|
Figure 1.
Swimming in response to headâskin stimulation
A, tadpole and video frames (150 frames sâ1) showing a tadpole viewed from above responding to touch on the left side of the head with a fine hair (arrow) by flexing to the left (frames 4â10) and then swimming off (frames 14â18). B, tracings from six similar videos after head stimulation on the left side (asterisk) plot head position to show flexion to left or right and swimming in a range of directions (arrows). C and D, an immobilized tadpole showing the brain, spinal cord and innervation of head skin by a sensory neuron in the trigeminal ganglion (tg). The positions of suction electrodes to record ventral root activity from swimming muscles (Rr, Rc, Lr, Lc) and to stimulate the right head skin are indicated. Fictive swimming can start on the stimulated side after touch (arrow, C) or current pulse (asterisk, D) stimulation to the head skin. E, histogram of latencies to the start of swimming in immobilized tadpoles following skin stimulation on the same (red bars) or opposite (black bars) side of the head. Bin size = 5 ms.
|
|
Figure 2.
Excitation and alternating firing of reticulospinal dINs on both sides after headâskin stimulation
A, positions of recording and stimulating electrodes. B, photograph of a recorded pair of dINs on either side of the hindbrain (dorsal view, rostral to the left; red rectangle in A) with soma, dendrites and both ascending and descending axons. The descending axon of the left side dIN was damaged during the dissection (at*). C and D, each panel shows three overlain responses from the right ventral root and the pair of dINs shown in B. In each case, swimming starts following a left headâskin stimulus (at arrowhead) and grey bars show the phase when spikes occur on the unstimulated side. C, the first dIN spike is at short latency on the left, stimulated side (red). D, there is a shortâlatency EPSP in the left dINs but the first spike is later on the unstimulated right side (black) and is followed by the first spike on the stimulated side.
|
|
Figure 3.
Alternating pattern of first and second reticulospinal dIN spike times on each side of the body following headâskin stimulation
A and B, responses from a pair of dINs different from the one shown in Fig. 2. In each case, activity starts following a headâskin stimulus (at arrowhead) and grey bars show the phase when spikes occur on the unstimulated side. A, the first dIN spike is at short latency on the stimulated side (red), whereas, in B, the first spike is later on the unstimulated side (black) and each is followed by the first spike on the other side. C, plots show spike occurrences on each side at different times after a stimulus to one side of the head (red and pink bars are the stimulated side; black and grey are the unstimulated side). Measures are from 230 responses in six dIN pairs. Bin size = 5 ms.
|
|
Figure 4.
Asymmetry in EPSP responses to lowâlevel skin stimulation in dINs on stimulated (red) and unstimulated (black) sides
AâC, example EPSPs in response to low level stimulation show differences between the two sides in recordings from three different dIN pairs. D, averages of 10 responses from records in C (continuous lines) show that the stimulated side EPSPs are earlier and rise faster. Dotted lines either side of curves are the SDs.
|
|
Figure 5.
Effects of removing the tIN population on one side on swimming responses to headâskin stimulation
A, stimulation and recording setâup and area where tINs were removed (blue). B, side of first ventral root response (mean ± SD) to electrical stimulation and light dimming in control (black) and treated animals with tINs removed on the stimulated side (blue). Responses to light were similar before and after treatment. C, delays to first ventral root response (median, IQR) on the stimulated and unstimulated side following headâskin stimulation. Delays increase significantly after removal of the tINs on the treated side only. Asterisks indicate a significant difference at P < 0.0001 (MannâWhitney test).
|
|
Figure 6.
Lesion experiments showing where trigeminal excitation can cross to the unstimulated side in the caudal hindbrain
A, summary of regions of the hindbrain where unilateral transection prevented swimming (pink) or where midâline cuts (green) still allowed the initiation of bilateral swimming following a headâskin stimulus applied to the right side (rhombomeres numbered). The trigeminal ganglia on the left side were severed to prevent contralateral sensory access (red line). B, example of a rostral unilateral transection (pink line). C, ventral root recordings show this lesion prevented swim initiation to a headâskin stimulus (arrowhead) but not to a trunkâskin stimulus (D). E, example of a midâline lesion (green line) that only left a short intact region of caudal hindbrain (arrow). F, this short intact region was sufficient to allow initiation of bilateral swimming following a headâskin stimulus to the right side. Ventral root recording positions on right (upper traces) and left (lower traces) sides are indicated in B and E.
|
|
Figure 7.
Sensory pathway rostral dorsolateral commissural neurons (rdlcs) respond to head and trunk skin stimulation
A, dorsal view showing the position of the stimulating and recording electrodes. B, rdlc neurobiotin filling showing the extent of the axon in hindbrain, with enlarged tracing of rdlc with soma, dendrites and axon crossing the midâline and ascending and descending on the contralateral side. C and D, traces showing rdlc response to ipsilateral trunk and headâskin stimulation (arrowhead) of increasing strength from black = no response, red = EPSP alone, blue = spike, green = response to strong stimulus. E, current injected into an rdlc evokes a single spike at threshold and fast multiple firing at higher levels. F, showing the position of the electrodes for paired recording and a record of responses of rdlc and dIN to right head stimulus that starts swimming, as well as low and higher magnification tracings of the recorded cells in the box. G, responses of another rdlc/dIN pair to headâskin stimuli, aligned by rdlc spikes to show EPSPs with variable delays (arrowheads) and shapes. Coupling artefact indicated by an asterisk (*).
|
|
Figure 8.
Swim initiating network
A, brain in dorsal view and the headâskin trigeminal pathways initiating swimming on the same or opposite sides. Dots show locations of the neuron populations, not their real numbers: touchâsensory trigeminal neurons (tSt, yellow), sensory pathway neurons (tIN, pale magenta; rdlc, red), electrically coupled reticulospinal excitatory neurons (dIN, brown). Fb, forebrain; hb, hindbrain; mb, midbrain; m, muscles; oc, otic capsule; sc, spinal cord; tg, trigeminal ganglion. B, functional diagram of the network including the central pattern generator (swim circuit) neurons: inhibitory interneurons (cIN, blue; aIN, purple) and motoneurons (mn, green). Continuous lines indicate evidence for a monosynaptic connection. Dashed lines indicate indirect connections (delay). Large circles represent populations of neurons. Small circles (inhibitory) and triangles (excitatory) are synapses and, when they contact a box, they connect to all neurons in the box.
|