B.J. Yates

Vestibular effects on respiratory outflow in the decerebrate cat.

Recordings were made from phrenic, abdominal and intercostal nerves following electrical stimulation of the vestibular nerve to test for the presence of vestibulo-respiratory connections in cats that were decerebrate, paralyzed, and artificially ventilated. Short stimulus trains (2 or 3 shocks) at intensities < or = 125 microA typically elicited responses bilaterally in all of the respiratory nerves; the onset latency of the evoked activity was < 15 ms from the effective shock. The mean peak-to-peak amplitudes of integrated vestibulo-respiratory responses were 15% of the average amplitude of spontaneous respiratory-related discharges in the case of the phrenic nerve and 100% in the case of the abdominal nerve. The vestibulo-respiratory reflexes, as well as vestibulo-sympathetic responses recorded from the splanchnic nerve, could be abolished by injections of the excitotoxin kainic acid confined primarily to the medial and adjacent inferior vestibular nuclei. The physiological role of vestibulo-respiratory connections is yet to be determined, but possible functions include adjustments of respiration during changes in posture, assistance in venous return to the heart during movements that might lead to orthostatic hypotension, and direct participation in the execution of specific movements and the maintenance of some postures.

The ventrolateral medulla of the cat mediates vestibulosympathetic reflexes

Extracellular recordings were made from 94 neurons located in the ventrolateral medulla (VLM) whose firing rate was affected by vestibular nerve (VN) stimulation; 50 of these units were in the subretrofacial (SRF) nucleus, which contains cells that make direct excitatory connections with sympathetic preganglionic neurons. The sample included 12 SRF cells which were antidromically driven from the upper thoracic spinal cord and had conduction velocities of 10 m/s or less; the effect of VN stimulation on all but one of these units was inhibition. The onset latency of the response to VN stimulation was long [20.3 +/- 3.7 (S.E.M.) ms, n = 9, for the antidromically activated neurons and 12.1 +/- 1.2 ms, n = 73, for the others], suggesting that the effects were predominantly polysynaptic. In addition, most of the spontaneously active units tested (33/36) received convergent inputs from the carotid sinus nerve (CSN), as would be expected for neurons which influence sympathetic outflow. Vestibular-elicited inhibition of SRF neurons with projections to the intermediolateral cell column could account for late, long duration inhibition of sympathetic discharges produced by labyrinth stimulation.

Responses of neurons in the rostral ventrolateral medulla of the cat to natural vestibular stimulation.

To investigate the neural substrate of vestibulo-sympathetic reflexes, we studied the responses of neurons in the rostral ventrolateral medulla (RVLM) of decerebrate cats to natural stimulation of the labyrinth in vertical and horizontal planes. The RVLM is a major source of excitatory inputs to sympathetic preganglionic neurons. The animals used in these studies were baroreceptor-denervated and vagotomized and had a cervical spinal transection so that inputs from tilt-sensitive receptors outside of the labyrinth did not influence the units we recorded. Of the 38 neurons whose type of vertical vestibular inputs could be classified, the majority (27) received signals mainly from otolith organs. Only 4 of the neurons received inputs predominantly from vertical semicircular canals, and 7 were classified as having convergent inputs from otoliths and canals that were spatially aligned (2 cells) or misaligned (5 cells). In addition, only 2 of 68 neurons tested responded to sinusoidal horizontal rotations in a manner typical of brainstem neurons that receive inputs from the horizontal semicircular canals. Thus, the vestibular inputs to the RVLM appear to come mainly from otolith receptors. In labyrinthectomized cats, we were unable to locate neurons with responses to tilt similar to those of cells recorded in labyrinth-intact cats, confirming that the responses attributed to vertical vestibular inputs were produced by signals from the labyrinth. In animals whose semicircular canals had been rendered dysfunctional by plugging, we only recorded responses similar to those of neurons classified as having mainly otolith inputs in canal-intact animals, indicating that the dynamic behavior of these cells does not depend upon canal inputs. The presence of otolith inputs to the RVLM is consistent with the hypothesis that this region mediates vestibulo-sympathetic reflexes involved in correcting posturally-related changes in blood pressure.

Responses of neurons in the caudal medullary raphe nuclei of the cat to stimulation of the vestibular nerve

SummaryIn the decerebrate cat, recordings were made from neurons in the caudal medullary raphe nuclei to determine if they responded to electrical stimulation of the vestibular nerve and thus might participate in vestibulo-sympathetic reflexes. Many of these cells projected to the upper thoracic spinal cord. The majority (20/28) of raphespinal neurons with conduction velocities between 1 and 4 m/s received vestibular inputs; 13 of the 20 were inhibited, and 7 were excited. Since many raphespinal neurons with similar slow conduction velocities are involved in the control of sympathetic outflow, as well as in other functions, these cells could potentially relay vestibular signals to sympathetic preganglionic neurons. The onset latency of the vestibular effects was long (median of 15 ms), indicating the inputs were polysynaptic. In addition, 34 of 42 raphespinal neurons with more rapid conduction velocities (6–78 m/s) also received long-latency (median of 10 ms) labyrinthine inputs; 26 were excited and 8 were inhibited. Although little is known about these rapidlyconducting cells, they do not appear to be involved in autonomic control, suggesting that the function of vestibular inputs to raphe neurons is not limited to production of vestibulosympathetic reflexes. One hypothesis is that raphe neurons are also involved in modulating the gain of vestibulocollic and vestibulospinal reflexes; this possibility remains to be tested.

Evaluation of role of upper cervical inspiratory neurons in respiration, emesis and cough

Upper cervical (C1-3) inspiratory (UCI) propriospinal neurons project to spinal segments containing respiratory motoneurons, but their functional significance is unknown. Bilateral kainic acid injections into this cell column in 12 decerebrate cats (11 paralyzed and artificially ventilated) had no major effect on phrenic, intercostal, and abdominal nerve discharge or EMG activity during (fictive) respiration, vomiting and coughing. Thus, UCI neurons are unessential for activation of major respiratory muscles during these behaviors.

Convergence of cardiovascular and vestibular inputs on neurons in the medullary paramedian reticular formation

Extracellular recordings were made from 47 neurons in the caudal medullary paramedian reticular formation of chloralose-urethan anesthetized cats. The firing rate of all of these neurons was affected by stimulation of the contralateral vestibular nerve. The onset latency of the response to vestibular nerve stimulation was accurately determined for 43 of these neurons. The large majority (36 of 43) received short-latency (onset latency less than 5 ms) inputs; 7 were driven within 2 ms and were classified as receiving disynaptic inputs. Twenty of the 47 neurons also received inputs from the ipsilateral carotid sinus nerve. The paramedian reticular formation thus appears to be an important site for the integration of cardiovascular afferent and vestibular signals. However, only a small fraction of the neurons (2 of 13 tested) which received such convergent inputs could be antidromically driven from the upper thoracic spinal cord.

Effects of muscle and cutaneous hindlimb afferents on L4 neurons whose activity is modulated by neck rotation

SummaryWe recorded extracellularly, in decerebrate, labyrinthectomized cats, from spontaneously active L4 neurons whose activity was modulated by head rotation, and studied the effects of stimulation of ipsilateral hindlimb nerves. Rotation of the head about the longitudinal (roll) axis was more effective than rotation about the transverse (pitch) axis or vertical (yaw) axis for this group of neurons. Most units received convergent excitatory or inhibitory inputs from several nerves, with excitation being more prominent. The most effective muscle nerves were quadriceps (37/43 neurons), sartorius (19/21) and tibialis anterior (17/ 35); stimulation of biceps posterior-semitendinosus, biceps anterior-semimembranosus, or gastrocnemius rarely influenced the firing of the neurons. Group I effects were present in only a small fraction of neurons; however, short latency (central latency ≤ 5 ms) group II effects were observed in almost one-third. Longer latency group II as well as group III inputs were also common. All neurons received inputs from mixed and cutaneous nerves which usually had low thresholds and central latencies > 5 ms. Most recording sites were in medial lamina VII or lamina VIII; some of the units were identified by antidromic stimulation as propriospinal neurons which projected to the lumbar enlargement.

Participation of Ia reciprocal inhibitory neurons in the spinal circuitry of the tonic neck reflex

SummaryAs part of our studies of the spinal circuitry of the tonic neck reflex, we have recorded extracellularly from Ia reciprocal inhibitory neurons of the decerebrate, labyrinthectomized cat. The activity of a majority of neurons driven by stimulation of the quadriceps nerve was modulated by sinusoidal rotation of the neck; such modulation was much less frequent in the case of neurons driven by stimulation of nerves to more distal muscles. The results suggest that some of the inhibition which is part of the tonic neck reflex is mediated by Ia reciprocal inhibitory neurons, but that other pathways must also play an important role.

Peripheral input to L4 neurons whose activity is modulated by neck rotation

We studied, in decerebrate cats, peripheral input from the ipsilateral hindlimb to L4 neurons whose activity was modulated by neck rotation (neck-modulated neurons). Most neurons received convergent input from muscle, cutaneous and mixed nerves. In about half the neurons muscle input consisted of short-latency group I or group II excitation or inhibition, with group II effects far more frequent. Such synaptic actions were produced almost entirely by stimulation of quadriceps or sartorius. In the other neurons muscle afferents produced only late, diffuse excitation or inhibition, with thresholds usually in the group III range.

Spatial transformation in the vertical vestibulocollic reflex.

Movement of the body in space results in vestibulospinal reflexes acting on the head, limbs, and trunk. These reflexes, which have been studied extensively in neck and forelimb muscles of the decerebrate cat, are produced by converging otolith and canal inputs. The former predominate at low, the latter at high, frequencies of sinusoidal stimulation. Reflexly activated muscles have an associated response vector orientation that characterizes the stimulus plane evoking the maximal response. In forelimb muscles of decerebrate cats, vector orientation remains stable as the frequency of sinusoidal tilt increases: the reflex is apparently produced by convergent inputs from otolith and canal receptors that are spatially aligned.’ The situation is more complex for the vestibulocollic reflex (VCR) evoked by stimuli in vertical planes.2 In contrast to the frequency-stable response vector orientation in limb muscles, the orientation of neck muscle vectors can shift considerably as frequency changes. Indeed, at frequencies around 0.1-0.2 Hz, it may be impossible to define a “best” stimulus direction as all directions may have comparable efficacy. This behavior of the VCR was called spatiotemporal convergence (STC behavior), as it was hypothesized to arise from canal and otolith influences (with inherently different temporal properties) having different spatial characteristia2 Because of STC behavior, response vector orientations of neck muscles have been measured only at higher stimulus frequencies, such as 0.5 Hz, at which they are mainly produced by canal inputs. The orientations of these vectors, some being near pitch, others near roll, others at planes in between: suggest that many of them are produced by convergent inputs from more than one vertical canal. The question arises, where does the convergence that produces neck muscle vectors and STC behavior take place? More specifically, to what extent does it take place in the brain stem, i.e., in the vestibular nuclei and pontomedullary reticular formation? Our recent experiments have addressed this question.