Orienting gaze shifts during muscimol inactivation of caudal fastigial nucleus in the cat. I. Gaze dysmetria

The cerebellar control of orienting behavior toward visual targets was studied in the head-unrestrained cat by analyzing the deficits of saccadic gaze shifts after unilateral injection of muscimol in the caudal part of the fastigial nucleus (cFN). Gaze shifts are rendered strongly inaccurate by muscimol cFN inactivation. The characteristics of gaze dysmetria are specific to the direction of the movement with respect to the inactivated cFN. Gaze shifts directed toward the injected side are hypermetric. Irrespective of their starting position, all these ipsiversive gaze shifts overshoot the target by a constant horizontal error (or bias) to terminate at a "shifted goal" location. In particular, when gaze is directed initially at the future target's location, a response with an amplitude corresponding to the bias moves gaze away from the actual target. Additionally, when gaze is initially in between the target and this shifted goal location, the response again is directed toward the latter. This deficit of ipsiversive gaze shifts is characterized by a consistent increase in the y intercept of the relationship between horizontal gaze amplitude and horizontal retinal error. Slight increases in the slope sometimes are observed as well. Contraversive gaze shifts are markedly hypometric and, in contrast to ipsiversive responses, they do not converge onto a shifted goal but rather underestimate target eccentricity in a proportional way. This is reflected by a decrease in the slope of the relationship between horizontal gaze amplitude and horizontal retinal error, with, for some experiments, a moderate change in the y-intercept value. The same deficits are observed in a different setup, which permits the control of initial gaze position. Correction saccades rarely are observed when visual feedback is eliminated on initiation of the primary orienting response; instead, they occur frequently when the target remains visible. Like the primary contraversive saccades, they are hypometric and the ever-decreasing series of three to five correction saccades reduces the gaze fixation error but often does not completely eliminate it. We measured the position of gaze after the final correction saccade and found that fixation of a visible target is still shifted toward the inactivated cFN by 4.9 +/- 2.4 degrees. This fixation offset is correlated to, but on average 54% smaller than, the hypermetric bias of ipsiversive responses measured in the same experiments. In conclusion, the cFN contributes to the control of saccadic shifts of the visual axis toward a visual target. The hypometria of contraversive gaze shifts suggests a cFN role in adjusting a gain in the translation of retinal signals into gaze motor commands. On the basis of the convergence of ipsiversive gaze shifts onto a shifted goal, the straightness of gaze trajectory during these responses and the production of misdirected or inappropriately initiated responses toward this shifted goal, we propose that the cFN influences the processes that specify the goal of ipsiversive gaze shifts.     

Contribution of the rostral fastigial nucleus to the control of orienting gaze shifts in the head-unrestrained cat

The implication of the caudal part of the fastigial nucleus (cFN) in the control of saccadic shifts of the visual axis is now well established. In contrast a possible involvement of the rostral part of the fastigial nuceus (rFN) remains unknown. In the current study we investigated in the head-unrestrained cat the contribution of the rFN to the control of visually triggered saccadic gaze shifts by measuring the deficits after unilateral muscimol injection in the rFN. A typical gaze dysmetria was observed: gaze saccades directed toward the inactivated side were hypermetric, whereas those with an opposite direction were hypometric. For both movement directions, gaze dysmetria was proportional to target retinal eccentricity and could be described as a modified gain in the translation of visual signals into eye and head motor commands. Correction saccades were triggered when the target remained visible and reduced the gaze fixation error to 2.7 +/- 1.3 degrees (mean +/- SD) on average. The hypermetria of ipsiversive gaze shifts resulted predominantly from a hypermetric response of the eyes, whereas the hypometria of contraversive gaze shifts resulted from hypometric responses of both eye and head. However, even in this latter case, the eye saccade was more affected than the motion of the head. As a consequence, for both directions of gaze shift the relative contributions of the eye and head to the overall gaze displacement were altered by muscimol injection. This was revealed by a decreased contribution of the head for ipsiversive gaze shifts and an increased head contribution for contraversive movements. These modifications were associated with slight changes in the delay between eye and head movement onsets. Inactivation of the rFN also affected the initiation of eye and head movements. Indeed, the latency of ipsiversive gaze and head movements decreased to 88 and 92% of normal, respectively, whereas the latency of contraversive ones increased to 149 and 145%. The deficits induced by rFN inactivation were then compared with those obtained after muscimol injection in the cFN of the same animals. Several deficits differed according to the site of injection within the fastigial nucleus (tonic orbital eye rotation, hypermetria of ipsiversive gaze shifts and fixation offset, relationship between dysmetria and latency of contraversive gaze shifts, postural deficit). In conclusion, the present study demonstrates that the rFN is involved in the initiation and the control of combined eye-head gaze shifts. In addition our findings support a functional distinction between the rFN and cFN for the control of orienting gaze shifts. This distinction is discussed with respect to the segregated fastigiofugal projections arising from the rFN and cFN.     

Rostral fastigial nucleus activity in the alert monkey during three-dimensional passive head movements

The fastigial nucleus (FN) receives vestibular information predominantly from Purkinje cells of the vermis. FN in the monkey can be divided in a rostral part, related to spinal mechanisms, and a caudal part with oculomotor functions. To understand the role of FN during movements in space, single-unit activity in alert monkeys was recorded during passive three-dimensional head movements from rostral FN. Seated monkeys were rotated sinusoidally around a horizontal earth-fixed axis (vertical stimulation) at different orientations 15 degrees apart (including roll, pitch, vertical canal plane and intermediate planes). In addition, sinusoidal rotations around an earth-vertical axis (yaw stimulus) included different roll and pitch positions (+/-10 degrees, +/-20 degrees). The latter positions were also used for static stimulation. One hundred fifty-eight neurons in two monkeys were modulated during the sinusoidal vertical search stimulation. The vast majority showed a uniform response pattern: a maximum at a specific head orientation (response vector orientation) and a null response 90 degrees apart. Detailed analysis was obtained from 111 neurons. On the basis of their phase relation during dynamic stimulation and their response to static tilt, these neurons were classified as vertical semicircular canal related (n = 79, 71.2%) or otolith related (n = 25; 22.5%). Only seven neurons did not follow the usual response pattern and were classified as complex neurons. For the vertical canal-related neurons (n = 79) all eight major response vector orientations (ipsilateral or contralateral anterior canal, posterior canal, roll, and nose-down and nose-up pitch) were found in Fn on one side. Neurons with ipsilateral orientations were more numerous and on average more sensitive than those with contralateral orientations. Twenty-eight percent of the vertical canal-related neurons also responded to horizontal canal stimulation. None of the vertical canal-related neurons responded to static tilt. Otolith-related neurons (n = 25) had a phase relation close to head position and were considerably less numerous than canal-related neurons. Except for pitch, all other response vector orientations were found. Seventy percent of these neurons responding during dynamic stimulation also responded during static tilt. The sensitivity during dynamic stimulation was always higher than during static stimulation. Sixty-one percent of the otolith-related neurons responded also to horizontal canal stimulation. These results show that in FN, robust vestibular signals are abundant. Canal-related responses are much more common than otolith-related responses. Although for many canal neurons the responses can be related to single canal planes, convergence between vertical canals but also with horizontal canals is common.     

Respiratory-related neurons of the fastigial nucleus in response to chemical and mechanical challenges

Responses of cerebellar respiratory-related neurons (CRRNs) within the rostral fastigial nucleus and the phrenic neurogram to activation of respiratory mechano- and chemoreceptors were recorded in anesthetized, paralyzed, and ventilated cats. Respiratory challenges included the following: 1 ) cessation of the ventilator for a single breath at the end of inspiration (lung inflation) or at functional residual capacity, 2) cessation of the ventilator for multiple breaths, and 3) exposure to hypercapnia. Nineteen CRRNs having spontaneous activity during control conditions were characterized as either independent (basic, n = 14) or dependent (pump, n = 5) on the ventilator movement. Thirteen recruited CRRNs showed no respiratory-related activity until breathing was stressed. Burst durations of expiratory CRRNs were prolonged by sustained lung inflation but were inhibited when the volume was sustained at functional residual capacity; it was vice versa for inspiratory CRRNs. Multiple-breath cessation of the ventilator and hypercapnia significantly increased the firing rate and/or burst duration concomitant with changes noted in the phrenic neurogram. We conclude that CRRNs respond to respiratory inputs from CO2 chemo- and pulmonary mechanoreceptors in the absence of skeletal muscle contraction.     

Tooth pulp neurons in the caudal medulla oblongata of the cat

The medulla oblongata caudal to the obex was explored for neurons responsive to tooth pulp (TP) stimulation in cats. Four different subclasses of TP neurons were found. The latter included TP specific (TPS) neurons, trigeminal wide dynamic range (trigeminal WDR) neurons with TP input, trigeminal subnucleus reticularis ventralis (trigeminal SRV) neurons with TP input and convergent reticular formation (convergent RF) neurons with TP input. TPS neurons were located in the dorsal marginal rim of the trigeminal subnucleus caudalis, i.e., in the marginal layer or the outer zone of substantia gelatinosa. WDR neurons with TP input were found in the neck region of medullary dorsal horn which corresponds to the lateral part of subnucleus reticularis dorsalis (SRD). Trigeminal SRV neurons with TP input were located in the lateral part of SRV. Convergent RF neurons with TP input were found in the middle third of the caudal bulbar RF consisting of SRD and SRV. Both TPS neurons and WDR neurons with TP input included trigeminothalamic neurons as evidenced by the antidromic activation from the nucleus ventralis posteromedialis of the contralateral thalamus. A significant proportion of both trigeminal SRV and convergent RF neurons with TP input were antidromically activated by stimulation of the nucleus centralis lateralis of the contralateral thalamus. The former two subclasses may subserve the sensory-discriminative aspect of toothache, while the latter two subclasses, the emotional-motivational aspect.     

Electrophysiologic analysis of efferent projections of the cerebellar fastigial nucleus of cats

Stimulation of ipsilateral gigantocellular reticular and lateral vestibular nuclei of medulla oblongata, pontine lateral and medial nuclei, reticular formation of the midbrain, as well as the ipsi- and contralateral sensomotor cortex induced antidromic ATs recorded intracellularly in Fastigial neurons. Conduction velocity of Fastigial cell axons projecting to various nuclei of the brainstem was 30--35 m/sec and to the cerebral cortex--43--52 m/sec. Organization of the nuclear efferent system is peculiar by the branching of some axons into various structures of the brainstem.     

Medullary respiratory neuronal activity modulated by stimulation of the fastigial nucleus of the cerebellum

The ability of the rostral fastigial nucleus (FNr) of the cerebellum to modulate medullary respiratory neuronal activity was examined in 17 anesthetized, paralyzed and ventilated cats. A bipolar stimulating electrode was positioned into the FNr and tungsten microelectrodes used to record units within the nucleus tractus solitarius (NTS), nucleus ambiguus (NA) and nucleus retroambigualis (NRA). Transient stimuli (< 150 microA, 5-200 Hz) were delivered during inspiration or expiration, and the effects noted on medullary neuronal activity and the phrenic neurogram. The results showed that FNr stimulation: (1) modulated inspiratory and expiratory neuronal (ramp-, early- and late-inspiratory and stage I and II expiratory) discharges recorded from the NTS, NA and NRA (n = 67, 14 and 28) when stimuli (> or = 20-50 Hz) were delivered during either the inspiratory or expiratory phases; (2) terminated the burst durations of inspiratory (77%) and expiratory (94%) neurons with stimulus-response latencies of 28.2 +/- 3.1 ms (inspiratory) and 29.4 +/- 3.6 ms (expiratory); (3) elicited changes in phrenic neurogram concomitant with the effects noted on medullary neuronal activities; (4) failed to change heart rate and arterial blood pressure; and (5) did not affect medullary neuronal and phrenic nerve activity following kainic acid injection into the FNr. We conclude that activation of the FNr (likely its cell bodies) can modulate the respiratory output via influences on medullary respiratory-related neurons. The primary cerebellar effect across all sub-types of respiratory neurons was early termination.     

The role of the pedunculopontine tegmental nucleus in relation to conditioned motor performance in the cat. II. Effects of reversible inactivation by intracerebral microinjections

OBJECTIVE: To describe lower-extremity functioning in community-dwelling older Mexican Americans and to examine its relationship with medical problems. DESIGN: Cross-sectional analyses of survey and performance-based data obtained in a population-based study employing area probability sampling. SETTING: Households within selected census tracts of five Southwestern states: Arizona, California, Colorado, New Mexico, and Texas. PARTICIPANTS: A total of 2873 Mexican Americans aged 65 years and older. MEASUREMENTS: A multidimensional questionnaire assessing demographic, sociocultural, and health variables. Standardized tests of lower-extremity physical functioning included measures of standing balance, repeated chair stands, walking, and an overall summary measure. RESULTS: Regression analyses revealed that being more than age 75 and female, having arthritis diabetes, visual impairments, or being obese or underweight were all significantly associated with performance on both individual and summary tests of lower-extremity functioning. In separate regression analyses, the total number of medical conditions was also associated with performance. CONCLUSIONS: The likelihood of predicting performance or inability to complete tests of lower-extremity functioning was greatest for those aged 80 and older, those with arthritis or diabetes, and those with three or more medical conditions. Because of the high prevalence of diabetes in Mexican Americans, documentation of the association of diabetes with performance-based tests of lower-extremity functioning may help guide early interventions targeted to prevent progression to more severe limitations or disability.     

Neural unit activity in the trigeminal complex with interpositus or red nucleus inactivation during classical eyeblink conditioning.

During classical conditioning, many neurons in the trigeminal complex of rabbits exhibit activity that is related to the conditioned stimulus (tone), the unconditioned stimulus (airpuff), or to the conditioned response (eyeblink). For these reasons the trigeminal complex has been hypothesized to be a brainstem locus for the neuronal plasticity associated with conditioning. In this experiment, the learning-related activity (unit activity associated with the conditioned response) in the trigeminal is abolished when either the red nucleus or interpositus nucleus of the cerebellum is temporarily inactivated by cooling, but the stimulus-evoked activity is unaffected by cooling. This study and previous results support the suggestion that the learning-related activity seen in the trigeminal is driven by the interpositus by way of the red nucleus. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Structural characteristics of the mediodorsal nucleus of the cat thalamus

By means of the light and electron microscopy, construction and architectonics of neurons in the cat thalamus medio-dorsal nucleus are described. Differences in the neuronal body dimentions, their form, number of deferent dendrites and character of axonal branching made it possible to define 3 types of neurons. Applying the method based on retrograde axonal transport of exogenous horse-radish peroxidase, it was demonstrated that neurons of type II, contrary to the existing opinion, send their axons into the prefrontal cortex and, hence, are associative-projective. Ultrastructural difference of axonal terminals, differences in the form and size of synaptic vesicles made it possible to reveal 5 types of presynapses.     

Effects of caudal traction of the spinal cord on evoked spinal cord potentials in the cat

This study attempts clarify the mechanism of neurological deficits in tethered cord syndrome using evoked spinal cord potentials (ESCPs). ESCPs in response to both sciatic nerve (SN-ESCP) and spinal cord stimulation (SC-DESCP) were recorded from the dorsal epidural space. With a fixed degree of caudal traction on the spinal cord in ten cats for 2-4 hours, ESCPs were increased in amplitude in the N1 and N2 deflections of the SC-DESCPs to 158% and 154% at L5 and decreased to 91% and 76% after transient augmentation at L3. On the other hand, the amplitude in the N1 deflection of the SN-ESCPs at L3 and L5 was decreased to 40% and 68%. These findings suggest that not only the force but also the duration of traction influence the degree of the spinal cord dysfunction. When the spinal cords of 17 cats received compression with traction and without traction, the SN-ESCPs of the former became positive earlier than that of the latter. The extent of the recovery in amplitude of both SC-DESCPs and SN-ESCPs propagated over compression site was far limited in the former than in the latter. These results would indicate that the spinal cord subjected to traction is vulnerable to compression.     

Reflex facilitation during eyeblink conditioning and subsequent interpositus nucleus inactivation in the rabbit (Orycotolagus cuniculus).

In eyeblink conditioning in the rabbit (Oryctolagus cuniculus), not only is a conditioned response (CR) acquired, but also the original reflex is modified as a function of training. In Experiment 1, by comparing unconditioned responses in unpaired and paired groups, 3 types of reflex facilitation were distinguished. One type was linked to exposure to the unconditioned stimuli (USs) and/or experimental setting. The 2nd type was related to the formation of the memory trace for conditioned eyeblink. The 3rd type was linked to the conditioned stimulus immediately preceding the US in the paired group. In Experiment 2, reversible inactivation of the interpositus nucleus (IPN) abolished the CR and reduced the CR-related reflex facilitation, indicating that the latter depends on the plasticity of the IPN. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Morphologic characteristics of physiologically defined neurons in the cat trigeminal nucleus principalis

Although the principalis nucleus (Vp) contains trigeminothalamic and internuclear tract cells, the functional and morphologic differences between the two kinds of neurons have remained unsettled. The present study was aimed to address these problems by using the intracellular horseradish peroxidase injection technique in the cat. Of 20 neurons stained, 7 and 13 were located in the dorsomedial subnucleus (Vpd) and ventrolateral subnucleus (Vpv) of Vp, respectively. The Vpd neurons received input from the intraoral structures only but the Vpv neurons from the intraoral or facial structures. Nineteen neurons could be divided as class I and class II, based on the branching pattern of their stem axons. Class I (eight neurons) had an ascending stem axon without branching. Class II was divided into two subclasses (IIa and IIb). Class IIa (eight neurons) had an ascending stem axon from which branches were given off. Their branches formed a local-circuit restricted to the lower brainstem. Class IIb (three neurons) had a stem axon that formed the local-circuit only. The dendritic morphology was indistinguishable between different classes of neurons and between the subdivisions. Although the dendritic arborization pattern was governed by the location of the somata, it was suggested to be also important elements for determining primary afferent arborizations. In the brainstem nuclei, the jaw-closing motor nucleus received the highest density of projections from class II neurons with the receptive field involving the periodontal ligaments. The present study provides new findings that Vp neurons could be divided into three distinct populations and suggests that each population exerts a distinct function with respect to sensory discrimination, sensorimotor reflexes, or both.     

Pharmacological characterisation of excitatory synaptic transmission in the cat red nucleus in vivo

Extracellular recordings were made from the magnocellular neurones of the red nucleus (mRN) in anaesthetised cats. A study was made of the effects of selective excitatory amino acid receptor antagonists on excitatory monosynaptic responses evoked from the sensorimotor cortex (SMC) and cerebellar interpositus nucleus (IPN). Iontophoretically applied CNQX and NBQX antagonised both SMC and IPN responses whereas, D-AP5 inhibited the SMC response but was ineffective to the IPN. At currents that selectively antagonised NMDA responses, CPPene had no effect on either SMC or IPN responses. 7-chlorokynurenate inhibited both SMC and IPN responses but required currents that antagonised both AMPA and NMDA responses and was therefore acting in a non-selective manner. Iontophoretically applied glycine was inhibitory to both agonist and synaptic responses, whilst D-serine potentiated NMDA responses but did not enhance monosynaptic responses of the SMC. However in the presence of either 7-chlorokynurenate or high currents of CNQX that reduced the SMC synaptic activation of the mRN neurones, D-serine attenuated the inhibitory action of these antagonists. It is concluded that monosynaptic responses from the SMC are mediated by both NMDA and non-NMDA receptors whereas the monosynaptic responses evoked from the IPN are mediated only by non-NMDA receptors. The lack of effect of CPPene is consistent with the postulate that two NMDA receptor subtypes are present on mRN neurones.     

Effects of saccades on the activity of neurons in the cat lateral geniculate nucleus

Effects of saccades on individual neurons in the cat lateral geniculate nucleus (LGN) were examined under two conditions: during spontaneous saccades in the dark and during stimulation by large, uniform flashes delivered at various times during and after rewarded saccades made to small visual targets. In the dark condition, a suppression of activity began 200-300 ms before saccade start, peaked approximately 100 ms before saccade start, and smoothly reversed to a facilitation of activity by saccade end. The facilitation peaked 70-130 ms after saccade end and decayed during the next several hundred milliseconds. The latency of the facilitation was related inversely to saccade velocity, reaching a minimum for saccades with peak velocity >70-80 degrees /s. Effects of saccades on visually evoked activity were remarkably similar: a facilitation began at saccade end and peaked 50-100 ms later. When matched for saccade velocity, the time courses and magnitudes of postsaccadic facilitation for activity in the dark and during visual stimulation were identical. The presaccadic suppression observed in the dark condition was similar for X and Y cells, whereas the postsaccadic facilitation was substantially stronger for X cells, both in the dark and for visually evoked responses. This saccade-related regulation of geniculate transmission appears to be independent of the conditions under which the saccade is evoked or the state of retinal input to the LGN. The change in activity from presaccadic suppression to postsaccadic facilitation amounted to an increase in gain of geniculate transmission of approximately 30%. This may promote rapid central registration of visual inputs by increasing the temporal contrast between activity evoked by an image near the end of a fixation and that evoked by the image immediately after a saccade.     

Differential spinal projections from the forelimb areas of the rostral and caudal subregions of primary motor cortex in the cat

We used anterograde transport of WGA-HRP to examine the topography of corticospinal projections from the forelimb areas within the rostral and caudal motor cortex subregions in the cat. We compared the pattern of these projections with those from the somatic sensory cortex. The principal finding of this study was that the laminar distribution of projections to the contralateral gray matter from the two motor cortex subregions was different. The rostral motor cortex projected preferentially to laminae VI-VIII, whereas caudal motor cortex projected primarily to laminae IV-VI. Confirming earlier findings, somatic sensory cortex projected predominantly to laminae I-VI inclusive. We found that only rostral motor cortex projected to territories in the rostral cervical cord containing propriospinal neurons of cervical spinal segments C3-4 and, in the cervical enlargement, to portions presumed to contain Ia inhibitory interneurons. We generated contour maps of labeling probability on averaged segmental distributions of anterograde labeling for all analyzed sections using the same algorithm. For rostral motor cortex, heaviest label in the dorsal part of lamina VII in the contralateral cord was consistently located in separate medial and lateral zones. In contrast, no consistent differences in the mediolateral location of label was noted for caudal motor cortex. To summarize, laminae I-III received input only from the somatic sensory cortex, while laminae IV-V received input from both somatic sensory and caudal motor cortex. Lamina VI received input from all cortical fields examined. Laminae VII-IX received input selectively from the rostral motor cortex. For motor cortex, our findings suggest that projections from the two subregions comprise separate descending pathways that could play distinct functional roles in movement control and sensorimotor integration.     

Differential effects of morphine on corneal-responsive neurons in rostral versus caudal regions of spinal trigeminal nucleus in the rat

The initial processing of corneal sensory input in the rat occurs in two distinct regions of the spinal trigeminal nucleus, at the subnucleus interpolaris/caudalis transition (Vi/Vc) and in laminae I-II at the subnucleus caudalis/spinal cord transition (Vc/C1). Extracellular recording was used to compare the effects of morphine on the evoked activity of corneal-responsive neurons located in these two regions. Neurons also were characterized by cutaneous receptive field properties and parabrachial area (PBA) projection status. Electrical corneal stimulation-evoked activity of most (10/13) neurons at the Vi/Vc transition region was increased [146 +/- 16% (mean +/- SE) of control, P < 0.025] after systemic morphine and reduced after naloxone. None of the Vi/Vc corneal units were inhibited by morphine. By contrast, all corneal neurons recorded at the Vc/C1 transition region displayed a naloxone-reversible decrease (55 +/- 10% of control, P < 0.001) in evoked activity after morphine. None of 13 Vi/Vc corneal units and 7 of 8 Vc/C1 corneal units tested projected to the PBA. To determine if the Vc/C1 transition acted as a relay for the effect of intravenous morphine on corneal stimulation-evoked activity of Vi/Vc units, morphine was applied topically to the dorsal brain stem surface overlying the Vc/C1 transition. Local microinjection of morphine at the Vc/C1 transition increased the evoked activity of 4 Vi/Vc neurons, inhibited that of 2 neurons, and did not affect the remaining 12 corneal neurons tested. In conclusion, the distinctive effects of morphine on Vi/Vc and Vc/C1 neurons support the hypothesis that these two neuronal groups contribute to different aspects of corneal sensory processing such as pain sensation, autonomic reflex responses, and recruitment of descending controls.     

Glutamate, gamma-aminobutyric acid and tachykinin-immunoreactive synapses in the cat nucleus tractus solitarii

Neurophysiological and pharmacological evidence suggests that glutamate, gamma-aminobutyric acid and tachykinins (substance P and neurokinin A) each have a role in cardiovascular regulation in the nucleus tractus solitarii. This study describes the ultrastructural relationships between nerve terminals immunoreactive for these substances in the nucleus tractus solitarii of the cat using post-embedding immunogold (single and double) labelling techniques on sections of tissue embedded in LR White resin. The technique combines a high specificity of labelling with good ultrastructural and antigenic preservation. Glutamate-immunoreactive terminals, recognized by their high density of gold particle labelling compared to the mean tissue level of labelling, accounted for about 40% of all synaptic terminals in the region of the nucleus tractus solitarii analysed (medial, dorsal, interstitial, gelatinosus and dorsolateral subnuclei). They appeared to comprise several morphological types, but formed mainly asymmetrical synapses, most often with dendrites of varying size, and contained spherical clear vesicles together with fewer dense-cored vesicles. Substance P- and neurokinin A-immunoreactive terminals were fewer in number (9% of all terminals) but similar in appearance, with the immunoreaction restricted to the dense-cored vesicles. Analysis of serial- and double-labelled sections showed a co-existence of substance P and neurokinin A-immunoreactivity in 21% of glutamate-immunoreactive terminals. Immunoreactivity for gamma-aminobutyric acid was found in 33% of all terminals in the nucleus tractus solitarii. These predominantly contained pleomorphic vesicles and formed symmetrical synapses on dendrites and somata. Possible sites of axo-axonic contact by gamma-aminobutyric acid-immunoreactive terminals onto glutamate-or tachykinin-immunoreactive terminals were rare, but examples of adjacent glutamate and gamma-aminobutyric acid-immunoreactive terminals synapsing on the same dendritic profile were frequent. These results provide an anatomical basis for a gamma-aminobutyric acid mediated inhibition of glutamatergic excitatory inputs to the nucleus tractus solitarii at a post-synaptic level.