HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: 93/6/3177    most recent 01248.2004v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (15) Citing Articles via Google Scholar Google Scholar Articles by Masino, M. A. Articles by Fetcho, J. R. Search for Related Content PubMed PubMed Citation Articles by Masino, M. A. Articles by Fetcho, J. R.

Fictive Swimming Motor Patterns in Wild Type and Mutant Larval Zebrafish

Department of Neurobiology and Behavior, Cornell University, Ithaca, New York

Submitted 6 December 2004; accepted in final form 19 January 2005

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Larval zebrafish provide a unique model for investigating the mechanisms involved in generating rhythmic patterns of behavior, such as swimming, due to the array of techniques available including genetics, optical imaging, and conventional electrophysiology. Because electrophysiological and imaging studies of rhythmic motor behaviors in paralyzed preparations depend on the ability to monitor the central motor pattern, we developed a fictive preparation in which the activity of axial motor neurons was monitored using extracellular recordings from peripheral nerves. We examined spontaneous and light induced fictive motor patterns in wild type and mutant larval zebrafish (4–6 days post-fertilization) paralyzed with curare. All spontaneous and light-induced preparations produced alternation of motor activity from side-to-side (mean contralateral phase = 50.7 ± 7.0%; mean burst frequency = 35.6 ± 4.7 Hz) and a progression of activity from head-to-tail (mean ipsilateral rostrocaudal delay = 0.8 ± 0.5 ms per segment), consistent with lateral undulation and forward propulsion during swimming, respectively. The basic properties of the motor pattern were similar in spontaneous and light-induced swimming. This fictive preparation can be used in combination with conventional electrophysiological and imaging methods to investigate normal circuit function as well as to elucidate functional deficits in mutant lines. Toward this end, we show that two accordion class mutants, accordion and bandoneon, have alternating activity on opposite sides of the body, contradicting the hypothesis that their deficit results from the absence of the reciprocal glycinergic inhibition that is typically found in the spinal cord of swimming vertebrates.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Neural circuits in the spinal cord play essential roles in vertebrate movements including the limb movements used for walking as well as axial movements involved in undulatory swimming. Central pattern generators located in the spinal cord can produce coordinated rhythmic motor output in the absence of sensory feedback. Zebrafish larvae are a good model for examining the mechanisms underlying rhythmic motor patterns because of the genetic and optical accessibility of the preparation (Fetcho and Liu 1998; Fetcho and O'Malley 1995; Fetcho et al. 1998; Granato et al. 1996; Higashijima et al. 2003, 2004; O'Malley et al. 1996; Ritter et al. 2001). We have been establishing the foundation for a combined genetic, conventional electrophysiological, and optical analysis of spinal circuits involved in motor behavior. One significant gap has been the lack of a fictive preparation in which spinal motor patterns can be monitored in a paralyzed animal. Such preparations are necessary to record the motor patterns in normal fish as well as to examine the motor pattern disruption in mutant lines. Here we report a fictive preparation using larval zebrafish that shows the features of the motor pattern for swimming. Our analysis of the pattern provides a background for future studies of the underlying circuits. We also use the fictive preparation to test an earlier hypothesis about the functional deficit in two accordion class mutant lines.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Electrophysiological recordings

All procedures were approved by the Institutional Animal Care and Use Committee at Cornell University. Wild type (local commercial supplier; Brian's wild type) or accordion class mutant [accordion (acc, allele tq206), bandoneon (beo, allele ta86d)] zebrafish larvae [4–6 days post-fertilization (dpf)] were anesthetized with 0.02% Tricaine-S (Western Chemical) in an extracellular recording solution that contained (in mM) 134 NaCl, 2.9 KCl, 1.2 MgCl₂, 2.1 CaCl₂, 10 HEPES buffer, and 10 glucose, adjusted to pH 7.8 with NaOH (Drapeau et al. 1999; Legendre and Korn 1994). The preparations were paralyzed with 0.01 mM d-tubocurarine (Sigma) added to the recording solution, which significantly reduced or abolished postsynaptic muscle activity based on patch recordings from muscle fibers (data not shown). Peripheral nerve recordings were observed even at much higher curare concentrations (0.05 mM), which completely abolished synaptic activity in the muscle fibers. The extracellular solution was bubbled with ambient air and superfused continuously at 22–26°C.

Larvae were pinned on their side to a Sylgard-lined glass bottom petri dish with short pieces (1 mm) of fine tungsten wire (0.001 in) pushed through the notochord—one pin placed near the air bladder and another near the anus. The skin between the two pins was removed with a pair of fine forceps. For paired bilateral extracellular recordings, larvae were reoriented into a dorsoventral posture and held in place using additional tungsten wires placed between the wires in the notochord and along the body wall (Fig. 1A). For whole cell patch recordings, collagenase (0.1%, Sigma) in recording solution was applied to the preparation for 3–5 min to prepare enzymatically the muscle fibers for removal. The collagenase solution was washed off and a large bore (15 µm diam) glass microelectrode attached to an extracellular suction electrode holder was used to aspirate individual muscle fibers overlying a small section (2–3 segments) of the spinal cord. All preparations were observed using a water immersion objective (x40, 0.80 NA, Olympus) on an upright microscope (BX51WI, Olympus) fitted with differential interference contrast (DIC) optics.

View larger version (30K): [in this window] [in a new window]   FIG. 1. Extracellular peripheral nerve recording setup and electrical activity in the fictive preparation. A: schematic diagram of a larval zebrafish at 4–6 dpf. Short horizontal lines beginning at the hindbrain and ending in the tail indicate segmentation. Fine tungsten wires are used to secure the fish in a dorso-ventral posture. Lateral wires (solid thick black lines) are pushed through the notochord—1 near the air bladder and 1 near the tail. An additional 4 wires (4 black circles represent the wires end-on) are positioned against the body wall and lateral wires and pushed into the Sylgard dish to produce tension along the body. Extracellular suction electrodes are placed at various points along the midbody axis to monitor activity in peripheral nerves from axial motor neurons. B: central motor output from a fictive preparation at a slow time base. Multiple synchronized bouts of episodic activity are observed in each of the peripheral nerve recordings. Note the fictive activity mimicked the pattern of swimming in unrestrained larvae—several episodes of activity each of which is followed by an interval of inactivity. Segment (S) number is indicated in parentheses. Gray box outlines a single episode of activity shown in Fig. 2A.

Standard whole cell patch recording techniques, modified from Drapeau et al. 1999, were used to monitor the activity of motor neurons in vivo. As described above, the fish were mounted on their side, and the skin and muscle overlying a portion of the spinal cord was removed. Patch electrodes (15 MOhm) pulled on a Flaming/Brown micropipette puller (P-97, Sutter Instruments) from borosilicate glass (1.5 mm OD, 0.86 mm ID, Warner Instruments) were filled with patch solution containing (in mM) 125 K gluconate, 2 MgCl₂, 10 HEPES buffer, 10 EGTA, and 4 Mg ATP, adjusted to pH 7.2 with KOH. We did not correct for junction potentials because we were only concerned with the timing of activity in the motor neurons relative to ventral root activity. Positive pressure (30–50 mmHg) was applied to the patch electrode as it approached the exposed surface of the spinal cord. The tip of the electrode was carefully lowered until it broke into the cord. Motor neurons were targeted for recording based on their size, shape, and position in the spinal cord. Once the tip of the patch electrode was directly apposed to a motor neuron, release of positive pressure allowed a gigaohm seal to form. Suction pulses were applied to break the seal for whole cell voltage recordings. Whole cell voltage was monitored with a MultiClamp 700A (Axon Instruments) amplifier at a gain of 100 (R_f = 5 GOhm) filtered at 30 kHz and digitized at 66 kHz. The recordings were accepted for data analysis if the resting membrane potential was more negative than –45 mV. Neurons were labeled with 0.1% Sulforhodamine B (Sigma) added to the patch solution, and fluorescent images were acquired with a CCD camera (C-72-CCD, Dage MTI, Michigan City, IN), a frame grabber (LG3, Scion, Frederick, MD) and imaging software (Scion National Institutes of Health Image, Scion) for morphological identification.

Data acquisition and analysis

Extracellular and whole cell voltage recordings were digitized using a digitizing board (DigiData series 1322A, Axon Instruments), acquired using pClamp 8.2 software (Axon Instruments) and analyzed off-line with a spike train analysis program written in Matlab 5.3 (Mathworks, Natick, MA).

In the analysis program, spikes were detected with a discrimination window. When voltage crossed a lower threshold value, but did not exceed an upper threshold, a spike event was detected and was indicated by a raster point above the spike (Fig. 2A). The upper threshold eliminated transient artifacts in the recording. To prevent multiple detection of the same spike, a refractory period (1 ms), during which spikes could not be recognized, was applied after each detected event. To ensure that all spikes were detected, the refractory period was considerably shorter than the shortest interspike interval (2 ms). Spikes were grouped into bursts as follows. After an interburst interval (10 ms) elapsed without any spikes detected, the next spike event was identified as the first spike of a burst. Subsequent spikes with interspike intervals less than the interburst interval were grouped into that burst. To eliminate the effects of stray spikes, single spike events were not considered as bursts. The median spike in each burst was indicated by a diamond above the burst (Fig. 2A). An episode of fictive activity was composed of a group of sequential bursts with interburst intervals >8 ms.

View larger version (45K): [in this window] [in a new window]   FIG. 2. Recording of a single episode of fictive activity. A: example of the spike train analysis. A raster point placed above each detected spike indicates individual events. Median spike in each burst is indicated by a black diamond above the burst. Open squares and open diamonds indicate the 1st and last spike in a burst, respectively. Episode duration (ED), cycle period (T), and the time difference between median spikes from phase-locked bursts (t) are measured. Phase () is calculated as = (t/T) x 100. B: phase diagrams display phase relationships and Ds of peripheral nerves from A. Black dashed vertical line (100/0%) indicates occurrence of average median spike of bursts in the right peripheral nerve in segment 11 [Right (S11)]. Phase relationships of bursts in the other peripheral nerves are plotted on the phase diagram relative to this "phase marker." Gray dashed vertical lines indicate occurrence of average median spike of bursts in the peripheral nerves in segments 11 [Left (S11); 50%] and 16 [Right (S16); 12%]. Occurrence of the average median spike to the right of the dashed line indicates the phase lag. Error bars indicate SD of the 1st and last spikes across bursts.

ED was defined as the interval in milliseconds from the first spike of the first burst to the last spike of the last burst for the phase marker peripheral nerve (see METHODS section dealing with phase measurements below; Fig. 2A) and T as the interval in msec from median spike to median spike of consecutive bursts (Fig. 2A). The mean cycle period (T_X) across an episode was determined for each nerve (X). BF (Hz) was defined as the inverse of the cycle period (1/T_X). The mean BF was determined for each episode. BD was defined as the portion of T occupied by spike activity. D was defined as the percentage of T occupied by BD [(D = BD/T) x 100)]. The mean D across an episode for each peripheral nerve was displayed as box plots (normalized BD) in the phase diagrams (Fig. 2B). To convey information regarding the variation of consecutive bursts within an episode, some figures plotted the dependent variable against burst position in the episode (BPE). BPE was defined on a burst-by-burst basis as the median spike time of a burst (MST) divided by ED and expressed as a percentage of the ED [BPE = (MST/ED) x 100].

A multivariate ANOVA (MANOVA) with repeated measures (SuperAnova, Abacus Concepts, Berkeley, CA) was used to compare the effects of spontaneous versus light-induced activity on ED, T, and BD. For comparisons, four episodes from both spontaneous and light-induced activities were selected from each of the six preparations examined. We picked episodes at the onset and offset of the activity as well as at various time-points in between (Fig. 1B; extracellular peripheral nerve recordings). Onset was defined as episodic activity that occurred following a significant period (>5 s) of inactivity. Offset, however, was more difficult to define. In some cases it was clear since a "final" episode of activity occurred followed by a period of inactivity. In others, the final episode in the electrophysiological record was used as the offset episode because the robust nature of the activity gave no clear indication that the episodic activity would terminate. Individual episodes were further divided into an "early" and "late" component. The early component consisted of bursts from the first half of each episode, whereas the late component consisted of bursts from the second half of each episode. The mean value for each component (early and late) was determined and used in the analysis. Specific contrasts were used to reveal embedded relationships within the MANOVA (see RESULTS). The P values from these contrasts are presented in Tables 1 and 2. Differences were considered significant at a level of P < 0.05. A model II principal axis regression analysis (Sokal and Rohlf 1995) was used to examine the slope of the relationships between various dependent variables (contralateral phase, Fig. 5C; rostrocaudal delay, Fig. 6D), and rostrocaudal phase (Fig. 6G) and T. We used Pearson product moment correlation to determine the intensity of the association between these same dependent variables and T.

View larger version (39K): [in this window] [in a new window]   FIG. 5. Alternating (side-to-side) pattern of activity during a fictive episode. A: extracellular recordings indicate a robust alternation of activity (C = 50%) between paired contralateral peripheral nerves from the same body segment (S9). Gray boxes overlay the rhythmic burst activities in the Right (S9) peripheral nerve to indicate the alternating pattern of activity. B: plot of contralateral phase against BPE within a single episode of activity. C: plot of contralateral phase against cycle period (T) for all preparations (Pearson product moment correlation = –0.09; P = 0.04). D: frequency histogram showing the variability of contralateral phase among preparations.

View larger version (38K): [in this window] [in a new window]   FIG. 6. Rostrocaudal progression of activity during a fictive episode. A: extracellular recordings indicate a head-to-tail delay between bursts in paired ipsilateral peripheral nerve recordings from different rostrocaudal points along the body. Gray boxes overlay the rhythmic burst activities in the Right (S8) peripheral nerve to indicate the head-to-tail delay of activity. B: plot of rostrocaudal delay against BPE within a single episode of activity. C: plot of mean rostrocaudal delay against the number of body segments separating the recording electrodes for 2 fish (each represented by a different symbol). Errors bars indicate the SD. D: plot of rostrocaudal delay per segment against cycle period (T) for a single representative preparation (slope = 0.02). E: plot of rostrocaudal phase against BPE within a single episode of activity. F: plot of mean rostrocaudal phase against the number of body segments separating the recording electrodes for 2 fish (same fish and symbols as in C). Error bars indicate SD. G: plot of rostrocaudal phase normalized per segment against cycle period (T) for all preparations (Pearson product moment correlation = –0.01; P = 0.6).

Phase diagrams were used to show phase differences between peripheral nerves (Fig. 2B). The beginning and end of each box plot indicated the average time of the first and last spike, respectively, in a series of bursts from a single episode relative to the median spike time of the bursts in the phase marker nerve. Error bars indicated the SD around the mean first and last spike in a burst. The average median spike time of the bursts in the phase marker nerve, indicated by a dashed vertical line that bisected the phase box near its midpoint, was positioned at 100/0% phase on the diagram. The mean median spike time for bursts in each nerve was plotted on the phase diagram with respect to the phase marker nerve. A shift of the average median spike to the right of the 100/0% position indicated a phase lag, whereas a shift of the average median spike time to the left of the 100/0% position indicated a phase lead.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Unrestrained swimming behavior reflected in the wild type fictive motor pattern

To compare the overall pattern of fictive motor activity with swimming behavior in unrestrained fish, we used standard extracellular recording techniques to simultaneously monitor motor output in peripheral nerves located at various body segments (Fig. 1A). Unrestrained larvae produce an episode of swimming followed by a period of inactivity. We observed discrete episodes of spontaneous and light-induced fictive activity separated by intervals of inactivity. At the developmental stages examined (4–6 dpf), this pattern of fictive activity (Fig. 1B) mimicked the overall pattern of periodic episodes of swimming in unrestrained larvae (Budick and O'Malley 2000; Muller and van Leeuwen 2004; Ritter et al. 2001).

Timing between peripheral nerve and ipsilateral motor neuron activity

To establish that peripheral nerve activity was synchronized with membrane potential changes in motor neurons, we examined the timing relationships between activity recorded from peripheral nerves (extracellular) and membrane potential recorded from a single ipsilateral primary motor neuron (whole cell patch) within one to three segments more caudal (n = 4 recordings from 4 fish). The identity of motor neurons was confirmed after recording by imaging the dye-filled cells. During an episode of swimming, the membrane potential of motor neurons depolarized coincident with an episode of activity in the peripheral nerves and remained depolarized throughout the entire episode (Fig. 3A). Riding on the depolarization were rhythmic events, some of which reached threshold (Fig. 3B). In paired recordings, each single spike or subthreshold event in the motor neuron occurred during a burst of activity in the nearby ipsilateral peripheral nerve (Fig. 3B). The T (30 ms) and thus BF (33 Hz) of the rhythmic depolarization in the motor neurons matched those recorded from peripheral nerves. All of these observations support the conclusion that the peripheral recordings were from axons of motor neurons.

View larger version (36K): [in this window] [in a new window]   FIG. 3. Synchronized activity between ipsilateral peripheral nerves and a motor neuron. A: paired whole cell [top trace; Left (S11)] and extracellular [bottom trace; Left (S15)] recordings show that the motor neuron is active during bursts of activity in an ipsilateral peripheral nerve. B: rhythmic bursting within an episode shows that motor neurons generate a single spike or subthreshold event (*) that occurs during a burst in an ipsilateral peripheral nerve.

To determine whether differences in the properties of the motor pattern were present between spontaneous and light-induced activity, we compared features of the motor patterns produced spontaneously or induced by light. A MANOVA with repeated measures (see METHODS) revealed there was no general effect of stimulus type (spontaneous or light-induced) on ED (P = 0.7), T (P = 1.0), or BD (P = 0.2; Table 1). Subsequently, contrasts were used to test the effects of stimulus type on T and BD between bursts that occurred early or late in episodes (see METHODS). Significant differences were found between early and late bursts for both T and BD in spontaneous (P_T = 0.02 and P_BD = 0.0001, respectively) and light-elicited (P_T = 0.03 and P_BD = 0.0001, respectively) activity (Table 2). In addition, a significant difference was evident between spontaneous and light-induced activity for burst duration of late bursts within episodes (P_BD = 0.004; Table 2). However, given the general overall similarities in the properties of spontaneous and light-induced activity, data from both types of activity were merged for all subsequent analyses.

Temporal characteristics of fictive activity

Swimming in unrestrained fish is characterized by the lateral undulation of the body wall, which is generated by the alternation of muscle contractions that originate at or near the head and progress caudally (Cohen and Wallen 1980; Fetcho and Svoboda 1993; Grillner and Kashin 1976; Grillner and Matsushima 1991; Grillner et al. 1991; Roberts 1990). To assess whether the fictive motor activity replicated the characteristics of unrestrained swimming, we analyzed paired peripheral nerve recordings with the spike train analysis program (see METHODS).

ED varied among different episodes within individual preparations (Fig. 4A), as well as across episodes from different preparations (Fig. 4B). Fictive EDs ranged from 91 to 967 ms (mean ED = 303 ± 137.3 ms, n = 199 episodes from 17 preparations; Fig. 4, B and C) and were comparable with, although somewhat longer than, swim episodes observed in unrestrained fish (Brustein et al. 2003, mean ED = 200 ms; Buss and Drapeau 2001, mean ED = 180 ± 20 ms; n = 12). The mean number of consecutive bursts within an episode was 10.1 ± 4.7 bursts (n = 199 episodes from 17 preparations, range = 3–30 bursts).

View larger version (38K): [in this window] [in a new window]   FIG. 4. Analysis of the general properties of fictive episodic activity. A: plot of ED of episodes from 2 fish (each represented by a different symbol). B: histogram plot of EDs measured for each wild type preparation (n = 17). C: frequency histogram showing the variability of ED among preparations. D: plot of burst duration against burst position in episode (BPE; see METHODS) within a single episode of activity for 2 fish (each represented by a different symbol). E: plot of burst duration against BPE for all preparations. F: frequency histogram showing the variability of burst duration among preparations. G: plot of burst frequency against BPE within a single episode of activity for 2 different fish (each preparation represented by a different symbol). H: plot of burst frequency against BPE for all preparations. I: frequency histogram showing the variability of burst frequency among preparations.

Burst frequencies were regular within an episode (Figs. 2A and 4G), but also could vary slightly from burst to burst (Fig. 4G). There was no consistent tendency for BF to increase or decrease as the episode proceeded (Fig. 4H). However, in some episodes (17%; 31 of 189 from episodes), the BF at the start of an episode was markedly different from the BF at the end of the episode. The BF increased as the episode proceeded in 11 episodes and decreased as the episode proceeded in 20 episodes. BF ranged from 20.3 to 63.1 Hz (mean BF = 35.6 ± 4.7 Hz, n = 2,000 bursts in 199 episodes from 17 preparations; Fig. 4I) and was comparable with the swim (tail-beat) frequencies observed in unrestrained animals (Budick and O'Malley 2000; Ritter et al. 2001, 15–70 Hz; Buss and Drapeau 2001, 25–63 Hz; Muller and van Leeuwen 2004, 30–100 Hz in 3 dpf fish). Accordingly, T ranged between 15.8 and 49.3 ms (mean T = 28.6 ± 3.7 ms, n = 2,000 bursts in 199 episodes from 17 preparations).

Phase relationships during fictive activity

To examine the side-to-side pattern of activity between bursts during fictive episodes, we used paired extracellular electrodes to record simultaneously from contralateral peripheral nerves within the same body segment (Fig. 1, A and B). A robust alternation of activity was observed (Figs. 2A and 5A) in each preparation examined (n = 11). There was no indication of a preference for activity to initiate on a particular side of the fish (data not shown). The contralateral phase difference was regular from burst-to-burst within an episode (Fig. 5B), showed a slight tendency to decrease as T increased (r = –0.09, P = 0.04; Fig. 5C), and was normally distributed across preparations (Fig. 5D; mean contralateral phase difference = 50.7 ± 7.0%, n = 537 bursts in 55 episodes from 11 preparations).

To examine the progression of activity along the rostrocaudal axis of the fish during fictive episodes, we used paired extracellular electrodes to record simultaneously from ipsilateral peripheral nerves located at different body segments (Fig. 1A). In all wild type preparations (n = 11), we observed a progression of activity from head-to-tail (Figs. 2, A and B, and 6A), consistent with swimming behavior. A reversal in the progression of activity (i.e., tail-to-head), as seen in struggling, was not observed (0 of 11) during spontaneous or light-elicited activity. However, struggling was observed when other forms of stimuli were applied, such as repetitive electrical stimulation (0.1- to 1.0-ms pulse width; 20 µA; see Soffe 1993) to the yolk sac or slight pressure on the dorsomedial aspect of the head (data not shown). The rostrocaudal delay observed during fictive activity was regular from burst-to-burst within episodes that produced regular Ts (Figs. 2A and 6B) and increased as the number of body segments between recording electrodes increased (Fig. 6C).

To compare the head-to-tail delays from different fish and to account for the different number of segments separating the recording electrodes in the preparations, we normalized the absolute rostrocaudal delay by dividing it by the number of segments separating the recording electrodes. A small positive correlation between the normalized rostrocaudal delay and T was found among pooled data from all preparations (r = 0.16, P < 0.001; n = 1,744 bouts in 178 episodes from 11 fish), indicating that rostrocaudal delay increased as T increased. The mean slope (0.03 ± 0.02; n = 11) of the regression lines fitted to the plots of normalized rostrocaudal delay versus T from individual preparations was determined using a model II principal axis regression. A representative example of such a fit is shown for a preparation that produced Ts over a broad range (15–45 ms) during the fictive activity (Fig. 6D). The 95% CI lines included the origin in 8 of the 11 cases, suggesting that the actual regression line could pass through, or very near to, the origin.

The mean rostrocaudal delay per segment (0.8 ± 0.5 ms; n = 1,744 bursts in 178 episodes from 11 preparations) was on average 2.8% of the mean T (28.6 ± 3.7 ms). Since zebrafish larvae have 33 body segments, a 2.8% delay translated into 92% of a wave of activity along the body at any point in time. Our measured average slope of 0.03 ms per segment for the regression of normalized rostrocaudal delay versus cycle time would produce 99% of a wave of activity along the body at any point in time. Both estimates showed that the fictive swimming activity represented 90–99% of a wave of activity along the body at any point in time. These estimates are consistent with high-speed video analyses of swimming behavior in which normal, unrestrained zebrafish larvae at 3–5 dpf have approximately one wave of bending along the body at any point in time (Budick and O'Malley 2000; Liu and Fetcho 1999).

In many systems that generate rhythmic swimming motor patterns, such as lamprey (Grillner 1974; Grillner and Kashin 1976; Grillner and Wallen 2002; Wallen and Williams 1984), crayfish swimmeret (Jones et al. 2003; Mulloney 1997), Xenopus tadpoles (Tunstall and Roberts 1991; Tunstall and Sillar 1993; Tunstall et al. 2002), leeches (Cang and Friesen 2002; Pearce and Friesen 1988), and fish (Buchanan 1992; Fetcho and Svoboda 1993; Sigvardt and Williams 1996; Wallen and Williams 1984), rostrocaudal delay scales proportionally with T to generate a constant rostrocaudal phase difference between adjacent segments across a range of cycle frequencies. To address this relationship in zebrafish larvae, we asked whether the rostrocaudal phase difference (normalized to a single body segment) within single episodes and among different episodes varied with T. The rostrocaudal phase difference observed was regular from burst-to-burst within individual episodes (Figs. 2, A and B, and 6E) and varied with the distance between recording sites (Fig. 6F). More importantly, rostrocaudal delay scaled proportionally with T when normalized to a single body segment, leading to a constant rostrocaudal phase (Fig. 6G; mean rostrocaudal phase difference per segment = 2.6 ± 1.7%; n = 1,744 bursts in 178 episodes from 11 preparations). The Pearson product moment correlation did not indicate a significant relationship between normalized rostrocaudal phase and T (r = –0.01; P = 0.6).

Examination of the fictive motor pattern in motor mutants

To determine whether the pattern of fictive activity in the accordion class of motor mutants suggested a central mechanism potentially responsible for the behavioral phenotype, we monitored fictive peripheral nerve activity in these mutants to determine if the general properties of the motor pattern were like wild type. Since these mutants [accordion (acc) and bandoneon (beo)] exhibited bilateral contractions during unrestrained, free swimming, we asked whether there was a loss of the normal side-to-side alternation in the mutants that might reflect a disruption of left/right coordination in the network. In all mutants examined, we observed rhythmic bursting within episodes as well as an antiphasic (50%) relationship of paired contralateral peripheral nerve recordings within the same body segment (Fig. 7, A and B, bottom traces). This indicated that there was not a wholesale disruption of left/right coordination. A head-to-tail progression of activity was observed in all acc mutant preparations examined with recording electrodes placed at different points along the rostrocaudal axis of the body (2 of 2; data not shown). There was no evidence for a reversal (tail-to-head) in the progression of activity, as seen in struggling, during spontaneous or light-elicited activity (0 of 2).

View larger version (33K): [in this window] [in a new window]   FIG. 7. Side-to-side alternation of activity observed during fictive swimming in wild type larvae is retained in the accordion class of mutants. A: acctq206 mutants show episodic activity, similar to wild type larvae, during fictive swimming (top pair of traces). At a faster sweep speed, there is a clear alternation of bursting activity between paired contralateral peripheral nerve recordings [bottom pair of traces; Left (S13) and Right (S13)]. Gray boxes overlay the rhythmic burst activities in the Right (S13) peripheral nerve to indicate the alternating pattern of activity. B: in beota86d mutants, episodic activity is less organized, and EDs tend to be longer than in both wild type and acc mutant fish. An alternation of bursting activity between paired contralateral peripheral nerve recordings is present [bottom pair of traces; Left (S9) and Right (S9)]; however, it is not as clear as that observed in either wild type or acc fish. Gray boxes overlay the rhythmic burst activities in the Left (S9) peripheral nerve to indicate the alternating pattern of activity.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
The accessibility of the larval zebrafish preparation to optical, genetic, and electrophysiological techniques provides a unique model for investigating the mechanisms that underlie motor control of rhythmic patterns of behavior, such as swimming. Previous studies of axial muscle activity in larval zebrafish presented recordings from muscle fibers because the peripheral nerves were thought to be too small and inaccessible for extracellular recording of fictive motor activity (Buss and Drapeau 2001). Here we describe such a fictive preparation, in which activity in axial motor neurons was monitored using extracellular recordings from peripheral nerves (Fig. 1) to provide a detailed description of the activity pattern observed during spontaneous and light elicited fictive swimming (Fig. 2). The work provides a necessary foundation both for future studies of the underlying circuits as well as for understanding motor pattern disruption in mutant lines. Some of the work was previously presented in an abstract (Masino and Fetcho 2003).

We are confident that the activity observed in our recordings was due to motor neuron activity based on several lines of evidence. First, curare works in a concentration-dependent manner to reduce and ultimately eliminate all postsynaptic activity in muscle fibers (data not shown). We could record peripheral motor activity at concentrations of curare that completely eliminated postsynaptic potentials in muscle fibers, indicating that the recordings were from nerves and not muscle fibers. Second, the activity in peripheral nerves coincides with activity in individual motor neurons during fictive swimming. Finally, if the activity in the peripheral nerve recordings was due to an afferent component, not only would the sensory input need to be rhythmic in the absence of movement, but it would also have to be rhythmic at a very high-frequency (30 Hz and greater) in synchrony with the ipsilateral motor neuron activity, both of which seem unlikely. Taken together, these observations suggest that the activity observed in our peripheral nerve recordings is the result of active motor neurons.

Normal unrestrained larval zebrafish produce both spontaneous and light-elicited swimming behavior. Using the extracellular recordings from peripheral nerves, we found that fictive preparations also produced spontaneous and light-elicited motor activity. Our analyses did not show any dramatic differences between the basic properties (e.g., ED, BF, BD) of the two forms of activity in fictive preparations (Tables 1 and 2). Consequently, we primarily used light to elicit rhythmic activity because it allowed control over the onset of the motor output.

The motor pattern showed several features of the periodic episodes of swimming produced by freely swimming larvae, which typically swim for a short time, pause, and swim again. The fictive motor pattern that occurred spontaneously or in response to light also consisted of discrete episodes of rhythmic bursts of activity in peripheral nerves (Figs. 1 and 2). The mean ED (303 ± 137.7 ms; n = 199 episodes) generated by fictive preparations (Fig. 4, A–C) was longer than the mean ED observed during swimming in unrestrained fish (Brustein et al. 2003, mean ED = 200 ms; Buss and Drapeau 2001, mean ED = 180 ± 20 ms, n = 12); however, the range of EDs overlapped with that of free-swimming fish. Additionally, the range of rhythmic BFs (20–63 Hz) recorded from individual nerves during episodes in fictive preparations (Fig. 4, G–I) was similar to the range of tail beat frequencies observed in free-swimming fish (Budick and O'Malley 2000; Ritter et al. 2001, 15–70 Hz; Buss and Drapeau 2001, 25–63 Hz; Muller and van Leeuwen 2004; 30–100 Hz in 3 dpf fish). These data suggested that the fictive motor pattern was that for swimming.

Curare was used in our experiments as a paralytic agent to block acetylcholine (ACh) receptors at the neuromuscular junction. Curare has been shown to function as a GABA antagonist (Caputi et al. 2003; Lebeda et al. 1982; Wotring and Yoon 1995). Since GABA can control or regulate the speed of locomotor activity (Cazalets et al. 1994, 1998; Krogsgaard-Larsen and Johnston 1975; Tegner et al. 1993), it is possible that curare could modify the fictive swimming motor pattern in zebrafish larvae. We cannot rule out subtle effects of curare on the fictive motor pattern; however, the overlap of the burst frequencies we observed with the tail-beat frequencies in freely swimming fish suggests that curare does not dramatically change the frequency of the normal rhythm.

To confirm that the pattern of motor activity observed in fictive preparations was indeed swimming, we compared the pattern of activity monitored in the peripheral nerves of fictive preparations with the features common to patterns of activity recorded in EMGs from freely swimming fish. Fishes (agnathans, cartilaginous, and bony fishes) and swimming salamanders and frog tadpoles generate a swimming electromyographic motor pattern with several common features (Cohen and Wallen 1980; Cohen et al. 1982; Fetcho and Svoboda 1993; Grillner 1974; Grillner and Kashin 1976; Mos et al. 1990; Roberts 1981; Williams et al. 1989). A side-to-side alternation of activity generates the lateral undulation of the body, with the BD occupying nearly one-half of the cycle time between successive bursts of activity in a segment. A similar, robust alternation of activity with an 50% contralateral phase difference was evident in all fictive zebrafish preparations examined (Fig. 5). This pattern of alternating activity is consistent with the alternating bending seen during swimming. In fishes and amphibians, a traveling wave of activity originates at or near the head and progresses along the body toward the tail during swimming. In all fictive preparations examined, activity initiated at a more rostral location and progressed caudally (Fig. 6), consistent with the pattern that usually produces the forward propulsion necessary for swimming.

In swimming fishes, because T varies with swim frequency, the BD remains proportional to period to generate a constant D of 50% (Grillner and Kashin 1976; Wallen and Williams 1984; Williams 1986). The BD scales with cycle time in our fictive preparations as well (Fig. 2B), but the Ds (mean D = 27.6 ± 13.7%) were lower than the expected 50%. This might result from an undersampling of the overall activity during swimming. Because the segmental ventral roots project out of the ventral spinal cord well medial to the lateral edge of the body wall, our superficially located recording electrodes probably sampled only a fraction of the motor axons from a particular segment.

Finally, in swimming animals, as the wave propagates from head to tail, the ipsilateral rostrocaudal delay scales proportionally with T to generate a constant rostrocaudal phase lag between adjacent segments across a range of cycle frequencies (Cang and Friesen 2000; Fetcho and Svoboda 1993; Grillner and Wallen 2002; Jones et al. 2003; Tunstall et al. 2002; Wallen and Williams 1984). The fictive preparations we studied also showed a constant rostrocaudal phase lag at different swimming frequencies (Fig. 6G).

Estimates of the relationship between rostrocaudal delay and T indicated that the fictive zebrafish preparation produced approximately one full wave of activity (90–99%) along the body at any point in time. This result is consistent with observations made during swimming in lampreys, crayfish, and leeches. In lamprey, the rostrocaudal delay is 1% of the cycle time (Grillner et al. 1991; Williams et al. 1989). Because lampreys have 100 body segments, there is about one complete wave along the body at any point in time. In crayfish, the rostrocaudal phase lag between the movements of each of the four pairs of swimmerets is 25%, thus generating a full wave of activity along the body at any point in time (Mulloney et al. 1998; Skinner and Mulloney 1998). Leeches also generate nearly one full wave of activity at any given time during dorsoventral undulatory swimming (Hill et al. 2003; Kristan et al. 1974). In contrast, however, adult goldfish generate less than a complete wave of activity (63%) along the body at any point in time (Fetcho and Svoboda 1993). This reduction in the extent of the wave of activity along the body may be due to the fact that adult goldfish are much less flexible than any of the other preparations mentioned above and thus mechanically are limited in their ability to generate a complete wave during swimming.

The baseline data from fictive preparations can be used to assess potential central deficits in different motor mutant lines. For example, a motor mutation originally identified by Granato et al. (1996) shows simultaneous bilateral contractions during swimming that result in the fish compressing along the rostrocaudal axis, leading to the mutant name accordion (acc). The authors suggested that the disruption in motor behavior was due to a loss of glycinergic reciprocal inhibition in the spinal network that produced swimming. Recordings from fictive acc preparations allowed a quick assay of the integrity of their pattern generating networks. Paired contralateral peripheral nerve recordings within the same body segment showed that these mutants retained the ability to generate rhythmic bursts within episodes as well as a robust antiphasic (50% contralateral phase difference) relationship (Fig. 7A). This refutes the hypothesis that a wholesale disruption of reciprocal inhibition is responsible for generating the acc mutant phenotype.

These data are consistent with recent reports of the cloning of the acc gene, which show that the phenotype is produced by a primary deficit in the periphery rather than in the CNS (Gleason et al. 2004; Hirata et al. 2005). The acc mutation is in a gene that encodes the sarco(endo)plasmic reticulum Ca²⁺-ATPase 1 (SERCA1). The mutation leads to an impaired Ca²⁺ reuptake in the sarcoplasmic reticulum of muscle and a slowed relaxation time of the muscle. The resulting overlap of muscle contractions on the two sides of the body leads to the accordion phenotype.

Bandoneon (beo), another accordion class mutant, identified by Granato et al. (1996), shows a more severe behavioral phenotype and a less well-organized pattern of central motor activity than acc mutants. Nonetheless, they retain some rhythmic bursting and a side-to-side alternation during fictive swimming (Fig. 7B). This suggests that they too do not have a wholesale disruption of left/right coordination. Similar relatively simple recordings might help to identify subclasses of mutants that show disruptions of pattern that might reflect true central pattern-generating deficits.

In conclusion, our development of a fictive preparation of larval zebrafish will allow for a more thorough exploration of the neural circuits involved in swimming, as well as in other rhythmic behaviors such as struggling. The activity of individual cells monitored by whole cell patch recordings can be linked to the fictive motor pattern recorded from peripheral nerves. This preparation also simplifies the assessment of motor deficits in mutant lines. It will be increasingly useful for relating patterns of activity imaged in cells (with calcium and/or voltage indicator dyes) to the motor output as well as for examining how genetic or optical perturbations of neurons affect the motor patterns.

	GRANTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
This work was supported by National Institute of Neurological Disorders and Stroke Grant NS-26539 to J. R. Fetcho and National Research Service Award postdoctoral fellowship NS-44758 to M. A. Masino.

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
We thank Dr. Andrew A. V. Hill for providing the Matlab scripts used for data analysis, Dr. David McLean for providing the schematic drawing of the larval zebrafish used in Fig. 1A, and Drs. Paul Brehm and Lonnie Wollmuth for assistance with developing the patch-clamp recording techniques.

	FOOTNOTES

 
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Address for reprint requests and other correspondence: M. A. Masino, Cornell Univ., Dept. of Neurobiology and Behavior, W101 Mudd Hall, Ithaca, NY 14853 (E-mail: mam287{at}cornell.edu)

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Brustein E, Saint-Amant L, Buss RR, Chong M, McDearmid JR, and Drapeau P. Steps during the development of the zebrafish locomotor network. J Physiol Paris 97: 77–86, 2003.[CrossRef][ISI][Medline]

Buchanan JT. Neural network simulations of coupled locomotor oscillators in the lamprey spinal cord. Biol Cybern 66: 367–374, 1992.[CrossRef][ISI][Medline]

Budick SA and O'Malley DM. Locomotor repertoire of the larval zebrafish: swimming, turning and prey capture. J Exp Biol 203: 2565–2579, 2000.[Abstract]

Buss RR and Drapeau P. Synaptic drive to motoneurons during fictive swimming in the developing zebrafish. J Neurophysiol 86: 197–210, 2001.[Abstract/Free Full Text]

Cang J and Friesen WO. Sensory modification of leech swimming: rhythmic activity of ventral stretch receptors can change intersegmental phase relationships. J Neurosci 20: 7822–7829, 2000.[Abstract/Free Full Text]

Cang J and Friesen WO. Model for intersegmental coordination of leech swimming: central and sensory mechanisms. J Neurophysiol 87: 2760–2769, 2002.[Abstract/Free Full Text]

Caputi L, Bengtson CP, Guatteo E, Bernardi G, and Mercuri NB. D-tubocurarine reduces GABA responses in rat substantia nigra dopamine neurons. Synapse 47: 236–239, 2003.[CrossRef][ISI][Medline]

Cazalets JR, Bertrand S, Sqalli-Houssaini Y, and Clarac F. GABAergic control of spinal locomotor networks in the neonatal rat. Ann NY Acad Sci 860: 168–180, 1998.[Abstract/Free Full Text]

Cazalets JR, Sqalli-Houssaini Y, and Clarac F. GABAergic inactivation of the central pattern generators for locomotion in isolated neonatal rat spinal cord. J Physiol 474: 173–181, 1994.[Abstract/Free Full Text]

Cohen AH, Holmes PJ, and Rand RH. The nature of the coupling between segmental oscillators of the lamprey spinal generator for locomotion: a mathematical model. J Math Biol 13: 345–369, 1982.[CrossRef][ISI][Medline]

Cohen AH and Wallen P. The neuronal correlate of locomotion in fish. "Fictive swimming" induced in an in vitro preparation of the lamprey spinal cord. Exp Brain Res 41: 11–18, 1980.[ISI][Medline]

Drapeau P, Ali DW, Buss RR, and Saint-Amant L. In vivo recording from identifiable neurons of the locomotor network in the developing zebrafish. J Neurosci Methods 88: 1–13, 1999.[CrossRef][ISI][Medline]

Fetcho JR, Cox KJ, and O'Malley DM. Monitoring activity in neuronal populations with single-cell resolution in a behaving vertebrate. Histochem J 30: 153–167, 1998.[CrossRef][ISI][Medline]

Fetcho JR and Liu KS. Zebrafish as a model system for studying neuronal circuits and behavior. Ann NY Acad Sci 860: 333–345, 1998.[Abstract/Free Full Text]

Fetcho JR and O'Malley DM. Visualization of active neural circuitry in the spinal cord of intact zebrafish. J Neurophysiol 73: 399–406, 1995.[Abstract/Free Full Text]

Fetcho JR and Svoboda KR. Fictive swimming elicited by electrical stimulation of the midbrain in goldfish. J Neurophysiol 70: 765–780, 1993.[Abstract/Free Full Text]

Gleason MR, Armisen R, Verdecia MA, Sirotkin H, Brehm P, and Mandel G. A mutation in serca underlies motility dysfunction in accordion zebrafish. Dev Biol 276: 441–451, 2004.[CrossRef][ISI][Medline]

Granato M, van Eeden FJ, Schach U, Trowe T, Brand M, Furutani-Seiki M, Haffter P, Hammerschmidt M, Heisenberg CP, Jiang YJ, Kane DA, Kelsh RN, Mullins MC, Odenthal J, and Nusslein-Volhard C. Genes controlling and mediating locomotion behavior of the zebrafish embryo and larva. Development 123: 399–413, 1996.[Abstract]

Grillner S. On the generation of locomotion in the spinal dogfish. Exp Brain Res 20: 459–470, 1974.[ISI][Medline]

Grillner S and Kashin S. On the generation and performance of swimming in fish. In: Neural Control of Locomotion. Advances in Behavioral Biology, edited by Hermann RM, Grillner S, Stein PSG, and Stuart DG. New York: Plenum, 1976, p. 181–201.

Grillner S and Matsushima T. The neural network underlying locomotion in lamprey–synaptic and cellular mechanisms. Neuron 7: 1–15, 1991.[CrossRef][ISI][Medline]

Grillner S and Wallen P. Cellular bases of a vertebrate locomotor system-steering, intersegmental and segmental co-ordination and sensory control. Brain Res Brain Res Rev 40: 92–106, 2002.[CrossRef][Medline]

Grillner S, Wallen P, Brodin L, and Lansner A. Neuronal network generating locomotor behavior in lamprey: circuitry, transmitters, membrane properties, and simulation. Annu Rev Neurosci 14: 169–199, 1991.[CrossRef][ISI][Medline]

Higashijima S, Masino MA, Mandel G, and Fetcho JR. Imaging neuronal activity during zebrafish behavior with a genetically encoded calcium indicator. J Neurophysiol 90: 3986–3997, 2003.[Abstract/Free Full Text]

Higashijima S, Masino MA, Mandel G, and Fetcho JR. Engrailed-1 expression marks a primitive class of inhibitory spinal interneuron. J Neurosci 24: 5827–5839, 2004.[Abstract/Free Full Text]

Hill AA, Masino MA, and Calabrese RL. Intersegmental coordination of rhythmic motor patterns. J Neurophysiol 90: 531–538, 2003.[Free Full Text]

Hirata H, Saint-Amant L, Waterbury J, Cui W, Zhou W, Li Q, Goldman D, Granato M, and Kuwada JY. accordion, a zebrafish behavioral mutant, has a muscle relaxation defect due to a mutation in the ATPase Ca²⁺ pump SERCA1. Development 13: 5457–5468, 2004.

Jones SR, Mulloney B, Kaper TJ, and Kopell N. Coordination of cellular pattern-generating circuits that control limb movements: the sources of stable differences in intersegmental phases. J Neurosci 23: 3457–3468, 2003.[Abstract/Free Full Text]

Kristan WB, Stent GS, and Ort CA. Neuronal control of swimming in the medicinal leech. J Comp Physiol 94: 97–119, 1974.[CrossRef]

Krogsgaard-Larsen P and Johnston GA. Inhibition of GABA uptake in rat brain slices by nipecotic acid, various isoxazoles and related compounds. J Neurochem 25: 797–802, 1975.[CrossRef][ISI][Medline]

Lebeda FJ, Hablitz JJ, and Johnston D. Antagonism of GABA-mediated responses by d-tubocurarine in hippocampal neurons. J Neurophysiol 48: 622–632, 1982.[Free Full Text]

Legendre P and Korn H. Glycinergic inhibitory synaptic currents and related receptor channels in the zebrafish brain. Eur J Neurosci 6: 1544–1557, 1994.[CrossRef][ISI][Medline]

Liu KS and Fetcho JR. Laser ablations reveal functional relationships of segmental hindbrain neurons in zebrafish. Neuron 23: 325–335, 1999.[CrossRef][ISI][Medline]

Masino MA and Fetcho JR. Fictive swimming motor patterns in wildtype and mutant larval zebrafish. Soc Neurosci Abstr 277: 272, 2003.

Mos W, Roberts BL, and Williamson R. Activity patterns of motoneurons in the spinal dogfish in relation to changing fictive locomotion. Philos Trans R Soc Lond Ser B Biol Sci 330: 329–339, 1990.[CrossRef]

Muller UK, and van Leeuwen JL. Swimming of larval zebrafish: ontogeny of body waves and implications for locomotory development. J Exp Biol 207: 853–868, 2004.[Abstract/Free Full Text]

Mulloney B. A test of the excitability-gradient hypothesis in the swimmeret system of crayfish. J Neurosci 17: 1860–1868, 1997.[Abstract/Free Full Text]

Mulloney B, Skinner FK, Namba H, and Hall WM. Intersegmental coordination of swimmeret movements: mathematical models and neural circuits. Ann NY Acad Sci 860: 266–280, 1998.[Abstract/Free Full Text]

O'Malley DM, Kao YH, and Fetcho JR. Imaging the functional organization of zebrafish hindbrain segments during escape behaviors. Neuron 17: 1145–1155, 1996.[CrossRef][ISI][Medline]

Pearce RA and Friesen WO. A model for intersegmental coordination in the leech nerve cord. Biol Cybern 58: 301–311, 1988.[CrossRef][ISI][Medline]

Ritter DA, Bhatt DH, and Fetcho JR. In vivo imaging of zebrafish reveals differences in the spinal networks for escape and swimming movements. J Neurosci 21: 8956–8965, 2001.[Abstract/Free Full Text]

Roberts A. How does a nervous system produce behaviour? A case study in neurobiology. Sci Prog 74: 31–51, 1990.[Medline]

Roberts BL. The organization of the nervous system of fishes in relation to locomotion. Symp Zool Soc Lond 48: 115–136, 1981.

Sigvardt KA and Williams TL. Effects of local oscillator frequency on intersegmental coordination in the lamprey locomotor CPG: theory and experiment. J Neurophysiol 76: 4094–4103, 1996.