In the CNS, central pattern generators (CPGs) produce the basic temporal pattern of
electrical motor activity that produces rhythmic behaviors such as locomotion. A major
issue in motor control is whether CPGs are organized into distinct modules that can
function autonomously or if the different CPG modules are interdependent (Figure 1
in Messina et al., 2017). Regarding the previous commentary (Cangiano and Grillner,
2018), there are at least two relevant questions concerning the locomotor rhythm generating
capabilities of CPG modules in right and left hemi-spinal cords in adult lampreys:
(a) Does the burst activity generated by hemi-spinal cords represent well-coordinated
locomotor activity? (b) Does generation of burst activity by hemi-spinal cords represent
a normal physiological capability of these spinal circuits, or is this capability
manifested under certain, possibly artificial, experiment conditions? In the text
below, all page numbers and figures refer to those in our previous study (Messina
et al., 2017).
For our previous study, a longitudinal midline (ML) lesion was performed along the
rostral spinal cord of adult lampreys, and rostral muscle activity (adjacent to ML
spinal lesion) and caudal muscle activity (below ML spinal lesion) were recorded in
response to sensory stimulation of the anterior head (Figure 3A). The rationale was
that initiation of spinal motor activity in whole animals in response to sensory stimulation
and descending brainstem activation would provide the most physiological situation
possible in which to test the rhythm generating capabilities of hemi-spinal cords
(also see Jackson et al., 2005). First, under these conditions, sensory stimulation
elicited coordinated rostral and caudal locomotor muscle burst activity (Figure 3B1),
but with somewhat higher than normal burst proportions (BP) and intersegmental rostrocaudal
phase lags (ΦINT) (Figure 3C). Second, following a subsequent spinal transection at
the caudal end of the ML lesion in the same animals (Figure 3A), thereby isolating
rostral right and left hemi-spinal cords from intact caudal cord, sensory stimulation
usually elicited tonic, unpatterned rostral muscle activity (Figure 3B2). However,
for ~30% of the trials, stimulation could elicit relatively high frequency rostral
ipsilateral “burstlet” activity (Figures 5–7, 9). Importantly, there are at least
six features of “burstlet” activity suggesting that it does not represent well-coordinated
locomotor activity (p. 14). Our muscle “burstlet” activity (Messina et al., 2017)
and the in vitro “high-frequency rhythm” that is assumed to represent locomotor activity
(Cangiano and Grillner, 2003, 2005) have some significant differences, and it is uncertain
if they are equivalent.
In the previous commentary (Cangiano and Grillner, 2018), there were several incomplete,
misleading, or incorrect statements. First, the commentary stated that stimulation
of the in vitro lamprey hemi-spinal cord initiates a “bout of locomotor burst activity”
consisting of “well-coordinated bursting.” However, the term “well-coordinated locomotor
activity” embodies more than just burst frequency, but also should include data for
BP and ΦINT (Wallén and Williams, 1984; Boyd and McClellan, 2002). Because BP and
ΦINT are obvious and easy parameters to measure, it is surprising that these parameters
were not reported in the previous studies (Cangiano and Grillner, 2003, 2005). For
the “burstlet” activity in our study, plots of BP vs. cycle time (CT) usually were
abnormal (Figures 6A2–C2), and phase lags usually were absent (Figure 9; p. 9, lower
right column).
Second, the previous commentary stated that prior to a more caudal spinal transection
in our study, the muscle burst activity produced by rostral hemi-spinal cords was
“driven from the intact [caudal] part of the spinal cord.” However, in our experiments
under these conditions, average ΦINT values were positive (Figures 3B1, 3C3) and did
not change significantly with CT (unpublished), suggesting that intact caudal spinal
circuits were not driving (activating) bursts in the rostral hemi-spinal cords. Also,
experimental evidence does not support this proposed activation mechanism in the lamprey
spinal cord. Lastly, BP and ΦINT values for rostral muscle burst activity (adjacent
to spinal ML lesion) were only moderately higher than those during normal swimming
(Figures 3C2, 8; p. 9), suggesting that the CPGs in rostral hemi-spinal cords were
functional when connected to intact caudal spinal locomotor circuitry.
Third, the previous commentary stated that in our study, the presence of rostral muscle
burst activity prior to a more caudal spinal transection (Figures 3A, 3B1) was not
a convincing test of the integrity of the CPGs in rostral hemi-spinal cords. They
suggested that the mini-scalpel blade we used to make ML spinal lesions might have
damaged the spinal cord, and they showed a misleading image of the “dorsal view” of
this blade (i.e., wider, non-cutting edge). (a) The spinal cords of animals in our
study were ~200–300 μm thick (~1 mm wide), only the very tip of this blade was used
for cutting, and right and left halves of the spinal cord often separated slightly
during ML lesioning. (b) For ~35% of the experiments in Cangiano and Grillner (2003),
ML spinal lesions were made with a 0.4 mm dia. needle, which is more blunt and wider
than the tip of the blade we used, yet in their previous study there were no reported
experimental differences using the needle compared to using an ophthalmic blade.
Fourth, the previous commentary stated that the frequencies of muscle “burstlet” activity
for our experiments (Messina et al., 2017) were “somewhat higher” than those obtained
for their in vitro experiments (Cangiano and Grillner, 2003, 2005). However, for our
experiments, the average (~25 Hz) and maximum (>60 Hz) “burstlet” frequencies were
about 4–5 times higher than those observed for their in vitro experiments (~6 and
~12–15 Hz, respectively). (Note: the somewhat higher burst frequencies in Cangiano
et al., 2012 presumably occurred because recordings were made within minutes after
performing ML lesions, when the excitability of hemi-spinal cords probably was very
high.) Also, the average and maximum “burstlet” frequencies in our study were ~6 and
~10 times higher than the average and maximum muscle burst frequencies, respectively,
during swimming for normal whole animals (McClellan et al., 2016). The relatively
high “burstlet” frequencies appear to be due to changes in the properties of hemi-spinal
cords, such as an increase in excitability because of the lack of left-right coupling
(p. 13) and possibly lesion-induced cellular and synaptic plasticity (Hoffman and
Parker, 2010).
Fifth, the previous commentary stated that the “isolated condition [in vitro spinal
cord] is a much cleaner situation” than in our whole-animal experiments. However,
we are not convinced that synchronous, non-specific, high-frequency (33 Hz) stimulation
of the surface of hyper-excitable in vitro hemi-spinal cords (Cangiano and Grillner,
2003, 2005) is cleaner or more physiological compared to sensory stimulation and descending
brainstem activation of hemi-spinal cords in whole animals (Messina et al., 2017;
also see Jackson et al., 2005).
Sixth, the previous commentary incorrectly stated that in our paper we claimed “that
the burst generation is crucially dependent on reciprocal inhibition.” On the contrary,
we stated that “reciprocal inhibition mainly regulates left-right phasing of [lamprey]
locomotor activity and is not critical for rhythmogenesis” (p. 14; also see Hagevik
and McClellan, 1994). In addition, we emphasized that right-left coupling involves
both reciprocal inhibition as well as reciprocal excitation (p. 14, 15; Figure 1).
Thus, blocking left-right reciprocal inhibition is not a definitive test for rhythmogenesis
of hemi-spinal cords because right and left motor networks would still be coupled
by reciprocal excitation.
Seventh, the previous commentary stated that the higher than normal BP values for
“burstlet” activity in our study (Figures 6A2–C2) are expected, thereby implying that
the “burstlet” activity represents lamprey locomotor activity. Indeed, the higher
than normal BPs are expected (p. 13), but this represents a small part of only 1 of
the 6 reasons we provided to suggest that our “burstlet” activity does not represent
well-coordinated swimming activity (p. 14). For example, for the majority of our experiments,
BP values for “burstlet” activity changed significantly with CT (Figures 6A2–C2),
which is not characteristic of swimming motor activity in the lamprey.
In conclusion, our experiments employed the most physiological conditions possible
for testing the rhythm generating capabilities of hemi-spinal cords in adult lampreys.
Our results strongly suggest that isolated right and left spinal cord modules are
not autonomous and, by themselves, do not generate coordinated ipsilateral locomotor
burst activity.
Author contributions
The author confirms being the sole contributor of this work and approved it for publication.
Conflict of interest statement
The author declares that the research was conducted in the absence of any commercial
or financial relationships that could be construed as a potential conflict of interest.