| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Laboratorio de Neurobiología de la Memoria, Departamento Fisiología, Biología Molecular y Celular, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Instituto de Fisiología, Biología Molecular y Neurociencias–Consejo de Investigaciones Científicas y Técnicas, Buenos Aires, Argentina
Submitted 19 July 2007; accepted in final form 16 August 2007
 |
ABSTRACT |
|---|
|
|
|
INTRODUCTION |
|---|
|
; Korn and Faber 2005; Rind and Simmons 1999). How fast an escape response must be performed depends on such factors as the sensory system detecting the threat and the imminence and nature of the risk signaled by the stimulus. For instance, a shockwave produced by the sudden strike of an aquatic predator causes crayfish and fishes to perform avoidance behaviors whose initial components are completed in about 30 ms after the stimulus onset (Herberholtz et al. 2001; Preuss et al. 2006). This last-chance type of behavior represents a straightforward reflex that implies a relatively basic level of decision making.
In contrast, visual systems permit the detection of danger stimuli at greater distances. In this case, elaborate decision-making processes can be used that identify the precise location of the stimulus in space, assess the level of risk posed by the stimulus, and choose the best strategy for avoidance. Consequently, behavioral responses elicited by visual danger stimuli seldom start <100 ms after the stimulus onset and usually take longer to evolve (Preuss et al. 2006). The variety of escape performances is reflected in the organization of the neural circuits underlying these behaviors. Extremely fast responses can be triggered by a single spike in an individual neuron, for instance the lateral giant neuron of the crayfish (Edwards et al. 1999) and the Mauthner cell of fishes (Korn and Faber 2005). In contrast, spike trains in similar or different sets of neurons are integrated to trigger and steer visually elicited escapes (Santer et al. 2006; Wu et al. 2005).
In the crab Chasmagnathus, the escape response to visual danger stimuli, in particular the learning and memory processes underlying its long-term change by repeated stimulation, has been the subject of intense investigation. Studies have ranged from behavioral analyses (e.g., Kaczer et al. 2007; Lozada et al. 1988; Tomsic et al. 1998) to pharmacology (e.g., Merlo et al. 2007; Pedreira and Maldonado 2003; Tomsic et al. 1991) and molecular biology (e.g., Feld et al. 2005; Freudenthal et al. 2000; Locatelli and Romano 2005). The visual nervous system of crustaceans consists of three ganglia, named from the periphery to the center: the lamina, the medulla, and the lobula. Each of these areas exhibits parallel retinotopic organization (Strausfeld and Nässel 1981). These so-called optic neuropils have been commonly assumed to be sensory centers devoted solely to visual processing. However, in recent years we have identified a generic group of large tangential neurons from the lobula of Chasmagnathus that integrates visual and mechanosensory information arriving from different parts of the animal's body (Berón de Astrada and Tomsic 2002). We found that the response of these neurons to visual danger stimuli highly correlates with the escape performance of the crab (Oliva et al. 2007; Tomsic et al. 2003). These neurons reflect the short- and long-term behavioral changes that follow repeated stimulation and are thought to play a key role in the formation and maintenance of visual memories in the crab (Tomsic 2002; Tomsic et al. 2003). Moreover, recent results indicate that the learning-induced neuronal changes persist even when the position of the stimulus within the animal's receptive field is altered. These data indicate that these neurons support the ability of crabs to generalize between memories acquired in distinct positions of the visual field (J Sztarker and D Tomsic, unpublished observations).
Mounting evidence suggests that the lobula is not simply a visual processing neuropil, but also processes features that are often ascribed to "higher centers" (Sztarker et al. 2005; Tomsic et al. 2003). In previous publications we generically termed the large tangential lobula elements "movement detector neurons" (MDNs). Now that we know they possess a much higher level of integration than previously suspected, it would be misleading to continue referring to them as simply "movement detectors." Throughout this paper, we use the term "lobula giant" (LG) neurons to refer to the cells previously called MDNs in the crab.
Ideally, the neuronal underpinnings of higher brain functions such as decision making and learning should be investigated in intact and behaving animals (Nichols and Newsome 1999). However, the study of neuronal function with intracellular recording and imaging techniques is rarely achieved in such circumstances. Crabs provide a robust subject for the neurophysiological study of awake, behaving animals. In particular, we study the crab's escape response to visual stimuli, which appears to be greatly determined by the activity of LG neurons that are readily accessible for in vivo electrophysiological recording (Oliva et al. 2007). In addition, learned responses to natural sensory stimulation can be assessed at the level of individual LG neurons through intracellular (Tomsic et al. 2003) and Ca2+ image recording (A Delorenzi, personal communication). Remarkably, the operation to obtain intracellular recordings from brain neurons is so noninvasive that crabs remain healthy and no subsequent behavioral differences are observed with respect to naive animals.
The general morphology of LGs is characterized by: 1) a wide-field tangential arborization in the lobula; 2) a somata located beneath this neuropil; and 3) an axon that leaves the optic lobe, projecting toward the midbrain. Despite the fact that all LGs studied to date share this general morphology, intracellular staining revealed the existence of neurons with single or bistratified dendritic organization in the lobula (Berón de Astrada and Tomsic 2002; Sztarker et al. 2005). Physiological characteristics also indicate that the group is formed by distinct neuronal classes. However, because studies on LGs were performed with a single visual stimulus (i.e., a black screen moving horizontally above the animal), we have yet to identify the different classes of neurons based on their preferences for visual stimuli and their correspondence with particular morphologies.
Here we investigate the response of LGs of the crab Chasmagnathus to a variety of computer-generated images moving across different parts of the animals' visual field. In vivo intracellular recording followed by dye injection was used to characterize visual preferences, intrinsic properties, and morphologies of the cells. We found that LGs invariably showed a preference for motion of single objects in comparison to flow field patterns. Physiological and morphological analyses allowed the identification of four neuronal subclasses that could be characterized in detail, as well as the recognition of additional elements that, due to infrequent recordings, could not be ascribed to any particular group. This study suggests that visual assessment of threatening objects—and thus the decision-making strategy for avoidance—operates through the cooperative activity of central neurons with distinct filter properties.
|
|
METHODS |
|---|
|
Animals were adult male Chasmagnathus granulatus crabs 2.7–3.0 cm across the carapace, weighing about 17 g, collected in the rías (narrow coastal inlets) of San Clemente del Tuyú, Argentina. The crabs were transported to the laboratory, where they were maintained in plastic tanks (35 x 48 x 27 cm) filled to 2-cm depth with artificial seawater to a density of 20 crabs per tank. Water used in tanks and other containers during the experiments was prepared using hw-Marinex (Winex, Hamburg, Germany), salinity 10–14%, at a pH of 7.4–7.6, and maintained within a range of 22–24°C. The holding and experimental rooms were kept on a 12-h light/dark cycle (lights on 7:00 am to 7:00 pm) and the experiments were run between 8:00 am and 7:00 pm. Experiments were performed within the first 2 wk of the animals' collection from the field. Crabs were fed rabbit pellets (Nutrients Argentina) every 3 days and after feeding the water was changed. Experiments were carried out year-round.
Visual stimuli
The experiments were performed by using computer-generated visual stimuli projected either simultaneously or alternately on four 17-in. flat-screen monitors (Phillips 107T, refreshing rate 60 Hz). The monitors were located 20 cm in front of, above, and on either side of the animal (Fig. 1), and were housed inside a Faraday cage completely covered to prevent outside visual stimuli from reaching the animal. Antiglare screens largely prevented reflections among the monitors. Three of the monitors stood on a vibration-damped table and the fourth hung from the ceiling of the cage. All visual stimuli were generated with a single PC, using commercial software (Presentation 5.3, Neurobehavioral Systems, Albany, CA). The stimulus image generated by the PC was first split and then sent to four selector switches. From each selector the video signal could be rapidly turned on and off. The selectors as well as other control systems used during the experiments were located outside the Faraday cage. In this way, the experimenter could choose which screen or set of screens showed any stimulus at any time without distressing the animal. This arrangement allowed us to deliver a variety of visual stimuli and to quickly switch the presentation of the stimuli across specific parts of the visual field while neuronal activity was being recorded.
|
), and no differences between the responses elicited by real objects or computer simulations could be found (V Medan and D Tomsic, unpublished observations). The visual stimuli used throughout the present study consisted of black objects moving over a white background, and were of three types: ). In all cases, stationary stimulus images were present at the screen for 30 s before they started to move.
Experiments included measurements in randomized order for the three types of stimuli. Thus a single series of stimulations consisted of 19 different stimuli, which varied in their type (1, 2, or 3), motion direction, and location in space. At least two complete series of stimulus presentations were recorded from each neuron. To curtail the habituation effect of the LGs, stimuli were presented every minute. Thus stimulation series in a single experiment took 
40 min to complete. During this time the animal occasionally moved its legs, which sometimes caused loss of the cell impalement before the end of the experiment. Data from incomplete series, however, provided valuable information and were included in some analyses.
Two additional stimuli were also used. One of them consisted of a 1-s pulse of light delivered through an optic fiber that reached the eye with an intensity of 130 Wm2. The second was a mild mechanical stimulation applied with a paintbrush to the dactyl of one of the animal's legs or another part of the body.
Electrophysiology
Intracellular recordings from interneurons in the right optic lobe were performed in the intact living animal according to methods previously described (Berón de Astrada et al. 2001). The crab was firmly held in an adjustable clamp. The eyestalks were cemented to the carapace at an angle of approximately 70° from horizontal, which corresponds to their normal resting position. A tangential cut performed with a sharp scalpel was made to remove a small section of cuticle (
500 µm in diameter) from the tip of the eyestalk without causing damage to the ommatidial area. A ground electrode consisting of a silver wire was inserted through a small hole in the dorsal carapace. The clamped crab was positioned in the center of the monitor configuration and held in position using a magnetic holding device. The entire setup was surrounded by a Faraday cage.
The glass microelectrode was then positioned and advanced through the opening in the cuticle. Microelectrodes (borosilicate glass; 1.2 mm OD, 0.68 mm ID) were pulled on a Brown-Flaming micropipette puller (P-77; Sutter Instrument, Novato, CA) yielding tip resistances of 40–60 M
when filled with 3 M KCl. A bridge balance amplifier was used for intracellular recordings (Axoclamp 2B; Axon Instruments, Burlingame, CA). The output of the amplifier was monitored on an analogue oscilloscope, digitized at 10 kHz (Digidata 1320; Axon Instruments), and recorded with Clampex (from pClamp9 suite, Axon Instruments) for subsequent analysis.
LGs are easy to recognize on the basis of their strong response to motion stimuli in comparison to stationary changes of illumination (Berón de Astrada and Tomsic 2002; Tomsic et al. 2003). Thus the cell identity was tested at the beginning of the experiment by moving a hand around the animal and by delivering a flash of light. A black curtain was then lowered in front of the cage and the animal was left undisturbed for 10 min before the experiment began. All intracellular recordings were performed at the membrane resting potential.
Cell morphology
Electrode tips were backfilled with 5% Neurobiotin, 50 mM Tris buffer in 500 mM KCl solution, backed up with 3 M KCl. Following characterization of their response to the visual stimuli, the cells were filled iontophoretically for 15–60 min using 1- to 4-nA positive current. Only one neuron per animal was injected. After iontophoresis, the trace was allowed to diffuse for 2–4 h in the living animal. Then, the crab was anesthetized on ice and the optic lobes and the midbrain were dissected and immersed in 4% paraformaldehyde in phosphate buffer (pH 7.2) to be fixed overnight. After five 20-min washes with PTA (PBS 0.1 M, Triton X 2% vol/vol, and sodium azide 0.1% wt/vol; pH 7.4), ganglia were incubated overnight with avidinrodamin (1/3,000 vol/vol in buffer PTA) at 4°C with constant shaking, after which they were again washed five times with PTA. The ganglia were then dehydrated in ethanol series and cleared in methyl salicylate.
Cleared ganglia were imaged as whole mounts and scanned at 2- to 4-µm intervals with a confocal microscope equipped with a Helium/Neon laser (Olympus Fluoview 1000). Images, saved as three-dimensional (3D) stacks, were adjusted for brightness and contrast, and illustrations were obtained by merging the individual serial sections with ImageJ 1.33U (National Institutes of Health, Bethesda, MD). The morphologies of filled neurons were reconstructed by tracing and hand drawing the neuronal structure from the series of individual optical sections.
Data analysis
All visual stimuli were triggered by the program used for electrophysiological recording, which allowed us to align the time course of neuronal response with that of the stimulus motion. Relative times of action potentials were evaluated off-line by the threshold-detection algorithm of pClamp9, and confirmed by visual inspection of the records. The spontaneous firing rate was estimated from periods immediately preceding the onset of the stimuli. Neurons showed a fairly stable membrane resting potential and rate of spontaneous firing; when fluctuations in these levels occurred, the experiment was terminated. The maximum firing rate was calculated as the highest instantaneous frequency (1/minimum interspike interval) elicited by tangential or looming stimuli. Receptive field and directional preferences were investigated using stimuli that moved with a constant speed (18 cm/s) for a known duration (1 s), allowing comparisons based on the number of spikes elicited during the entire period of stimulation. Responses to the same stimuli obtained in the two stimulation series were averaged, and the mean value was used in the analysis. Polar plots for individual cells were built by normalizing the response value at each screen and direction of motion to the maximum obtained value.
Because LGs are wide-field elements, a quality estimation of the size and location of their receptive field could be obtained by comparing the responses across the screens. To this aim, a magnitude of response at each screen was calculated by adding the response values corresponding to the four directions of motion. For quantitative assessment of preferred direction, each cell was assigned a direction index (DI), which compared the response in the preferred direction of movement with the response in the direction 180° opposite. The index was defined as follows. DI = 1 – np/p, where np = (number of spikes in nonpreferred direction) – (spontaneous spikes) and p = (number of spikes in preferred direction) – (spontaneous spikes). The DI varies from 0 (equal responses in the two directions) to 1 (no spikes in the null direction). Directional preference was considered to exist when neurons responded with 50% more spikes in the preferred direction than in the null direction and thus had a DI of 0.33. To evaluate the preference for a particular axis of motion, we compared the percentage of responses (in preferred + nonpreferred direction) obtained along the horizontal and the vertical axes. To assess the cell preference for a single object motion or a wide-field motion, we used a preference index (PI) defined as: PI = 1 – om/fm, where om = (number of spikes to object motion) – (spontaneous spikes) and fm = (number of spikes to field motion) – (spontaneous spikes). Whenever possible we used parametric statistics (t-test or ANOVA followed by multiple comparisons); otherwise, we used nonparametric tests (Kruskal–Wallis followed by Dunn's comparisons).
|
|
RESULTS |
|---|
|
1 h. As illustrated throughout this paper, all stained neurons showed a general pattern consisting of an extensive tangential arborization in the lobula, with the somata located in the cell body cluster between the lobula and the lateral protocerebrum, and the axon exiting the optic lobe through the protocerebral tract. Despite these commonalties, four distinguishable cell morphologies were repeatedly observed. An analysis of physiological characteristics of the stained cells revealed clear relationships between each morphological identity and certain intrinsic cell properties and response characteristics. Thus we were able to recognize any one of the four types based solely on its electrophysiological behavior. Before outlining the particular characteristics of each of the different cell types, we describe their common features. Response preference for single objects versus wide-field motion
Crabs are highly sensitive to both wide-field pattern and object motion. The former is clear from their strong optomotor responses to panoramic optic flows (e.g., Johnson et al. 2004), whereas the latter is evident from their directional escape responses to visual danger stimuli (e.g., Nalbach 1990). We began our investigation of the large tangential neurons of the lobula by analyzing whether these neurons have a response preference for wide-field or single object motion. Figure 2A illustrates the responses of two LG neurons of different classes to 1 s of motion with each type of stimuli. Despite the fact that the wide-field pattern included the single object, responses to this stimulus were clearly weaker than those to the object moving alone. Indeed, recordings obtained from 33 neurons in 33 animals showed that the vast majority of LGs have a response preference for the single object compared with the wide-field pattern. Of these 33, 29 neurons (88%) showed PI values <1, whereas only four neurons had indexes between 1 and 1.6 (Fig. 2B). On average, LG neurons responded with about 50% more spikes to the single object than to the pattern (Fig. 2C; Student's paired test, t = 6.03; P < 0.001; df = 32). From the neurons sampled in this experiment, 17 were successfully stained and recognized as members of one of the groups subsequently described (Fig. 5–8). No obvious relationship, however, was observed between the neuronal classes and the level of object-to-pattern preference revealed by the distribution of index values. The three higher indexes corresponded to unstained cells, whose identities could not be established from their response properties.
|
|
|
|
|
). The response of LG neurons to wide-field motion was clearly weaker than that to single object motion. Still, the response was noticeable and is, in principle, capable of driving optomotor behavior. To further investigate the possibility that LG neurons may participate in the optomotor behavior, we recorded responses of LG neurons to sustained motion of the wide-field pattern for 8 s (Fig. 3A). The two illustrative examples in Fig. 3A demonstrate that, after an initial reaction elicited during the first few seconds of motion, the response rapidly wanes. Similar results were obtained in experiments from 27 additional cells (Fig. 3B). Altogether, these results indicate that large tangential neurons from the lobula of the crab are more likely involved in object detection than in flow field analysis.
Responses to oriented bars and squares
When drifting bars are used as visual stimuli, information about orientation specificity and direction specificity is inherently confounded. Several methods have been used to distinguish between directional and orientational information in the response of visual neurons (Zhang 1990), but all of these methods require a large number of repeated measurements to be reliable. The tendency for LG neurons to modify their response on repeated stimulation (i.e., habituate) precludes the evaluation of repeated responses using short intertrial intervals. Consequently, the capacity to distinguish between orientational and directional preference is limited. For this reason, we decided to evaluate the directional sensitivity and forgo the assessment of orientational preference by using a moving square. Despite its smaller size, this moving square elicited neuronal responses as readily as the bar stimulus (Fig. 4).
MLG1 neurons
Thirty neurons recorded in this study could be identified as corresponding to a single cell class that we named "monostratified lobula giant 1" (MLG1). Thirteen of these neurons were successfully stained. The morphological differences between MLG1 and other neuronal classes subsequently described largely relate to differences in the arborizations of the cells within the complex retinotopic organization of the lobula. Therefore a brief anatomical description of this neuropil is warranted.
Because different cell types have their tangential processes oriented either along the anteroposterior or the lateromedial axis, the fibro-architectural appearance of the Chasmagnathus lobula depends on the section orientation. Transverse sections show four strata of tangential processes oriented lateromedially. From the periphery to the center, these strata are referred to as the first through the fourth lateromedial tangential layers LMT1–LMT4 (Sztarker et al. 2005). These strata are separated by regions containing arborization profiles belonging to columnar elements and local interneurons, as well as the profiles of tangential processes running anteroposteriorly. LMT1 and LMT4 consist of relatively thin tangential processes, whereas LMT2 and LMT3 contain long, wide-diameter tangential fibers that increase in girth toward the medial side of the neuropil. Longitudinal sections of the lobula reveal five strata composed of tangential processes oriented anteroposteriorly. From the periphery to the center, these are called the anteroposterior tangential levels APT1–APT5. The morphology and cell ensemble of the MLG1 class, whose physiological characteristics we analyze here, were first described in an anatomical study performed by Sztarker et al. (2005). They revealed that level APT4 contains a set of 14 easily recognizable neurons, each of which possesses an exceptionally wide-diameter primary branch (8- to 10-µm diameter) oriented along the anteroposterior axis (Fig. 5, A, D, and E; see also Figs. 8 and 9 in Sztarker et al. 2005). Primary branches from 14 MLG1 cells are arranged in parallel, separated from one another by about 35 µm. Each primary branch gives rise to several secondary processes that arise at right angles and thus extend lateromedially within the LMT3 stratum in close proximity to tangential processes of the other classes of LGs as subsequently described. Each primary branch is connected to a prominent axon that extends centrally for a short distance before being reduced in diameter, after which it descends through the lateral protocerebrum with few arborizations and reaches the optic tract. The descending axons of the 14 neurons converge to form a discrete bundle. The large perikarya of these neurons are clustered beneath the lobula; each is connected by a neurite to its primary branch.
|
). Although each individual element processes information from a restricted sector of the animal's visual field, the entire assembly of MLG1 neurons enables the assessment of the whole panorama.
|
50% more spikes in one direction than the other (compare top and bottom traces on Fig. 5H). Therefore MLG1 neurons are directionally sensitive elements. In addition, a comparison of the responses to vertical and to horizontal motion showed that these neurons are significantly more responsive to objects moving horizontally than vertically (Fig. 5J; Student's paired test, t = 3.6, P < 0.002, df = 23). Finally, MLG1 cells are highly sensitive to images of approaching objects (i.e., looming stimuli). They were shown to respond to these stimuli by gradually increasing their rate of firing, encoding the velocity of the expanding image on the retina (Fig. 5K; Oliva et al. 2007). MLG2 neurons
Thirty neurons recorded in different animals were identified as members of the class herein called monostratified lobula giant 2 (MLG2), from which nine units were successfully stained (Fig. 6, A–C). The dendritic tree consists of several branches that run parallel to each other along the lateromedial axis of the lobula. These branches originate multiple slender branches that project laterally and then run in parallel to the main branches in such a way that they spread over the entire retinotopic mosaic within the LMT3 layer of the lobula. In the medial side of the lobula the branches converge into a single trunk that descends toward the lateral protocerebrum, where it gives rise to a second large tree that penetrates several protocerebral regions. The soma is located above the lateral protocerebrum and is connected to the main trunk by a thin and rather long neurite. An axonal fiber that departs from the main trunk descends toward the midbrain through the protocerebral tract. The large size of this neuron, when considered alongside the similar morphologies of all stained units, suggests that there may be only one element of this class in each optic lobe.
|
The polar plots of Fig. 6D show that the neurons responded to stimuli presented at any screen, suggesting that their receptive field encompasses the entire visual field of the animal. Moreover, responses obtained from the four screens were very similar; in other words, motion sensitivity appears to be rather uniform throughout the visual receptive field. Equivalent results were obtained in 26 MLG2 neurons (Fig. 6E). On average, responses from the most effective screen accounted for only 27% of the total responses, whereas the sum of responses in the second and third most preferred screens was 52 and 77%, respectively (an even distribution would have rendered 25, 50, and 75%). Analysis of the sensitivity to motion directions of 28 MLG2 neurons from different animals revealed a mean (±SE) index of directionality of 0.21 ± 0.04, which is below the criterion to assume directional sensitivity. In fact, only 5 of 28 cells (<18%) showed index values >0.33 (Fig. 6K). Therefore MLG2 neurons appear not to be sensitive to the direction of motion. Furthermore, no difference could be found between responses to objects moving along the vertical and the horizontal axes (Fig. 6L).
BLG1 neurons
We recorded only 11 units of a class herein termed bistratified lobula giant 1 (BLG1) neurons, 7 of which were sufficiently stained to allow the morphological description of the class. The low rate of recording success with elements of this class is likely related to the small diameter of the neurites of these cells in the lobula. The morphology is shown in Fig. 7, A–C. In the lobula the cells possess a clear bistratified organization. A pair of long neurites traverses the neuropil lateromedially, running parallel along two different strata. According to the terminology for lobula layers established by Sztarker et al. (2005), the distal and the proximal neurites run along strata LMT2 and LMT3, respectively. Within the lateral part of the lobula, each of the two neurites branches into a tree of arborizations that expands horizontally within the same layer until reaching the lateral pole of the neuropil. On the opposite side of the lobula, the two neurites converge into an axonal fiber, which projects undivided through the lateral protocerebrum and descends by the protocerebral tract. The cell body could be observed in only three of the seven stained cells. In all cases it was located just above the lateral protocerebrum and connected to the axonal fiber by a short neurite.
|
BLG2 neurons
A morphological description of neurons of the class here named bistratified lobula giant 2 (BLG2) was first provided by Berón de Astrada and Tomsic (2002). In the present study we recorded 22 of these units in different animals, 8 of which were suitably stained. The cell morphology is shown in Fig. 8, A and B. In the lobula the neurons are organized in a bistratified manner, with their distal and proximal dendrites staggered at layer LMT2 and layer LMT3, respectively. At each level, the dendritic tree is formed by several branches that run parallel to each other all along the lateromedial axis of the neuropil. These branches converge toward the medial side of the neuropil into a thicker single trunk that descends to the lateral protocerebrum. There, it further divides into several branches that arborize in different protocerebral regions. The main trunk extends as an axon that, after traversing through the neuropil, projects into the protocerebral tract. The soma, which is located above the lateral protocerebrum, is connected to the main trunk of the neuron by a thin fiber.
Figure 8 shows representative responses of the BLG2 type of neuron. These cells are spontaneously active. Even when the animal is still, a BLG2 cell shows bursts of two to five spikes occurring at unpredictable times (Fig. 8E). On average, the spontaneous firing frequency is 3 ± 0.5 Hz, whereas the mean of the maximum instantaneous spike rate is 150 ± 9 Hz (Table 1). The response to a pulse of light consists of brief hyperpolarizations typically to both the initiation and the termination of light (Fig. 8F). The response of a BLG2 cell to motion usually begins with a hyperpolarization that is followed by several irregular bursts of spikes (Fig. 8G). The depth and duration of the hyperpolarization, as well as the bursts of activity, vary among cells. However, for any single cell, responses to a given stimulus are fairly consistent. This consistency allowed us to distinguish truly evoked responses from random spontaneous bursts during a stimulus sweep. When faced with the looming stimulus, BLG2 neurons showed poor responses, which failed to reflected the dynamic of the image expansion (Fig. 8H). Analyses of responses to stimuli in the four screens showed that these neurons possess broad receptive fields. The mean percentage of responses elicited from the most effective screen was 34%. Only 2 of 21 neurons concentrated >50% of the response in a single screen. Thus as in MLG2 neurons, the sensitivity to motion is quite homogeneously distributed across the receptive field of BLG2 cells (Fig. 8D). The mean (±SE) index of directionality was 0.25 ± 0.05, indicating the cells have no directional preferences (Fig. 8I). Responses to horizontal and vertical motion were very similar (Fig. 8J). These neurons are extremely sensitive to mechanical stimulations applied with the paintbrush (Fig. 8K), and also show bursts of spikes when the animal moves its legs.
Additional properties of the four cell groups
Because our recordings were obtained in cells that were likely impaled at a variety of sites across the neuritic arborization, the preceding physiological characterizations were made by taking into account only the number of elicited spikes, a parameter that is unaffected by variation in the recording site. We then analyzed some intrinsic cell properties of the identified classes and found that, despite any inconsistency resulting from variation in recording sites, further distinctions between the elements of these groups can be made (Table 1). Representative examples in Fig. 9 illustrate the differences between typical spikes of each cell type (see also Figs. 5–8). Of note, the large spikes (corresponding to recordings taken near the spike-initiation zone) from MLG2 normally have an afterhyperpolarization component that was rarely observed in the other neurons. On the other hand, only spikes in BLG1 neurons exhibit longer repolarizing phases than their depolarizing counterparts. Consequently, properties such as the shape of the spike are also helpful for recognizing the cell identity. A comprehensive summary of characteristics that allow differentiation among the four cell classes is provided in Table 2.
|
|
|
DISCUSSION |
|---|
|
; Marder and Calabrese 1996), command neurons (Krasne and Lee 1988; Shirinyan et al. 2006), neurobiological correlates of social behavior (Heberholz et al. 2001; Yeh et al. 1996, 1997), polarization sensitivity, motion detection, and flow field analysis (Glantz and Schroeter 2004; Miller et al. 2003), and on learning and memory (Merlo and Romano 2007; Pedreira and Maldonado 2003; Shirinyan et al. 2007). Most of these behaviors are guided or strongly affected by visual information. However, despite the fact that Crustacea represents a major phylogenetic group, knowledge of the visual nervous system of these animals is still scarce. In pioneering works, Wiersma and coworkers recorded movement-sensitive fibers from the protocerebral tract of different crustacean species (e.g., Waterman et al. 1964; Wiersma 1966; Wiersma and Yamaguchi 1967a,b; York and Wiersma 1975). Several different types of motion fibers were distinguished according to the receptive field, velocity sensitivity, response to mechanical stimulation, speed of habituation, and other factors (reviewed in Wiersma et al. 1982). However, because the recordings were taken extracellularly, the location and morphology of these cells remained unknown.
We studied a generic group of motion-sensitive cells of the crab Chasmagnathus that had been previously implicated in the escape of the animal to visual danger stimuli (Oliva et al. 2007). These neurons are also involved with short- and long-term memories related to repeated visual stimulation (Berón de Astrada and Tomsic 2002; Tomsic et al. 2003). By performing in vivo intracellular recording and staining, we demonstrate that this group is indeed composed of several classes of neurons, four of which could be repeatedly recorded and characterized both morphologically and physiologically. Cells of all classes showed strong responses to motion, had wide dendrites in the lobula, and projected their axons toward the midbrain. Moreover, we found that neurons of the four classes showed much higher sensitivity for single object than for wide-field motion; they behave as object detectors rather than flow field analyzers.
Despite these commonalties, the four classes clearly differed in their morphology and response preferences. Analysis of physiological properties in morphologically identified units allowed us to establish functional features of each cell group. Our data combine information about the size of the receptive field, the response to a flash of light, the sensitivity to motion features (e.g., motion direction), the rate and type of spontaneous activity, the sensitivity for mechanical stimulation, and the shape of the spikes (see Tables 1 and 2 and Fig. 9). With these traits, we are often able to reliably predict the cell type in the absence of morphological data.
Nevertheless, throughout this study we recorded from a number of unstained motion-sensitive neurons that could not be ascribed with certainty to any of the four classes. Further investigations varying the size, velocity, and other stimulus parameters might reveal more and perhaps deeper differences among LG neurons. Results from the current study pose specific questions about each particular type of LG neuron. Even so, our data are sufficient to confidently hypothesize that the different classes of LG neurons represent matched filters that contribute to visual assessment of moving objects.
Flow field versus object motion
Activities like walking, swimming, or flying generate a flow field across the retina that informs the animal about its translational and rotational motion. Alternatively, a discrete image moving across the retina tells the animal about the motion of an external object. In arthropods, visually elicited behaviors in response to flow field and to object motion are widely documented (Borst and Haag 2002; Tammero and Dickinson 2002; Zeil and Hemmi 2006). In fact, much of what we know about the computational neuronal processing underlying the optokinetic response is derived from studies in flies (Grewe et al. 2006). In these animals, flow field analysis is performed by large tangential cells in a neuropil separate from the lobula called the lobula plate (e.g., Gauck and Borst 1999). Crabs show strong optokinetic responses and, even though their visual system is homologous to that of insects (Strausfeld 2005), they are thought to lack a lobula plate. Therefore we searched for large tangential cells in the lobula of the crab that might support the optokinetic response observed in crabs. Clearly no such neurons could be found through our experiments. This raises the question of which part of the crab's nervous system performs the analysis of the optic flow field. Recently, we found the existence of a small neuropil located next to the lobula of the crab. On the basis of anatomical evidence, we proposed that it may be homologous to the lobula plate of the fly (Sztarker et al. 2005). We have not yet recorded from neurons in this neuropil. However, our inability to find neurons tuned to flow field stimulation in the lobula circumstantially supports the hypothesis that optic flow information could be analyzed in the newly discovered neuropil. If this is true, the visual processing of self-motion and external motion would be processed through segregated pathways within the crab's brain.
What are different tangential lobula neurons good for?
Adaptive behavior requires that crabs, like other animals, extract information from the visual scene. Among other things, visual information allows crabs to distinguish whether a moving object is a charming conspecific or a perilous predator. In addition to shape and size, visual assessment requires information about where and how (in which direction, velocity, etc.) the movement occurred with respect to the observer. It is not plausible that all this information is processed in a single neuron, nor is it likely that the system contains a limitless number of neurons with each specifically tuned to a unique stimulus. A more parsimonious manner of classifying the stimulus and translating the information into adaptive motor responses would be to use sets of neurons with different feature sensitivities, which through distinct combinations of their relative activation may encode separate motor programs (Ewert 1987, 1997).
There are some indications that this may be the case in the crab. In previous studies, we compared the time course of activity of the large tangential lobula neurons with that of the crab escape response when challenged with a complex stimulus moving overhead. The comparison revealed a remarkable correlation between the temporal profiles of the neuronal and the behavioral responses. Similar analyses comparing response habituation on repeated stimulations at different frequencies rendered even more striking similarities between the neuronal and the behavioral responses (Tomsic et al. 2003). Because we had not yet identified the different classes of lobula tangential neurons, we did not take into account the particular class of cell in our analysis of the escape response. We think, however, that the similarity in the time course of the behavioral and the neuronal activity resulted from the cumulative response of the cell groups. In other words, the escape appeared to be better explained by the combined activity of the group, rather than by the activity of a particular class of tangential lobula neuron.
Still, each class may play a unique role. For example, MLG1, MLG2, and BLG1 neurons all show robust responses to approaching objects (Figs. 5–7) that trigger strong escapes in the crab (Oliva et al. 2007). However, they may serve different purposes. With their wide receptive fields, the activation of MLG2 cells may provide a warning signal that an object is approaching from anywhere in space. Meanwhile, the ensemble of 14 MLG1 neurons that we recorded, each one tuned to a different segment of the visual scene, may serve to locate the object in space and to calculate the precise directionality observed in the escape response of the animal (Oliva et al. 2007). In addition, the clear preference of MLG1 neurons for objects moving in the horizontal plane (Fig. 5J) may function as a mechanism for the judgment of interobject distance in fiddler crabs (Hemmi and Zeil 2003; Zeil and Hemmi 2006). Although speculative, these ideas are congruent with those proposed by Ewert (1987, 1997) to explain the neurophysiological mechanisms that mediate prey-catching and avoidance behaviors in toads.
A major caveat of our study is that we were unable to track the axonal projection of the neurons beyond the protocerebral tract, probably due to the great distance between the optic lobe and the midbrain (20 mm). We observed the axon fiber descending for a short distance along the protocerebral tract, at which point the trace gradually vanished. In several experiments we allowed the trace to diffuse 
8 h in the living animal. Although the mark progressed a little farther into the tract, it never reached the midbrain. Longer diffusion times (>12 h) resulted in the complete loss of staining. Therefore we know that the neurons convey their information directly toward the midbrain, but we are unsure exactly where and how far they project into the system.
In locusts, a lobula giant movement detector neuron (LGMD), which in many respects resembles MLG1 neurons of the crab (Oliva et al. 2007), projects its axon to protocerebral regions. In the protocerebrum it synapses with a contralateral descending neuron (DCMD) that reaches the thoracic ganglia and activates the motor pattern responsible for an avoidance behavior (Rind 1989; Santer et al. 2006). Taking into account that, as in the locust (Matheson et al. 2004), the lag between visual input and the onset of a behavioral response is quite short (100 ms), some of the large lobula neurons of the crab should also convey their signals fairly directly to motor centers in the thoracic ganglia. Studies aimed at investigating the capacity of LG neurons for binocular integration, using stimuli directed at the dorsal rim, rendered quite remarkable results. Responses obtained by stimuli presented exclusively to the ipsilateral eye (the eye from where recording was taken), or to the contralateral eye, were almost identical (Sztarker and Tomsic 2004). One possibility to explain the extraordinary similarity in the responses (so far tested only in the dorsal field of view) is to postulate that one or more classes of LG neurons are reciprocally connected with their counterparts in the opposite lobula (Sztarker 2006). Current experiments aimed at elucidating the neuronal projections into the midbrain and the contralateral optic lobe are being attempted using different dyes and staining protocols.
In previous publications we referred to cells in the large tangential lobula of the crab as movement detector neurons. In light of our recent findings, this nomenclature appears to be rather misleading. In fact, these neurons integrate information from both eyes (Sztarker and Tomsic 2004), from different sensory modalities, and from different sensory fields (Berón de Astrada and Tomsic 2002 and results in this paper). Furthermore, they undergo changes supporting long-term memory (Tomsic et al. 2003). Indeed, rather than simple motion detectors, these neurons appear well suited to play a multifarious functional role in the decision-making processes that underlie visually guided behaviors in the crab.
|
|
GRANTS |
|---|
|
|
|
ACKNOWLEDGMENTS |
|---|
|
|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: D. Tomsic, Laboratorio de Neurobiología de la Memoria, Depto. Fisiología, Biología Molecular y Celular, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Pabellón 2 Ciudad Universitaria (1428), Buenos Aires, Argentina (E-mail: tomsic{at}fbmc.fcen.uba.ar)
|
|
REFERENCES |
|---|
|
Berón de Astrada M, Tomsic D. Physiology and morphology of visual movement detector neurons in a crab (Decapoda: Brachyura). J Comp Physiol A Sens Neural Behav Physiol 188: 539–551, 2002.[CrossRef][Web of Science][Medline]
Borst A, Haag J. Neural networks in the cockpit of the fly. J Comp Physiol A Sens Neural Behav Physiol 118: 419–437, 2002.
Edwards DH, Heitler WJ, Krasne FB. Fifty years of a command neuron: the neurobiology of escape behavior in the crayfish. Trends Neurosci 22: 153–161, 1999.[CrossRef][Web of Science][Medline]
Ewert JP. Neuroethology of releasing mechanism; prey-catching in toads. Behav Brain Sci 10: 337–405, 1987.[Web of Science]
Ewert JP. Neural correlates of key stimulus and releasing mechanism: a case study and two concepts. Trends Neurosci 20: 332–339, 1997.[CrossRef][Web of Science][Medline]
Feld M, Dimant B, Delorenzi A, Coso O, Romano A. Phosphorylation of extra-nuclear ERK/MAPK is required for long-term memory consolidation in the crab Chasmagnathus. Behav Brain Res 158: 251–261, 2005.[CrossRef][Web of Science][Medline]
Fort TJ, García-Crescioni K, Agricola HJ, Brezina V, Miller MW. Regulation of the crab heartbeat by crustacean cardioactive peptide (CCAP): central and peripheral actions. J Neurophysiol 97: 3407–3420, 2007.
Freudenthal R, Romano A. Participation of Rel/NF-kappaB transcription factors in long-term memory in the crab Chasmagnathus. Brain Res 855: 274–281, 2000.[CrossRef][Web of Science][Medline]
Gauck V, Borst A. Spatial response properties of contralateral inhibited lobula plate tangential cells in the fly visual system. J Comp Neurol 406: 51–71, 1999.[CrossRef][Web of Science][Medline]
Glantz RM, Schroeter JP. Analysis and simulation of gain control and precision in crayfish visual interneurons. J Neurophysiol 92: 2747–2761, 2004.
Grewe J, Matos N, Egelhaaf M, Warzecha AK. Implications of functionally different synaptic inputs for neuronal gain and computational properties of fly visual interneurons. J Neurophysiol 96: 1838–1847, 2006.
Hemmi JM, Zeil J. Robust judgement of inter-object distance by an arthropod. Nature 421: 160–163, 2003.[CrossRef][Medline]
Herberholz J, Issa FA, Edwards DH. Patterns of neural circuit activation and behavior during dominance hierarchy formation in freely behaving crayfish. J Neurosci 21: 2759–2767, 2001.
Johnson AP, Barnes WJ, Macauley MW. Effects of light intensity and pattern contrast on the ability of the land crab, Cardisoma guanhumi, to separate optic flow-field components. Vis Neurosci 21: 895–904, 2004.[CrossRef][Web of Science][Medline]
Kaczer L, Pedetta S, Maldonado H. Aggressiveness and memory: subordinate crabs present higher memory ability than dominants after an agonistic experience. Neurobiol Learn Mem 87: 140–148, 2007.[CrossRef][Web of Science][Medline]
Korn H, Faber DS. The Mauthner cell half a century later: a neurobiological model for decision-making? Neuron 47: 13–28, 2005.[CrossRef][Web of Science][Medline]
Krasne FB, Lee SC. Response-dedicated trigger neurons as control points for behavioral actions: selective inhibition of lateral giant command neurons during feeding in crayfish. J Neurosci 8: 3703–3712, 1988.[Abstract]
Locatelli F, Romano A. Differential activity profile of cAMP-dependent protein kinase isoforms during long-term memory consolidation in the crab Chasmagnathus. Neurobiol Learn Mem 83: 232–242, 2005.[CrossRef][Web of Science][Medline]
Lozada M, Romano A, Maldonado H. Effect of morphine and naloxone on a defensive response of the crab Chasmagnathus granulatus. Pharmacol Biochem Behav 30: 635–640, 1988.[CrossRef][Web of Science][Medline]
Marder E, Calabrese RL. Principles of rhythmic motor pattern generation. Physiol Rev 76: 687–717, 1996.
Matheson T, Rogers SM, Krapp HG. Plasticity in the visual system is correlated with a change in lifestyle of solitarious and gregarious locusts. J Neurophysiol 91: 1–12, 2004.
Merlo E, Romano A. Long-term memory consolidation depends on proteasome activity in the crab Chasmagnathus. Neuroscience 147: 46–52, 2007.[CrossRef][Web of Science][Medline]
Miller CS, Johnson DH, Schroeter JP, Myint L, Glantz RM. Visual responses of crayfish ocular motoneurons: an information theoretical analysis. J Comput Neurosci 15: 247–269, 2003.[CrossRef][Web of Science][Medline]
Nalbach HO. Visually elicited escape in crabs. In: Frontiers in Crustacean Neurobiology, edited by Wiese K. Basel: Birkhäuser-Verlag, 1990, p. 165–172.
Nichols MJ, Newsome WT. The neurobiology of cognition. Nature 402: C35–C38, 1999.[CrossRef][Medline]
Oliva D, Medan V, Tomsic D. Escape behavior and neuronal responses to looming stimuli in the crab Chasmagnathus granulatus (Decapoda: Grapsidae). J Exp Biol 210: 865–880, 2007.
Pedreira ME, Maldonado H. Protein synthesis subserves reconsolidation or extinction depending on reminder duration. Neuron 38: 863–869, 2003.[CrossRef][Web of Science][Medline]
Pérez Sáez JM. Origen de la Variabilidad en Respuestas a Estímulos Visuales en el Cangrejo Chasmagnathus. Alteraciones, Ajuste y Posible Reorganización Neural del Sistema Visual (MS thesis). Buenos Aires: Univ. de Buenos Aires, Facultad de Ciencias Exactas y Naturales, 2003.
Preuss T, Osei-Bonsu PE, Weiss SA, Wang C, Faber DS. Neural representation of object approach in a decision-making motor circuit. J Neurosci 26: 3454–3464, 2006.
Rind FC. Identification of directionally selective motion-detecting neurones in the locust lobula and their synaptic connections with an identified descending neurone. J Exp Biol 149: 21–43, 1989.[Web of Science]
Rind FC, Simmons PJ. Seeing what is coming: building collision-sensitive neurones. Trends Neurosci 22: 215–220, 1999.[CrossRef][Web of Science][Medline]
Santer RD, Rind FC, Stafford R, Simmons PJ. Role of an identified looming-sensitive neuron in triggering a flying locust's escape. J Neurophysiol 95: 3391–3400, 2006.
Shirinyan D, Teshiba T, Taylor K, O'Neill P, Lee SC, Krasne FB. Rostral ganglia are required for induction but not expression of crayfish escape reflex habituation: role of higher centers in reprogramming low-level circuits. J Neurophysiol 95: 2721–2724, 2007.[Web of Science]
Strausfeld N, Nässel DR. Neuroarchitectures serving compound eyes of Crustacea and insects. In: Vision in Invertebrates. Handbook of Sensory Physiology, edited by Autrum H. Berlin: GB Springer, 1981, p. 1–65.
Strausfeld NJ. The evolution of crustacean and insect optic lobes and the origins of chiasmata. Arthropod Struct Dev 34: 235–256, 2005.[CrossRef]
Sztarker J. Organización Neuroanatómica y Funcional de Neuronas Involucradas en Diferentes Aspectos del Aprendizaje Visual de Chasmagnathus (PhD thesis). Buenos Aires: Univ. de Buenos Aires, Facultad de Ciencias Exactas y Naturales, 2006.
Sztarker J, Strausfeld NJ, Tomsic D. Organization of optic lobes that support motion detection in a semiterrestrial crab. J Comp Neurol 493: 396–411, 2005.[CrossRef][Web of Science][Medline]
Sztarker J, Tomsic D. Binocular visual integration in the crustacean nervous system. J Comp Physiol A Sens Neural Behav Physiol 190: 951–962, 2004.[Web of Science][Medline]
Tammero LF, Dickinson MH. Collision-avoidance and landing responses are mediated by separate pathways in the fruit fly, Drosophila melanogaster. J Exp Biol 205: 2785–2798, 2002.
Tomsic D. Visual learning in crabs investigated by intracellular recordings in vivo. In: The Crustacean Nervous System, edited by Wiese K. Berlin: Springer, 2002, p. 328–335.
Tomsic D, Berón de Astrada M, Sztarker J. Identification of individual neurons reflecting short- and long-term visual memory in an arthropod. J Neurosci 23: 8539–8546, 2003.
Tomsic D, Maldonado H, Rakitin A. Morphine and GABA: effects on perception, escape response and long-term habituation to a danger stimulus in the crab Chasmagnathus. Brain Res Bull 26: 699–706, 1991.[CrossRef][Web of Science][Medline]
Tomsic D, Pedreira ME, Romano A, Hermite G, Maldonado H. Context-US association as a determinant of long-term habituation in the crab Chasmagnathus. Anim Learn Behav 26: 196–209, 1998.[Web of Science]
Waterman TH, Wiersma CAG, Bush BM. Afferent visual responses in the optic nerve of the crab Podophthalmus. J Cell Comp Physiol 63: 135–155, 1964.[CrossRef][Web of Science]
Wiersma CAG. Integration in the visual pathway of crustacea. Symp Soc Exp Biol 20: 151–177, 1966.[Medline]
Wiersma CAG, Roach JLM, Glantz RM. Neural integration in the optic system. In: The Biology of the Crustacea: Neural Integration and Behavior, edited by Sandeman DC, Atwood HL. New York: Academic Press, 1982, vol. 4, p. 1–31.
Wiersma CAG, Yamaguchi T. The integration of visual stimuli in the rock lobster. Vision Res 7: 197–204, 1967a.[CrossRef][Web of Science][Medline]
Wiersma CAG, Yamaguchi T. Integration of visual stimuli by the crayfish central nervous system. J Exp Biol 47: 409–431, 1967b.
Wu LQ, Niu YQ, Yang J, Wang SR. Tectal neurons signal impending collision of looming objects in the pigeon. Eur J Neurosci 22: 2325–2331, 2005.[CrossRef][Web of Science][Medline]
Yeh SR, Fricke RA, Edwards DH. The effect of social experience on serotonergic modulation of the escape circuit of crayfish. Science 271: 366–369, 1996.[Abstract]
Yeh SR, Musolf BE, Edwards DH. Neuronal adaptations to changes in the social dominance status of crayfish. J Neurosci 17: 697–708, 1997.
York B, Wiersma CAG. Visual processing in the rock lobster crustacea. Prog Neurobiol 5: 127–166, 1975.[CrossRef][Medline]
Zeil J, Hemmi JM. The visual ecology of fiddler crabs. J Comp Physiol A Sens Neural Behav Physiol 192: 1–25, 2006.[Web of Science][Medline]
Zhang J. How to unconfound the directional and orientational information in visual neuron's response. Biol Cybern 63: 135–142, 1990.[CrossRef][Web of Science][Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
TOP
ABSTRACT
INTRODUCTION
