Electron-microscopic analysis of synaptic input from the perigeniculate nucleus to the A-laminae of the lateral geniculate nucleus in cats.

The perigeniculate nucleus of carnivores is thought to be a part of the thalamic reticular nucleus related to visual centers of the thalamus. Physiological studies show that perigeniculate neurons, which are primarily GABAergic, provide feedback inhibition onto neurons in the lateral geniculate nucleus. However, little is known about the anatomical organization of this feedback pathway. To address this, we used two complementary tracing methods to label perigeniculate axons for electron microscopic study in the geniculate A-laminae: intracellular injection of horseradish peroxidase (HRP) to fill an individual perigeniculate cell and its axon; and anterograde transport of Phaseolus vulgaris leucoagglutinin to label a population of perigeniculate axons. Labeled perigeniculate terminals display features of F1 terminals in the geniculate neuropil: they are small, contain dark mitochondria, and form symmetric synaptic contacts. We found that most of the perigeniculate terminals (greater than 90%) contact geniculate cell dendrites in regions that also receive a rich innervation from terminals deriving from visual cortex (e.g., "cortico-recipient" dendrites). The remainder of the perigeniculate synapses (10%) contacted dendrites in regions that also received direct retinal input (e.g., "retino-recipient" dendrites). Serial reconstruction of segments of dendrites postsynaptic to perigeniculate terminals suggests that these terminals contact both classes of relay cell in the A-laminae (X and Y), although our preliminary conclusion is that an individual perigeniculate cell contacts only one class. Finally, our quantitative comparison between labeled perigeniculate terminals and unlabeled F1 terminals indicates that these perigeniculate terminals form a distinct subset of F1 terminals. We quantitatively compared the labeled perigeniculate terminals to unlabeled F1 terminals. Although the parameters of the perigeniculate terminals fell entirely within the range of those for the unlabeled F1 terminals, as populations, we found consistent differences between these two groups. We thus conclude that, as populations, other sources of F1 terminals are morphologically distinct from perigeniculate terminals and innervate different targets.     

Topographic organization of projections from the thalamic reticular nucleus to the anterior thalamic nuclei in the rat

We have investigated connections between the thalamic reticular nucleus (TRN) and the anterior thalamic nuclei (ATN) in the rat, following injections of horseradish peroxidase (HRP) into subnuclei of the ATN and different regions of the rostral TRN. Three nonoverlapping groups of neurons in the dorsal part of the ipsilateral rostral TRN project to, and receive reciprocal projections from, specific subnuclei of the ATN. A vertical sheet of neurons in the most dorsal part of the rostral TRN projects to the dorsal half of the posterior subdivision of the anteroventral thalamic nucleus (AVp), the dorsal region of the medial subdivision of the anteroventral thalamic nucleus (AVm), and the dorsolateral part of the rostral anterodorsal thalamic nucleus (AD). Immediately ventral to this part of TRN, but still within its dorsal portion, are a lateral cluster of neurons and a medially located vertical sheet of neurons. The lateral cluster projects to the ventral part of AVp and to the dorsomedial part of rostral AD. The medial sheet projects to the ventral part of AVm, the ventral part of rostral AD, and to the caudal portions of both AV and AD. There appears to be no input to the anteromedial thalamic nucleus (AM) from the TRN. These findings shed new light on the anatomy of the rostral TRN, the ATN, and the connections between the two, and are relevant to emerging hypotheses about the functional organization of the TRN and reticulo-thalamic projections.     

Organization of diencephalic projections from the medullary subnucleus reticularis dorsalis and the adjacent cuneate nucleus: A retrograde and anterograde tracer study in the rat

The distribution and organization of diencephalic projections from the subnucleus reticularis dorsalis (SRD) and the neighbouring cuneate nucleus (Cu) were studied in the rat by using microinjections of Phaseolus vulgaris leucoagglutinin in SRD and Cu and wheat germ agglutinin-apo horseradish peroxidase-gold in some selected thalamic areas. As previously reported, the efferent projections from the Cu were essentially contralateral and terminated mainly in the ventroposterolateral thalamic nucleus. Less dense terminals from the Cu were also observed in the posterior thalamic group, the ventral aspect of the zona incerta and the caudal and dorsal portion of the reuniens area. Retrograde tracer injections in the medial ventroposterolateral thalamic nucleus labeled numerous cells in the contralateral Cu, with a smaller number in the gracile nucleus. From the SRD, terminals were observed in the lateral aspect of the ventromedial thalamic nucleus, the lateral parafascicular area and, to a lesser extent, in the ventral aspect of the zona incerta and the core of the reuniens area. Retrograde tracer injections in the lateral part of the ventromedial thalamic nucleus labeled cells in the caudal medulla, many of which were located in the dorsal-most aspect of the SRD throughout its caudo-rostral extent. The existence of SRD-thalamic connections reinforces the idea that the caudal reticular formation is an important nociceptive relay to the thalamus. Our data shed new light on old hypotheses suggesting that, in addition to spino-thalamic pathways, spino-reticulo-thalamic pathways may play an important role in distributing pain signals to the forebrain. J. Comp. Neurol. 390:133–160, 1998. © 1998 Wiley-Liss, Inc.     

Differential projections from the mediodorsal and centrolateral thalamic nuclei to the frontal cortex in rats

The aim of the present study was to investigate afferent projections from the medial thalamic nuclei (MT) to the frontal cortical areas using a single small iontophoretic injection of biotinylated dextran amine (BDA) and analysis of the anterogradely labeled fibers and varicosities. Projections from the mediodorsal (MD) nuclei were found primarily and extensively in the anterior cingulate cortex (ACC), whereas those from the centrolateral (CL) thalamic nucleus were found in the frontal motor cortex. The density of terminals in the ACC was high in layers II and III and sparse in layer I. The majority of projected fibers from the CL were found at a high density in layer V, with a moderate density in the superficial layers. The differential projection patterns were topographically organized in the medial prefrontal cortex and sensory motor cortex. These findings support the results of our previous electrophysiological studies suggesting that neurons in the medial thalamic nuclei relay nociceptive information to the limbic or sensory motor cortical areas. The present results agree with the current notion that the medial thalamo-frontal cortical network circuitry plays an important role in processing the emotional aspect of nociception.     

The integrative role of the rat medullary subnucleus reticularis dorsalis in nociception

Neurons within the medullary subnucleus reticularis dorsalis (SRD) of the rat convey selectively nociceptive information from all parts of the body. We have sought to define the neuronal networks that convey information from widespread noxious stimuli to the diffuse thalamocortical system and also modulate spinal outflow. The experiments, which were performed in rats, were designed to determine whether efferents from the SRD issue collaterals to the thalamus and spinal cord. Injections of the tracers fluorogold and tetramethylrhodamine-labelled dextran were centred stereotaxically in two areas that receive dense projections from the SRD: the cervical spinal cord and the lateral ventromedial thalamus (VMl), respectively. In other experimental series, SRD neurons were characterized electrophysiologically and individually labelled in a Golgi-like manner following juxtacellular iontophoresis of biotin-dextran. More than half reticulothalamic neurons within the SRD provided monosynaptic connections to the spinal cord. SRD neurons that responded to Adelta- or Adelta- and C-fibre activation from any area of the body had axons that gave both ascending and descending collaterals. Because the SRD innervates several areas involved in motor processing and receives strong, direct influences from several cortical regions, it could provide a structural basis for the processing of nociceptive and motor activities.     

Inhibitory pathway from the frontal cortex to the hypothalamic ventromedial nucleus in the rat

Stimulation of the ventromedial nucleus (VMH) evoked a short latency negative wave with two peaks cN₁ and cN₂ followed by a small positive wave (cP) at the ipsilateral dorsal frontal cortex (area 10) in the rat. The maximum response was observed from the lateral edge of the frontal pole. From the depth profiles of recordings, cN₁ changed polarity at a depth of about 4 mm and the cN₁−cN₂ changed into a large compound action potential at the medioventral part of the frontal pole at a depth of about 6 mm. Since the surface evoked potential and the compound action potential followed high frequency stimulation, these respective potentials are concluded to be due to antidromic and monosynaptic activation of the cortical neurons. This was verified by unit recording experiments. The cP was concluded to be produced by the initial rise of the monosynaptic EPSP.     

Monosynaptic facilitatory pathway from the hypothalamic ventromedial nucleus to the frontal cortex in the rat

Stimulation of the ventromedial nucleus (VMH) evoked a short latency negative wave with two peaks cN₁ and cN₂ followed by a small positive wave (cP) at the ipsilateral dorsal frontal cortex (area 10) in the rat. The maximum response was observed from the lateral edge of the frontal pole. From the depth profiles of recordings, cN₁ changed polarity at a depth of about 4 mm and the cN₁?cN₂ changed into a large compound action potential at the medioventral part of the frontal pole at a depth of about 6 mm. Since the surface evoked potential and the compound action potential followed high frequency stimulation, these respective potentials are concluded to be due to antidromic and monosynaptic activation of the cortical neurons. This was verified by unit recording experiments. The cP was concluded to be produced by the initial rise of the monosynaptic EPSP.     

Organization of the projections from the pericruciate cortex to the pontomedullary brainstem of the cat: A study using the anterograde tracer Phaseolus vulgaris-leucoagglutinin

The anterograde tracer Phaseolus vulgaris-leucoagglutinin (PHA-L) was used to study the distribution and density of the projections that originate from four identified subdivisions of the pericruciate cortex (namely, the forelimb and hind limb representations of area 4, area 6aβ, and area 6aγ) and that terminate in the pontomedullary brainstem in the cat. Injections of PHA-L in all areas of the pericruciate cortex labelled numerous fibers and their terminal swellings in the brainstem. The major target regions of all four cortical areas were the pontine nuclei and the pontomedullary reticular formation (PMRF). Injections into both the forelimb and hind limb representations of area 4 and into area 6aβ resulted in a dense pattern of terminal labelling in restricted regions of the medial and lateral parts of the ipsilateral pontine nuclei. The labelling following the area 6aβ injection was spatially distinct from that seen following the area 4 injections. Injections into the forelimb representation of area 4 as well as into area 6aβ and 6aγ resulted in the labelling of numerous terminal swellings bilaterally in the PMRF; in contrast, there were few labelled terminal swellings in the PMRF following injections into the hind limb representation of area 4. Terminal swellings on individual corticoreticular fibers were far less densely aggregated than those in the pontine nuclei. The dense pattern of innervation to restricted regions of the pontine nuclei supports previous suggestions that the corticopontine projections retain a high degree of topographical specificity that could be used in the control of discrete voluntary movements. In contrast, the more diffuse pattern of the projections to the PMRF may facilitate the selection and activation of the complex postural patterns that accompany voluntary movement. J. Comp. Neurol. 389:617–641, 1997. © 1997 Wiley-Liss, Inc.     

Branching projections from mesopontine nuclei to the nucleus reticularis and related thalamic nuclei: A double labelling study in the rat

Branching projections from pedunculopontine and laterodorsal tegmental nuclei to different thalamic targets were studied by means of a double retrograde tracing technique. The results show a topographic distribution of mesopontine neurons projecting to different thalamic targets. In addition, the present data demonstrate that a small percentage (≤ 5%) of mesopontine neurons projecting to the intralaminar nuclei or to the rostral pole of the reticular nucleus innervate both these areas by means of branching axons. By contrast, a large number of mesopontine neurons projecting to the sensorimotor thalamic nuclei send axon collaterals to the caudal part of the reticular nucleus. The present findings support the hypothesis of an inhomogeneity of different sectors of the thalamic reticular nucleus. Thus, this nucleus can be differentiated into two functional areas, in accordance with their connections with functionally different cortical fields and thalamic districts. The possibility that these two areas of the thalamic reticular nucleus subserve different mechanisms during sleep phenomena is discussed. © 1993 Wiley-Liss, Inc.     

Organization of subcortical pathways for sensory projections to the limbic cortex. II. Afferent projections to the thalamic lateral dorsal nucleus in the rat

Afferent projections to the thalamic lateral dorsal nucleus were examined in the rat by the use of retrograde axonal transport techniques. Small iontophoretic injections of horseradish peroxidase were placed at various locations within the lateral dorsal nucleus, and the location and morphology of cells of origin of afferent projections were identified by retrograde labeling. For all cases examined, subcortical retrogradely labeled neurons were most prominent in the pretectal complex, the intermediate layers of the superior colliculus, and the ventral lateral geniculate nucleus. Labeled cells were also seen in the thalamic reticular nucleus and the zona incerta. Within the cerebral cortex, labeled cells were prominent in the retrosplenial areas (areas 29b, 29c, and 29d) and the presubiculum. Labeled cells were also seen in areas 17 and 18 of occipital cortex. Peroxidase injections in the dorsal lateral part of the lateral dorsal nucleus result in labeled neurons in all of the ipsilateral pretectal nuclei, but especially those that receive direct retinal afferents. Labeled cells were also seen in the ventral lateral geniculate nucleus and the rostral tip of laminae IV-VI of the superior colliculus. In contrast, peroxidase injections in ventral medial portions of the lateral dorsal nucleus result in fewer labeled pretectal cells, and these labeled cells are found exclusively in the pretectal nuclei that do not receive retinal afferents. Other labeled cells following injections in the rostral and medial portions of the lateral dorsal nucleus are seen contralaterally in the medial pretectal region and nucleus of the posterior commissure, and bilaterally in the rostral tips of laminae IV and V of the superior colliculus. Camera lucida drawings of HRP labeled cells reveal that projecting cells in each pretectal nucleus have a characteristic soma size and dendritic branching pattern. These results are discussed with regard to the type of sensory information that may reach the lateral dorsal nucleus and then be relayed on to the medial limbic cortex.     

Mode of termination of afferents from the thalamic reticular nucleus in the dorsal lateral geniculate nucleus of the rat

The terminals of axons projecting to the dorsal lateral geniculate nucleus from the thalamic reticular nucleus were identified by electron microscopy 8–24 h after placing small lesions in the ipsilateral reticular nucleus. The terminals contained flattened synaptic vesicles and made Gray type II axo-dendritic synaptic contacts with geniculate neurons. Their identification as F-axons accords well with physiological evidence for a powerful monosynaptic inhibitory input to geniculocortical projection cells from reticular nucleus neurons.     

Contribution of brainstem GABAergic circuitry to descending antinociceptive controls: I. GABA-immunoreactive projection neurons in the periaqueductal gray and nucleus raphe magnus

The fact that GABA receptor agonists and antagonists influence nociceptive thresholds when microinjected into the rostroventral medulla or in the spinal cord may reflect the involvement of GABAergic neuronal elements in endogenous antinociceptive pathways. In the present study we used immunocytochemistry and retrograde tract tracing to investigate the contribution of GABAergic projection neurons to the antinociceptive network linking the midbrain periaqueductal gray matter (PAG), the nucleus raphe magnus (NRM), and the spinal cord dorsal horn. The tracer, WGAapoHRP-Au was injected into either the NRM or the spinal cord and the distribution of labeled neurons in sections of the PAG and medulla, respectively, was studied. The same sections were immunostained to demonstrate GABA-immunoreactive neurons. Although GABA-immunoreactive neurons were abundant in the PAG, only 1.5% were retrogradely labeled from the NRM. Similarly, very few GABA-immunoreactive neurons within the cytoarchitectural boundaries of the NRM were retrogradely labeled from the spinal cord. A much higher proportion of GABA-immunoreactive neurons in the region lateral to the NRM, however, were retrogradely labeled from the spinal cord. Eighteen percent of GABA-immunoreactive neurons were retrogradely labeled in the nucleus reticularis paragigantocellularis; conversely, 15% of the retrogradely labeled neurons in this region were GABA-immunoreactive. These results indicate that GABAergic projections constitute a very minor component of the PAG-NRM-spinal cord pathway; however, there is a significant contribution of GABAergic neurons to the spinal projections that originate lateral to the NRM. The majority of GABAergic neurons in the PAG and NRM are presumed to be inhibitory interneurons that directly or indirectly regulate activity in efferent pathways from these regions.     

Axon collaterals of pontine taste area neurons project to the posterior ventromedial thalamic nucleus and to the gustatory neocortex

Retrograde axonal transport of fluorescent dyes was used to demonstrate collateral projections from neurons of the pontine taste area (PTA) to gustatory-responsive areas of the posterior ventromedial thalamic nucleus (VPM), and to the gustatory neocortex (GN) of the rat. Dual-labeled PTA neurons were reliably observed following application of two different fluorescent dyes to the GN and to VPM thalamus. Dye injections into the GN and into thalamic regions surrounding the VPM nucleus, the bed nucleus of stria terminalis or the infralimbic neocortex, did not result in dual-labeled cells within the PTA. This finding suggests that gustatory information may be relayed simultaneously and specifically to VPM thalamus and to the GN via collateral axons of PTA neurons.     

Postnatal development of the afferent projections from the dorsomedial thalamic nucleus to the frontal cortex in mice

The postnatal development of mediodorsal thalamic projections to the dorsomedial frontal cortex of mice was assessed by means of the retrograde peroxidase-colloidal gold complex tract tracing system. The tracer was injected into the dorsomedial frontal cortex from the day of birth (P0) to 60 days of postnatal age (P60). Since birth, a dense retrograde labeling has been found in the mediodorsal nucleus, which increased progressively from P4 to P8 and began to decrease at P10 until P13 (67.37% vs. the maximal average, P4). After P16, the mean average remains stable up to P60.     

Mediodorsal thalamic afferents to layer III of the rat prefrontal cortex: synaptic relationships to subclasses of interneurons

The mediodorsal nucleus of the thalamus (MD) represents the main subcortical structure that projects to the prefrontal cortex (PFC) and it regulates key aspects of the cognitive functions of this region. Within the PFC, GABA local circuit neurons shape the activity patterns and hence the "memory fields" of pyramidal cells. Although the connections between the MD and PFC are well established, the ultrastructural relationships between projecting fibers from the MD and different subclasses of GABA cells in the PFC are not known. In order to address this issue in the rat, we examined MD axons labeled by tract-tracing in combination with immunogold-silver to identify different calcium-binding proteins localized within separate populations of interneurons. Electron micrographic examination of PFC sections from these animals revealed that MD terminals made primarily asymmetric synapses onto dendritic spines and less commonly onto dendritic shafts. Most of the dendrites receiving MD synaptic input were immunoreactive for parvalbumin (ParV), whereas MD synapses onto dendrites labeled for calretinin or calbindin were less frequent. We also observed that some MD terminals were themselves immunoreactive for calcium-binding proteins, again more commonly for ParV. These results suggest that the MD exerts a dual influence on PFC pyramidal cells: direct inputs onto spines and an indirect influence mediated via synapses onto each subclass of interneurons. The apparent preferential input to ParV cells endows MD afferents with a strong indirect inhibitory influence on pyramidal neuron activity by virtue of ParV cell synapses onto soma, proximal dendrites, and axon initial segments.     

An ascending serotonergic pain modulation pathway from the dorsal raphe nucleus to the parafascicularis nucleus of the thalamus

(1) Three types of spontaneously active neurons were found in the parafascicularis (PF) nucleus of the thalamus of the rat: slow firing units (0.5–10 spikes/s), bursting units (2–5 spikes/burst in 10–20 ms, one burst every 1–2 s) and fast firing units (15–40 spikes/s). A similar population of neurons was found in the PF of rats treated with 5,7-dihydroxytryptamine (5,7-DHT), a serotonin neurotoxin.

(2) Noxious tail pinch (TP) caused 68% of the PF neurons to increase their firing rates to 242% of their initial baseline activity, while non-noxious touch stimulation failed to induce a response. In the 5,7-DHT-treated rats, TP caused 85% of the neurons in the PF to increase their firing rates to 581% of their initial baseline activity and 22% of the neurons increased their firing in response to touching the tail. Both the number of cells responding (P < 0.05) and the percentage increase (P < 0.001) were statistically greater in serotonin-depleted rats than in controls. This indicates that serotonin (5-HT) has a tonic inhibitory influence on responses to both noxious and non-noxious sensory stimuli.

(3) In control rats, electrical stimulation of the dorsal raphe nucleus (DR) decreased the firing rates of PF neurons. In contrast, the same DR stimulation induced an increase in PF firing rates during stimulation in serotonin-depleted rats and this increase in firing rates remained several seconds after cessation of stimulation. This indicates that the DR may use at least two different neurotransmitters in its projections to forebrain structures.

(4) In control rats, the TP stimulation induced an increase in firing rates of PF neurons while DR stimulation attenuated the excitation induced by TP stimulation. In serotonin-depleted rats, DR stimulation and TP both caused an increase in firing rates. This effect was not additive indicating that there may be a serotonergic projection from the DR to the PF which modifies responses to somatosensory stimuli.

(5) The inhibitory effects elicited by electrical stimulation were limited to the immediate area of the DR. Stimulation of the adjacent reticular formation 1 mm lateral to the DR produced the opposite effect, an increase in firing rate often accompanied by driven spike activity in the PF.

Projection of brain stem neurons to the perigeniculate nucleus and the lateral geniculate nucleus in the cat

Small horseradish peroxidase injections in the perigeniculate nucleus (PGN) or the lateral geniculate nucleus (LGN) gave retrograde labeling of many cells in the pontomesencephalic reticular formation (RF), the nuclei raphe dorsalis and centralis linearis, locus coeruleus, nucleus of the optic tract and nucleus parabigeminalis. Antidromic stimulation was used to identify neurons in the RF projecting to the PGN-LGN complex. Threshold mapping through the PGN and the LGN shows separate projection from the reticular formation to the PGN and the LGN.     

The effect of morphine on responses of mediodorsal thalamic nuclei and nucleus submedius neurons to colorectal distension in the rat

In halothane-anesthetized rats, we characterized the responses of single neurons in the nuclei of medial thalamus (MT), specifically the mediodorsal thalamic nucleus (MD) and the nucleus submedius (Sm), to a noxious visceral stimulus (colorectal balloon distension, CRD), and studied the effects of intravenous morphine (Mor) on these responses using standard extracellular microelectrode recording techniques. 62 MD and 46 Sm neurons were isolated on the basis of spontaneous activity. 47 of the MD neurons (76%) responded to CRD, of which 70% had excitatory and 30% had inhibitory responses. 38 of the Sm neurons (83%) responded to CRD, of which 89% had excitatory and 11% had inhibitory responses. Responses of MD and Sm neurons excited by CRD were related significantly to distension pressure (20–100 mmHg), with maximum excitation occurring at 60 and 100 mmHg, respectively. MD neurons inhibited by CRD also had graded responses to graded CRD, with maximum inhibition occurring at 80 mmHg. The responses to noxious (pinch, heat) and nonnoxious (tap, brush) cutaneous stimuli were studied in 59 of the MD and 44 of the Sm neurons isolated. 22 of the MD neurons (37%) studied had cutaneous receptive fields, of which 59% were large and bilateral, 41% were small and usually contralateral receptive fields. 55% of these neurons were nociceptive-specific, 45% responded to both noxious and nonnoxious cutaneous stimulation. 29 of the Sm neurons (66%) studied had cutaneous receptive fields, of which 72% were large and usually bilateral, 14% were small and bilateral, 14% were small and contralateral receptive fields. 90% of these neurons were nociceptive-specific, 10% responded to both noxious and nonnoxious stimulation. No MD or Sm neurons responded exclusively to nonnoxious cutaneous stimulation. Mor (0.125, 0.25, 0.5 and 1 mg/kg IV) attenuated MD and Sm neuronal excitatory responses to CRD in a dose-dependent fashion, abolishing evoked activity with a dose of 0.5 mg/kg (p<0.05) and 1 mg/kg (p<0.05), respectively. Naloxone (0.4 mg/kg IV) reversed the effects of Mor. Mor and naloxone had no effects on spontaneous activity. These data support the involvement of MD and Sm neurons in visceral nociception, and are consistent with a role of Sm in affective-motivational, and MD in both sensory-discriminative and affective-motivational aspects of nociception.     

The morphology of trigeminal nociceptive neurons in the caudal bulbar lateral reticular formation of the cat

Two types of trigeminal nociceptive neurons, i.e. subnucleus reticularis ventralis (SRV) and wide dynamic range (WDR) neurons were identified in the caudal bulbar lateral reticular formation (LRF) and intracellularly stained with horseradish peroxidase. SRV neurons were large neurons characteristic of the subnucleus reticularis ventralis. Their dendrites were confined to the LRF. WDR neurons were situated in the subnucleus reticularis dorsalis. Their dendrites penetrated into the magnocellular layer, but did not reach the substantia gelatinosa.     

All-or-nothing and graded slow waves in the lateral geniculate nucleus to stimulation of the midbrain reticular formation

The occurrence of slow negative potential changes in the lateral geniculate nucleus of the cat in response to stimulation of the midbrain reticular formation was studied. A method for distinguishing all-or-nothing from graded potentials in the presence of noise is described. Graded potential changes were obtained in the conscious and anesthetized states. All-or-nothing potential changes were obtained in the anesthetized animal only at times of spontaneous all-or-nothing waves (similar to ponto-geniculo-occipital waves). All-or-nothing waves could also be elicited in both lateral geniculate nuclei when a transection was made between the lateral geniculate nucleus and the stimulating site in the midbrain reticular formation on one side; this suggests the existence of a descending pathway to the ponto-geniculo-occipital generator in the pontine tegmentum. All-or-nothing waves could also be evoked in both lateral geniculate nuclei after a transection caudal to the stimulating site on one side. This suggests either that such stimulation excites a pathway which crosses to the other side of the brain and then descends to the ponto-geniculo-occipital generator or that the all-or-nothing mechanism is more rostral and stimulation of the midbrain reticular formation either excites this region directly or excites pathways projecting to this region.