Spatiotemporal Neural Activities Involved in the Olfactory Processing of the Land Slug using Fluorescent-Imaging Technique

The brain or central nervous system forms a network composed of so many neurons and their function is based on complex interactions among electric neural activities, intracellular calcium signals, intercellular communications by neurotransmitter, and so on. For multipoint measurement of neural activities, fluorescent-imaging technique using voltagesensitive dyes or calcium-sensitive dyes can be a powerful technique. This technique has been applied to measure spatiotemporal neural activities involved in the olfactory processing of the land slug Limax. In Limax, the procerebral (PC) lobe, which is the olfactory center located in the lateral part of each cerebral ganglion, spontaneously produces a periodic oscillation of local field potential (LFP) of about 1 Hz, and the phase of the LFP oscillation is advanced at the distal region, resulting in periodic propagating waves of neural activity from the distal to proximal regions. The previous studies showed that odor stimuli change the LFP frequency and the wave propagation speed. In this article, we review the previous studies, as well as our recent studies, on the spatiotemporal neural activities of the land slug Limax using fluorescent-imaging technique.


Introduction
The conventional electrophysiological methods have been well used for the measurement of neural activities of vertebrate and invertebrate. The direct measurement from single neurons and neural areas is possible by the conventional methods using microelectrodes. However, the brain or central nervous system forms a network composed of so many neurons and their function is based on complex interactions among electric neural activities, intracellular calcium signals, intercellular communications by neurotransmitter, etc. The electrophysiological methods are not suitable for the simultaneous measurement from many neurons or neural areas. For multi-point measurement of neural activities, fluorescent-imaging technique using voltage-sensitive dyes or calcium-sensitive dyes can be a powerful technique. In this technique, the dyes, which change their fluorescent intensities due to the voltage or calcium ion concentration change [1], are loaded into the cells, and their fluorescence changes are acquired into a computer as a series of images.
The multi-point measurement using fluorescent-imaging technique has been reported in vertebrates [2][3][4][5] and invertebrates [6][7][8][9][10][11][12][13][14][15][16]. In invertebrates, for example, this technique has been applied to measure spatiotemporal neural activities involved in the olfactory processing of the land slug Limax [6,7,[9][10][11][13][14][15][16]. In Limax, the procerebral (PC) lobe, which is the olfactory center located in the lateral part of each cerebral ganglion, spontaneously produces a periodic oscillation of local field potential (LFP) of about 1 Hz. The LFP oscillation is well synchronized over the entire PC lobe, but the phase of oscillation is advanced at the distal region, resulting in periodic propagating waves of neural activity from the distal to proximal regions. The previous studies showed that odor stimuli change the LFP frequency and the wave propagation speed.
In this article, we review the previous studies, as well as our recent studies, on the spatiotemporal neural activities of the land slug Limax using fluorescent-imaging technique.

Fluorescent-imaging technique
In fluorescent-imaging technique, the dyes, which change their fluorescence intensities due to the voltage or calcium ion concentration change [1], are loaded into cells, and their fluorescence changes are acquired into a computer as a series of images. However, the voltage-sensitive dyes generally exhibit a relatively small change in the fluorescence intensities, resulting in a low signal-to-noise (S/N) ratio. On the other hand, the calcium-sensitive dyes exhibit a larger change in the fluorescence intensities than that of the voltage-sensitive dyes. Therefore, their fluorescence changes can be detected easily, which enables us to indirectly measure neural activities because the intracellular calcium concentration often increases with neural activities.
As the calcium-sensitive dyes, those bonded with the acetoxymethyl (AM) group are commonly used because they are permeable into cells. After the AM-bonded dyes permeate into cells, the AM group is dissociated by intracellular esterase, and then the dyes can be nonpermeable and loaded into cells. Vertebrate neurons are well known to be easily loaded with the AM-bonded calcium-sensitive dyes. In invertebrates, however, the AM-bonded dyes cannot be easily loaded into neurons because the AM group is difficult to be dissociated due to their weak activity of intracellular esterase.
For these reasons, it has not been determined yet which type of dyes is better for fluorescent imaging of neural activities in invertebrates. In Section 4, we introduce some studies on the spatiotemporal neural activities of the land slug Limax using fluorescent-imaging technique with both types of dyes.

Neural systems for olfactory processing of the land slug
The land slug Limax has two pairs of tentacles, the superior and inferior tentacles. Figure 1 shows a photograph of the central nervous system of Limax and the schematic illustration of the cerebral ganglion and superior tentacle. Each tentacle has an olfactory epithelium, which receives odors, at its tip. Olfactory receptor neurons are located at the tip regions of the digitate extensions of the tentacular ganglion (TG) (referring as "digitate ganglion" in Figure 1B) beneath the epithelium and terminate to synapse projection neurons in the TG that is the primary center of olfactory information processing ( Figure 1B). The TG contains interneurons, as well as projection neurons that project to the cerebral ganglion [17]. The TG shows a spontaneous oscillation of LFP at about 2 Hz, which is usually measured by extracellular recording [18]. The oscillation is modulated by odor and air stimuli [19]. Several neurotransmitters such as serotonin, acetylcholine, glutamate, and gamma-aminobutyric acid (GABA), exist in the TG and modulate the spontaneous oscillation [20][21][22]. In addition, the olfactory epithelium also shows an oscillatory electro-olfactogram (EOG) in response to odor and air stimuli. The EOG oscillation interacts with the LFP oscillation in the TG [19].
TG neurons project to the cerebral ganglion via the tentacle or olfactory nerves. The major target of projection is the procerebral lobe, which is the lateral part of the cerebral ganglion (Figure 1). The PC lobe is a division of the cerebral ganglion unique to terrestrial slugs and snails that is specialized for the processing of olfactory information [23][24][25]. The PC lobe consists of three layers, the cell mass (CM), terminal mass (TM), and internal mass (IM) (Figure 2). The CM contains a large number of cell bodies of neurons, which are generally  [7] with permission. small (5-8 μm) [23]. The TM and IM are neuropile layers. The TM receives input from the tentacle nerves [9]. The PC lobe spontaneously produces a periodic slow oscillation of LFP at about 1 Hz, which is usually measured by extracellular recording [26]. The LFP oscillation is well synchronized over the entire PC, but the phase of the oscillation is advanced at the distal region. This results in periodic propagating waves of neural activity from the distal to proximal regions, which has been clearly shown by fluorescent-imaging technique [6,7,[9][10][11][13][14][15][16].
Patch-clamp recording from single neurons showed that the neurons in the PC lobe are categorized as either bursting (B) or nonbursting (NB) neurons [6,7] (Figure 2). B neurons are characterized by periodic bursting activity, and NB neurons are silent or fire at a low frequency. NB neurons are more numerous (about 90%) and have projection of neurites to the TM and IM [27]. NB neurons receive olfactory input from the tentacle nerves in the TM [28].  The propagating waves of neural activity have also been found in mammalian cortex. However, the function and computational role are still unclear even in the land slug Limax. To elucidate it, fluorescent-imaging technique is expected to be a powerful technique as described in the subsequent section.

Fluorescent imaging of spatiotemporal neural activities involved in the olfactory processing of the land slug
As mentioned earlier, periodic waves of neural activity propagate from the distal to proximal regions in the PC lobe of the land slug Limax. Pioneer works of a group clearly showed this phenomenon for the first time in the PC lobe of L. maximus by fluorescent voltage imaging using a voltage-sensitive dye [6,7]. They used di-4-ANEPPS, a substituted aminonaphthylethenylpyridinium [29], as the fluorescent voltage-sensitive dye. The fluorescence was detected with either an array of photodiodes or a cooled charge-coupled device (CCD) camera. Figure 3 shows the experimental apparatus that they used. Figure 4 shows the propagating waves of fluorescence change measured by the CCD camera. A band of depolarization, followed by a band of hyperpolarization, began at the distal tip and moved along the length of the PC lobe, resulting in a phase difference between the distal and proximal regions (Figure 4A and C ). The detailed shape of the optical signal as a function of time also varied between different regions ( Figure 4B). Simultaneous recording of the optical signal and the LFP by extracellular recording showed that the oscillatory changes in the fluorescent intensity were synchronized to the LFP oscillation ( Figure 5). Additionally, from simultaneous recording of the optical signal and the intracellular potential by whole-cell patch technique (Figure 6), the authors concluded that the major part of the optical signal can be interpreted as a superposition of the intracellular signals arising from the B and NB neurons.
They next examined the responses of spatiotemporal neural activity in the PC lobe to odor stimuli (Figure 7). The stimuli were applied to the epithelium of the superior tentacle. The phase difference of neural activity between the distal and proximal regions was not changed in response to the moist air ( Figure 7A). However, in response to a component of natural potato odor (2-ethyl-3-methoxypyrazine; EMOP), a known appetitive stimulus to the naïve slugs, the phase difference was dramatically changed (Figure 7B-F). The stimulus caused a collapse of the phase difference, that is, an initial difference of Δt > 250 ms was reduced to Δt < 50 ms. Repeated stimulus also caused a collapse of the phase difference, although it recovered faster after cessation of the stimulus. The transient switch from the states with propagating waves to one with spatially uniform oscillations is hypothesized to allow nearest-neighbor interactions suppressed by the propagating wave activity [30]. This study suggested that olfactory information is encoded in the spatiotemporal neural activities in the PC lobe.
Thereafter, other groups have also studied the spatiotemporal neural activities of the PC lobe using fluorescent voltage imaging. Kimura et al. examined the responses of the PC lobe of L. marginatus to aversively conditioned odors [10]. In this study, di-4-ANEPPS was also used as the voltage-sensitive dye. First, they performed aversive conditioning procedures to the slugs using two odors (carrot and cucumber, which strongly attract the naïve slugs) as conditioned stimuli (CS) and quinidine sulfate solution, which induces a strong aversive effect, as an unconditioned stimulus (UCS). The odor stimuli were applied to the epithelium of the inferior tentacle. When the conditioning is completed, the slugs avoid the CS odor and exhibit eating behavior at a lower frequency [31]. In the PC of the naïve slugs, they observed the propagating waves of fluorescence change similar to those reported in the previous study [6,7] (Figure 8). In the PC lobe of the slugs aversively conditioned by carrot odor, the instantaneous frequency of the oscillatory fluorescence change was decreased from about 4 s after the onset of conditioned carrot odor presentation (Figure 9a). The optical signals recorded in an area of the PC lobe (area 3 in the upper image of Figure 9) showed depolarization in the basal potential (bottom level of each cycle of the oscillation) immediately after the onset of the odor presentation (Figure 9d). In another area of the PC lobe (area 4 in the upper image of Figure 9), hyperpolarization in the basal potential occurred from about 5 s after the carrot odor presentation (Figure 9e). On the other hand, the control cucumber odor did not significantly induce those changes (Figures 9f-j). Figure 10 shows temporally average images of optical signals calculated in 10 different periods of two recordings shown in Figure 9 (periods a-j in the upper graphs of Figure 10). This result shows that depolarization occurred especially in the neurons located in a belt-shaped region of the middle of the PC lobe (red-colored region in image b in Figure 10) immediately after the onset of conditioned carrot odor presentation. The belt-shaped region corresponded to area 3 in the upper image of Figure 9, in which depolarization in the basal potential occurred in the time profile, while the early phase depolarization was not observed for the control odor cucumber. After the depolarization was diminished, hyperpolarization occurred in the neurons in a relatively wide area of the PC lobe (blue-colored region in images c and d in Figure 10). Similar results were obtained in the PC lobe of the slugs aversively conditioned by cucumber odor. From these results, the authors concluded that in the PC lobe, learned odors are represented as spatiotemporal activity patterns, which is supported by activity-dependent dye uptake study [32].   (Figure 9), together with the PC lobe by fluorescent voltage imaging, and they analyzed the relationship between them using the correlation analysis [13]. In this study, di-4-ANEPPS was also used as the voltage-sensitive dye. Figure 11 shows the fluorescent optical recording of spontaneous activity of the cerebral ganglion of L. valentianus. The image field covered the proximal part of the PC and the entire ipsilateral MC region. In all preparations, they observed oscillatory neural activity in the PC. Additionally, they found clear oscillatory activity in the MC with the same frequency as the PC oscillation in some preparations, although the amplitude of the MC oscillation was smaller than that of the PC oscillation. Their phase relationship was variable. In some preparations, the two oscillations were antiphase, while in the other preparations the oscillations were inphase. This study indicated that the oscillatory activity in the PC lobe, in which olfactory information is encoded as mentioned earlier, propagates to the MC that plays a role in motor command. The spatiotemporal neural activities in the PC lobe were also measured by fluorescent calcium imaging using calcium-sensitive dyes. Inoue et al. used Oregon Green-1, rhod-2, or Ca Orange bonded with AM group as the calcium-sensitive dye, and compared the fluorescent calcium imaging using each of them with the florescent voltage imaging using di-4-ANEPPS [11]. In fluorescent calcium imaging, propagation waves of fluorescence change were also observed in the PC lobe of L marginatus (Figure 12). The PC lobe was well stained by the calcium-sensitive dyes, even though AM group-bonded dyes were used. The calcium signals originated predominantly from the cell body layer, CM, while the voltage signals originated mainly from the neurophil layer, TM. The calcium-sensitive dyes gave 1-3% fluorescence change, which increased rapidly and decreased slowly, with a good signal-to-noise ratio (Figure 13). Among the calcium-sensitive dyes, rhod-2 gave the largest fluorescence oscillation. On the  other hand, the voltage-sensitive dye gave 0.1-1% fluorescence change, which increased slowly and decreased rapidly (Figure 13). In this study, the authors concluded that the oscillatory signals of the calcium-sensitive dyes arise from B neurons of the CM and the neural activity of the CM can be monitored with them because the oscillatory signals were synchronized with the LFP (Figure 13).
Fluorescent calcium imaging has been limited with respect to the selectivity of staining of the dyes. Bath application of the dyes stains both B neurons and NB neurons. Even when we intend to observe the activity of NB neurons, which is thought to be involved in odor coding [33], it is obscured by stronger calcium signals from B neurons that exhibit periodic bursting activity. To observe the activity of NB neurons, Watanabe et al. attempted selective fluorescent calcium imaging for the PC lobe of L. valentianus by retrograde staining of NB neurons with rhod-2 [14]. When the sheath was removed only from the surface of the IM and the PC lobe was immersed in the dye solution, the NB neurons were expected to be stained retrogradely through neurites projected to the IM (Figure 14). After nonselective staining, in which the sheath was removed from the entire surface of the PC lobe, a strong spontaneous oscillation was observed, which presumably arose from calcium oscillations of B neurons ( Figure 15). On the other hand, the PC lobe showed a much smaller spontaneous oscillation after the selective staining, although a large number of cells were clearly stained (Figure 15). This result suggested that NB neurons are mostly silent in contrast to B neurons. However, NB neurons responded to electric stimulation of the superior tentacle nerve as shown in the inset of Figure 15D.
As described above, fluorescent-imaging technique is applicable for the measurement of the spatiotemporal neural activities involved in the olfactory processing of the land slug. Through the studies using it, especially, the function and computational roles of the propagation waves of neural activity in the PC lobe on odor discrimination and learning have been elucidated. Meanwhile, the experimental apparatus has also been improved. In the earlier reports of the fluorescent voltage imaging, consecutive images were acquired at a flame rate of about 100 ms   for about 100 × 100 pixels in size [6,7], while they can be acquired at a flame rate of about 10 ms for about 1000 × 1000 pixels in size in our recent apparatus as described in Section 5.

Our recent fluorescent-imaging apparatus and its application
We have also measured the spatiotemporal neural activities in the PC lobe of the land slug Limax by fluorescent voltage imaging [15,16]. Figure 16 shows our recent apparatus. We also use di-4-ANEPPS as the voltage-sensitive dye. The recording chamber, in which the stained preparation is placed, is mounted on the stage of a microscope (E-FN1, Nikon, Japan). The dye is excited by LED of 530 nm with a half width of 25 nm (LEX2-G, Brain Vision, Japan) through an excitation filter (EX510-560). The emitted fluorescence of 705 nm is detected through a dichroic   Figure 17). Here, a 10× objective was used and the space resolution was about 0.6 μm/pixel. An aversive odor, isoamyl acetate, was applied to the epithelium of   the superior tentacle. Figure 18 shows the fluorescence changes in four ROIs. The time lag between two peaks at the distal and proximal regions significantly decreased, that is, the propagation speed significantly increased when the odor was applied. In our experiments, this phenomenon occurred only for the aversive odors, not for the appetitive odors.
We also examined the involvement of nitric oxide (NO) in the odor-induced changes in wave propagation by our apparatus. NO, which is a gaseous neurotransmitter, has been reported to function specifically in the PC lobe. Gelperin et al. showed that the excitability of B neurons is modulated by NO [34] and suppression of NO synthesis blocks the LFP oscillation and wave propagation in the PC lobe [35]. In our study [36], the stimulation of another aversive odor, hexanol, to the epithelium of the superior tentacle increased NO concentration measured by an NO electrode applied to the PC lobe of L. valentianus. This was blocked by NO synthase inhibitor L-NAME (3.7 mM), although it did not block the LFP oscillation and wave propagation in the PC lobe at this concentration. Additionally, the LFP frequency and wave propagation speed were increased by the odor (hexanol) stimulation or electrical stimulation of the superior tentacle nerve, and these phenomena also disappeared by L-NAME, but not by its inactive enantiomer D-NAME (Figure 19). On the other hand, L-NAME did not block synaptic transmission from the superior tentacle nerve to NB neurons in the PC lobe. These results suggested that NO is released from NB neurons diffuses into the CM to depolarize B neurons, which modifies the network activity, presumably the main effect of suppressing NB neurons (Figure 20). Based on this study as well as previous studies, the authors speculated that the effects of NO released during olfactory perception underlie the mechanism of precise odor discrimination and learning in the PC lobe of the land slug. Average and individual data are shown. L-NAME significantly reduced the decrease in the lag (*p < 0.05, N = 10), whereas D-NAME did not significantly change the response (NS, not significant; N = 8). These figures are reproduced from Ref. [36] with permission.

Conclusions
Fluorescent voltage imaging is applicable for the measurement of the spatiotemporal neural activities involved in the olfactory processing of the land slug Limax. Fluorescent calcium imaging is also applicable for it, even though generally the AM group-bonded calciumsensitive dyes cannot be easily loaded into invertebrate neurons. Through the studies using the fluorescent-imaging technique, especially, the function and computational roles of the propagation waves of neural activity in the PC lobe on odor discrimination and learning have been elucidated. Meanwhile, the experimental apparatus has also been improved. By our recent apparatus, consecutive images can be acquired at a flame rate of about 10 ms for about 1000 × 1000 pixels in size, and some new findings on the olfactory processing were obtained. Such an improvement of the experimental apparatus will enable us to obtain further information not only on the olfactory processing of the land slug but also on various biological functions.