Activities of antennal and ocellar interneurones in the protocerebrum of the honey-bee

The role of the insect dorsal ocellus is not so well understood as the role of the compound eye. Behavioural experiments have shown that the ocellus is a stimulatory orean: locomotor activities, orientatory movements, and flight speeds are affected by illumination of the ocelli (see reviews of Tateda, 1968; Goodman, 1970). Achievement of the conditioned reflex of the honey-bee is also affected by the ocellus (Takeda, 1961). Electrophysiologically, Goodman (1970) showed that neurones (presumably the descending neurones) in the ventral nerve cord of the locust were excited when the ocelli were stimulated, indicating that the activities of the motoneurones were regulated by light falling on the ocellus, as is seen in the central ganglia of a marine mollusc (Gotow, Tateda & Kuwabara, 1973). Mimura et al. (1969, 1970) showed that unit responses to antennal stimuli in the brain of the fly are changed when illumination of the ocelli was turned off, and they concluded that the ocellus regulates brain activity.

Although the insect central nervous system has been studied extensively (see Bullock & Horridge, 1965; Huber, 1965), studies on converging and integrating regions of sensory information in the brain are few. Olfactory inputs entering the deutocerebrum are conveyed to the calyx of the mushroom body by passing through the antenno-cerebral tracts, and then reach the protocerebral lobe in the honey-bee (Suzuki & Tateda, 1974). Histological investigations (see Bullock & Horridge, 1965), and behavioural experiments involving brain lesions (Vowles, 1964), have revealed another tract which runs between the deutocerebrum and the a-lobe in hymen-opterans. Convergence of olfactory inputs occurs in the deutocerebrum (Boeckh, 1974), whereas convergence of inputs from the antennae is in the protocerebrum of the honey-bee (Suzuki, 1975 a). Bimodal neurones receiving outputs from the antennae and the ocelli are mainly found in the protocerebrum of the fly (Mimura et al. 1969). These findings suggest the localization of an integration centre in the protocerebrum.

The present study is an investigation of ocellar effects upon responses to antennal stimuli, and the distribution of the neurones involved, in the protocerebrum of the honey-bee.

Workers of the honey-bee (Apis mellifera L.) were used. Experimental methods were the same as those of Suzuki and Tateda (1974), and Suzuki (1975 a), with the exception of the method for stimulating the ocelli with light. The ocelli were illuminated with light focused upon them with the object lens of a microscope. The light was fed, by means of a light guide, from a microscope lamp with an adiabatic filter, so that each of the three ocelli were illuminated from the same angle. The light spot illuminating the ocelli was about 1·5 mm in diameter, which was just enough to illuminate three ocelli. The intensity of the light spot was constant, at about 50 lux, throughout the experiment. Experiments were performed in the dark at room temperature.

Electrical recordings of units in the protocerebrum of the honey-bee were made extracellularly. The activity of the units was recorded during illumination of the ocelli, during exposure of the antennae to odour (methyl-ethyl ketone), and during mechanical stimulation of the antennae. Duration of odorous stimulation was 500−1000 ms. Mechanical stimulation of the antennae was a bending, caused as an artifact of olfactory stimulation. Characterization of a positive response to antennal stimulation was as described previously (Suzuki & Tateda, 1974). Usually the right antenna of the bee was stimulated and recordings were carried out in the right hemisphere of the brain. The recording site in the brain was localized as described previously (Suzuki, 1975 a).

The units reported here were recorded in the ipsilateral side of the protocerebrum of the worker honey-bee during olfactory or mechanical stimulation (duration about 500 ms) of an antenna and illumination of the ocelli. During stimulation some of the units showed an increase in impulse frequency (i.e. excitation), while others showed a decrease (i.e. inhibition). In the excited units, some showed background discharge (Suzuki & Tateda, 1974), but others did not.

Among the 118 units studied in these experiments, there were 32 which responded only to antennal stimulation (antennal units) and 27 which responded only to ocellar stimulation (ocellar units). The remaining 59 units responded both to antennal and to ocellar stimulation (bimodal units).

Units which responded to antennal stimulation were 91 in number (including antennal and bimodal units). Of these, 69 responded to olfactory stimuli (olfactory units) and 22 responded to mechanical stimuli (mechanical units). As reported previously (Suzuki, 1975 a), unit activities were checked as to whether they were responsive to odorous or to mechanical stimulation of the antenna; obvious differences in magnitude of responses to odorous and to mechanical stimuli were not observed. One of the units, which showed the largest difference in magnitude between responses to olfactory and to mechanical stimuli of the antenna, produced (mean ± s.D.) 11·0±2,35 (n = 10) and 9·5 ± 1·6 (n = 8) spikes in response to olfactory and mechanical stimulation, respectively. These two responses showed little statistical difference (P> 0·1). We were unable to divide mechanical units into pure mechanical and olfactory-mechanical units. Therefore the units which responded either to odorous and to mechanical stimuli were referred to as mechanical units.

There were on-responses, on-off-responses, and off-responses. In the on-responses, the pattern of impulse frequency could be classed as phasic, phasic-tonic, and tonic. The number of the units which showed these pattern of responses in constant illumination of the ocelli are shown in Fig. 1. Characteristics of the olfactory interneurones in the protocerebrum were almost the same as those reported previously (Suzuki & Tateda, 1974), that is, the dominant pattern was phasic-tonic during excitation, and tonic during inhibition; and rebound and long-lasting responses were observed in some of the units. Among the mechanical units, the phasic pattern was predominant during both excitation and inhibition. Off-responses were given only to mechanical stimulation.

Since the light was focused on the ocelli and the compound eyes were covered with black lacquer as reported previously (Suzuki, 1975 a), the responses to light were not due to stimulation of the compound eyes. The responses evoked by turning on or off the illumination are caused by ocellar stimulation.

There were 86 units which responded to illumination of the ocelli (including ocellar and bimodal units). They showed excitation or inhibition at the start and/or end of illumination of the ocelli. Slow potentials often accompanied these responses, and antennal responses, but are not considered here. The responses, determined from changes in impulse frequency, were phasic and were classified into six types; ⁠, ⁠, ⁠, and ⁠. In this notation L_on and L_off mean light on and off, respectively, and + and - mean excited and inhibited, respectively. For example, the type of response means that there was excitation at the start of the illumination, but there was no response at the end of illumination. Units exhibiting response only at the end of illumination were not found. Most of the units showed (30 units) or (42 units) type of responses. Few inhibition responses were found. There were 6 units which showed different types of response to light on and light off (that is, and type). No difference in the response to ocellar stimulation was observed between ocellar and bimodal units.

Bimodal units were 59 in number. The relationship between the type of response given to antennal and to ocellar stimulation is shown in Fig. 2. Most units showed excitation in response both to antennal and to ocellar stimulation. An example of such a unit is presented in Fig. 3. This unit showed a phasic response to odour (A, C, and E) but no response to mechanical stimulation of the antenna (F). A burst of impulses was evoked when light illuminating the ocelli was turned on (D) or turned off (B). More impulses appeared during ocellar stimulation than during odour stimulation; it was found to be characteristic of most units which showed an excitation to both ocellar and antennal stimuli that there were more impulses during the response to the former than to the latter.

Most of the units which showed inhibition in response to ocellar stimulation showed excitation in response to antennal stimulation. As shown in Fig. 2 there were 7 units which showed or responses to light stimuli, among which 5 units showed excitation in response to olfactory stimuli. The example presented in Fig. 4 showed excitation in responses to odour in the illuminated (A and E) as well as in the non-illuminated states of the ocelli (C). Mechanical stimulation of the antenna did not evoke responses (F). The background discharge frequency of about 6 impulses/s was suppressed when the ocelli were illuminated (D).

The units which showed inhibition in response to antennal stimuli showed dominantly excitation in response to ocellar stimulation (Fig. 2). The discharges of the unit presented in Fig. 5 were suppressed tonically by odour in the illuminated state (A and E). In the non-illuminated state, however, inhibition in response produced by odour was small (C), that is, the response to odour decreased in magnitude. Difference in magnitude of the responses in the illuminated and non-illuminated states will be described below (§ B). When illumination of the ocelli was begun, there was an increase in frequency of impulses, which was accompanied by the appearance of a slow potential. The slow potential was not accompanied by a change in amplitude of the impulses, indicating that these impulses were recorded from a single neurone. There was a slight increase in impulse frequency at the cessation of ocellar illumination (B).

The location of the units was determined for about half of those investigated (Table 1). All of the units were found in the neuropiles of the protocerebrum. Antennal and ocellar units were widely distributed. Some of the bimodal units were found within the mushroom body, but most of the units were located mainly around the a-lobe of the mushroom body, about 100 microns below the front of the protocerebrum.

The background discharge frequency and response pattern of some of the units which responded to antennal stimuli were found to change as illumination of the ocelli was turned on or off (as shown in the fly by Mimura et al. 1969).

Most of the units showed background discharge. The discharge varied in frequency from unit to unit, but did not exceed 40 impulses/s. Mean impulse frequency was Measured at least 5 min after the beginning or end of illumination, with an integration period of several seconds. The discharge frequency during illumination (L) was divided by the frequency during darkness (D) (except when D was greater than L, when D/L was calculated), and the result was plotted against L (Fig. 6). The difference between discharge frequency in the illuminated and non-illuminated states was larger in the units which showed a lower frequency of background discharge. Most units had a higher frequency in the non-illuminated state than in the illuminated state.

Fig. 7 shows D/L or L/D for the different classes of unit: antennal, ocellar, and bimodal (units showing no background discharge were excluded). Most of the antennal units showed a higher frequency of background discharge when the ocelli were not illuminated than when they were illuminated. Background discharge in the ocellar units did not appear to be affected, while the bimodal units showed a wide variation in response.

The units shown in Fig. 8 showed a phasic response to olfactory stimulation with the ocelli illuminated, which changed to a phasic-tonic response when the illumination was turned off. The response to ocellar stimulation was of the type. Background discharge frequency was (mean ± s.D.) 13·0 ± 2·4/5 (n = 15) in the illuminated and 18·7 ± 1·2/s (n = 8) in the non-illuminated state; an increase of about 1·4 times. The rate of response was 80·8 in the illuminated state and 97·4 in the nonilluminated state R₁), where A and B were obtained with a 200 ms counting period. When A and B were obtained with a 500 ms counting period, R₁ and R_d became 33·3 and 85·3, respectively. To sum up, this unit showed an increase in background discharge frequency, a change in the response pattern to odour, and change in the magnitude of that response when the illumination of the ocelli was turned off.

A change of response from tonic into phasic, upon turning off the illumination of the ocelli, was observed in the unit shown in Fig. 5. In this unit, the background discharge frequency was (mean ± s.D.) 8·4±1·8/S (n = 10) in the illuminated and 12·8 ±3·3/3 (n = 8) in the non-illuminated state. Absolute values of the R₁ and R_d were 87·5 and 66·7, respectively. That is, this unit showed an increase in the background discharge frequency when the illumination of the ocelli was turned off, but the magnitude of the response to odour decreased.

These units were divided into three groups based on the effects of change in illumination of the ocelli upon responses to antennal stimuli. The first group included the units which showed increase in magnitude of response to odour when the ocelli were illuminated (i.e. R_d< R₁). The second group included those in which the magnitude of the response to odour was increased when the illumination of the ocelli was turned off (i.e. R_d>R₁). The third group contained the units which did not show statistical differences between the magnitude of responses (P>0·1) in the illuminated and in the non-illuminated states. In Table 2 49 units responding to antennal stimuli are classified in this way. About a half of the units investigated showed a larger magnitude of response to antennal stimuli when the ocelli were illuminated than when they were not. Effects of change in illumination of the ocelli on response to antennal stimuli were mainly observed in the bimodal units.

A typical example of the units which showed increase in magnitude of response to antennal stimuli when the ocelli were non-illuminated is presented in Fig. 9. This unit showed the type of response to ocellar stimuli. The background discharge frequency (measured during 500 ms before starting stimulation) was about 2 times higher with the ocelli in the dark than with them illuminated, and the magnitude of the response to antennal stimuli (measured during 500 ms after stimulation) increased markedly when the illumination was turned off. Illumination of the ocelli did not alter the response pattern to odour.

Some of the units which showed differences in magnitude of response to antennal stimuli when the illumination was turned on or turned off required several minutes illumination of the ocelli before full development of their response. A typical example is shown in Fig. 10. The magnitude of the phasic-tonic response to odour, which decreased markedly when the illumination of the ocelli was turned off, did not fully recover until ten minutes after re-illumination was begun.

The units showing effects of the ocelli upon response to antennal stimuli were found mainly in the surface region of the neuropiles in the protocerebral lobe (Table 3). The units which showed no differences in magnitude of response to antennal stimuli when the illumination was turned on or turned off were widely distributed in the protocerebrum.

The above results suggest that convergence of antennal and ocellar inputs might occur in the protocerebrum and that the activities of the neurones responding to antennal stimuli in the protocerebrum were regulated by the ocelli. Conversely, there were units whose response to ocellar stimulation was suppressed or enhanced by antennal inputs. To demonstrate the effects of antennal inputs on response to ocellar stimuli, the ocelli were stimulated with or without presentation of odour.

The unit shown in Fig. 11 showed inhibition in response to odour (though magnitude was small) when the ocelli were illuminated, but no response when they were not. Turning the illumination of the ocelli on or off evoked excitation. If an odour was present, there was less response when the illumination was turned on, and a facilitated response when the light was turned off.

The findings presented in this section were obtained from only a few units, but it was clearly shown that activities of the neurones responding to ocellar stimuli in the protocerebrum were also altered by the antennal inputs.

The response of units in the brain of the fly to stimulation of the antennae was found by Mimura et al. (1969, 1970) to be changed by illumination of the ocelli. In the present study of the honey-bee, units of this type were found to be located mainly in the surface region of the protocerebral lobe.

Illumination of the ocelli was shown to give on-responses and off-responses in brain units in the locust by Horridge et al. (1965) and in the fly by Mimura et al. (1969, 1970). Units showing continuous discharges with the ocelli in the dark, as are seen in the ocellar nerves (see review by Goodman, 1970), were not found in the present study, but the background discharge frequency was altered by illumination of the ocelli. Most units showed excitation in response to starting or to stopping the illumination of the ocelli. The responses to ocellar stimuli of the units studied here were different from the responses that have been found in the ocellar nerves. The afferent pathways of the ocellar nerves contain a small neurone which shows phasic responses to light, and a small neurone and giant neurones which both show continuous discharges which decrease in frequency when the ocelli are illuminated (Rosser, 1974). It is possible that only the former unit influences the units in the brain since most of the ocellar units showed the or type of response. The information carried by the other two units may go to the ventral nerve cord, where neurones responding to light illuminating the ocelli were reported by Goodman (1970).

Multimodal interneurones which respond to both light and mechanical stimuli (Dingle & Cardwell, 1967), and to both illumination of the compound eye and acoustic stimuli (Horridge et al. 1965) are known. Neurones, like those reported here, which respond both to antennal and to ocellar stimuli have been found in the fly by Mimura et al. (1969). Most of these neurones found in the honey-bee showed excitation in response to both stimuli. But about one-third of the units showed the opposite type of response to odour as to light, that is, excitation by one and inhibition by the other.

This may be due to the differences in synaptic connexions of the neurones arriving from the antenna or ocellus to the recording neurone. For example, the units which showed excitation in response both to antennal and to ocellar stimuli may receive excitatory inputs from the antenna and from the ocelli. In the antennal units, there were units which showed increase in magnitude of response to antennal stimuli when illumination of the ocelli was turned on. Synaptic input of subthreshold level may arrive from the ocelli to these units.

The present study showed that inputs from the antennae and from the ocelli interact with each other in the brain. The second order neurones of the ocellus (giant ocellar neurones) and the primary neurones of the antenna are closely located in the circum-oesophageal connective and in the ventro-posterial region of the protocerebrum (Pareto, 1972; Suzuki, 1975b). The integration area may be located in the neuropiles of the protocerebral lobe, since the bimodal units and the light-dependent units were distributed mainly in this region; that is convergence may be at the level of the third or more higher order ocellar nerves. Convergence of inputs from both antennae was also found in this region (Suzuki, 1975a), and presumably interaction between other sensory inputs, such as light and sound, sound and smell, and smell and touch, may occur in this region.