Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Feb 27.
Published in final edited form as: Dev Psychobiol. 2017 Apr;59(3):357–366. doi: 10.1002/dev.21500

Differences in flexor and extensor activity during locomotor-related leg movements in chick embryos

Soo Yeon Sun 1, Nina S Bradley 1
PMCID: PMC9969848  NIHMSID: NIHMS1876139  PMID: 28323348

Abstract

Prior to hatching, chick embryos spontaneously produce repetitive limb movements (RLMs), a developmental precursor to walking. During RLMs, flexor and extensor muscles are alternately active as during stance and swing phases of gait. However, previous studies of RLMs observed that flexor muscles were rhythmically active for many cycles, whereas extensors often failed to be recruited. Thus, we asked if flexor muscles are preferentially recruited during RLMs in chick embryos 1 day before hatching and onset of walking. Using a within-subject design, we compared EMG burst parameters for flexor and extensor muscles acting at the hip or ankle. Findings indicated that flexor burst count exceeded extensor count. Also, flexor muscles were consistently recruited at the lowest levels of neural drive. We conclude that there is a bias favoring flexor muscle recruitment and drive during spontaneously produced RLMs. Potential neural mechanisms and developmental implications of our findings are discussed.

Keywords: electromyography, motor development, recruitment bias, stepping, tibialis anterior

1 |. INTRODUCTION

When an animal is walking, leg flexor and extensor muscles are repeatedly activated to produce the alternating swing and stance phases of gait. The reciprocal activation of flexor and extensor muscles is also observed during repetitive locomotor-related leg movements spontaneously generated in chick embryos during the final week before hatching (Bradley, Ryu, & Lin, 2008; Ryu & Bradley, 2009). During these repetitive limb movements (RLMs), flexor and extensor muscle activity is similar to activity during posthatching locomotion. For example, leg flexor and extensor muscles burst within the same frequencies as during walking, swimming and airstepping (Johnston & Bekoff, 1996; Ryu & Bradley, 2009). Left and right homologous muscles burst alternately, as observed during the stepping of left and right legs. Interestingly, previous studies have observed that leg flexor muscles, compared to extensors, were more likely to be rhythmically active during multiple cycles of an RLM sequence (Bradley et al., 2008; Ryu & Bradley, 2009; Sindhurakar & Bradley, 2012). These studies also observed that extensors frequently dropped out or failed to burst rhythmically during RLMs. Further, left-right alternation of leg muscle activity was more commonly observed in flexors than extensors 1 day prior to hatching and onset of walking (Sindhurakar & Bradley, 2012). Collectively, these observations may indicate that there is a greater net excitatory drive to flexor than extensor motor pools during locomotor-related leg movements in late embryogenesis.

A small body of literature has identified a fundamental bias in spinal locomotor circuitry that favors flexor recruitment in neurologically reduced preparations. For example, it was reported that a subset of rhythm-generating excitatory interneurons exhibited greater synaptic connection with flexor than extensor motor neurons (Dougherty et al., 2013). Further, it was also shown that premotor neurons exerted greater excitatory inputs to flexor motor neurons than extensor motor neurons (Endo & Kiehn, 2008). However, the significance of flexor-biased neural circuitry has received little attention. It is possible, for example, that flexor-biased circuits are responsible for the more common failure of extensor motor neuron bursting during locomotor-like activities in neurologically reduced preparations (Duysens, 1977a; Lafreniere-Roula & McCrea, 2005; Zhong, Shevtsova, Rybak, & Harris-Warrick, 2012). However, these failures of motor neuron bursting do not appear to occur during locomotion in a neurologically intact system (Martinez, Tuznik, Delivet-Mongrain, & Rossignol, 2013).

In this study, we hypothesized that there is a bias in excitatory drive favoring flexor motor pool recruitment during spontaneously generated RLMs in chick embryos 1 day prior to hatching. Electromyography (EMG) and kinematic recordings captured leg flexor and extensor muscle activity during RLMs. We performed two sets of experiments with different recording configurations to determine if there was a bias in excitatory drive at multiple joints. We also recorded activity from left and right ankle flexors and extensors to control for the possibility that postural asymmetry might influence a bias in excitatory drive because late-stage chick embryos assume an asymmetric posture of the head, neck, and wing (Hamburger & Oppenheim, 1967). Previous studies have shown that this postural asymmetry contributed to structural laterality of the brain and a behavioral bias during locomotion posthatching (Casey & Martino, 2000; Rogers, 1982). However, to our knowledge, the impact of postural asymmetry on prenatal locomotor-related leg movements has not been previously examined.

In this study, we provide evidence indicating that there is a bias favoring leg flexor muscle recruitment during RLMs in late-stage chick embryos. Our results provide the first behavioral evidence of a greater net excitatory drive to flexor than extensor motor pools in an intact neurological system. We propose that the bias in flexor recruitment is a characteristic of neural circuits that either generate or adapt locomotor-related leg movements in ovo during late-stage embryo-genesis. We discuss features of spinal locomotor circuitry that may contribute to a flexor recruitment bias during RLMs. We also discuss developmental implications of flexor recruitment bias related to the extremely flexed posture chicks assume during late embryogenesis.

2 |. METHODS

2.1 |. Subjects

Fertile Leghorn chicken (Gallus gallus) eggs were obtained from a local hatchery and incubated in a commercial incubator at 37.5 °C, 62% humidity. Eggs were automatically rotated at 2 hr intervals and exposed to light (426–2160 lux) 12 hr daily. Embryos were prepared for recording at embryonic day 20 (E20). They were placed in a humidified, heated chamber during both surgical procedures and data collection. Data were collected from a total of 29 embryos. At the end of data collection, embryos were euthanized by intraperitoneal injection of Euthasol® (240 mg/kg). All procedures were approved by the University Institutional Animal Care and Use Committee.

2.2 |. Electromyographic preparation and recording

E20 embryos were prepared for electromyographic (EMG) recording during spontaneous leg movements in the egg. The shell wall and membranes were removed to expose one or both legs for EMG implantation. An analgesic (Buprenex®, 0.01 mg/kg) was dripped onto the allantoic membrane. Two pairs of muscles were implanted with bipolar silver wire electrodes (o.d. 50 μm). Each pair consisted of a flexor and extensor antagonist acting at the same joint. In a portion of experiments, ipsilateral hip and ankle muscles were implanted (unilateral configuration): sartorius (SA, hip flexor), caudilioflexorius (CF, hip extensor), tibialis anterior (TA, ankle flexor) and lateral gastrocnemius (LG, ankle extensor). Alternatively, the TA and LG of the left and right ankle were implanted (bilateral configuration).

EMG was recorded continuously for 2–8 hr to capture all spontaneous activity. EMG signals were band-pass filtered (100–1000 Hz), notch-filtered (60 Hz), amplified (×2000), and sampled at 4000 Hz (Datapac 2K2, Run Technologies). At the end of recordings, embryos were euthanized. A cauterizing current was applied through each pair of electrodes and the muscles were dissected to verify electrode tip locations. In three embryos, electrode tips were in the extensor digitorum longus (EDL, ankle flexor). Nonetheless, these data were retained for analyses because EDL activity during RLMs is very similar to TA activity (Bradley et al., 2008).

2.3 |. Video recording

Leg movements were recorded by a high-speed video camera located directly above the chamber. Images were captured at 99 fps. Markers were placed laterally at the knee to track right leg displacements during unilateral EMG recordings, and on the posterior surface of both ankles during bilateral recordings. For offline synchronization of video and EMG recordings, LED lights were placed in the video field and manually triggered at 5 min intervals. The voltage output was written to disk as an additional channel in EMG files.

2.4 |. Identification of RLM sequences

Markers were digitized (198 pictures/s) in video sequences of apparent leg movement to obtain the 2D coordinates (x, y) of marker displacements (Datapac 2K2, Run Technologies). The coordinates were filtered (5th order Butterworth, cut-off frequency 10 Hz) and plotted as a time series to identify repetitive oscillations in the vertical direction indicative of repetitive leg movement (Figure 1, Sindhurakar & Bradley, 2012).

FIGURE 1.

FIGURE 1

Open in a new tab

Quantification of EMG burst parameters. EMG bursts were detected and analyzed based on previously established criteria (Bradley et al., 2008; Bradley, Ryu, & Yeseta, 2014; Ryu & Bradley, 2009; Sindhurakar & Bradley, 2012). In this example, tibialis anterior (TA) and lateral gastrocnemius (LG) are rectified and plotted for one RLM sequence. A burst threshold was separately established for each EMG channel to detect onsets and offsets of activity in each muscle. In this example, threshold was identified as 2× baseline amplitude for TA and LG (horizontal lines). Upward and downward deflections of the horizontal lines identify onsets and offsets of bursts. Burst duration (BD) was defined as the interval of time between burst onset and offset, and cycle duration (CD) was defined as the interval between consecutive burst onsets. The bottom trace (ankle) represents vertical displacements of the ankle during the RLM sequence. See text for further details. Vertical scale equals 62.5 μV

Detection of muscle bursts from rectified EMG was automated employing four criteria: baseline amplitude, burst threshold, burst duration, and interburst interval (Datapac, Run Technologies). Baseline amplitude was estimated for each EMG channel during 100 ms of muscle inactivity and burst threshold was set at 2–3× baseline amplitude. Burst duration, EMG activity exceeding threshold for 20–1000 ms, and interburst interval, activity dropping below threshold for ≥20 ms, were used to locate burst activity characteristic of RLMs (Bradley et al., 2008; Ryu & Bradley, 2009). Only EMG sequences exhibiting all three key features of RLMs were retained for analyses: (1) a minimum of four rhythmically repetitive bursts in one or more EMG channels; (2) a burst frequency range of 1–10 Hz; and (3) leg movements documented in the time series of marker displacements (Figure 1). These RLM criteria also excluded hatching EMG sequences. During prehatching at E20 (type III motility), leg muscles are recruited in shorter sequences of only 1–3 cycles at frequencies typically less than 1 Hz (Bekoff, 1976; Bekoff & Kauer, 1984).

2.5 |. Data analyses

Recruitment of flexor and extensor muscles during RLMs was quantified by four burst parameters: burst count, burst duration, cycle duration, and integrated burst amplitude. Burst count, the total number of bursts produced, was determined for each EMG channel. Burst duration was the interval of time between burst onset and offset, and cycle duration was the length of time between two consecutive burst onsets in the same channel (Figure 1). Integrated burst amplitude was equal to the area of the rectified burst, that is, the product of burst duration and amplitude (Eq. 1). Integrated burst amplitudes were normalized relative to the maximum burst amplitude within channel (Eq. 2), to control for variations in electrode sampling of muscle activity in comparisons between muscles.

Integrated amplitude=∫burst onset burst offsetEMG(t)dt (1)
Normalized integrated amplitude(%)= Mean integrated amplitudeMaximum integrated amplitude during RLMS×100 (2)

Three inclusion criteria were established for analyses of paired recordings. Both the flexor and extensor muscles were successfully implanted. At least one of the two muscles generated ≥100 RLM bursts during 1 hr of the experiment. Plus, at least one of the muscles generated ≥150 RLM bursts over the duration of the experiment.

We employed a within-subject design to compare flexor and extensor recruitment in paired recordings of muscles acting at a common joint during RLMs. To test for differences in flexor and extensor recruitment, burst count for each muscle was normalized by hour to control for variations in recording time. Normalized burst counts were compared across four joints (unilateral: hip, ankle; bilateral: left ankle, right ankle), using a two-factor mixed ANOVA. To test for differences in flexor and extensor burst parameters (burst duration, cycle duration, normalized integrated amplitude), subject means were calculated for each paired recording and compared across four leg joints using a mixed-factor MANOVA. For post hoc comparisons, a repeated measures ANOVA and paired t-tests were performed (SPSS, Version 22.0). Significance was set at p < 0.05, one-tailed, and Bonferroni corrections were applied. Post hoc p values were also adjusted to account for the potential interdependence of burst parameters (burst duration, cycle duration and integrated amplitude). Data summaries report the grand mean ± SD.

3 |. RESULTS

We report RLM burst analyses for flexor and extensor paired recordings (e.g., Figure 1) drawn from 23 of 29 embryos that met inclusion criteria. Data sets were nearly equally drawn from the two recording configurations: 12 unilateral recordings and 11 bilateral recordings. Unilateral recordings yielded 1,786 RLMs that included 1,320 burst sequences for paired recordings at the hip and 1,175 sequences at the ankle. Bilateral recordings yielded 926 RLMs that included 741 burst sequences for the left ankle and 728 for the right. We first describe the general features of the burst sequences. We then report analyses comparing recruitment of flexor and extensor muscles. We also report analyses comparing the neural drive of flexor and extensor muscles during RLMs. Finally, we report analyses for muscle activity when recruitment of the extensor antagonist varied.

3.1 |. General features of RLM EMG

Both the flexor and extensor muscles acting at the same joint were recruited during the majority of RLM sequences. For example, in Figure 2A, the left ankle flexor (l-TA) and its extensor antagonist (l-LG) produced ten alternating bursts. However, in many sequences, one muscle was consistently active, and the antagonist muscle failed to produce one or more bursts. Typically, the flexor muscle was recruited consistently across all cycles of the RLM, whereas the extensor antagonist was not recruited in one or more cycles. For example, in Figure 2B, the hip flexor (SA) produced 11 bursts (10 cycles), and its antagonist (CF) failed to produce bursts that met detection criteria in four cycles (noted by asterisks). Similarly, at the ankle, the TA was recruited consistently, but the LG was not recruited during this particular RLM. Occasionally, an extensor muscle was recruited consistently, and the flexor antagonist was not recruited during one or more cycles. In Figure 2C, for example, the right LG (r-LG) produced five rhythmic bursts, and the right TA (r-TA) failed to produce bursts during three of the RLM cycles (asterisks).

FIGURE 2.

FIGURE 2

Open in a new tab

Variations in EMG recordings during RLM sequences. EMG traces are shown for flexor and extensor antagonist muscles acting at the same joint during either unilateral (hip, ankle) or bilateral recordings (left ankle, right ankle). (A) In this RLM from a bilateral recording, muscles for both the left ankle (l-TA, l-LG) and right ankle (r-TA, r-LG) were consistently recruited. The alternating bursts of the TA and LG were accompanied by repetitive vertical displacement of the left and right ankle (l-ankle, r-ankle). (B) In this RLM from a unilateral recording, both the hip flexor (SA) and ankle flexor (TA), were consistently recruited. The hip extensor (CF) produced activity that met burst detection criteria in only six of the SA cycles. Asterisks note cycles in which CF activity failed to meet burst criteria. The ankle extensor (LG) was not recruited in this RLM sequence, but was active during sequences before and after this RLM, indicating the silence was not due to technical problems. The EMG bursts were accompanied by repetitive vertical displacements of the right knee (r-knee). (C) Consistent bursting in the l-TA trace and inconsistent bursting in the l-LG illustrate a common pattern of paired flexor and extensor recordings. The inconsistent bursting in the r-TA trace and consistent bursting in the r-LG, observed in the bottom two traces were rarely observed in paired flexor and extensor recordings. Vertical scales equal 125 μV

3.2 |. Comparisons of flexor and extensor recruitment parameters

We first tested for a possible muscle recruitment bias during RLMs by comparing burst counts for paired muscle recordings and predicted that the number of bursts would be greater for flexors than extensors. After normalizing burst counts for variations in recording duration, a two-factor mixed ANOVA was used to test for differences in flexor and extensor burst counts across four leg joints (unilateral: hip, ankle; bilateral: left ankle, right ankle). Results indicated that the flexor burst count (137 ± 46 bursts/hr) was greater than the extensor count (71 ± 44 bursts/hr), F (1, 37) = 96.91, p < 0.001, and that there was no main effect across joints, F (3, 37) = 0.71, p > 0.5. However, there was a significant interaction, F (3, 37) = 2.93, p < 0.05, and the post hoc comparisons (Bonferroni p < 0.013) indicated that flexor burst count exceeded extensor count at each of the four joints (Figure 3).

FIGURE 3.

FIGURE 3

Open in a new tab

Burst counts for flexor and extensor muscles acting at the same joint. Bars represent the grand mean ± SD for flexor (dark gray) and extensor burst counts (gray) normalized by time (hr). Average flexor burst count exceeded extensor count by 50–100 bursts/hr, both in unilateral (A) and bilateral recordings (B). All differences were significant (paired t-test, Bonferroni p < 0.013, see text for details)

Analyses indicated flexor bursts outnumbered extensor bursts because, in the majority of paired EMG sequences, flexor muscles were the first recruited, and extensor muscles failed to be recruited during one or more cycles. For example, the hip flexor (SA) and ankle flexor (TA) initiated the first cycle of the RLM in Figure 2B. RLMs began with flexor bursts in 74 ± 16% of the paired sequences and extensor bursts in 26 ± 16%, and the difference was significant, t (40) = 9.07, p < 0.0001. Extensor recruitment failure was a common characteristic of most EMG recordings as observed in the hip extensor (CF) and ankle extensor (LG) in Figure 2B and C, respectively. Extensor muscles were inconsistently recruited in 37 ± 12% of paired sequences, whereas flexors were inconsistently recruited in only 8 ± 7%, t (40) = 12.74, p < 0.0001. Further, extensors failed to be recruited during any cycles in 32 ± 22% of paired sequences, whereas flexors failed to be recruited in only 2 ± 3%, t (40) = 7.89, p < 0.0001. Additionally, extensor muscles were consistently recruited across all cycles in only 31 ± 22% of paired sequences, whereas flexors were consistently recruited across all cycles in 90 ± 10%, t (40) = 12.83, p < 0.0001.

3.3 |. Comparisons of flexor and extensor drive parameters

We next tested for a difference in drive of flexor and extensor muscles, and predicted that integrated burst amplitude, in specific, would be greater for flexors because burst amplitude is an established indicator of neural drive (Farina, Merletti, & Enoka, 2004; Staudenmann, van Dieen, Stegeman, & Enoka, 2014). In contrast, we predicted that neither burst duration or cycle duration would differ between antagonists, regardless of differences in drive, based on findings in previous studies (Bradley, Ryu, & Yeseta, 2014; Ryu & Bradley, 2009; Sindhurakar & Bradley, 2012). Contrary to our predictions, analyses did not provide clear evidence of a bias in flexor drive. A mixed-factor MANOVA was used to compare three burst parameters (burst duration, cycle duration, integrated amplitude) for flexor and extensor bursts across four joints. Results for main effects indicated that collectively, the three burst parameters for flexor activity differed from those for extensor activity, F (1, 111) = 12.47, p < 0.002, and that there was no main effect across the four joints, F (3, 111) = 0.14, p > 0.9. However, there was a significant interaction between muscles and burst parameters, F (2, 111) = 5.10, p < 0.009. Contrary to our prediction, post hoc paired t-tests (Bonferroni p < 0.0083, 0.05/6) indicated that normalized integrated burst amplitude did not differ between flexor (11 ± 6%) and extensor muscles (10 ± 5%), t(40) = 0.56, p > 0.2. Also contrary to our prediction, flexor cycle duration (252 ± 57 ms) was longer than extensor cycle duration (226 ± 45 ms), t (40) = 2.76, p < 0.005, yet as predicted, burst durations were similar for flexor (87 ± 23 ms) and extensor activity (78 ± 17 ms), t (40) = 2.35, p > 0.018.

3.4 |. Flexor burst parameters varied with presence or absence of extensor recruitment

Though subject means for normalized flexor and extensor integrated amplitudes did not suggest that drive was greater to flexor muscles than extensors, variations in flexor burst amplitude during RLM sequences were visually apparent when the presence or absence of activity in the extensor antagonist was also considered. Note, for example, in Figure 4 the variations in LG recruitment across three RLMs from a single experiment. During some RLM sequences, LG was not recruited during any of the cycles (Figure 4A1, “no extensor bursting”). During many sequences, LG was inconsistently recruited, producing a burst in some but not all of the cycles (Figure 4B1, “inconsistent extensor bursting”). Occasionally, LG was consistently recruited during the RLM, producing a burst in every cycle of the sequence (Figure 4C1, “consistent extensor bursting”). TA burst amplitude appeared to vary systematically across these differences in LG recruitment. In specific, average TA amplitude for an RLM appeared to be least when LG was not recruited during any of the cycles in the sequence (Figure 4A2), somewhat greater when LG was inconsistently recruited across cycles of the RLM (Figure 4B2), and greatest when LG was consistently recruited across all cycles of the RLM sequence (Figure 4C2).

FIGURE 4.

FIGURE 4

Open in a new tab

Three patterns of extensor recruitment produced during RLMs within experiments. In the left column, paired EMG samples of TA and LG activity for three RLMs selected from a single experiment are plotted. (A1) TA was consistently recruited across all RLM cycles in the absence of LG recruitment. (B1) TA was consistently recruited and LG was inconsistently recruited. (C1) Both TA and LG were consistently recruited across all cycles. In the middle column, TA and LG bursts were plotted relative to TA burst onset and averaged across nine TA cycles. Averaged TA burst amplitude was least in the absence of LG recruitment (A2), greater when LG was inconsistently recruited (B2), and greatest when LG was consistently recruited (C2). In the right column, TA and LG bursts were plotted relative to the LG burst onset and averaged across 10 LG cycles. Averaged LG burst amplitude was low when LG was inconsistently recruited (B3), and greater when LG was consistently recruited (C3). Vertical scales left column 125 μV (A1, B1, C1); middle column 16 μV (A2, B2, C2); right column 8 μV (B3, C3)

Based on the above observations, we predicted that flexor burst amplitude, and therefore drive, scaled with the recruitment of the extensor antagonist. To test the prediction, 3,623 paired EMG sequences in which the flexor muscle was consistently recruited were sorted into one of the three patterns of extensor recruitment described above: “no extensor bursting” (Figure 4A1), “inconsistent” (Figure 4B1) or “consistent extensor bursting” (Figure 4C1). Analyses indicated that normalized flexor integrated amplitude differed across the three extensor recruitment patterns. A mixed-factor MANOVA for three burst parameters (integrated burst amplitude, burst duration, cycle duration) and three extensor recruitment patterns indicated that burst parameters were similar across patterns, F (2, 116) = 1.04, p > 0.3. However, the interaction was significant, F (4, 234) = 4.99, p < 0.002, and post hoc repeated measures ANOVAs (Bonferroni p < 0.0083) indicated that only integrated burst amplitude differed across the three extensor recruitment patterns, F (2, 38) = 17.97, p < 0.001 (see Supplementary Material for details). All pairwise comparisons for flexor integrated amplitude were significant (p < 0.0005, Bonferroni p < 0.0028), indicating that flexor amplitude was least during no extensor bursting sequences, and greatest during consistent extensor bursting (Figure 5A).

FIGURE 5.

FIGURE 5

Open in a new tab

Variations in flexor burst parameters associated with the presence or absence of extensor recruitment. Paired samples of flexor and extensor activity were sorted into one of three patterns of extensor recruitment: “no extensor recruitment” (light gray), “inconsistent extensor recruitment” (gray), “consistent extensor recruitment” (dark gray). Grand mean ± SD for flexor burst parameters are plotted for each of the three patterns. Significant post hoc paired t-tests are indicated (Bonferroni p < 0.0028). (A) Normalized integrated burst amplitude differed across patterns. All post hoc paired t-tests were significant, indicating that flexor integrated amplitude positively varied with increased recruitment of extensor antagonist. (B) Flexor burst count per sample differed across the three patterns. Flexor burst count was least in the absence of extensor recruitment and greatest in samples of inconsistent extensor recruitment

Given the apparent scaling of flexor burst amplitude, we asked if flexor burst count also varied with extensor recruitment. Comparison of flexor burst count across the three extensor recruitment patterns using a repeated measures ANOVA indicated the count varied significantly, F (2, 38) = 19.30, p < 0.001. Further, all post hoc paired t-tests (Bonferroni p < 0.017) were also significant (Figure 5B), indicating that flexor burst count was least during sequences of no extensor bursting, and greatest during inconsistent extensor bursting.

3.5 |. Extensor burst parameters varied with consistency of extensor recruitment

Given the significant scaling of flexor burst amplitude across extensor recruitment patterns, we asked if extensor integrated amplitude also scaled with extensor recruitment. For example, in Figure 4, average LG burst amplitude was greater when LG was consistently recruited (Figure 4C3) than when LG was recruited in only a portion of the RLM cycles (Figure 4B3). Using a two-factor mixed ANOVA for three burst parameters and two extensor recruitment patterns, results indicated that burst parameters were similar for the two patterns, F (1, 115) = 1.00, p > 0.3. However, the interaction was significant, F (2, 115) = 5.92, p < 0.005, and post hoc paired t-tests (Bonferroni p < 0.0083) indicated that normalized extensor integrated burst amplitude, t (39) = −5.86, p < 0.0005 (Figure 6A) and burst duration, t(39) = −5.22, p < 0.0005 (Figure 6B), were greater during consistent extensor bursting than inconsistent bursting. The increase in burst duration partially contributed to differences in integrated amplitude. Mean integrated burst amplitude increased by 80 ± 104%, and burst duration increased by 32 ± 43%, and the percent change was significantly greater for integrated amplitude, t (39) = −4.22, p < 0.0005.

FIGURE 6.

FIGURE 6

Open in a new tab

Variations in extensor burst parameters associated with inconsistent and consistent extensor recruitment. Lighter gray bars represent inconsistent extensor recruitment, and darker gray bars represent consistent extensor recruitment. Significant post hoc paired t-tests are indicated (Bonferroni p < 0.0083). (A) Normalized extensor integrated burst amplitude was greatest during consistent extensor recruitment. (B) Extensor burst duration was longest during consistent extensor recruitment

4 |. DISCUSSION

In this study we hypothesized that there is a bias in recruitment and drive favoring flexor motor pools during locomotor-related movements in chick embryos. We tested this hypothesis by comparing EMG activity for paired recordings of flexor and extensor muscles acting at the same joint. We also compared EMG activity for left and right ankle muscles to control for potential differences in activity due to the asymmetric posture assumed by chick embryos in late embryogenesis. We discuss how the results were consistent with our hypothesis. We also consider how these findings extend our knowledge of locomotor control. Finally, we consider the developmental implications of our findings.

4.1 |. RLMs exhibit bias in recruitment and drive

The greater incidence of bursts in flexor than extensor muscles acting at the same joint was a key finding in our study that quantitatively confirmed and advanced previous observations of prominent flexor activity during RLMs (Bradley et al., 2008; Ryu & Bradley, 2009). The significantly greater flexor burst count, flexor burst initiation of RLMs and sustained repetitive bursting of flexors, even in the absence of extensor recruitment, strongly indicate that there is a bias favoring flexor recruitment during RLMs. Further, differences in recruitment of flexor and extensor antagonists were observed ipsilaterally at both the hip and ankle when an embryo was lying on its side, and at the ankles bilaterally when an embryo was lying prone. These findings further lead us to conclude that flexor recruitment bias is neither joint nor posture specific, and therefore a fundamental attribute of RLM circuits in the final days before onset of precocious locomotor behavior. Nonetheless, on occasion during most recordings, an extensor muscle was more readily recruited, producing an RLM with few or no flexor bursts. Thus, despite the significant bias in flexor recruitment, there remains a flexible neural mechanism responsible for recruitment of motor pools during spontaneous rhythmic leg movements prior to hatching. Although our study does not point to any particular mechanism, we surmise that taken together, the bias and flexibility in recruitment are important indicators of a mechanism that is responsive to shifting demands from adaptive behavior in ovo to posthatching. For example, the flexor recruitment bias may be an adaption to physical constraint in ovo. Whereas, the flexibility in recruitment bias may be an adaptive window in circuit function required for hatching and precocious transition to weight-supported locomotion within hours after hatching.

The more consistent recruitment of flexor than extensor muscles across cycles of an RLM and scaling of flexor bursts across the three extensor recruitment patterns (no extensor recruitment, inconsistent, and consistent extensor recruitment) also strongly indicate there is a flexor-biased drive during RLMs. In previous work, it was established that increases in EMG amplitude are attributable to increases in excitatory drive to motor pools (Farina et al., 2004; Staudenmann et al., 2014). Thus, our results indicated that during RLMs, flexor motor pools were selectively recruited at the lowest levels of neural drive. A flexor-biased drive has been identified also in the isolated spinal cord of neonatal mice during chemically-induced locomotor activity. For example, recordings in the lumbar cord indicated that flexor premotor networks generated stronger and more robust rhythmic activity in flexor motor neurons than extensor networks produced in extensor motor neurons (Endo & Kiehn, 2008). Similarly, in another study, it was found that most rhythmically active Shox2-expressing interneurons were synchronized with flexor motor neuron activity, and that they made far more synaptic contact with flexor than extensor motor neurons (Dougherty et al., 2013). Thus, the consistency of flexor recruitment and scaling of EMG amplitude during spontaneously generated RLMs extends these findings of reduced, chemically-induced activity to intact embryonic behavior in the chick and leads us to conclude that flexor recruitment bias during RLMs is at least partially attributable to spinal locomotor circuitry.

4.2 |. Separate rhythm-generating modules during locomotor-related behavior

It has long been proposed that during locomotor activity, flexor motor circuits (e.g., flexor half-centers) form the rhythm-generating core and reciprocally inhibit extensor circuits to produce rhythmic extensor activity (Duysens, 1977b; Zhong et al., 2012). In our study, the rhythmic recruitment of flexor muscles during a third of RLMs in the absence of extensor activity appears consistent with the flexor half-center model. However, we also observed instances when the extensor was consistently recruited across all cycles of an RLM but the antagonist flexor was not; and in a small percent of RLMs, the extensor was rhythmically recruited in the absence of bursting in its antagonist flexor. Thus, the singular rhythmic recruitment of flexor or extensor muscles during a portion of RLMs challenges the flexor half-center model in providing evidence that extensor motor pools do not require reciprocal inhibition to generate RLMs. Rather, these findings support the proposal that there are separate rhythm-generating modules for recruitment of flexor and extensor motor pools in chicks (Bradley et al., 2008). Further, our findings may generalize to developing vertebrates more broadly, for both independent control of flexor and extensor motor neurons, and more readily evoked locomotor activity in flexor motor neurons have been observed in the isolated spinal cord of neonatal mice (Hagglund et al., 2013).

4.3 |. Significance of flexor bias during development

A flexed posture prior to birth is commonly observed in many species, including chicks (Hamburger & Oppenheim, 1967), crocodiles (Vieira et al., 2011), cats (Windle & Griffin, 1931), and humans (Casaer, 1979), but the mechanisms responsible for this posture are not well understood. In chick embryos, the flexed posture of the neck, spine, wings, and legs becomes progressively more extreme in the final week before hatching (plates 1–3, Hamburger & Oppenheim, 1967). By E21, the ankle is so extremely dorsiflexed that the foot can touch the anterior surface of the lower leg. This flexed posture may be mechanically imposed by spatial constraints, as when the embryo or fetus begins to outgrow the volume of the prenatal environment and there is no longer room to extend the limbs. Kinematic study of RLMs in ovo indicated that at E20, hip, knee, and ankle excursions averaged 3°, compared to an average of approximately 8° at E18 (Bradley et al., 2014) and almost 50° at E12 (Bradley, Solanki, & Zhao, 2005).

It is also possible that mechanical constraint is but one of several different mechanisms contributing to the flexed prenatal posture and that a combination of mechanisms may variably contribute across the prenatal to postnatal transition. For example, flexion posture may be passively assumed due to insufficient oppositional force generation by extensor antagonist muscles. In rats, extensor forces for postural support and locomotion appear to gradually emerge over the first three postnatal weeks as dopaminergic pathways mature (Navarrete, Slawinska, & Vrbova, 2002). Based on our observations over the course of several studies, it appears that extensor forces may be very limited during RLMs at E18–E20. For example, we have consistently noted that at E20, the leg remains deeply flexed throughout RLM sequences and the foot makes little or no contact with the shell wall to load the limb. Yet, during brief bouts of prehatching movements, the legs may press sufficiently against the wall to produce large amplitude, tonic bursts in extensor muscles and alter the embryo’s posture (Bradley et al., 2014; see also Bekoff & Kauer, 1984). Also, extensor muscle forces may be less than that of flexors due to the slower maturation of membrane properties in extensor motor neurons. Interestingly, in rats, extensor membrane properties do not achieve equivalence with those for flexor motor neurons until the time point weight-supported locomotion is attained (Vinay, Brocard, & Clarac, 2000). Extensor motor units are also slower to mature in rats, as noted by the slower postnatal resolution of polyneuronal innervation in ankle extensor muscles (Balice-Gordon & Thompson, 1988; Brown, Jansen, & Van Essen, 1976). However, the extent to which such maturational difference may account for flexor postures in chick embryos at E20 is less certain given that newly hatched chicks are motorically precocious, and walk within 6 to 8 hrs after hatching (Muir, 2000).

The extensive evidence of flexor bias during RLMS leads us to propose that a central neural component also actively contributes to the flexed posture of chicks in ovo. RLMs are frequently produced in the final 2–3 days before hatching. In EMG and kinematic recordings at E18 and E20, RLMs were expressed for 20 min or longer per hour, for as many as 4–9 hr (Bradley et al., 2008), and in 24 hr force recordings between E18–E19, an average of 40 RLMs were generated every hour (Bradley & Jahng, 2003). During these periods, many RLMs composed of four or more cycles were produced in quick succession, just seconds apart, and some RLMs contained as many as 12–27 cycles of uninterrupted flexor bursting accompanied by repetitive joint excursions (Bradley et al., 2005, 2008). Further, RLMs typically began with a flexor burst, a characteristic typical of movement initiation during grooming behaviors, such as scratching and wiping to position the limb prior to onset of a power stroke directed a cutaneous site (Deliagina, Feldman, Gelfand, & Orlovsky, 1975; Giszter, McIntyre, & Bizzi, 1989; Robertson, Mortin, Keifer, & Stein, 1985). We speculate that flexors may be recruited first also during RLMs so as to retract the foot from the shell wall and avoid excessive loading of recruited extensor muscles already lengthened by extreme dorsiflexion. It is our experience that when not producing RLMs, embryos are generally quiet. Although we have yet to systematically examine these intervals, short sequences of one or two prolonged tonic extensor bursts, presumably related to hatching, may be generated. However, this activity appears to be much less intense than the 45–90 min of actual hatching, when extensors produce one or two prolonged bursts (800 ms) every 10–30 s until escape is complete (Bekoff & Kauer, 1984).

In conclusion, results of our study of spontaneously generated limb movements provide evidence of a flexor bias in recruitment and drive of leg muscle activity during locomotor-related behavior in the intact chick embryo 1 day prior to hatching. Our results also suggest that although flexor muscle recruitment initiates and dominates most RLM activity, there is sufficient flexibility for extensors to produce RLMs absent flexor participation, a finding that may be important in the transition to hatching and precious locomotor skill soon after hatching. Further, the highly stable expression of flexor activity during RLMs inspires us to propose that normally occurring flexor bias is fundamentally important to the prenatal developmental continuum. We propose that in the final days before hatching, RLMs are a principal driver of movement in ovo and that the biased, repetitive recruitment of flexor muscles contributes in a synergistic manner with environmental constraints and neuromaturational properties to the flexed posture assumed by chick embryos. To what extent central neural mechanisms, as well as sensory inputs, directly contribute to the flexor posture common in early motor development remains to be explored.

Supplementary Material

supinfo

ACKNOWLEDGMENT

This work was funded by National Institute of Child Health and Human Development Contract (ROI HD 053367).

Footnotes

SUPPORTING INFORMATION

Additional Supporting Information may be found online in the supporting information tab for this article.

REFERENCES

  1. Balice-Gordon RJ, & Thompson WJ (1988). Synaptic rearrangements and alterations in motor unit properties in neonatal rat extensor digitorum longus muscle. Journal of Physiology, 398, 191–210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bekoff A (1976). Ontogeny of leg motor output in the chick embryo: A neural analysis. Brain Research, 106, 271–291. [DOI] [PubMed] [Google Scholar]
  3. Bekoff A, & Kauer JA (1984). Neural control of hatching: Fate of the pattern generator for the leg movements of hatching in post-hatching chicks. Journal of Neuroscience, 4, 2659–2666. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bradley NS, & Jahng DY (2003). Selective effects of light exposure on distribution of motility in the chick embryo at E18. Journal of Neurophysiology, 90, 1408–1417. doi: 10.1152/jn.00393.2003 [DOI] [PubMed] [Google Scholar]
  5. Bradley NS, Ryu YU, & Lin J (2008). Fast locomotor burst generation in late stage embryonic motility. Journal of Neurophysiology, 99, 1733–1742. doi: 10.1152/jn.01393.2007 [DOI] [PubMed] [Google Scholar]
  6. Bradley NS, Ryu YU, & Yeseta MC (2014). Spontaneous locomotor activity in late-stage chicken embryos is modified by stretch of leg muscles. Journal of Experimental Biology, 217, 896–907. doi: 10.1242/jeb.093567 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bradley NS, Solanki D, & Zhao D (2005). Limb movements during embryonic development in the chick: Evidence for a continuum in limb motor control antecedent to locomotion. Journal of Neurophysiology, 94, 4401–4411. doi: 10.1152/jn.00804.2005 [DOI] [PubMed] [Google Scholar]
  8. Brown M, Jansen J, & Van Essen D (1976). Polyneuronal innervation of skeletal muscle in new-born rats and its elimination during maturation. Journal of Physiology, 261, 387–422. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Casaer P (1979). Postural behaviour in newborn infants. London: William Heinemann Medical Books. [Google Scholar]
  10. Casey MB, & Martino CM (2000). Asymmetrical hatching behaviors influence the development of postnatal laterality in domestic chicks (Gallus gallus). Developmental Psychobiology, 37, 13–24. [DOI] [PubMed] [Google Scholar]
  11. Deliagina TG, Feldman AG, Gelfand IM, & Orlovsky GN (1975). Role of central program and afferent inflow in control of scratching movements in cat. Brain Research, 100, 297–313. doi: 10.1016/0006-8993(75)90484-9 [DOI] [PubMed] [Google Scholar]
  12. Dougherty KJ, Zagoraiou L, Satoh D, Rozani I, Doobar S, Arber S … Kiehn O (2013). Locomotor rhythm generation linked to the output of spinal shox2 excitatory interneurons. Neuron, 80, 920–933. doi: 10.1016/j.neuron.2013.08.015 [DOI] [PubMed] [Google Scholar]
  13. Duysens J (1977a). Fluctuations in sensitivity to rhythm resetting effects during the cat’s step cycle. Brain Research, 133, 190–195. [DOI] [PubMed] [Google Scholar]
  14. Duysens J (1977b). Reflex control of locomotion as revealed by stimulation of cutaneous afferents in spontaneously walking premammillary cats. Journal of Neurophysiology, 40, 737–751. [DOI] [PubMed] [Google Scholar]
  15. Endo T, & Kiehn O (2008). Asymmetric operation of the locomotor central pattern generator in the neonatal mouse spinal cord. Journal of Neurophysiology, 100, 3043–3054. doi: 10.1152/jn.90729.2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Farina D, Merletti R, & Enoka RM (2004). The extraction of neural strategies from the surface EMG. Journal of Applied Physiology, 96, 1486–1495. doi: 10.1152/japplphysiol.01070.2003 [DOI] [PubMed] [Google Scholar]
  17. Giszter SF, McIntyre J, & Bizzi E (1989). Kinematic strategies and sensorimotor transformations in the wiping movements of frogs. Journal of Neurophysiology, 62, 750–767. [DOI] [PubMed] [Google Scholar]
  18. Hagglund M, Dougherty KJ, Borgius L, Itohara S, Iwasato T, & Kiehn O (2013). Optogenetic dissection reveals multiple rhythmogenic modules underlying locomotion. Proceedings of the National Academy of Science of the United States of America, 110, 11589–11594. doi: 10.1073/pnas.1304365110 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hamburger V, & Oppenheim R (1967). Prehatching motility and hatching behavior in the chick. Jounal of Experimental Zoology, 166, 171–203. doi: 10.1002/jez.1401660203 [DOI] [PubMed] [Google Scholar]
  20. Johnston RM, & Bekoff A (1996). Patterns of muscle activity during different behaviors in chicks: Implications for neural control. Journal of Comparative Physiology, 179, 169–184. [DOI] [PubMed] [Google Scholar]
  21. Lafreniere-Roula M, & McCrea DA (2005). Deletions of rhythmic motoneuron activity during fictive locomotion and scratch provide clues to the organization of the mammalian central pattern generator. Journal of Neurophysiology, 94, 1120–1132. [DOI] [PubMed] [Google Scholar]
  22. Martinez M, Tuznik M, Delivet-Mongrain H, & Rossignol S (2013). Emergence of deletions during treadmill locomotion as a function of supraspinal and sensory inputs. Journal of Neuroscience, 33, 11599–11605. doi: 10.1523/JNEUROSCI1126-13.2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Muir GD (2000). Early ontogeny of locomotor behaviour: A comparison between altricial and precocial animals. Brain Research Bulletin, 53, 719–726. [DOI] [PubMed] [Google Scholar]
  24. Navarrete R, Slawinska U, & Vrbova G (2002). Electromyographic activity patterns of ankle flexor and extensor muscles during spontaneous and L-DOPA-induced locomotion in freely moving neonatal rats. Experimental Neurology, 173, 256–265. doi: 10.1006/exnr.2001.7791 [DOI] [PubMed] [Google Scholar]
  25. Robertson GA, Mortin LI, Keifer J, & Stein PS (1985). Three forms of the scratch reflex in the spinal turtle: Central generation of motor patterns. Journal of Neurophysiology, 53, 1517–1534. [DOI] [PubMed] [Google Scholar]
  26. Rogers LJ (1982). Light experience and asymmetry of brain function in chickens. Nature, 297, 223–225. [DOI] [PubMed] [Google Scholar]
  27. Ryu YU, & Bradley NS (2009). Precocious locomotor behavior begins in the egg: Development of leg muscle patterns for stepping in the chick. Public Library Of Science, 4, e6111. doi: 10.1371/journal.pone.0006111 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Sindhurakar A, & Bradley NS (2012). Light accelerates morphogenesis and acquisition of interlimb stepping in chick embryos. Public Library Of Science, 7, e51348. doi: 10.1371/journal.pone.0051348 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Staudenmann D, van Dieen JH, Stegeman DF, & Enoka RM (2014). Increase in heterogeneity of biceps brachii activation during isometric submaximal fatiguing contractions: A multichannel surface EMG study. Journal of Neurophysiology, 111, 984–990. doi: 10.1152/jn.00354.2013 [DOI] [PubMed] [Google Scholar]
  30. Vieira L, Lima F, Santos A, Mendonça S, Moura L, Iasbeck J, & Sebben A (2011). Description of embryonic stages in Melanosuchus niger (Spix, 1825) (Crocodylia: Alligatoridae). Brazilian Journal of Morphological Sciences, 28, 11–22. [Google Scholar]
  31. Vinay L, Brocard F, & Clarac F (2000). Differential maturation of motoneurons innervating ankle flexor and extensor muscles in the neonatal rat. European Journal of Neuroscience, 12, 4562–4566. [DOI] [PubMed] [Google Scholar]
  32. Windle WF, & Griffin AM (1931). Observations on embryonic and fetal movements of the cat. Journal of Comparative Neurology, 52, 149–188. doi: 10.1002/cne.900520103 [DOI] [Google Scholar]
  33. Zhong G, Shevtsova NA, Rybak IA, & Harris-Warrick RM (2012). Neuronal activity in the isolated mouse spinal cord during spontaneous deletions in fictive locomotion: Insights into locomotor central pattern generator organization. Journal of Physiology, 590, 4735–4759. doi: 10.1113/jphysiol.2012.240895 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supinfo
NIHMS1876139-supplement-supinfo.docx (45.8KB, docx)

RESOURCES