Semi-intact tethered preparations were used to characterize neuronal activity patterns in midbody ganglia of the medicinal leech during crawling. Extra- and intracellular recordings were obtained from identified interneurons and from motor neurons of the longitudinal and circular muscles during crawling episodes. Coordinated activities of nine excitatory and inhibitory motor neurons of the longitudinal and circular muscles were recorded during the appropriate phases of crawling. Thus, during crawling, the leech uses motor output components known to contribute to other types of behavior, such as swimming or the shortening/local bending reflex. Interneurons with identified functions in these other types of behavior exhibit membrane potential oscillations that are in phase with the behavior pattern. Therefore, the recruitment of neuronal network elements during several types of behavior occurs not only at the motor neuron level but also involves interneurons. This applies even to some interneurons that were previously thought to have dedicated functions (such as cells 204 and 208 and the S cell). The function of neuronal circuitries in producing different types of behavior with a limited number of neurons is discussed.

Locomotion is in principle produced by alternating, rhythmic activation of antagonistic muscle groups. Most types of locomotory activity are driven by central pattern generators (CPGs) consisting of neuronal network elements that can act together to pace rhythmic motor output activity without sensory feedback (Delcomyn, 1980; Roberts and Roberts, 1983). In addition, these CPGs are augmented, sometimes strongly, by cyclic sensory feedback (Grillner, 1975; Robertson and Pearson, 1985; Pearson, 1985). The medicinal leech Hirudo medicinalis exhibits two different locomotory activities, swimming and crawling. The two behaviors differ both in cycle period and in the degree of control by sensory feedback. The cyclic undulatory movements that constitute swimming have a period length of 0.4–2.0 s (Kristan et al. 1974a), whereas the period of a crawling step ranges from 3 s to more than 10 s (Stern-Tomlinson et al. 1986). In addition, crawling is strongly influenced by sensory feedback (Baader and Kristan, 1995), whereas swimming is not (Kristan and Stent, 1976).

In general, the leech uses three distinctive muscle layer groups for locomotion, the longitudinal, circular and dorsoventral muscles. Alternating contractions of the dorsal and ventral longitudinal muscles produce undulatory swimming movements (Kristan et al. 1974a,b; Friesen, 1989). Waves of contractions, which are intersegmentally delayed, travel along the animal and propel it through the water. The dorsoventral muscles are quasi-tonically active throughout swimming (Ort et al. 1974; Kristan et al. 1974a) and flatten the animal in the dorsoventral plane. In contrast to swimming, each crawling step requires alternating contractions of two different muscle groups: circular muscles to produce extension and longitudinal muscles to produce contraction (Gray et al. 1938; Stern-Tomlinson et al. 1986; Baader and Kristan, 1995). The two types of locomotory behavior both involve the longitudinal motor neurons but in different activation modes.

The neuronal network generating swimming is well understood (for reviews, see Friesen, 1989; Brodfuehrer et al. 1995) but little is known about the neuronal circuitry that produces crawling, which is a multi-component behavior influenced by sensory feedback (Stern-Tomlinson et al. 1986). Surgical manipulation studies have revealed that the neuronal elements responsible for triggering crawling cycles are not located at the level of individual midbody ganglia but are in the head brain and tail brain (Baader and Kristan, 1995). In addition, these studies indicated a distributed organization of crawling pathways within the central nervous system. Information on whether network components behave in a distributed way or are dedicated to mediating just one type of behavior is crucial to understanding the organization of neuronal networks. In the leech, many interneurons contribute to both the local shortening and the local bending reflex (Lockery and Kristan, 1990b; Wittenberg and Kristan, 1992b). Moreover, many of these interneurons activate a variety of different muscles in a distributed manner and in patterns that appear inappropriate to the observed activity of the animal (Kristan et al. 1995). In contrast, some of the swim interneurons seem to be dedicated to serve only one function, namely swimming. To obtain more information about the neuronal architecture in the leech, it would be very helpful to test previously identified interneurons for involvement in other types of behavior. Consequently, the aim of the current study was to characterize some of the neuronal elements activated during leech crawling in minimally restrained animal preparations by (i) examining the activity patterns of interneurons during crawling of restrained animals, especially the role of previously identified neurons known to participate in other patterns of movement, and (ii) assessing the contribution of the circular and longitudinal motor output to crawling behavior.

It will be shown that the activity patterns of most of the motor neurons supplying circular and longitudinal muscles match the phase of activity associated with crawling, that excitatory input to circular muscles and inhibitory input to longitudinal muscles occur during extension, while inhibitory input to the circular muscles and excitatory input to the longitudinal muscles occur during contraction. At the motor neuron level, there are qualitative and quantitative differences in the control of muscles between crawling and swimming: one longitudinal muscle motor neuron, the L cell, which is inhibited during leech swimming (Ort et al. 1974), is active during the contraction phase of crawling. Several interneurons with well-defined roles during swimming and shortening are also involved in crawling. Swim interneuron cells 204 and 208 are hyperpolarized during the contraction phase of crawling, and they are excited when the leech elongates or swims. Cell 159 and the S cell, two neurons participating in the local bending/shortening reflex, are also excited when the animal shortens during crawling. These results will be discussed in the context of the efficacy of neuronal networks in producing different behavior patterns.

Behavioral preparation

Leeches weighing between 2.5 and 6 g were obtained from laboratory cultures or from commercial suppliers (Leeches USA). A detailed description of the apparatus used to record from active semi-intact preparations is given elsewhere (Baader and Kristan, 1992). In brief, the body wall from 2–5 midbody segments was removed, exposing the ventral nerve cord, which was pinned down onto a small platform. The animal was suspended in a tank filled with saline (Muller et al. 1981), in which its front sucker could attach to a floating ball and its back sucker to a horizontal stage (see Fig. 2A). Recordings from neurons were made using glass electrodes filled with 3 mol l−1 potassium acetate and with resistances ranging from 20 to 40 MΩ. Some electrodes were filled with 5 % Lucifer Yellow (Stewart, 1978) to help visualize the structure of the neuronal processes and identify the neurons. While the tank precluded the illumination of the ganglion with dark-field optics to visualize single neurons, the positioning of the end of a light guide (250 μm in diameter) close to the side of the ganglion enabled many individual neurons to be seen and penetrated. Depending on the vigor of the crawling movements, the preparation also allowed paired intracellular recordings and electrical stimulations to be used to assess connections between neurons. A video camcorder system (Panasonic PV535) recorded both the electrophysiological events and the different types of movement of the animal.

Fig. 1.

Behavioral components of a leech crawling step and the temporal periods used to evaluate neuronal responses. The elongation phase consists of front sucker release (fs−), body elongation and front sucker attachment (fs+); the contraction phase then begins, with segmental shortening, rear sucker release (rs−) and rear sucker re-attachment (rs+). The final component is the interstep interval (Pause). Neuronal responses were measured either as action potential frequencies, calculated during eight 1 s periods (A), or as membrane voltage fluctuations at six fixed time points (B). The color coding for the elongation phase (light gray) and contraction phase (dark gray) and the abbreviations used apply to all the figures in this study.

Fig. 1.

Behavioral components of a leech crawling step and the temporal periods used to evaluate neuronal responses. The elongation phase consists of front sucker release (fs−), body elongation and front sucker attachment (fs+); the contraction phase then begins, with segmental shortening, rear sucker release (rs−) and rear sucker re-attachment (rs+). The final component is the interstep interval (Pause). Neuronal responses were measured either as action potential frequencies, calculated during eight 1 s periods (A), or as membrane voltage fluctuations at six fixed time points (B). The color coding for the elongation phase (light gray) and contraction phase (dark gray) and the abbreviations used apply to all the figures in this study.

Fig. 2.

Video recording showing the crawling step of a tethered leech in the saline-filled tank recorded from above at the times indicated. The insets show simultaneous intracellular recordings from a longitudinal excitor motor neuron (L cell) in midbody ganglion 12. (A) At the beginning of the step, the animal lifted its head sucker from the substratum (i.e. the white ball, marked with 2 cm pieces of black tape) and started to elongate (B). At the end of the elongation phase (C), it re-attached its front sucker (D) and began to contract (E). The horizontal platform under the rear sucker was moved backwards (arrows pointing to the left) just before the contraction phase, which caused tail sucker release (F). As the animal contracted further, the platform was returned to its original position (G, arrows pointing to the right). The animal re-attached the rear sucker to complete the step (H). During this step, the L cell was increasingly inhibited during the elongation phase and maximally excited during the final phase of contraction. L, light guide for illuminating the ganglion; H, holder for the animal; R, recording electrode holder with the portion of the electrode outlined with a dashed line. The hatched area in A indicates the horizontal platform (movable by the experimenter) which served as an attachment point for the back sucker.

Fig. 2.

Video recording showing the crawling step of a tethered leech in the saline-filled tank recorded from above at the times indicated. The insets show simultaneous intracellular recordings from a longitudinal excitor motor neuron (L cell) in midbody ganglion 12. (A) At the beginning of the step, the animal lifted its head sucker from the substratum (i.e. the white ball, marked with 2 cm pieces of black tape) and started to elongate (B). At the end of the elongation phase (C), it re-attached its front sucker (D) and began to contract (E). The horizontal platform under the rear sucker was moved backwards (arrows pointing to the left) just before the contraction phase, which caused tail sucker release (F). As the animal contracted further, the platform was returned to its original position (G, arrows pointing to the right). The animal re-attached the rear sucker to complete the step (H). During this step, the L cell was increasingly inhibited during the elongation phase and maximally excited during the final phase of contraction. L, light guide for illuminating the ganglion; H, holder for the animal; R, recording electrode holder with the portion of the electrode outlined with a dashed line. The hatched area in A indicates the horizontal platform (movable by the experimenter) which served as an attachment point for the back sucker.

Evaluation of neuronal activity patterns

The individual behavioral components of crawling cycles show large temporal variations (Baader and Kristan, 1992). To compare the firing patterns of different neurons, each crawling step was divided into eight temporal components (Fig. 1A). The elongation phase was divided into three parts: the start of elongation covering the 1 s period after front sucker release (fs−), the 1 s period just before the re-attachment of the front sucker (fs+, end of elongation), and a 1 s period at the midpoint between fs− and fs+ (middle elongation). The contraction phase was divided into four components: the 1 s period just after fs+, a 1 s period at the midpoint between fs+ and rear sucker release (rs−), the 1 s period just after rs−, and a further 1 s period before rear sucker re-attachment (rs+). The eighth component was the interstep interval (pause). Depending on the speed of the steps,these periods could partially overlap. The firing frequency of each neuron was calculated during each 1 s window, and the mean frequency over several steps (+S.E.M.) formed the ‘activity profiles’. Where the amplitudes of motor neuron action potentials recorded intracellularly from somata were too small (below 1–2 mV) to be distinguished from background noise, the membrane potential fluctuations of the neurons were measured at six fixed time points during each step cycle instead. These measurements were made at fs−, fs+, the midpoint between fs− and fs+, rs−, rs+, and the midpoint between rs− and fs+ as shown in Fig. 1B. The mean membrane potential fluctuations during the steps (+S.E.M.) formed the activity profiles. A comparison of both kinds of evaluation, action potential frequency and membrane potential, performed using the same neuron (a cell 3 motor neuron, data not shown), revealed no difference in the principal shape of the curves. To ease readability of the profiles, their values were connected by dashed lines.

Evaluation of recordings and identification of neurons

Electrophysiological data were stored on tape (model 4000, Vetter) and evaluated with commercially available software (Superscope, GW Instruments) on a computer system. Some recordings were digitized from video image sequences (Video Vision Studio, Radius, San Jose; Photoshop, Adobe, Mountain View).

All motor neurons were identified by their position in the appropriate neuromere within the ganglion, by their corresponding spikes in the nerve roots and by their effects, following intracellular stimulation, on motor units in other nerves recorded extracellularly. Interneurons in midbody ganglia were identified by the position of their somata in the ganglion and by their known physiological and morphological properties. All motor neurons to circular and longitudinal muscles and all interneurons investigated in this study were additionally characterized by injection with the fluorescent dye Lucifer Yellow. Their major neurites were traced using a camera lucida.

Leech anatomy

Each leech midbody ganglion contains approximately 400 neurons. It is divided into six ‘packets’ on the dorsal and ventral side (see Muller et al. 1981, for a ganglionic map). Adjacent ganglia are linked by a pair of connectives, each containing several thousand axons, and an additional unpaired connective (Faivre’s nerve) containing approximately 100 axons (Wilkinson and Coggeshall, 1975). Nerve roots connect the central nervous system to the periphery. Two lateral roots on each side of each ganglion divide into four main branches: the medial branch of the anterior (MA) nerve, the anterior branch of the anterior (AA) nerve, the dorsal branch of the posterior nerve (DP) and the posterior branch of the posterior nerve (PP) (Ort et al. 1974).

The motor pattern for crawling

Gross characterization of the crawling motor output

Tethered animals produced crawling episodes either spontaneously or following slight tactile or electrical stimulation of the body wall. A video sequence of a typical step, together with the intracellular recording from a longitudinal motor neuron of midbody segment 12, the L cell, is shown in Fig. 2. A step began with the release and lifting of the front sucker (Fig. 2A). When the animal elongated (Fig. 2B,C), activity in the L cell decreased until it was inactive at the point of full elongation of the animal. After the front sucker had re-attached to the substratum (Fig. 2D), the animal started to contract. At this point, the horizontal platform was slowly moved backwards (i.e. to the left in Fig. 2D,E; see arrows pointing to the left) to mimic a forward movement of the animal (Baader and Kristan, 1992) and passively stretching the rear part of the animal. The firing frequency in the L cell increased to its maximum as the contraction wave arrived at the midbody and the rear sucker released (Fig. 2F). The platform was returned to its original position (arrows pointing to the right, Fig. 2G) while the whole animal contracted further and finally re-attached its tail sucker to the platform (Fig. 2H) at the end of the crawling cycle.

Crawling steps are produced by contractions of circular and longitudinal muscles (Gray et al. 1938; Stern-Tomlinson et al. 1986; Baader and Kristan, 1992). To determine the activity patterns of many motor neurons simultaneously, extracellular recordings were obtained from the four main nerves of the midbody segments as tethered leeches crawled (Fig. 3). The recordings were obtained from segments with their body wall removed, which ensured that only efferent neuronal activity was recorded. All motor neurons to the longitudinal muscles run in the DP, PP and/or AA nerves, and none extends into the MA nerve (Ort et al. 1974). No motor neurons to circular muscles and only two excitatory motor neurons to longitudinal muscles, cell 3 and the L cell, connect to their muscles via the DP nerve. Conversely, all circular muscle motor neurons recorded in this study projected in the MA nerve (see Fig. 5) but never in the DP or PP nerves. In 16 intracellular recordings of the ventral circular (CV) cell and simultaneous extracellular recordings of the contralateral MA nerve, the CV cell produced the largest spike in the MA nerve. The only other known motor neurons which have axons running in the MA nerve are those innervating dorsoventral muscles, and only one of them, cell 109, produces large spikes in the MA nerve (Ort et al. 1974). This cell is not active during crawling (W. B. Kristan Jr, personal communication). Therefore, motor activity in neurons supplying longitudinal and circular muscles could be distinguished by recording from the MA and DP nerves, with the former carrying mostly axons to circular and flattener muscles and the latter bearing the spikes of cell 3, usually the largest unit in a DP recording (Kristan et al. 1974a), and the L cell. In the two step cycles shown in Fig. 3A, neuronal activity is correlated with all the phases of the crawling process as determined from video recordings. With the exception of the DP nerve (compare Fig. 3A and Fig. 3B), all preparations showed similar firing patterns in most of the units displayed in Fig. 3. In the MA nerve, several spike types were usually visible, most of them occurring from the middle of elongation through early contraction. One of the large units belonged to a ventral circular (CV) motor neuron (see Fig. 5B). For the three other nerves, the probable identity of large and small units was determined using the detailed description by Ort et al. (1974) for reference. The larger spikes in the AA nerve recording occurred during the late contraction phase. Ort et al. (1974, Figs 12, 14) identified a motor neuron to the longitudinal muscles (dorsomedial excitor cell 107) producing large spikes in the AA nerve. The activity patterns of units in the DP nerve varied considerably. In the recording from DP during the crawling sequence depicted in Fig. 3A, a single large spike type, the dorsal longitudinal muscle motor neuron cell 3, started to fire in the middle of the contraction cycle and continued firing to the end of the step cycle. No L cell activity was detected in this nerve recording. A somewhat different firing pattern is shown in a second DP nerve recording (Fig. 3B) during crawling. This pattern was observed in approximately 50 % of the preparations. In addition to the expected burst during contraction, a longer burst was observed during elongation. These additional discharges were often, but not always, associated with very strong crawling episodes and did not appear to be correlated with any other change of behavior in the crawling animal. In the PP nerve recording (Fig. 3A), bursts of impulses from a number of different neurons are visible. Larger-amplitude spikes reached a maximum during the contraction phase, while most smaller units were visible during elongation. According to the detailed characterization by Ort et al. (1974; Table 1), three inhibitory longitudinal muscle motor neurons (the dorsal inhibitor cell 1 and cell 102 and the ventral inhibitor cell 2) produce small units in the PP nerve, while the longitudinal muscle ventromedial (cell 4) and dorsal (cell 5) excitor neurons produce larger impulses in PP.

Fig. 3.

Extracellular recordings of lateral ganglionic nerves during crawling steps. The stage of movement of the animal displayed beneath each set of recordings was obtained from simultaneous video recordings. (A) Recordings from the four main branches. The majority of the excitatory motor bursts of the circular neurons [large units in the MA (12R) nerve] occurred when the elongation wave passed through the twelfth segment (at mid-elongation), and a corresponding activation of dorsolongitudinal motor neuron activity [cell 3, source of the largest spikes in DP (11L)] occurred during contraction. For a description of other units in the AA nerve and in the PP nerve, see text. (B) A frequently observed variant of the firing pattern of cell 3. In addition to the expected burst during contractions, there was an additional longer burst during elongations.

Fig. 3.

Extracellular recordings of lateral ganglionic nerves during crawling steps. The stage of movement of the animal displayed beneath each set of recordings was obtained from simultaneous video recordings. (A) Recordings from the four main branches. The majority of the excitatory motor bursts of the circular neurons [large units in the MA (12R) nerve] occurred when the elongation wave passed through the twelfth segment (at mid-elongation), and a corresponding activation of dorsolongitudinal motor neuron activity [cell 3, source of the largest spikes in DP (11L)] occurred during contraction. For a description of other units in the AA nerve and in the PP nerve, see text. (B) A frequently observed variant of the firing pattern of cell 3. In addition to the expected burst during contractions, there was an additional longer burst during elongations.

Fig. 4.

A comparison of the neuronal responses from selected excitatory and inhibitory motor neurons during crawling. Mean firing frequencies or mean membrane potential (Vm) fluctuations (+S.E.M.) are shown for the number of steps indicated.

Fig. 4.

A comparison of the neuronal responses from selected excitatory and inhibitory motor neurons during crawling. Mean firing frequencies or mean membrane potential (Vm) fluctuations (+S.E.M.) are shown for the number of steps indicated.

Fig. 5.

(A) Camera lucida drawings of Lucifer-Yellow-stained circular muscle neurons of ventral midbody ganglia. The scale bar applies to all drawings. (B) Intracellular recordings from cells 166 and CV on the same side of a midbody ganglion, together with an extracellular recording from the contralateral MA nerve. Depolarizing cell 166 (solid bar) inhibited the CV cell and turned off the largest spikes in the MA recording, which were those of the CV cell (see black dots). The high-frequency burst of somewhat smaller impulses was from cell 166, which was being depolarized. (C) Averaged inhibitory response of a CV cell to eight depolarizing stimulations of cell 166 in the same half-ganglion. Pairwise recordings were performed in six different preparations. (D) Activity of the CV cell during swimming. When the animal elongated (Elo, see bottom trace), the DP nerve was depolarizing, triggering a bout of swimming (curved line at bottom trace) and depolarizing the CV motor neuron. Units in both the DP and MA nerve (see black dots) were activated in swimming rhythm together with the CV cell. (E) Activity of the CV cell during shortening. The head of the animal was stimulated mechanically (stimulus applied at dashed line) as the animal elongated during crawling (Elo, at bottom trace), triggering a shortening response. The CV cell was inhibited during this behavior (Beh) and depolarized again during the subsequent elongation phase.

Fig. 5.

(A) Camera lucida drawings of Lucifer-Yellow-stained circular muscle neurons of ventral midbody ganglia. The scale bar applies to all drawings. (B) Intracellular recordings from cells 166 and CV on the same side of a midbody ganglion, together with an extracellular recording from the contralateral MA nerve. Depolarizing cell 166 (solid bar) inhibited the CV cell and turned off the largest spikes in the MA recording, which were those of the CV cell (see black dots). The high-frequency burst of somewhat smaller impulses was from cell 166, which was being depolarized. (C) Averaged inhibitory response of a CV cell to eight depolarizing stimulations of cell 166 in the same half-ganglion. Pairwise recordings were performed in six different preparations. (D) Activity of the CV cell during swimming. When the animal elongated (Elo, see bottom trace), the DP nerve was depolarizing, triggering a bout of swimming (curved line at bottom trace) and depolarizing the CV motor neuron. Units in both the DP and MA nerve (see black dots) were activated in swimming rhythm together with the CV cell. (E) Activity of the CV cell during shortening. The head of the animal was stimulated mechanically (stimulus applied at dashed line) as the animal elongated during crawling (Elo, at bottom trace), triggering a shortening response. The CV cell was inhibited during this behavior (Beh) and depolarized again during the subsequent elongation phase.

Activity profiles of motor neurons during crawling behavior

For a more thorough characterization of motor neurons during crawling behavior, intracellular recordings of representative longitudinal motor units in crawling animals were obtained. Each profile in Fig. 4 was generated from one animal (L cell profile, three animals) and was confirmed in at least two other preparations. Recordings were performed in midbody ganglia 10–13 and were verified by simultaneous extracellular recordings. The upper four profiles in Fig. 4 show excitatory longitudinal muscle motor neurons (cells 3, 4, 5 and the L cell). The two dorsal excitors (cells 3 and 5) showed a similar decrease in activity as the animal elongated, with a minimum during fs+, followed by an increase to maximum membrane potential between the middle and end of contraction. The same basic pattern but with a longer activation throughout the whole contraction phase was exhibited by the L cell, which excites all longitudinal muscles. The ventral longitudinal muscle excitor neuron (cell 4, fourth panel in Fig. 4) can be easily identified by its ipsilateral arborizations (Ort et al. 1974; Norris and Calabrese, 1987). It fired somewhat out of phase with the dorsal excitors, with maximal firing frequency being recorded before fs+ and with activity falling to a minimum after rs−. Two inhibitors of the dorsal and ventral longitudinal muscles (cell 1 and cell 2, respectively) had similar activity profiles, with oscillations in phase with cell 4 and in anti-phase with cells 3 and 5 and the L cell. Thus, the longitudinal muscle excitors (cells 3, 5 and the L cell) fired maximally during contraction and were inhibited during or slightly before front sucker attachment, which is when the longitudinal muscle inhibitors were showing maximal activity.

Four pairs of motor neurons have been reported to innervate the circular muscles of each segment (Stuart, 1970). Their axons run through the anterior nerve roots of the ganglion and divide up each segment into dorsolateral (cell 112) and ventrolateral (cells 11, 12 and CV) motor fields. The motor fields in each segment overlap with those of adjacent segments. Two additional neurons are described in this study as candidate motor neurons: cell 152 is a putative circular muscle excitor, while cell 166 inhibits circular muscle activity (see below). Their morphologies are shown together with that of the CV cell in Fig. 5A. The neurons share common features since each sends an axon out of the ganglion contralaterally through the medial anterior (MA) nerve; they each have two fields of dendritic arborizations, with the ipsilateral branches having fine endings, while the contralateral dendrites are coarser and sometimes end in blebs. The soma of the CV cell was always in the medial anterior packet on the ventral surface. The somata of cells 152 and 166 were both found in the lateral anterior packet on the ventral surface and, while the cell body of 152 was often near the nociceptive mechanosensory neuron, N2, the soma of cell 166 was either medially or laterally sited. Both the circular muscle excitor CV cell and cell 152 increased their activity as elongation progressed and were shut off during contraction (Fig. 4). The activity profile of cell 166 (Fig. 4, bottom panel) showed a broad minimum during elongation and a maximum during early contraction.

The two larger units in the extracellular recording of the MA nerve (14L) in Fig. 5B are produced by to the CV cell (14R, see second trace) and the ipsilateral cell 166 (14R, third trace). Recordings were obtained from a tethered, non-walking (quiescent) preparation. Cell 166 inhibited the CV cell (in the eight preparations examined) but was not affected by electrical depolarization of the CV cell itself (data not shown). Depolarization of cell 166 in isolated ganglia chains with pieces of body wall attached (Fig. 5C) had the same effect as in crawling semi-intact preparations, resulting in hyperpolarization of the CV cell. The efficacy of these cells in producing muscle contractions was tested in six body wall preparations. Intracellular stimulation of cell 152 caused circular contractions of the body wall from the dorsal midline to its lateral edge. Maximum contractions in response to electrical stimulation of cell 152 were achieved at firing frequencies between 12 and 20 Hz. Intracellular stimulation of cell 166 reduced circular muscle contractions caused by stimulation of the CV cell (data not shown). In conclusion, both cell 152 and cell 166 shared many of the morphological and physiological features of leech motor neurons, but further experiments are needed to prove their identity.

In addition to crawling, the circular muscle excitor CV is also active during other types of behavior. During swimming (Fig. 5D), the CV cell (shown together with other units in the recording from the MA nerve) was excited and produced rhythmic bursts of action potentials matching impulse bursts of the dorsal longitudinal muscle excitors in the DP nerve. These bursts had maximal firing frequencies of 13–18 Hz, which were similar to those seen during the elongation phase of crawling and which produced visible contractions of circular muscles in body wall preparations. When the front of the animal was stimulated mechanically to induce shortening, the CV cell was inhibited. In Fig. 5E, the animal was in the process of elongating (the degree of elongation is indicated by the gray shaded areas) when the stimulus was delivered. Hence, the CV motor neuron participates in at least two different types of behavior (crawling and swimming) and is actively inhibited during a third.

Responses of identified interneurons during crawling behavior

Swim interneurons

The swim-initiating interneuron cell 204 (Weeks and Kristan, 1978) and the oscillator neuron cell 208 (Weeks, 1982b) are two well-identified elements of the swim-generating circuitry. The membrane potential of both cells varied rhythmically in crawling animals. The action potential frequency of cell 208 (Fig. 6A) was elevated during the elongation phase and decreased when the animal contracted. The response of cell 204 (Fig. 6B) was similar, with inhibition during the contraction phase and excitation during elongation. The firing frequency of cell 204 when the animal was not crawling was between 2 and 5 Hz in four preparations. A comparison of the activity of cell 204 during crawling and swimming revealed a higher mean firing frequency during swimming (13.2±1.4 Hz, mean ± S.D., N=10 bouts of swimming) than when the animal crawled (9.2±2.6 Hz, N=13 steps). It has been shown that cell 204 receives excitation from tactile stimulation of the body wall of the same segment (Weeks and Kristan, 1978). Fig. 6C shows the changes in firing frequency of cell 204 in ganglion 12 in response to gently touching the head or the tail with a glass rod.

Fig. 6.

Activity of swim interneurons during crawling. (A) The firing frequency in cell 208 decreased during the contraction phase. (B) Cell 204 was activated during elongation and inhibited during contraction. Values are means + S.E.M. (C) Slight mechanical stimulation of the body wall with a glass rod resulted in an excitation of cell 204 in ganglion 12 when the tail was stimulated, but an inhibition when the head was stimulated. Note that, owing to the sampling intervals (250 ms), changes in the averaged frequency responses are partly visible before the stimulus occurred.

Fig. 6.

Activity of swim interneurons during crawling. (A) The firing frequency in cell 208 decreased during the contraction phase. (B) Cell 204 was activated during elongation and inhibited during contraction. Values are means + S.E.M. (C) Slight mechanical stimulation of the body wall with a glass rod resulted in an excitation of cell 204 in ganglion 12 when the tail was stimulated, but an inhibition when the head was stimulated. Note that, owing to the sampling intervals (250 ms), changes in the averaged frequency responses are partly visible before the stimulus occurred.

The responses to four stimulations of the head and of the tail are shown. Spike frequencies were counted at 250 ms intervals using a frequency counter, and the curves represent averaged frequency responses. The cell was depolarized with 0.5 nA throughout this experiment to induce spike activities comparable to those observed during crawling (see Discussion). After stimulation of the tip of the head, cell 204 was transiently inhibited, while stimulation of the dorsal part of the tail sucker produced an excitation.

Other neurons of midbody ganglia contributing to the crawling pattern

In searching for crawling-correlated neuronal activity in midbody ganglia, many identified neurons were found that were apparently unaffected by the crawling rhythm. Among these were the Leydig cells and the HE motor neuron of the heartbeat circuitry.

While crawling and swimming are incompatible behavior patterns, crawling and the shortening reflex (Kristan et al. 1982; Wittenberg and Kristan, 1992a,b), or crawling and local bending (Kristan, 1982; Lockery and Kristan, 1990a), are not. Both reflexes are highly compatible and show considerable overlap in their neuronal design. The neuronal circuitry producing the two withdrawal reflexes could, at least in part, be used during the contraction phase of crawling, e.g. during very rapid crawling cycles. The S cell (Gardner-Medwin et al. 1973; Peterson, 1984) is an interneuron which contributes to the shortening response of the leech (Magni and Pellegrino, 1978; Kramer, 1981; Sahley et al. 1994) without being by itself sufficient to induce the behavior (A. P. Baader, personal observation; Shaw and Kristan, 1995). The response of the S cell during crawling is shown in Fig. 7. The cell was maximally excited during crawling contractions after front sucker attachment. The anterior pagoda (AP cell) is a large neuron in the anterior lateral packet (Sunderland, 1980) that receives sensory monosynaptic input from the medial pressure-sensitive (P) cell (Gu et al. 1991) and is excited by the contralateral lateral touch-sensitive (T) cell (Sunderland, 1980). During crawling, the AP cell had a distinct response with no action potentials during elongation and tonic low-frequency spikes during the contraction phase, just until the animal initiated each new step (Fig. 7).

Fig. 7.

Activity of identified interneurons during crawling behavior. In the S cell, burst-like activity occurred during body contractions, but not during elongations. Similarly, AP neuron activity was only expressed during the contraction phase. The activity of a local bending interneuron, cell 159, during crawling was characterized by a minimum during the elongation phase and increasing excitation as the animal shortened. The amplitude of the membrane fluctuations in cell 151 had its maximum at the end of the elongation phase and a minimum when the animal was fully contracted. Membrane potential (Vm) changes in cell 151 were measured relative to the value at the beginning of each step (Vm at fs−=0 mV). Cells 258 and 213 were maximally excited during the elongation phase. This excitation decreased as soon as fs+ had occurred. Values are means + S.E.M.

Fig. 7.

Activity of identified interneurons during crawling behavior. In the S cell, burst-like activity occurred during body contractions, but not during elongations. Similarly, AP neuron activity was only expressed during the contraction phase. The activity of a local bending interneuron, cell 159, during crawling was characterized by a minimum during the elongation phase and increasing excitation as the animal shortened. The amplitude of the membrane fluctuations in cell 151 had its maximum at the end of the elongation phase and a minimum when the animal was fully contracted. Membrane potential (Vm) changes in cell 151 were measured relative to the value at the beginning of each step (Vm at fs−=0 mV). Cells 258 and 213 were maximally excited during the elongation phase. This excitation decreased as soon as fs+ had occurred. Values are means + S.E.M.

Cell 159 was previously identified as one of the local bending interneurons (Lockery and Kristan, 1990b). It receives inputs from all four pressure-sensitive (P) mechanoreceptors of the same ganglion and excites dorsolongitudinal muscles (Lockery and Kristan, 1990b). It can be easily identified by its ventral soma in the anterior lateral packet and its ipsilaterally descending axon (Fig. 8, see also Lockery and Kristan, 1990b), and it was recorded and dye-filled in two crawling preparations. The response of the neuron was modulated during crawling (Fig. 7), showing an increasing firing rate as the animal contracted, to reach a maximum after the completion of each step, and a decrease in firing rate during elongation.

Fig. 8.

Camera lucida drawings of three Lucifer-Yellow-filled interneurons described in Fig. 7. The cell body of a local bending interneuron, cell 159, is located on the ventral ganglion aspect in the lateral anterior packet. Projections ramify within the ipsilateral half-ganglion and the axon descends ipsilaterally. Cell 258 is located in the posterior packet of the ventral side, has mainly dorsally running arborizations and has two axons projecting into the anterior and posterior roots. The cell body of cell 213 is in the medial anterior packet of the ventral side and has an ipsilaterally ascending axon.

Fig. 8.

Camera lucida drawings of three Lucifer-Yellow-filled interneurons described in Fig. 7. The cell body of a local bending interneuron, cell 159, is located on the ventral ganglion aspect in the lateral anterior packet. Projections ramify within the ipsilateral half-ganglion and the axon descends ipsilaterally. Cell 258 is located in the posterior packet of the ventral side, has mainly dorsally running arborizations and has two axons projecting into the anterior and posterior roots. The cell body of cell 213 is in the medial anterior packet of the ventral side and has an ipsilaterally ascending axon.

One previously identified neuron which received modulatory drive during crawling was cell 151, a neuron that has extensive ramifications within the ganglion and into all roots and connectives (Wahdepuhl, 1989). This cell produces no action potentials. The membrane potential is driven by barrages of postsynaptic potentials leading to continuous membrane fluctuations during recording. When electrically hyperpolarized, the neuron inhibits virtually all efferent activity in the ipsilateral and contralateral roots (Wahdepuhl, 1989). In three crawling preparations, the membrane fluctuations showed the pattern displayed in Fig. 7, with a maximal depolarization during fs+ and a minimal depolarization at the end of each step. The paired cell 258 (Figs 7, 8) has a large cell body in the ventral posterior packet and two ipsilaterally running axons in the anterior and posterior roots. This neuron was maximally excited during the elongation phase of crawling (Fig. 7). The ascending interneuron cell 213 was found in the medial anterior packet (Fig. 8); it was recorded from and dye-filled in two preparations. Its instantaneous action potential frequency increased with the elongation of the animal and decreased again immediately after each front sucker attachment (Fig. 7). All cells were stimulated during crawling by injecting depolarizing and hyperpolarizing current into their somata. In no case was a visible change in the behavior pattern observed.

Neuronal activity during behavioral switches

In a few preparations, the animal switched spontaneously from crawling to swimming and back again during intra- and extracellular recordings. Fig. 9 shows an intracellular recording of a longitudinal muscle dorsal inhibitor motor neuron (cell 1) together with the extracellular recordings of the MA nerve and the DP nerve. The spike in the DP recording was identified as that of the dorsolongitudinal excitor motor neuron cell 3 by simultaneously recording from its cell body (not shown) just before the recording shown in Fig. 9. The MA nerve recording contains excitatory circular muscle neurons (see Figs 3, 5B). The bottom trace in Fig. 9 illustrates the behavior of the animal taken from video recordings. The transition between the two types of locomotion occurred consistently at defined fixed points of behavioral activity: the animal initiated a swimming episode at the end of an elongation phase and switched back to crawling by contracting following the swim. Cell 1 was increasingly depolarized during elongations and less depolarized when the animal contracted. The response of the neuron during these switches in behavior reflects exactly what would have been expected from its known activities during each type of behavior, swimming and crawling.

Fig. 9.

Changes in the neuronal firing pattern during switches in behavior. Intracellular recording of a dorsal inhibitor of longitudinal muscles (cell 1) in ganglion 11, together with extracellular recordings of the MA nerve and the DP nerve, and a record of the behavioral states of the animal. Excitation in cell 1 coincided with excitation of circular units in the MA nerve, with depression in the DP nerve (units represent longitudinal muscle excitor cell 3) and with the elongation (Elo) of the animal. The animal started a swimming episode from a crawling elongation, and switched back to crawling by producing a body contraction (Contr). During swimming, excitation in the MA nerve remained elevated, while rhythmic swim oscillations occurred in cell 1 and in the DP nerve.

Fig. 9.

Changes in the neuronal firing pattern during switches in behavior. Intracellular recording of a dorsal inhibitor of longitudinal muscles (cell 1) in ganglion 11, together with extracellular recordings of the MA nerve and the DP nerve, and a record of the behavioral states of the animal. Excitation in cell 1 coincided with excitation of circular units in the MA nerve, with depression in the DP nerve (units represent longitudinal muscle excitor cell 3) and with the elongation (Elo) of the animal. The animal started a swimming episode from a crawling elongation, and switched back to crawling by producing a body contraction (Contr). During swimming, excitation in the MA nerve remained elevated, while rhythmic swim oscillations occurred in cell 1 and in the DP nerve.

Characteristics of the motor output

The motor output for crawling is characterized by several interesting features: all of the recorded motor neuron types respond at the appropriate phases, and a concerted activity of excitatory and inhibitory motor neurons shapes the behavioral elements. There is a significant portion of inhibitory motor output during the elongation and contraction phases of each crawling step, in the same way that longitudinal muscle inhibitors provide the major inhibitory drive onto the longitudinal muscle excitors during swimming (Granzow and Kristan, 1986) and local bending (Lockery and Kristan, 1990a). In the longitudinal muscle motor neuron network, the bilaterally homologous cells have electrical connections (Ort et al. 1974) which aid the co-activation of bilateral longitudinal muscles during contractions. At least two different types of behavior, swimming and the shortening reflex, make use of these symmetrical excitatory connections. During both types of behavior, contractions occur straight along the horizontal plane, which requires bilateral activation of longitudinal muscles. This is especially important for rapid, whole-body shortening. During crawling it ensures that the released rear sucker is returned as quickly as possible to the re-attached state. The efficacy of straight body contractions during crawling is increased by the L cell, which activates all longitudinal muscles via strong electrical connections to longitudinal muscle motor neurons (Stuart, 1970). It was, therefore, not surprising to find that the L cell, which is inhibited during swimming (Ort et al. 1974), is strongly activated during whole-body shortening (Shaw and Kristan, 1995). Unlike contractions, body elongations are often not directed in a straight line forwards; the animal curves laterally and searches around before finally re-attaching its front sucker to the substratum. This necessitates differential activation of right and left circular muscles by premotor interneurons to coordinate bilateral motor output to the circular muscles where necessary. In the stick insect, the organization of the motor output is even more strongly separated bilaterally in that the legs are driven by individual oscillating units which are variably coupled depending on the phase range within the stepping cycle (Cruse, 1990).

A plausible explanation for the observed increasing excitation of longitudinal muscle ventromedial excitor cell 4 (Fig. 4) during elongations is that it could well tend to lower the body of the animal and enable the attachment of the front sucker at the end of the elongation phase. If true, a consequence of this should be a corresponding activation of the appropriate muscles during the elongation phase to keep the body off the ground. Indeed, this is exactly what was observed in many preparations in the biphasic response of dorsomedial excitor cell 3 (Fig. 3B).

Cells 152 and 166 share all the morphological features of previously described motor neurons: they are active during crawling at the appropriate behavioral phases and produce muscle contractions in body wall preparations. For these reasons, they have been tentatively added to the list of motor neurons. Future experiments will reveal which muscles they innervate. Stuart (1970) described several other motor neurons in the anterior ganglion that innervate dorsoventral and oblique muscles, but none of them was located in the same packet as cells 152 and 166.

Neuronal activity during different types of behavior

In a number of simple systems, comprising a limited number of neurons, the neuronal circuits control different types of behavior by activating the same interneurons in different patterns (Getting, 1989; Wu et al. 1994). In the leech, interneurons involved in the local bending response (Kristan, 1982; Lockery and Kristan, 1990b) also contribute to the shortening reflex (Wittenberg and Kristan, 1992b). In the present study, neurons involved in one behavior pattern (swimming) are recruited during components of another (the contraction phase of crawling) with respect to the activity of longitudinal and circular muscles. The example of the CV cell shows that circular muscle motor neuron activity is expressed during swimming. The strength of this activity can be compared with that observed during crawling and contradicts past findings that circular muscle motor neurons play no role in undulatory swimming movements (Ort et al. 1974).

Although many behavioral patterns in the leech are produced by distributed processing, some are controlled by dedicated pathways (for a review, see Kristan et al. 1995). The results of this study confirm these findings, but reduce the number of neurons with possibly dedicated functions since both cell 204 and cell 208 have been shown to fire with different rhythms during swimming and crawling. The S cell (Peterson, 1984) is activated during at least two types of behavior, crawling and shortening. Its actual role in crawling, however, remains unclear. The neuron is inhibited during swimming (Weeks, 1982c) and makes strong electrical connections to L motor neurons (Gardner-Medwin et al. 1973). It receives mechanosensory inputs (Gardner-Medwin et al. 1973) which, in S cells of median midbody segments, are stronger when the posterior body parts are stimulated than with tactile stimulation of the anterior body wall. This cell, although morphologically a candidate well-suited to produce longitudinal contractions, does not seem to be directly involved in contractions during crawling since appropriately timed depolarizing current injections into this cell never enhanced the crawling contractions. In any case, an activation of the S cell during contractions affecting the whole ventral nerve cord simultaneously appears to be incompatible with a temporally organized contraction wave travelling from the anterior to the posterior end. In fact, Magni and Pellegrino (1978) did not see S cell activity in slowly crawling unrestrained animals chronically implanted with an extracellular electrode.

Moreover, leeches with a transected Faivre’s nerve (in which the S cell projects) produced normal contractions during crawling (Baader and Kristan, 1995). In contrast, as leeches produce faster crawling sequences, they must contract their body segments at shorter time intervals. It may well be that the S cell pathway is gradually recruited by the contraction motor as more rigidity and speed are required to execute this type of behavior.

Cells 151, 213 and 258 and the AP neuron produced specific responses during crawling. The effects of cell 258 and the AP cell on postsynaptic elements are not known. The inhibitory postsynaptic effects described for cell 151 (Wahdepuhl, 1989) and the way it is modulated during crawling (the present study) correlate well with the observed behavior of crawling leeches: cell 151 repolarizes at the end of each crawling cycle and, provided that this repolarization is of sufficient strength, could be effective in blocking all motor activity, thus terminating a given crawling episode. Activity in the ascending cell 213 of the anterior medial packet has not been described in any other behavioral context. The relative complexity of crawling in comparison with swimming or shortening would predict the occurrence of such additional elements not shared by other types of behavior. Certainly, the number of identified interneurons described here can represent only a small fraction of all the neuronal elements that form the crawling circuitry. More information about the central processing of crawling behavior is needed. Surgical manipulations of the central nervous system of the leech have localized possible centers that could generate crawling in the head and tail ganglia (Baader and Kristan, 1995), i.e. spatially separated from the central pattern generator for crawling in midbody ganglia. Locomotion in mammals is similarly produced by neuronal circuits in the brainstem, distinct regions of which activate spinal CPGs in the motor centers (Arshavsky et al. 1991). In the snail Clione limacina, both oscillator and motor output for locomotion are located in the pedal ganglia (Arshavsky et al. 1985), which are partially under the command of interneurons in cerebral and other ganglia (Arschavsky et al. 1991). In the flight system of the locust, large descending interneurons from the brain control both the wing motor and the steering output of the abdomen (Rowell, 1989; Baader et al. 1992). It appears, therefore, that similar neuronal principles organize locomotion in many systems in both vertebrates and invertebrates.

A possible role of cell 204 during behavioral decisions

The activity of swim-initiating neuron 204 during crawling and shortening has been studied previously, and its possible role in influencing the initiation of the different types of behavior in the leech has been emphasized (Kristan et al. 1988). Cell 204 is able to initiate swimming at firing rates of 10–30 Hz (Weeks and Kristan, 1978). Below this swim-initiating threshold, cell 204 is excited at rates which can be attributed to at least two other types of behavior: (1) firing frequencies just insufficient to elicit swimming cause dorsoventral motor neurons (e.g. cell 109, Weeks, 1982a) to flatten the animal in preparation for a swimming episode; and (2) the same firing rates occur during the elongation phase of crawling. The results shown in Fig. 9 provide some additional constraints which could affect the behavioral switch between crawling and swimming (and vice versa): during crawling, the probability of starting a swimming episode is greatly increased with the elongation of the animal. In contrast, a contraction regularly follows a swim, before the animal initiates a new step. The firing pattern in cell 204 can now be taken as a monitor of the behavioral state of the animal to illustrate the possible scenario. An increase in excitation of cell 204 during elongation (Fig. 6 and Kristan et al. 1988) may activate the dorsoventral musculature to flatten the animal. If a decision is taken to swim, the activity of cell 204 becomes suprathreshold, thus driving the swim oscillator. However, if the decision is taken to carry on crawling (i.e. to shorten), the animal re-attaches its front sucker and this is correlated with a decrease in the excitation of cell 204, reaching a minimum during the contraction phase, which helps to prevent a switch to swimming during this behavioral state. The fact that swimming can be induced by stimulation of the DP nerve even though cell 204 is kept hyperpolarized (Weeks and Kristan, 1978) is not incompatible with its suggested role, but supports the idea that several pathways can parallel the activity of cell 204 (Weeks, 1982a). Two observations suggest that cell 204 is probably not itself the behavioral switch but is only activated by the proposed switch: (1) electrical stimulation of cell 204 during crawling never affected the behavior of the animal; and (2) it is very difficult to induce swimming through electrical stimulation of cell 204 in complete ganglia chains (Brodfuehrer and Friesen, 1986c). In the leech, one possible way of influencing the decision to swim or crawl can be inferred from the results presented in Fig. 6C. Mechanical head stimulation simulating what the animal would experience when, for example, hitting an obstacle during extension would tend to inhibit cell 204 activity, which would favor a decision to crawl and lead to re-attachment of the front sucker rather than to swimming. However, excitatory drive from tactile inputs from the posterior sucker (Fig. 6C) could produce additional excitation in cell 204, thus initiating swimming at the end of an elongation phase. The crawling-modulated oscillations of the swim oscillator cell 208 (Fig. 6A) are in phase with the pattern produced by cell 204. Cell 208 receives monosynaptic input from cell 204 and makes excitatory connections with several excitatory and inhibitory longitudinal and dorsoventral motor neurons (Weeks, 1982b). It also excites cell 152 intraganglionically and produces circular contractions in body wall preparations when electrically depolarized intracellularly (data not shown). The co-activation of seemingly antagonistic muscle groups to produce behavioral patterns in the leech is a feature which was previously discussed for the output connections of local bending and shortening interneurons (Lockery and Kristan, 1990b; Kristan et al. 1995). The results support the proposed function of cell 204 as a neuronal pathway influencing both crawling and swimming in the leech (Kristan et al. 1988). In another invertebrate, the green shore crab Carcinus maenas, one interneuron controls different modes of ventilation and can itself switch between different behavioral states (DiCaprio, 1990). Some invertebrate systems also utilize neuromodulatory factors to modify given connections within central pattern generators and to produce different behavioral patterns (Hooper and Moulins, 1989; Katz et al. 1994).

As mentioned above, the ability of cell 204 to initiate swimming is greatly reduced when the brain in the head is connected to the ventral nerve cord (Brodfuehrer and Friesen, 1986c; Brodfuehrer and Burns, 1995). Descending excitatory (Brodfuehrer and Friesen, 1986a,b) and inhibitory (Brodfuehrer and Burns, 1995) pathways from the brain regulate the decision of the animal to swim or not to swim. It would be interesting to see whether these pathways are also recruited by the crawling circuitry to decide between swimming and crawling. The preparation used in this study did not permit electrophysiological recordings from the head brain. However, it recently became possible to record the crawling motor program from isolated nerve cords (Eisenhart et al. 1995), which will probably make it easier to track down further elements of the crawling circuit.

In conclusion, the concerted activity of an extended ensemble of motor units and of various interneurons to produce rhythmic muscle contractions is appropriately correlated to the different states of the animal during two types of locomotory behavior, crawling and swimming. Several of these neurons are also recruited during shortening and local bending. Thus, crawling can now be added to the list of leech behavior patterns in which neural correlates are widely distributed in segmental ganglia.

I am very grateful to W. B. Kristan Jr for his encouragement and support throughout this project. Supported by a NIH grant to W. B. Kristan Jr and a DFG grant to A.P.B.

Arshavsky
,
Y. I.
,
Grillner
,
S.
,
Orlovsky
,
G. N.
and
Panchin
,
Y. V.
(
1991
).
Central generators and the spatio-temporal pattern of movements
. In
Temporal Organization in Coordinated Action
(ed.
J.
Fagard
and
P. H.
Wolff
), pp.
93
115
.
Amsterdam
:
Elsevier Science Publication BV
.
Arshavsky
,
Y. I.
,
Orlovsky
,
G. N.
and
Panchin
,
Y. V.
(
1985
).
Control of locomotion in the marine mollusc Clione limacina
.
Exp. Brain Res
.
59
,
203
205
.
Baader
,
A. P.
and
Kristan
,
W. B.
Jr
,
(
1992
).
Monitoring neuronal activity during discrete behaviors: a crawling, swimming and shortening device for tethered leeches
.
J. Neurosci. Meth
.
43
,
215
223
.
Baader
,
A. P.
and
Kristan
,
W. B.
Jr
,
(
1995
).
Parallel pathways coordinate crawling in the medicinal leech, Hirudo medicinalis
.
J. comp. Physiol. A
176
,
715
726
.
Baader
,
A. P.
,
Schaefer
,
M.
and
Rowell
,
C. H. F.
(
1992
).
The perception of the visual flow field by flying locusts: a behavioral and neuronal analysis
.
J. exp. Biol
.
165
,
137
160
.
Brodfueher
,
P. D.
and
Burns
,
A.
(
1995
).
Neuronal factors influencing the decision to swim in the medicinal leech
.
Neurobiol. Learning Memory
63
,
192
199
.
Brodfuehrer
,
P. D.
,
Debski
,
E. A.
,
O’gara
,
B. A.
and
Friesen
,
W. O.
(
1995
).
Neuronal control of leech swimming
.
J. Neurobiol
.
27
,
403
418
.
Brodfuehrer
,
P. D.
and
Friesen
,
W. O.
(
1986a
).
Control of leech swimming activity by cephalic ganglia
.
J. Neurobiol
.
17
,
697
705
.
Brodfuehrer
,
P. D.
and
Friesen
,
W. O.
(
1986b
).
Initiation of leech swimming by trigger neurons in the leech subesophageal ganglia. I Output connections of Tr1 and Tr2
.
J. comp. Physiol. A
159
,
489
502
.
Brodfuehrer
,
P. D.
and
Friesen
,
W. O.
(
1986c
).
Initiation of leech swimming by trigger neurons in the leech subesophageal ganglia. II Role of segmental swim-initiating interneurons
.
J. comp. Physiol. A
159
,
503
510
.
Cruse
,
H.
(
1990
).
What mechanisms coordinate leg movement in walking arthropods?
Trends Neurosci.
13
,
15
21
.
Delcomyn
,
F.
(
1980
).
Neural basis of rhythmic behavior in animals
.
Science
210
,
492
498
.
Dicaprio
,
R. A.
(
1990
).
An interneurone mediating motor programme switching in the ventilatory system of the crab
.
J. exp. Biol
.
154
,
517
535
.
Eisenhart
,
F. J.
,
Cacciatore
,
T. W.
,
Kristan
,
W. B.
, Jr
and
Wessel
,
R.
(
1995
).
Crawling in the medicinal leech is produced by a central pattern generator
.
Soc. Neurosci. Abstr
.
21
,
1385
.
Friesen
,
W. O.
(
1989
).
Neuronal control of leech swimming movements
. In
Neuronal and Cellular Oscillators
(ed.
J. W.
Jacklet
), pp.
269
316
.
Basel
:
Marcel Dekker
.
Gardner-Medwin
,
A. R.
,
Jansen
,
J. K. S.
and
Taxt
,
T.
(
1973
).
The ‘giant’ axon of the leech
.
Acta physiol. scand
.
87
,
30A
31A
.
Getting
,
P. A.
(
1989
).
Emerging principles governing the operation of neural networks
.
A. Rev. Neurosci
.
12
,
185
204
.
Granzow
,
B AND
,
Kristan
,
W. B.
Jr
,
(
1986
).
Inhibitory connections between motor neurons modify a centrally generated motor pattern in the leech nervous system
.
Brain Res
.
369
,
321
325
.
Gray
,
J.
,
Lissmann
,
H. W.
and
Pumphrey
,
R. J.
(
1938
).
The mechanism of locomotion in the leech (Hirudo medicinalis Ray)
.
J. exp. Biol
.
15
,
408
430
.
Grillner
,
S.
(
1975
).
Locomotion in vertebrates: central mechanisms and reflex interaction
.
Physiol. Rev
.
55
,
247
304
.
Gu
,
X. N.
,
Muller
,
K. J.
and
Young
,
S. R.
(
1991
).
Synaptic integration at a sensory-motor reflex in the leech
.
J. Physiol., Lond
.
441
,
733
754
.
Hooper
,
S. L.
and
Moulins
,
M.
(
1989
).
Switching of a neuron from one network to another by sensory-induced changes in membrane potential
.
Science
244
,
1587
1589
.
Katz
,
P. S.
,
Getting
,
P. A.
and
Frost
,
W. N.
(
1994
).
Dynamic neuromodulation of synaptic strength intrinsic to a central pattern generator circuit
.
Nature
367
,
729
731
.
Kramer
,
A. P.
(
1981
).
The nervous system of the glossiphoniid leech Haementeria ghiliani. II. Synaptic pathways controlling body wall shortening
.
J. comp. Physiol
.
144
,
449
458
.
Kristan
,
W. B.
Jr
,
(
1982
).
Sensory and motor neurons responsible for the local bending response in leeches
.
J. exp. Biol
.
96
,
161
180
.
Kristan
,
W. B.
, Jr
,
Lockery
,
S. R.
and
Lewis
,
J. E.
(
1995
).
Using reflexive behaviors of the medicinal leech to study information processing
.
J. Neurobiol
.
27
,
380
389
.
Kristan
,
W. B.
, Jr
,
Mcgirr
,
S. J.
and
Simpson
,
G. V.
(
1982
).
Behavioural and mechanosensory neurone responses to skin stimulation in leeches
.
J. exp. Biol
.
96
,
143
160
.
Kristan
,
W. B.
, Jr
and
Stent
,
G. S.
(
1976
).
Peripheral feedback in the leech swimming rhythm
.
Cold Spring Harbor Symp. quant. Biol
.
40
,
663
673
.
Kristan
,
W. B.
, Jr
,
Stent
,
G. S.
and
Ort
,
C. A.
(
1974a
).
Neuronal control of swimming in the medicinal leech. I. Dynamics of the swimming rhythm
.
J. comp. Physiol
.
94
,
97
119
.
Kristan
,
W. B.
, Jr
,
Stent
,
G. S.
and
Ort
,
C. A.
(
1974b
).
Neuronal control of swimming in the medicinal leech. III. Impulse patterns of motor neurons
.
J. comp. Physiol
.
94
,
166
176
.
Kristan
,
W. B.
, Jr
,
Wittenberg
,
G.
,
Nusbaum
,
M. P.
and
Stern Tomlinson
,
W.
(
1988
).
Multifunctional interneurons in behavioral circuits of the medicinal leech
.
Experientia
44
,
383
389
.
Lockery
,
S. R.
and
Kristan
,
W. B.
Jr
,
(
1990a
).
Distributed processing of sensory information in the leech. I. Input–output relations of the local bending reflex
.
J. Neurosci
.
10
,
1811
1815
.
Lockery
,
S. R.
and
Kristan
,
W. B.
Jr
,
(
1990b
).
Distributed processing of sensory information in the leech. II. Identification of interneurons contributing to the local bending reflex
.
J. Neurosci
.
10
,
1816
1829
.
Magni
,
F.
and
Pellegrino
,
M.
(
1978
).
Patterns of activity and the effects of activation of the fast conducting system on the behaviour of unrestrained leeches
.
J. exp. Biol
.
76
,
123
135
.
Muller
,
K.
,
Nicholls
,
J. G.
and
Stent
,
G. S.
(
1981
).
Neurobiology of the Leech
.
Cold Spring Harbor: Cold Spring Harbor Laboratories
.
Norris
,
B. J.
and
Calabrese
,
R. L.
(
1987
).
Identification of motor neurons that contain a FMRFamide-like peptide and the effects of FMRFamide on longitudinal muscle in the medicinal leech, Hirudo medicinalis
.
J. comp. Neurol
.
266
,
95
111
.
Ort
,
C. A.
,
Kristan
,
W. B.
, Jr
, and
Stent
,
G. S.
(
1974
).
Neuronal control of swimming in the medicinal leech. II. Identification and connections of motor neurons
.
J. comp. Physiol
.
94
,
121
154
.
Pearson
,
K. G.
(
1985
).
Are there central pattern generators for walking and flight in insects?
In
Feedback and Motor Control in Invertebrates and Vertebrates
(ed.
W. P. J.
Barnes
and
M.
Gladden
), pp.
307
316
. London: Croom Helm.
Peterson
,
E. L.
(
1984
).
The fast conducting system of the leech: a network of 93 dye-coupled interneurons
.
J. comp. Physiol
.
154
,
781
788
.
Roberts
,
A.
and
Roberts
,
B. L.
(
1983
).
Neural Origins of Rhythmic Movements
.
Cambridge
:
Cambridge University Press. 501pp
.
Robertson
,
R. M.
and
Pearson
,
K. G.
(
1985
).
Neural circuits in the flight system of the locust
.
J. Neurophysiol
.
53
,
110
128
.
Rowell
,
C. H. F.
(
1989
).
Descending interneurones of the locust reporting deviation from flight course: what is their role in steering?
J. exp. Biol
.
146
,
177
194
.
Sahley
,
C. L.
,
Modney
,
B. K.
,
Boulis
,
N. M.
and
Muller
,
K. J.
(
1994
).
The S cell: an interneuron essential for sensitization and full dishabituation of leech shortening
.
J. Neurosci
.
14
,
6715
6721
.
Shaw
,
B. K.
and
Kristan
,
W. B.
Jr
,
(
1995
).
The whole-body shortening reflex of the medicinal leech: Motor pattern, sensory basis and interneuronal pathways
.
J. comp. Physiol. A
177
,
667
681
.
Stern-Tomlinson
,
W.
,
Nusbaum
,
M. P.
,
Perez
,
L. E.
and
Kristan
,
W. B.
Jr
,
(
1986
).
A kinematic study of crawling behavior in the leech, Hirudo medicinalis
.
J. comp. Physiol. A
158
,
593
603
.
Stewart
,
W. W.
(
1978
).
Functional connections between cells as revealed by dye-coupling with a highly fluorescent naphthalimide tracer
.
Cell
14
,
741
759
.
Stuart
,
A. E.
(
1970
).
Physiological and morphological properties of motoneurones in the central nervous system of the leech
.
J. Physiol., Lond.
209
,
627
646
.
Sunderland
,
A. J.
(
1980
).
A hitherto undocumented pair of neurons in the segmental ganglion of the leech which receive synaptic input from mechanosensory cells
.
Comp. Biochem. Physiol
.
67A
,
299
302
.
Wahdepuhl
,
M.
(
1989
).
Depression of excitatory motor neurones by a single neurone in the leech central nervous system
.
J. exp. Biol
.
143
,
509
527
.
Weeks
,
J. C.
(
1982a
).
Synaptic basis of swim initiation in the leech. I Connections of a swim-initiating neuron (cell 204) with motor neurons and pattern-generating ‘oscillator’ neurons
.
J. comp. Physiol
.
148
,
253
263
.
Weeks
,
J. C.
(
1982b
).
Synaptic basis of swim initiation in the leech. I A pattern-generating neuron (cell 208) which mediates motor effects of swim-initiating neurons
.
J. comp. Physiol
.
148
,
265
279
.
Weeks
,
J. C.
(
1982c
).
Segmental specialization of a leech swim-initiating interneuron, cell 205
.
J. Neurosci
.
2
,
972
985
.
Weeks
,
J. C.
and
Kristan
,
W. B.
Jr
,
(
1978
).
Initiation, maintenance and modulation of swimming in the medicinal leech by the activity of a single neuron
.
J. exp. Biol
.
77
,
71
88
.
Wilkinson
,
J. M.
and
Coggeshall
,
R. E.
(
1975
).
Axonal numbers and sizes in the connectives and peripheral nerves of the leech
.
J. comp. Neurol
.
162
,
387
396
.
Wittenberg
,
G.
and
Kristan
,
W. B.
Jr
,
(
1992a
).
Analysis and modeling of the multisegmental coordination of shortening behavior in the medicinal leech. II. Motor output pattern
.
J. Neurophysiol
.
68
,
1683
1792
.
Wittenberg
,
G.
and
Kristan
,
W. B.
Jr
,
(
1992b
).
Analysis and modeling of the multisegmental coordination of shortening behavior in the medicinal leech. II. Role of identified interneurons
.
J. Neurophysiol
.
68
,
1693
1707
.
Wu
,
J. Y.
,
Cohen
,
L. B.
and
Falk
,
C. X.
(
1994
).
Neuronal activity between different behaviors in Aplysia: a distributed organization?
Science
263
,
820
823
.