Fig 1.
CCAP neuron and motoneuron activity.
(a) Single-plane calcium imaging of GCaMP3.2-expressing CCAP neurons. (b) Projection of 5 images from different planes, of GCaMP3.2-expressing CCAP neurons and motoneurons. The same scale bar from panel (a) applies. (c) Time series of signals from calcium sensor in AN1-AN4 α CCAP neurons and motoneurons recorded in a single experiment; time zero corresponds to moment of ETH-stimulation. The letter indicates the side (left [L] and right [R]), while the number indicates the abdominal segment of the neurons. Vertical marks denote the onset of oscillatory activity, as detected by the procedure indicated in Methods. (d) Mean time of onset of α CCAP neuron and motoneuron activity, for each of 9 separate experiments, showing temporally close values between populations. “MN” and “CCAP” indicate motoneurons and α CCAP neurons, respectively. Each MN-CCAP pair corresponds to a single experiment, with its 2 MN and 8 CCAP recordings. Box plots indicate intra-experiment median and quartiles.
Fig 2.
Oscillation period of CCAP neurons and motoneurons.
(a, b) Average scaleogram for all α CCAP neuron (a) and motoneuron time series (b). (c) Mean oscillation periods of CCAP neurons and motoneurons for all 9 experiments. For CCAP neurons, only the data for neurons that passed the selection criteria are shown (see text). In some experiments none were accepted, and the data are missing. By contrast, in all experiments motoneurons of the left and right regions showed a dominant oscillation period.
Fig 3.
(a) Pearson’s correlations pairs are divided into 3 groups: contralateral neurons (“C”), ipsilateral neurons (“I”), and “others” (“O”). As an example, the graph shows pairs that include the left AN2 neuron. (b-j) Correlation coefficients between the time series of CCAP neurons, shown as violin plots with their minimum, maximum and mean values. (b) α CCAP neuron correlations in the time domain (T). (c) β CCAP neuron correlations in the time domain. (d) αβ CCAP neuron correlations in the time domain. (e) Correlations of α CCAP neuron amplitude of oscillations (A) in the time-frequency domain. (f) Correlation of α CCAP neuron amplitude and phase (Aϕ) in the time-frequency domain. (g-j) Correlation between ipsilateral pairs with different segmental separation, “I1”, “I2” and “I3” groups are contiguous, separated by 1 and separated by 2 segments, respectively. The plots use the same notation as the plots in (b-f). Data were compared using a non-parametric Mann-Whitney U test. *: p-value < 0.05, **: p-value < 0.01, ***: p-value < 0.001.
Fig 4.
Coordination between motoneurons.
(a) Correlation coefficients between left and right motoneuron time series, calculated over a sliding window of 100 s. Time series for left (“MN L”) and right (“MN R”) motoneuronal regions are shown in orange and purple, respectively, and the Pearson’s correlation coefficient (“r”) of the sliding window is shown in black. Gray horizontal lines indicate correlations of −1, 0 and +1. (b) Example of the method used to compute the mean phase difference of an experimental recording. Blue and green points represent the phase difference at every instant of the experiment; their amplitude is scaled to the mean amplitude of the oscillations of the left and right region. The red point represents the mean vector, whose phase represents the mean phase difference of the experiment. (c) Mean phase difference for 9 experiments (blue) and mean for all experiments (red).
Fig 5.
Correlation between α CCAP neuron and motoneuron activity.
(a) Orange and purple lines show the activity of the left (“MN L”) and right (“MN R”) motoneuron regions, respectively, and the black line (labeled “Activity”) shows the amplitude of the motoneuron signal. (b) Correlation coefficients between α CCAP neurons and the amplitude of the motoneuronal time series, with individual data as points and boxes indicating median and quartiles. (c) Analogous to (b), but with the pre-ecdysis phase removed. Significance of the within-experiment correlations was tested using a one-tailed Mann-Whitney U test comparing to null cross-experiment correlations *: p-value < 0.05, **: p-value < 0.01, ***: p-value < 0.001.
Fig 6.
Fit of the logistic model to the experimental data.
. (a, b) Time series of α CCAP neurons (red, blue), motoneurons (orange, purple), binarized oscillatory activity of motoneurons (black), and probability of oscillation predicted by the model (green). CCAP activity traces are shown in red with their corresponding weight value if it is positive (“CCAP (Wi > 0)”), or in blue with no value if it is zero (“CCAP (Wi = 0)”). All weight values were computed through multi-weight model fitting. (a) Example of a good match between the model p(t) and the oscillatory state of the motoneurons. (b) Example of a poor match during the post-ecdysis phase, as a result of the lack of α CCAP activity.
Table 1.
AIC values of single- and multi-weight model fit for every experiment.
Fig 7.
Simulation of fluorescence spikes.
(a) Model circuit structure, showing the 8 α CCAP neurons that release the peptidergic signal (black circles) and activate the antiphase oscillatory behavior of the left and right motoneurons. (b) Simulated fluorescence (top) and voltage (bottom) time series of left (“Sim L”) and right (“Sim R”) motoneurons. (c) Magnification of a small segment of (b). (d, e) Simulation of two different experiments. The gray grid marks probabilities of 0, 0.5 and 1. “MN L”, “MN R”, “Sim L” and “Sim R” indicate experimental left, experimental right, simulated left, and simulated right motoneurons, respectively. p(t) indicates the probability of oscillation.
Fig 8.
Simulations of the Logistic model linked to a conductance-based bursting model.
(a) Experimental and simulated time series for 3 experiments. “MN L”, “MN R”, “Sim L” and “Sim R” indicate the left and right experimental, and left and right simulated motoneuron activity respectively. (b) Corresponding scaleograms, except that the scaleogram for the simulations is the average of 10 simulations. Each row shows the experimental (bottom) and the simulated (top) motoneuron activity, respectively.
Fig 9.
Sequence of image processing steps used to compute the midline of the pupa for every frame of the video. A RGB video frame is extracted (a) and converted to a grayscale image (b), then thresholded (c), and holes are then removed (d). Blob borders are softened by applying a gaussian filter (e) and thresholded again (f). Small blobs are discarded (g) and the left (green) and right (red) borders are computed (h). The mean of the two borders represents the midline (white line).
Fig 10.
(a) Time-space diagram pattern that allows the identification of three different motor routines. The “AP axis” represents the anteroposterior axis, which is oriented so that the top of the diagram corresponds to the anterior side of the pupa. The color indicates the position of the midline in the left-right axis along the anteroposterior axis, with the top and bottom of the diagrams corresponding to the anterior and posterior sections of the midline, respectively. Darker colors indicate that the midline section is closer to the left side, whereas lighter colors indicate that it is closer to the right side. (b) Filtered time-space diagrams for 6 pupal recordings aligned to the time when the ecdysis phase began. White spaces in the top diagram correspond to times when the pupa moved outside of the microscope viewing field. (c) Time series of the mid-section of (b) (“LR axis” represent the left-right axis).
Fig 11.
Metrics and comparison of motoneuronal activity (n = 9) and pupal behavior (n = 6). (a) Period of the characteristic motor patterns of each of the ecdysial phases. (b) Duration of the ecdysis and of the fast post-ecdysis phase.
Table 2.
Values of τK and τf used for the fit of each experiments.