The brain matters: effects of descending signals on motor control

Abstract

The ability of nerve cords and spinal cords to exhibit fictive rhythmic locomotion in the absence of the brain is well-documented in numerous species. Although the brain is important for modulating the fictive motor output, it is broadly assumed that the functional properties of neuronal circuits identified in simplified preparations are conserved with the brain attached. We tested this assumption by examining the properties of a novel interneuron recently identified in the leech (Hirudo verbana) nerve cord. This neuron, cell E21, initiates and drives stereotyped fictive swimming activity in preparations of the isolated leech nerve cord deprived of the head brain. We report that, contrary to expectation, the motor output generated when cell E21 is stimulated in preparations with the brain attached is highly variable. Swim frequency and episode duration are increased in some of these preparations and decreased in others. Cell E21 controls swimming, in part, via excitatory synaptic interactions with cells 204, previously identified gating neurons that reliably initiate and strongly enhance leech swimming activity when the brain is absent. We found that in preparations with the brain present, the magnitude of the synaptic interaction from cell E21 to cell 204 is reduced by 50% and that cell 204-evoked responses also were highly variable. Intriguingly, most of this variability disappeared in semi-intact preparations. We conclude that neuronal circuit properties identified in reduced preparations might be fundamentally altered from those that occur in more physiological conditions.

Fig. 1.

Cell E21 excitation enhances swimming in midbody ganglion 2 through tail brain (M2-T) isolated preparations. A: schematic of M2-T isolated nerve cord. Head (H)-M1 is removed, and suction electrodes placed on dorsal-posterior (DP) nerves monitor swimming (M10) or deliver a shock for swim initiation (M18). A microelectrode is used to record from cell E21 (M21). B: control swim in a M2-T preparation (top 2 traces). Current injection into E21 (beginning at arrow, 3rd trace) during an ongoing swim extends swim duration (bottom trace). Bursts in DP traces here and elsewhere comprise spikes from motor neuron cell DE-3. Note that the time scale is 5 s, hence bursts are condensed in appearance. C: stimulation of cell E21 (top trace, at arrow) during an ongoing swim episode (2nd trace) decreases cycle period in a M2-T preparation. Graph shows the explicit periods of the cycles displayed above, with the “0” swim cycle indicating the cycle during which current injection was initiated. Gray shading indicates swim cycles when cell E21 was stimulated. The swim cycle at stimulus initiation is not shaded. Thick dashed line approximates cycle period before E21 stimulation.

Fig. 2.

Swim duration modulation by excitatory stimulation of cell E21 in H-T preparations. A: preparation. Extracellular nerve recordings sites vary in location. B–D: examples of the swim response to cell E21 stimulation (E21-stim) in H-T preparations. The top trace in each pair of traces is the control swim (no current injection); in the bottom traces, depolarizing current was injected into cell E21 (indicated by dashed lines) during fictive swimming. Swims were initiated by shock applied to a DP nerve (gray bars). These examples were taken from preparations in which cell E21 stimulation significantly decreased swim duration (B), increased swim duration (C), or had no significant effect on swim duration (D). E: averaged results from control and current-injected (stim) trials across all experiments. In M2-T, brain-removed preparations, current injection reliably increased swim duration, but effects were variable in H-T, brain-attached preparations, with no overall trend. F: effect on individual preparations. E21 stimulation increased swim duration in every M2-T preparation (n = 7). Swim duration was significantly increased (n = 3/11), decreased (n = 2/11), or unchanged (n = 6/11) by E21 stimulation in H-T preparations. Short bars indicate 0 values. Data in B–D are from 3 different leeches. BPE, bursts per swim episode. 2 s Applies to all scales. **P < 0.01.

Fig. 3.

Effects of cell E21 stimulation on cycle period. A: decreased cycle period. Current injection into cell E21 (top trace, at arrow) decreased cycle period (extracellular recording, bottom trace) in a H-T preparation. Explicit values of cycle period are shown in the graph below the traces. Gray shading indicates cycles during cell E21 stimulation; thick dashed line approximates the periods before stimulation. B: increased cycle period in a H-T preparation. Data are as in Fig. 2A. Data from A and B are from different preparations. C: stimulation of cell E21 during swimming (stim) decreased cycle period in M2-T preparations. The overall mean cycle periods were not significantly different in H-T preparations between the groups (P = 0.081). Cycle period is normalized. D: effect on individual preparations. E21 stimulation decreased cycle period in every M2-T preparation (n = 7) but only in 2/9 H-T preparations. In 5/9 H-T preparations, this stimulation increased cycle periods. In 2 preparations, there were no statistically significant effects. ***P < 0.01.

Fig. 4.

Swim response variability to E21 stimulation is greatly reduced by sensory environment. A: semi-intact preparation. Anterior portion of the leech is partially intact and suspended in a well. The posterior end is isolated. B1: control of swim expression by cell E21 in an H-T semi-intact leech suspended in deep water. Top 2 traces show a control swim episode. When cell E21 is stimulated by current injection (3rd trace, at arrow), swim duration increases (bottom trace). B2: bar graph demonstrating the increase in swim duration from E21 stimulation (dark gray bars) in all semi-intact M2-T and H-T preparations compared with control swims (light gray bars). M2-T, n = 3; H-T, n = 9. C1: current injection into cell E21 (top trace, at arrow) during an ongoing swim episode (2nd trace) decreases the cycle period. Graph shows explicit periods of cycles in the above trace; those that occurred completely during cell E21 stimulation are shaded in gray. C2: bar graph demonstrating changes in cycle period from E21 stimulation, as in B2. Data from 1 H-T preparation in which cell E21 stimulation significantly increased cycle period are not included. M2-T, n = 3; H-T, n = 11. **P < 0.01, ***P < 0.001. Bars are SE.

Fig. 5.

Summary of E21 effects on swimming in all preparations. A: fold increase in swim duration in E21-depolarized trials normalized to control trials in M2-T (top row) and H-T (bottom row) and isolated (1st column) and semi-intact (2nd column) preparations. Each pair of points connected with a line represents 1 experiment, and values >1.0 represent an increase in swim duration. Unlike the other conditions, E21 excitation sometimes decreased swim duration in isolated H-T preparations. B: fold increase in cycle period due to E21 excitation. Data are plotted as in A. Values <1.0 indicate a decrease in cycle period. E21 excitation often increased cycle period in H-T isolated preparations but decreased cycle period in all other conditions, with 1 outlier in the semi-intact H-T condition.

Fig. 6.

Some of the variable effects of E21 on swimming in H-T preparations are mediated by cell 204. A: an example of cell 204 depolarization increasing cycle period. Cell 204 was hyperpolarized during the remaining portions of the swim episode. B: comparison of the effects of 204 depolarization on swim duration and cycle period in individual H-T and M2-T preparations. Number of preparations analyzed is as follows. H-T: swim duration, n = 8; cycle period, n = 5. M2-T: swim duration, n = 9; cycle period, n = 3. L, left.

Fig. 7.

The cell E21-to-204 interaction is altered by the head brain. A: representative response of cell 204 (top trace, located in M10) to stimulation of E21 (middle trace, at arrow) in an M2-T preparation (inset). This excitation initiated swimming (bottom trace). B: representative response of cell 204 (top trace, located in M11) to E21 stimulation (middle trace) in an H-T preparation (inset). Despite the slightly higher E21 firing frequency and longer current pulse than in A, the increase in impulse frequency in cell 204 is less than in the M2-T preparation. Here, excitation of cell E21 did not initiate swimming (bottom trace). A and B are from the same nerve cord. C and D: plot of cell 204 vs. cell E21 impulse frequency in M2-T (n = 4 leeches, 96 trials; C) and H-T (n = 5 leeches, 94 trials; D) preparations. The M2-T slope, 0.65, is significantly steeper than the H-T slope, 0.32 (P < 0.01), and the r² values are significantly different (0.80 vs. 0.56, respectively, P = 0.038). Insets at top refer to data in the entire column underneath each respective nerve cord.

Fig. 8.

Cell E21 excitation drives multiple locomotor responses dependent on the sensory environment. A: schematic of setup. Semi-intact H-T leeches were placed in the dish shown in Fig. 4A; here, the well of the dish was filled with pebbles on which the intact portion of the leech lay. Arrows approximate the “low,” “medium,” and “high” fluid levels. In low levels, the fluid did not rise above the pebbles. In medium levels, the fluid level approximately just covered the leech body at rest. For high levels, the fluid was ∼1 cm above the leech body. B and C: cell E21 stimulation drives crawling in low fluid levels. B: the long DP nerve bursts before E21 excitation show spontaneous crawling activity; this stimulation reduced cycle period. C: stimulating cell E21 during a period of no locomotor activity in low fluid initiated and maintained a crawl episode. D: with a medium fluid level, a crawling episode initiated and maintained by E21 excitation was interrupted by 2 swim episodes, marked by an “s.” E: despite the presence of a substrate, E21 excitation almost exclusively drove swimming in deep fluid. F: here, in low fluid, E21 excitation initiated a shortening response. The animal remained shortened for the entire 41-s stimulation and for another 90 s following current termination. Motor neuron DE-3 fires tonically; the largest spikes in this recording are those of the shortener motor neuron (L cell; Ort et al. 1974). B and F are from 1 preparation; Note difference in time scales in the traces. Arrows indicate start of depolarizing current injection.

Fig. 9.

Proposed mechanism underlying variability. Sensory cell input to E21 elicits activity in 204, which leads to swimming. A: in the absence of the brain, depolarization of E21 or 204 during swimming simply leads to an enhancement of the behavior. B and C: we propose that in the presence of the brain, cell E21 and 204 directly or indirectly stimulate cephalic neurons that suppress swimming (SSN), perhaps through inhibition of the swim oscillator interneurons, creating simultaneous competing activity in swim-activating and -inactivating systems. The expression of behavior depends on the totality of sensory inputs and the internal state of the system, which is altered by neuromodulators. B: in state a, SSNs are inhibited by sensory input and E21 and 204 stimulation activate and enhance swimming. Only dominant interactions are shown. C: in state b, SSNs are activated by sensory input and by E21 and 204. The SSNs then compete with excitation from cells 204 and inhibit swimming. SSNs may also inhibit 204 through either direct circuit connections or neuromodulatory action. Gray interactions are hypothetical.