the area of brightness in the upper visual field (Fig. 7E). This procedure would also, potentially, introduce more stationary edges into the animal’s field of view. We therefore conclude that under the translucent drum the distribution of light over the eye is responsible for the observed effects, that these are proportional to the number of ommatadia stimulated, and that optokinetic responses to any remaining edges in the field of view play no significant part. The involvement of this second visual factor may underlie other observed effects: thus when the response to rapidly oscillating stripes has faded, the eye response to simultaneously imposed platform movement under the legs remains suppressed (Fig. 11B).

The two visual inputs appeared to interact destructively. By oscillating the striped drum at a high frequency which sustained the optokinetic response, in combination with a platform tilt at lower frequency, it was found that the inhibition of the proprioceptive eye reflex was less than that produced by a translucent drum transmitting the same illuminance (Fig. 8).

The effect of irradiance level on the strength of the second visual factor was measured in experiments under the translucent drum. At the higher light intensities (from 100 lx to 0·5 lx) the eye response to platform oscillation was reduced, although not completely abolished (Fig. 9). At 0·1 lx, however, there was a difference between the response on immediate presentation, when it appeared in a virtually complete form (Fig. 9, solid line), and after a period of several minutes, when it was again markedly suppressed (Fig. 9, dashed line).

When movements of the striped drum and platform were presented in a phase relationship promoting synergistic reflex effects, a major influence on the eye response was the gain-limiting property of the optokinetic system. This was expressed in the reduced gain of the reflex, compared to that obtained in the dark, in the low frequency range (Fig. 10A, C).

When the two stimuli were presented in an antiphase combination, the amplitude of the eye response showed a general reduction as frequency was increased, compared to that elicited by platform movement in the dark (Fig. 10C). The nature of these responses is made more apparent when their phase positions relative to the stimuli are taken into account. Below 0 06 Hz the response had a phase close to that of the drum, while in the high frequency range it was in phase with the platform, but with a much reduced amplitude (Fig. 10D). The dominance of the optokinetic system at low frequency is therefore confirmed, since it now induces a large countermovement of the eye against the direction dictated by the leg proprioceptor input.

With different frequencies and different relative strengths for the separate stimuli, interaction at all relative phase positions could be studied. We used an intermediate frequency of platform oscillation in combination with drum movements at both higher and lower frequencies. When the frequency difference was large, clear-cut effects were seen: the eye followed the drum if its frequency was low (Fig. 11A), but failed to follow either drum or platform if drum frequency was high (Fig. 11B). When the frequency difference between the drum and platform was small (Figs 8, 12A), the compound waveform of the eye response curve was difficult to interpret. However, Fast Fourier Transform (FFT) analysis of these waveforms revealed that they were composed exclusively of periodic components corresponding to the two input frequencies, and no others (Fig. 12B). An estimate of the relative weighting of the two inputs was provided by the area of each peak in the FFT plot, and was separately obtained by generating a matched waveform from the two input monitoring signals (Fig. 12C). Overall, these results indicate that at the chosen platform frequency, 0·02 Hz, weighting changes steadily with drum frequency from completely to the drum at 0·01 Hz to completely to the platform at 0·25 Hz.

The major features of the optokinetic reflex system of the spiny lobster in the vertical plane are consistent with those found in the other decapod crustaceans so far investigated (reviewed by Neil, 1982). Coupling between the two eyes is stronger than in crayfish (Hisada et al. 1969) but comparable to that for horizontal movements in the crab Carcinus (Barnes & Horridge, 1969; Sandeman, Erber & Kien, 1975). The strength of this bilateral coupling may vary with stimulus conditions, however, for York et al. (1972) report that the eyes of the spiny lobster Panulirus may move with a high degree of independence in a less uniform visual environment. The closed-loop frequency response (Fig. 5) typifies a system which normally operates with negative feedback control (Horridge, 1966). The lower threshold for movement detection by the visual system was not reached in our experiments, but must lie below 0·01 °s⁻¹ (Fig. 5). The performance of the optokinetic tracking system in Palinurus may therefore approach that found in Carcinus (Horridge, 1966).

The identification of a second visual response in Palinurus to the illuminance per se (Schôneei al. 1983) is confirmed here (Fig. 10). In addition, we have demonstrated that there is an interaction between this visual system and the optokinetic system, which under some circumstances can be antagonistic. The separate existence in crustaceans of visual systems attuned to overall light distribution and wide-field movement has been demonstrated both by behavioural measures (Wehner, 1979; Schône, 1980) and in certain species by electrophysiological techniques (Panulirus-. York & Wiersma, 1975,Procambarus: Wiersma & Oberjat, 1968; Higuchi, 1973; Higuchi & Hisada, 1973). In the crayfish, light entering the eye causes reflex excitation of muscles which lift the eyestalk (Higuchi, 1973) through a motor pathway which may be separate from that for other eye reflexes (Mellon & Lorton, 1977).

In considering the apparent dominance of vision over the proprioceptive inputs (Figs 6, 8, 10, 11) it must be recognized that the irradiance gradient contributes a strong stabilizing influence on eye position (Figs 7, 8, 9). This is probably not through a general inhibition, but is more likely to be due to its action as a stable directional cue. Thus, under appropriate conditions this visual stimulus can enhance the eye response to mechanosensory stimulation, rather than suppress it (Schône et al. 1983, Fig. 8).

In order to reveal the nature of the interaction between the optokinetic component of the visual system and leg proprioceptive inputs it is necessary either to create a uniform field of illumination around the optokinetic stimulus (Hisada et al. 1969; Neil, 1975), or, as here, to produce movements of both the optokinetic and platform stimuli relative to a stationary light gradient (Figs 10, 11, 12). This experimental arrangement is, of course, completely artificial, the natural relationship consisting of in phase stimulation of both visual subsystems, and of leg proprioceptor and statocyst inputs. However it demonstrates essential features of the combined control system, the most obvious of which is an extension of the linear range of the eye response through the individual contributions of the optokinetic and leg proprioceptive systems tuned to low-frequency and high-frequency ranges respectively. This is a common feature of related interactions in many different animals (crab: Sandeman, 1977; crayfish: Olivo & Mellon, 1980; rabbit: Collewijn, 1981; monkey: Waespe, Henn & Isoviita, 1980).

It has often been suggested that multisensory interaction in the control of crustacean eye movements is based upon a simple algebraic summation of the inputs from the different sensory systems involved (Fay, 1973; Mellon & Lorton, 1977; Olivo & Jazak, 1980). However, our results obtained over a wide frequency range clearly show that this is an inadequate explanation, which ignores the differences between the operation of the closed-loop optokinetic system and the open-loop proprioceptive system. Responses to low-frequency wide-field optokinetic stimuli are dominant, in the face of both antiphase platform movement and a conflicting stable light gradient (Figs 10, 11, 12). A similar effect has recently been described by Varju & Sandeman (1982) in the crab Leptograpsus.

Significantly, we find that although clearly distinct, the separate visual subsystems in Palinurus are affected in similar ways by reduced illuminance. They both demonstrate a threshold sensitivity and subsequent adaptation which have closely similar characteristics (Figs 6, 9), and we interpret this common effect as a strong indication that the adaptation process takes place at a site peripheral to the separate neural integration networks. The time course of visual adaptation is consistent with that reported for screening pigment migration in a number of different crustacean species (de Bruin & Crisp, 1957; Olivo & Larsen, 1978). Such pigment movements would have the effect of increasing the sensitivity of the eye, since the exposed tapetum would reflect light back onto the rhabdoms, increasing quantum capture. The necessary concomitant of this, a loss of acuity, may well not have limited the responses under our experimental conditions with wide-field stripes, although this has not been directly tested. The behaviour of screening pigments in crustacean eyes is complex (Autrum, 1979), being dominated in many cases by intrinsic circadian rhythms (Leggett & Stavenga, 1981). In the spiny lobster, Panulirus interruptus, it is not possible to cause light adaptation of the night eye, or dark adaptation of the day eye (Meyer-Rochow, 1975). Our results suggest that forPalinurus this question must remain open until a more thorough analysis of screening pigment movement in relation to both the day/night cycle, and to the behavioural responses at different light intensities has been performed.

DMN was supported by grants from the British Council and the Max-Planck Gesellschaft. JAM received a travel award from the British S.E.R.C.