INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Trigeminal sensory information that arises in orofacial organs facilitates exploratory, ingestive, and defensive behaviors that are essential to overall fitness and survival (Dubner et al. 1978; Feindel 1956; Geppetti et al. 1988; Guyton 1971; Lund and Dellow 1971). In the rat, tactile signals that arise in orofacial organs such as the vibrissae, nose, lips, and tongue contribute to the perception of subtle external cues that assist in the execution of feeding-related activities such as chewing, swallowing, licking, and suckling, while painful sensation that originates in these organs can alert the animal to potential dangers and allow for preservation of structures that subserve additional sensory modalities including vision, olfaction, and audition.

Because organs such as the mouth, nose, and eyes serve functions that must be performed continuously for survival, pain that originates in these structures is repeatedly aggravated and is therefore one of the most commonly cited sources of discomfort in human patients. Injuries such as facial lacerations (Bakay and Glasauer 1980) and diseases such as trigeminal neuralgia (Kugelberg and Lindblom 1959), headache (Olesen et al. 1993), sinusitis (Saunte and Soyka 1994), toothache (Sharav 1994), and temporomandibular joint pain syndrome (Sessle and Hu 1991) are believed to activate trigeminal nociceptors and consequently second-order trigeminal brain stem nuclear complex (TBNC) neurons in nucleus caudalis and C_1-2. Until now, however, most studies of TBNC neurons that process nociceptive and nonnociceptive information that arises in the cornea, nasal mucosa, tongue, tooth pulp, vibrissae, temporomandibular joint, facial muscles, and skin focused on local or thalamic projecting neurons that are likely to play a role in the sensory-discriminative aspect of trigeminal sensation and pain (reviewed in Malick and Burstein 1998). The purpose of this study was therefore to characterize TBNC neurons that project to brain areas that regulate behavioral, rather than sensory-discriminative, responses to pain (i.e., to determine what kind of information they convey and how they reach their targets). Because sensations that originate in orofacial organs often change hypothalamic-mediated behaviors such as food intake and sleep, we chose to study how orofacial sensory signals reach the hypothalamus.

Complex hypothalamic-mediated functions are commonly influenced by sensory and physiological signals arising from the body and cognitive signals arising from cortical and subcortical brain regions. The integration of sensory, physiological, and cognitive signals by hypothalamic neurons that regulate both hormonal secretion and the activity of brain stem and spinal cord neurons that mediate autonomic responses could provide a partial answer to the question of how sensory signals produce endocrine, autonomic, and affective responses. To be in a position to integrate somatosensory and visceral information with endocrine and autonomic responses, hypothalamic neurons must receive somatosensory and visceral inputs. The afferent inputs that the hypothalamus receives from brain stem nuclei, such as the parabrachial nuclei (Cechetto et al. 1985; Saper and Loewy 1980; Slugg and Light 1994), nucleus of the solitary tract (Menetrey and Basbaum 1987; Ricardo and Koh 1978), periaqueductal gray (Beitz 1982; Eberhart et al. 1985; Lima and Coimbra 1989; Liu 1983), and caudal ventrolateral medulla (Lima et al. 1991; Sawchenko and Swanson 1981), and the identification of neurons in these nuclei that respond to noxious and innocuous somatosensory and visceral stimulation (Bernard and Besson 1990; Kannan et al. 1986; Pan et al. 1999; Person 1989; Zhang et al. 1992) contributed to the notion that somatosensory signals reach the hypothalamus through several polysynaptic pathways.

Recent anatomical studies showed that somatosensory and visceral information can also reach the hypothalamus through monosynaptic pathways that originate in spinal cord and medullary dorsal horn neurons. The electrophysiological studies that followed described the course of sacral, lumbar, thoracic, and lower cervical spinohypothalamic tract (SHT) axons that convey to the hypothalamus sensory signals originating in the perineum and colorectal canal (Katter et al. 1996a), lower limbs (Burstein et al. 1991), abdominal region and bile duct (Zhang et al. 1999b), upper limbs (Dado et al. 1994a; Kostarczyk et al. 1997; Zhang et al. 1995) and cervical dermatomes (Dado et al. 1994b). Currently, however, no information is available on axons of TBNC neurons that carry to the hypothalamus sensory information that originates in orofacial organs innervated by the trigeminal nerve (Burstein et al. 1990; Carstens et al. 1990; Iwata et al. 1992; Li et al. 1997; Malick and Burstein 1998; Newman et al. 1996; Ring and Ganchrow 1983). To fill this gap, we sought to identify and physiologically characterize hypothalamic projection neurons in the caudal medulla and upper cervical spinal cord using antidromic microstimulation and single-unit recording techniques. Our hypothesis was that the trigeminohypothalamic tract (THT) is capable of transferring to the hypothalamus nociceptive and nonnociceptive information that arises in all orofacial organs innervated by the trigeminal nerve.

During the search for THT neurons, we occasionally encountered neurons that projected to the hypothalamus and responded to noxious stimulation of both orofacial and extracephalic receptive fields. Later anatomical analysis revealed that these neurons were located along the poorly defined ventromedial border of lamina V, in an area previously identified as the lateral reticular formation (Nord and Kyler 1968). The input that neurons in the lateral reticular formation (LRF) receive from nociceptive neurons in the spinal cord (Westlund and Craig 1996) and their well-documented projections to the hypothalamus (Cunningham and Sawchenko 1991; Loewy et al. 1981; McKellar and Loewy 1981; Sawchenko and Swanson 1981) make them additional candidates to transmit nociceptive information to the hypothalamus. Because our aim is to understand as completely as possible how sensory trigeminal information reaches the hypothalamus, these reticulohypothalamic tract (RHT) neurons were also studied.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Surgical preparations, neuronal recording, and identification of hypothalamic projecting neurons

Male Sprague-Dawley rats weighing 400-600 g were anesthetized with urethan (1.2 g/kg). A metal tube was inserted into the trachea for artificial ventilation, and the rat was mounted in a stereotaxic apparatus. Core temperature was maintained at 37°C by a feedback-controlled heating pad, and end-tidal CO₂ was monitored and kept at 4.0-4.5%. A laminectomy was carried out to expose the first cervical segment of the spinal cord (C₁), and portions of the occipital bone were removed to allow complete access to nucleus caudalis (Vc) in the caudal medulla. The dura was retracted, the pia removed, and a pool of warm mineral oil formed over the exposed area. Large portions of the frontal and parietal bones were removed on both sides to allow introduction of stimulating electrodes into the hypothalamus, basal ganglia and midbrain. Rats were then paralyzed with gallamine triethiodide (1 g/kg) and artificially ventilated.

Using stainless steel (8-12 M) or tungsten (4-6 M) microelectrodes (FHC), single units were recorded within the dorsal horn of C₁, the medullary dorsal horn of Vc, and the lateral reticular formation (LRF). To search for neurons that project to the hypothalamus, one or two monopolar stimulating electrodes were lowered into the contralateral hypothalamus and cathodal current pulses were delivered (500 µA, 200 µs, 10 Hz). When one stimulating electrode was used, it was placed in the anterior-lateral hypothalamus. When two stimulating electrodes were used, the second was placed in the dorsal-medial area of the posterior hypothalamus. After isolating the spikes of an antidromically activated neuron, the stimulating electrode from which the unit was antidromically activated was moved systematically through the hypothalamus (as described in Burstein et al. 1991; Dado et al. 1994a) until a point was found from which a current of 50 µA was capable of inducing consistent antidromic spikes in the neuron. Criteria for antidromic activation included constant latency (total variation 0.2 ms), ability to follow trains of high-frequency stimuli (>333 impulses/s), and collision of antidromically induced spikes with those induced orthodromically (Lipski 1981). All neurons described in this study were antidromically activated from at least one point in the hypothalamus by a current of 50 µA. Locations from which neurons were activated antidromically by currents of 50 µA were defined as low-threshold points. Action potentials were amplified, sent to a window discriminator, collected by computer, analyzed quantitatively by Neuro-spike software (Pearson Technical Software), and presented as peristimulus time histograms (500-ms binwidth).

Receptive-field mapping

Following the identification of neurons that project to the hypothalamus their cutaneous and intraoral receptive fields were mapped by applying brief innocuous (vibrissae and hair deflection, air puff, and brush) and noxious (pressure, pinprick, and pinch) mechanical stimuli to the nose, vibrissal pad, upper and lower lips, tongue, skin areas above the eye (ophthalmic) and below the eye (maxillary), on the ventral surface of the face (mandibular), and on the entire body. An area was considered outside the neuron's cutaneous receptive field if no stimulus was capable of producing a response in 50% of the trials. Neurons exhibiting restricted orofacial receptive fields were classified as "trigeminal" neurons (i.e., trigeminohypothalamic tract units) and neurons exhibiting both orofacial and extracephalic receptive fields (e.g., abdomen, limbs, tail) were classified as "non-trigeminal" neurons (i.e., reticulohypothalamic tract units). Noncutaneous receptive fields such as the cornea and intracranial dura were mapped by sliding a brush over these organs and by indenting them with calibrated von Frey hairs.

Physiological characterization

Neurons were then physiologically characterized according to their responsiveness to a series of brief (10 s) innocuous and noxious mechanical stimuli applied to the most sensitive portion of their cutaneous receptive field. Innocuous stimuli consisted of slowly passing a soft bristled brush across the cutaneous receptive field and pressure applied with a loose arterial clip. Noxious stimuli consisted of pinch with a strong arterial clip and crush with nonserrated forceps. To avoid inducing prolonged changes in spontaneous neuronal discharge or response properties, more intense or prolonged stimuli were not used. Neurons classified as low threshold (LT) responded maximally or exclusively to innocuous mechanical stimulation. Neurons classified as wide dynamic range (WDR) responded to brush and also to noxious mechanical stimulation in a graded fashion. Neurons designated as high threshold (HT) did not respond to brush but responded to more intense mechanical stimuli (pressure, pinch, and crush) of their cutaneous receptive fields (Dado et al. 1994b; Palecek et al. 1992). To further characterize the neurons, their responses to thermal stimulation were determined following the application of thermally conductive paste to the skin. In most cases, thermal responses were determined by rapidly (10°C/s) heating (to 39, 41, 46, 50, and 55°C) or cooling (in most cases to 20, 10, and 0°C, in several cases to 30, 25, 20, 15, 10, 5, 0, and 10°C) the skin with a 9 × 9 mm contact thermal stimulator (Yale University) for 30 s. The data obtained from the rapid-ramp thermal stimuli were used for the quantitative analyses of the response magnitude. Thermal responses were also determined by slowly heating (35-55°C at a rate of 2.4°C/s) or cooling (35-0°C at a rate of 2.0°C/s) the skin (Burstein et al. 1998) for 10-30 s. The data obtained from the slow-ramp thermal stimuli were used to determine response thresholds (Burstein et al. 1998). The skin surface was maintained at 35°C during the periods between stimuli. This period is defined as the interstimulus interval, which was 180 s. Because we examined only a small number of lamina I neurons and because not all were examined for their responses to heat, cold, and mechanical stimuli, we opted not to use the classification of thermoreceptive-specific (cold), and polymodal nociceptive (HPC) neurons (Craig and Dostrovsky 1991; Craig and Serrano 1994; Dostrovsky and Craig 1996; Han et al. 1998).

Physiologically characterized units were further classified according to whether they process sensory information that arises in the vibrissae, tongue, cornea, intracranial dura, or the entire body. To identify neurons that respond to vibrissal stimulation, we deflected individual vibrissae within the neuron's receptive field. The vibrissa that induced the largest response was then manually deflected in four orthogonal directions (in 90° increments relative to the horizontal alignment of the whisker row) at 10-s intervals. Each deflection lasted 5 s. To identify neurons that respond to intraoral stimulation, the tongue was gently pulled out and exposed to the same mechanical and thermal stimuli that were applied to the skin. Cornea-sensitive neurons were identified by gently sliding a brush over the corneal surface; activating LT A rapidly adapting mechanosensitive receptors (Giraldez et al. 1979; MacIver and Tanelian 1993a,b). To activate other corneal receptors and nociceptors (i.e., C-fiber cold receptors, A high-threshold mechano-heat nociceptors, and C-fiber chemosensitive receptors) a small portion of gelfoam dipped in 0.1 M of nicotine (temp = 35°C, pH =7.4) was laid on top of the corneal surface for 30 s (MacIver and Tanelian 1993a,b; Tanelian and Bisla 1992), then rinsed with physiological saline. Dura-sensitive neurons were identified by applying single shocks (0.8 ms, 0.5-4.0 mA, 1 Hz) through a bipolar stimulating electrode placed on the dura overlying the ipsilateral transverse sinus (Burstein et al. 1998). These stimulus parameters were capable of activating both A and C fibers that innervate the dura (Strassman et al. 1996). The sinus area from which the lowest current was capable of activating the neuron was then explored by mechanical stimuli such as dural indentation with calibrated von Frey hairs (Stoelting, flat and round tip shape, diameter range = 0.15-0.38 mm) and gentle rubbing with a brush.

When neurons were found to have both orofacial and extracephalic receptive fields, identical series of mechanical and thermal stimuli (described above) were applied to each area (on the face, limbs, etc.) to determine whether the information they process from different skin regions is qualitatively and/or quantitatively similar. Each series of stimuli was separated by 3 min. These neurons were classified separately for their responses to orofacial versus extracephalic stimulation.

Axonal mapping in the midbrain, hypothalamus, and basal ganglia

Once physiological characterization of THT and RHT neurons was completed, we mapped the course of their axons in the midbrain, hypothalamus, and basal ganglia by using the antidromic microstimulation mapping technique (for detailed description, see Burstein et al. 1991; Dado et al. 1994a,c; Fields et al. 1995; Zhang et al. 1995). To determine the course of the axon, the hypothalamic-stimulating electrode had to be repositioned. Before moving this electrode, a second stimulating electrode was inserted into the contralateral midbrain and placed ~1 mm lateral to the periaqueductal gray at the level of the superior colliculus. The position of the midbrain stimulating electrode was adjusted until the same neuron was activated using a current of 50 µA. To avoid damage to the parent axon, no attempt was made to lower the current by placing the electrode closer to the axon. The midbrain stimulating electrode was used to ensure that the neuron was not lost during the mapping of its axon when we were unable to activate it from other hypothalamic areas, to recognize the changes that occurred in the amplitude of the spike during the search, and to confirm that the spike elicited from the two stimulating electrodes propagated in the same axon. This last task was achieved by demonstrating that stimulation at the midbrain and hypothalamus (or any other more rostral point) induced spikes that were similar in shape, duration and amplitude, and antidromic spikes that collided with each other when the interspike interval was shorter than the time required for the spike to travel between the two stimulating electrodes.

To determine whether the axon of the examined neuron terminated in the contralateral hypothalamus, the hypothalamic stimulating electrode was moved as far rostral as the craniotomy allowed and reinserted into the brain in a systematic way that enabled us to determine thresholds for antidromicity at points separated by 200 µm dorsoventrally and 300-500 µm mediolaterally. If the neuron was not activated from any point within the most anterior level, the stimulating electrode was moved 500-1,000 µm posteriorly, and a similar search was made. At each anteroposterior level, the presence of the axon was indicated by a shift in latency of the antidromic spike to a value longer than that recorded from the contralateral midbrain. If a low-threshold point in the contralateral hypothalamus could be surrounded anteriorly, medially, laterally, dorsally, and ventrally by points from which higher currents were required to activate the neuron and if the spikes elicited from that point collided with spikes elicited from the midbrain, the axon was considered to terminate in the contralateral hypothalamus.

To determine whether the axon of the examined neuron crossed the midline, the stimulating electrode was moved to the ipsilateral side (1 mm from the midline) and repeatedly inserted along the midline (intervals of 500 µm) from the anterior diencephalon to the midbrain. In cases in which the neuron was antidromically activated from the ipsilateral side, the systematic mapping of the axon continued on both sides of the brain. In cases in which the neuron was not activated from the ipsilateral side, attempts were made to determine whether the parent axon issued collateral branches in the hypothalamus. Detailed description of collateral branches mapping with antidromic stimulation technique and their limitations are given in our recent paper (Fields et al. 1995). Briefly; the presence of a branch was indicated by a shift in latency of the antidromic spike to a value longer than that of the parent axon at the same anteroposterior level. The criteria used to confirm that the longer-latency spike was elicited from a branch of the parent axon were that the position of the low-threshold point for the putative branch be in one of the hypothalamic nuclei and at a clear distance from the parent axon in the supraoptic decussation (where most parent axons are found), and that the minimum current sufficient to activate the branch be too low to activate the parent axon by current spread.

Anatomical analysis

At the conclusion of each experiment, the recording site and the low-threshold points for antidromic activation were marked with electrolytic lesions (anodal DC of 25 µA for 20 s). Only one neuron was studied in each animal. In cases in which multiple low-threshold points were found, lesions were made at those points from which a clear shift in latency could be demonstrated. Conduction distances were measured between the recording site and midbrain by placing the midbrain stimulating electrode over the recording site and then calculating the differences from the anteroposterior, dorsoventral, and mediolateral stereotaxic coordinates as the shortest distance between the two points. Similar measurements were made between the midbrain low-threshold point and each of the hypothalamic low-threshold points. Rats were perfused with 1% potassium ferrocyanide in 10% formalin. The brain, brain stem, and upper cervical spinal cord were removed and postfixed for 5 days, during which time they were also reacted for Prussian blue stain of ferric ions. The tissue was cut transversely on a freezing microtome (50 µm) and examined under dark field illumination, which allowed clear identification of laminar borders in C₁ and Vc. The tissue was then stained for Nissl substance, and the sections were reexamined under bright field illumination that shows the cytoarchitectonic organization of the different brain stem nuclei and dorsal horn laminae. Detailed descriptions of the nuclear (C₁, Vc, Vi) and laminar (I-V) borders is given in our anatomical paper on the THT (Malick and Burstein 1998). Briefly, the rostral border of Vc was defined as the point at which substantia gelatinosa is "displaced" and becomes contiguous with the spinal tract of V. We considered Vc to extend 2.0 mm caudal to this point and C₁ to extend 1.6 mm beyond the caudal border of Vc (Falls 1984a,b; Gobel et al. 1981; Jacquin et al. 1986a). The transition zone between Vc and C₁ usually coincides with the caudal end of the pyramidal decussation.

Laminar borders in Vc and C₁ were determined using previously described criteria (Gobel et al. 1977; Jacquin et al. 1990; Strassman and Vos 1993). While the identification of the boundaries between laminae I and II, II and III, and IV and V is relatively straightforward, the precise identification of the inner border of lamina V is difficult to determine. Since there is no obvious difference between lamina V and the lateral reticular formation, the medial border of lamina V was defined according to the functional properties of the neurons in this area. As stated in the preceding text, all neurons were classified as either THT or LRF-RHT according to the location of their cutaneous receptive field. The rationale for this classification is based on the knowledge that lamina V but not LRF neurons receive direct input from trigeminal primary afferent fibers (Arvidsson and Rice 1991; Clarke and Bowsher 1962; Jacquin et al. 1982; Marfurt and Rajchert 1991) and that the LRF receives input from nociceptive dorsal horn neurons located in laminae I and V throughout the length of the spinal cord (Lima et al. 1991; Marfurt and Rajchert 1991).

Data and statistical analysis

The database consisted of measurements of the number of spikes/second (response) recorded in C1-THT, Vc-THT, and LRF-RHT neurons. Data organization and analysis were done on the Prophet System (release 4.1), a national computing resource for life science research sponsored by the National Institutes of Health, Division of Research Resources. Response magnitude to each stimulus was calculated by subtracting the mean ongoing activity occurring before the first stimulus (10 s for mechanical, 30 s for thermal and chemical) from the mean firing frequency that occurred throughout the duration of each stimulus. The means of the measurements were plotted against the mechanical (brush < pressure < pinch < crush, expressed as scale data 1, 2, 3, 4), heat, and cold stimuli. The resulting distributions were tested for normality using the D'Agostino test (D'Agostino 1986), and their central measures were computed. Comparison of responses among the respective levels of mechanical, heat, and cold stimuli were performed using appropriate multiple sample comparison procedures [Newman-Keuls if the data were normally distributed (parametric), Kruskal-Wallis if nonparametric]. The means of the measurements were also subjected to trend analysis using the Spearman rank correlation (r_s) for mechanical stimuli and regression analysis for heat and cold. The responses of LRF-RHT neurons to mechanical stimulation of their trigeminal and non-trigeminal receptive fields were subjected to unpaired two-sample comparison tests (unpaired t-test for parametric data, Mann-Whitney rank-sum test for nonparametric data). Neuronal responses to thermal stimuli were analyzed in two ways: during the dynamic phase and during the static phase. The dynamic phase (heating or cooling ramp) was defined as the time during which skin temperature is increasing or decreasing and the static phase as the time during which the temperature was maintained at constant temperature.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Physiological characterization

IDENTIFICATION OF HYPOTHALAMIC-PROJECTING NEURONS. Eighty-one neurons (success rate of ~1:3) were antidromically activated from the contralateral hypothalamus with currents of 50 µA (mean ± SE was 19 ± 11.8 µA). An example of the localization of a low-threshold point for antidromic activation of a THT neuron from the contralateral hypothalamus is shown in Fig. 1. In the first track from which the neuron was antidromically activated (the most medial track in the hypothalamus), the lowest threshold was 260 µA. After antidromic thresholds were determined every 200 µm throughout the track, the electrode was removed and reinserted 300 µm lateral to the first track. The lowest antidromic threshold in the second track was 70 µA. The electrode was again removed and reinserted 300 µm lateral to the second track. The lowest antidromic threshold in the third track was 8 µA. In the next two tracks (made 300 and 600 µm lateral to the 3rd track), the lowest antidromic thresholds were 32 and 224 µA, respectively. Since the lowest threshold point in the third track was surrounded medially, laterally, dorsally, and ventrally by points from which higher current was required to activate the neuron, it was considered as the lowest threshold point at this anterior-posterior level. This point was located in the supraoptic decussation (SOD) within the lateral hypothalamus (Fig. 1A). Antidromic action potentials elicited from this and all other low-threshold points in the hypothalamus for this and all other neurons included in the study fulfilled the standard criteria for antidromic activation: they occurred at constant latency, 6.2 ms in this case (Fig. 1B1), collided with orthodromic action potentials elicited by stimulating the cutaneous receptive field (Fig. 1B2), and followed a train of high-frequency stimulation (Fig. 1B3). The recording site of this HT-THT neuron was found in laminae I-II of Vc (C), and the receptive field was mostly within the territory of the maxillary branch of the trigeminal nerve (D).

View larger version (15K): [in this window] [in a new window]   Fig. 1. Antidromic activation of a trigeminohypothalamic tract neuron from the hypothalamus. A: locations of electrode penetrations. The amount of current required to activate the neuron antidromically from each track is indicated. Note that the lowest threshold point (circled) was located in the supraoptic decussation (SOD). B: digitized oscillographic traces illustrating the constant latency of antidromic action potentials (B1, 3 overlapping traces), a collision with an orthodromic action potential elicited by stimulating the peripheral receptive field (B2), and the ability to follow a train of high-frequency stimulation (B3, 500 Hz). C: reconstruction of the recording site in laminae I and II of nucleus caudalis. D: receptive field. Hyp, hypothalamus; IC, internal capsule; GP, globus pallidus, LRN, lateral reticulated nucleus, NTS, nucleus of the solitary tract; OT, optic tract; PVN, paraventricular hypothalamic nucleus; Px, pyramidal decussation; I-V, laminae of the medullary dorsal horn.

RECORDING SITES. Thirty-two neurons were recorded in Vc, 22 in C₁, 18 in LRF, and the locations of 9 neurons were not identified. Photomicrographs of lesions made in laminae I-II and IV-V, and in the LRF are shown in Fig. 2. Reconstructions of the locations of electrolytic lesions marking the recording sites of 72 neurons are illustrated in Fig. 3. Because many lesions were found at the border between laminae II and III and between laminae IV and V, it was difficult to assign each lesion to a particular lamina with certainty. Based on the center of the lesions that were made in Vc and C₁, however, it appears that 12 of the neurons were recorded in laminae I-II (20%), 7 in laminae III-IV (15%), and 35 in lamina V (65%). As explained in the preceding text, these 54 neurons were considered trigeminal because their receptive fields were restricted to skin areas innervated by the trigeminal nerve. The lesions of 18 additional recording sites were found in the medullary LRF. Although there is no easy way to differentiate between lamina V and the LRF, neurons assigned to the LRF were found deeper than the lamina V THT neurons and exhibited extracephalic, in addition to orofacial, receptive fields. They were therefore considered non-trigeminal neurons. The recording locations of physiologically characterized neurons in C1 and Vc were distributed as follows: laminae I-II contained 4 HT, 4 WDR, and 1 LT neurons; laminae III-IV contained 1 HT, 3 WDR, and 2 LT neurons; lamina V contained 11 HT, 12 WDR, and 8 LT neurons; and the LRF contained 7 HT, 7 WDR, and 2 LT neurons (LRF classification was based on responses to facial stimulation). Of the unclassified neurons, three were in laminae I-II, one in laminae III-IV, four in lamina V, and two in the LRF.

View larger version (124K): [in this window] [in a new window]   Fig. 2. Photomicrographs of electrolytic lesions (arrowheads) marking lowest threshold points for antidromic activation and recording sites. A and B: darkfield photomicrographs of unstained sections showing lesions in the optic chiasm and supraoptic decussation. C-F: bright field photomicrographs of Nissl-stained sections showing lesions in the ventrolateral region of caudate-putamen, dorsomedial hypothalamus, lateral hypothalamus and supraoptic decussation. G and H: darkfield photomicrographs of unstained sections showing lesions in the cuneiform nucleus and superior colliculus. I: bright field photomicrographs of a Nissl-stained section showing lesion in the brachium of the inferior colliculus. J-L: recording site lesions in laminae I and II of C1-V, laminae IV and V of Vc, and the lateral reticular formation. Scale bar in I =100 µm for A-F and I, 200 µm for G and H. Scale bar in L = 100 µm for A and C, 40 µm for B. AH, anterior hypothalamic area; Aq, cerebral aqueduct; BIC, brachium of the inferior colliculus; CG, central gray; Cnf, cuneiform nucleus; CPu, caudate putamen; ctf, central tegmental field; DMH, dorsomedial hypothalamic nucleus; F, fornix; internal capsule; LH, lateral hypothalamic area; mfb, medial forebrain bundle; MPO, medial preoptic area; OX, optic chiasm; Par, parietal cortex; SC, superior colliculus; SM, stria medullaris; SOD, supraoptic decussation; VL, ventrolateral thalamic nucleus; VMH, ventromedial hypothalamus; ZI, zona incerta.

View larger version (32K): [in this window] [in a new window]   Fig. 3. Locations of 72 lesions marking the recording sites in upper cervical spinal cord (C1, 22), nucleus caudalis (Vc, 32), and lateral reticular formation (LRF, 18). A: recording sites of trigeminal (i.e., neurons exhibiting cephalic receptive fields) and non-trigeminal (i.e., neurons exhibiting cephalic and extracephalic receptive fields) neurons. Note the anatomical segregation between neurons that receive orofacial input and those that receive orofacial and extracephalic inputs. B: recording sites of functionally identified [i.e., high threshold (HT), wide dynamic range (WDR), and low threshold (LT)] neurons. Note that lamina I contained HT and WDR neurons, laminae III and IV contained LT and WDR neurons, and lamina V contained all 3 classes of neurons. I-V, laminae of the medullary dorsal horn and C1.

RECEPTIVE FIELDS. Twenty of the 22 C₁-THT neurons had restricted ipsilateral orofacial receptive fields, and 2 receptive fields were restricted to the ipsilateral neck (Fig. 4A). As shown in the figure, all laminae I and II C₁-THT neurons exhibited small to medium receptive fields that extended over facial skin areas innervated by one or two branches of the trigeminal nerve, while many laminae III-V neurons exhibited medium to large receptive fields that extended over facial skin areas innervated by two to three branches of the trigeminal nerve. Similar receptive fields were mapped for the 32 Vc-THT neurons (26 of which are shown in Fig. 4B); laminae I-II neurons had primarily small receptive fields, whereas those located in deeper laminae exhibited large receptive fields as well. In general, most HT-THT neurons had small or medium receptive fields, and most WDR-THT neurons had medium or large receptive fields. This tendency, however, was influenced by their location in the different laminae; both HT and WDR neurons had smaller receptive fields if they were recorded in laminae I-II and larger receptive fields if they were recorded in laminae III-V.

View larger version (55K): [in this window] [in a new window]   Fig. 4. Orofacial receptive fields of 22 THT neurons in the 1st cervical segment of the spinal cord (A), 26 THT neurons in the trigeminal nucleus caudalis (B), and complex orofacial-extracephalic receptive fields of 16 representative RHT neurons in the lateral reticular formation (C). Black areas indicate most sensitive regions of the receptive fields (defined as the site from which highest response magnitudes were elicited by mechanical stimuli), and gray areas indicate less sensitive regions. Response categories and laminar locations of recording sites are indicated. Note that most LRF-RHT neurons responded to both innocuous and noxious stimulation of orofacial skin areas, but only to noxious stimulation of extracephalic regions.

HIGH THRESHOLD NEURONS. Twenty-four of the 64 (38%) physiologically characterized neurons responded exclusively to noxious mechanical stimuli and were therefore classified as HT. Examples of the responses of two HT-THT neurons are illustrated in Fig. 5. The neuron on the left (Fig. 5A) was recorded in the most dorsomedial portion of lamina V, exhibited a mandibular/maxillary receptive field, and was antidromically activated from the contralateral hypothalamus. It responded to noxious but not innocuous mechanical stimuli (Fig. 5C). The neuron on the right (Fig. 5B) was recorded in the most ventrolateral portion of lamina I, exhibited an ophthalmic receptive field, and was antidromically activated from the lateral hypothalamus. It also responded exclusively to the noxious mechanical stimuli (Fig. 5D). The mechanical response profiles of 23 HT neurons are illustrated in Fig. 5E. Their mean (± SE) firing rates to brush, pressure, pinch and crush were 0.2 ± 0.1, 5.0 ± 1.6, 29.0 ± 5.5, and 36.0 ± 4.8 spikes/s, respectively.

View larger version (29K): [in this window] [in a new window]   Fig. 5. Response properties of 2 THT units classified as HT neurons. A and B: locations of low-threshold points for antidromic activation in the hypothalamus, recording sites in dorsomedial lamina V (A) and ventrolateral lamina I (B), and the receptive fields. C and D: peristimulus time histograms and records of window discriminator output (below histograms) illustrating their responses to mechanical stimuli (applied to most sensitive receptive field areas). E: stimulus response curves illustrating the response profiles of all HT neurons to mechanical stimulation. Numbers in parentheses depict mean response (spikes/s) to each stimulus. Thick line denotes means of all HT neurons. Note correlation between the locations of recording sites and receptive fields (A and B). Br, brush; Cr, crush; Pi, pinch; Pr, pressure.

WIDE-DYNAMIC RANGE NEURONS. Twenty-seven of the 64 (42%) physiologically characterized neurons responded to innocuous and noxious stimuli in a graded fashion and were therefore classified as WDR. Examples of the responses of two WDR-THT neurons are illustrated in Fig. 6. The neuron on the left (Fig. 6A) was recorded in the dorsomedial portion of lamina V, exhibited a large mandibular/maxillary/ophthalmic receptive field, and was antidromically activated from the contralateral hypothalamus. It responded most vigorously to innocuous and noxious mechanical stimulation of its mandibular receptive field; stimulation of its maxillary and ophthalmic receptive field produced smaller responses (Fig. 6C). The neuron on the right (Fig. 6B) was recorded in the most ventrolateral portion of lamina V, exhibited a large ophthalmic/maxillary/mandibular receptive field, and was antidromically activated from the contralateral hypothalamus. It responded most vigorously to innocuous and noxious mechanical stimulation of its ophthalmic receptive field; stimulation of its maxillary and mandibular receptive field produced smaller responses (Fig. 6D). The mechanical response profiles of 27 WDR neurons are illustrated in Fig. 6E. Their mean (± SE) firing rate to brush, pressure, pinch, and crush were 8.3 ± 1.2, 26.6 ± 3.6, 40.2 ± 3.8, and 40.5 ± 2.7 spikes/s, respectively.

View larger version (32K): [in this window] [in a new window]   Fig. 6. Response properties of 2 THT units classified as WDR neurons. A and B: locations of low-threshold points for antidromic activation in the hypothalamus, recording sites in dorsomedial lamina V (A) and ventrolateral lamina V (B), and receptive fields. C and D: peristimulus time histograms and records of window discriminator output illustrating responses to mechanical stimulation of their ophthalmic, maxillary and mandibular receptive fields. E: stimulus response curves illustrating the response profiles of all WDR neurons to mechanical stimulation. Numbers in parentheses depict mean response (spikes/s) to each stimulus. Thick line denotes means of all WDR neurons. Note correlation between the locations of recording sites and the most sensitive area of the cutaneous receptive fields. Dorsomedially located THT neurons responded more vigorously to mandibular stimulation (A), and ventrolaterally located neurons responded more vigorously to ophthalmic stimulation (B).

LOW THRESHOLD (VIBRISSA-SENSITIVE) NEURONS. Thirteen of the 64 (20%) physiologically characterized neurons responded more vigorously to innocuous than to noxious stimuli and were therefore classified as LT. Most of these LT neurons responded to deflection of a single hair follicle or vibrissa. Examples of the responses of a vibrissa-sensitive LT-THT neuron are illustrated in Fig. 7. This neuron was recorded in the ventrolateral portion of laminae III-IV, exhibited a small receptive field, and was antidromically activated from the contralateral hypothalamus and ventromedial posterior thalamic nucleus (Fig. 7A). It responded maximally to the deflection of a single vibrissa in all four directions (Fig. 7B) and to brushing its receptive field (not shown). The mechanical response profiles of 13 LT neurons are illustrated in Fig. 7C. Their mean (± SE) firing rates to brush, pressure, pinch, and crush were 34.0 ± 3.0, 21.3 ± 5.5, 22.0 ± 4.7, and 24.5 ± 5.0 spikes/s, respectively.

View larger version (22K): [in this window] [in a new window]   Fig. 7. Response properties of a vibrissa-sensitive LT-THT unit. A: locations of low-threshold points for antidromic activation in the hypothalamus and thalamus and the recording site in lamina III. B: peristimulus time histograms and records of window discriminator output illustrating responses to a 4-way deflection of a single vibrissa. C: stimulus response curves illustrating the response profiles of all LT neurons to mechanical stimulation of their cutaneous receptive fields. Numbers in parentheses depict mean response (spikes/s) to each stimulus. Thick line denotes means of all LT neurons. Note that this THT neuron also projects to the ventroposterior medial thalamic nucleus.

RESPONSES TO THERMAL STIMULI. Heat. Innocuous and noxious heat stimuli were applied to the receptive fields of 29 neurons. Twenty-seven of these (93%) responded incrementally to graded increases in heat stimuli. Of the 27 heat-sensitive neurons, 5, 3, and 9 were recorded in laminae I-II, III-IV, and V of C₁-Vc, respectively, and 10 were recorded in the LRF. Figure 8 illustrates two different response types to heat stimuli: at left; a "static" response, defined as maximal discharge during the steady-state phase of the stimulus, and at right; a "dynamic" response, defined as maximal discharge during the heating ramp.

View larger version (27K): [in this window] [in a new window]   Fig. 8. Static and dynamic response patterns of 2 laminae I THT neurons to innocuous and noxious heat stimuli. A and B: locations of low-threshold points for antidromic activation, recording sites, and receptive fields. C and D: peristimulus time histograms illustrating responses to heat stimuli. Stimulus temperature and temperature tracing are shown above histograms. Histograms showing responses to 50 and 55°C in D are truncated to maintain identical y axis. Note that the neuron on the left responded maximally during the steady-state phase of the stimulus (i.e., the static phase) and the neuron on the right during the heating ramp (i.e., the dynamic phase). E: stimulus-response curves illustrating mean responses of individual HT, WDR, and LT neurons during the static phase of the stimulus. Mean responses during the dynamic phase of the heat stimulus are shown only for LT neurons because HT neurons did not respond during this phase and WDR neurons responded rarely. Thick lines show the overall mean responses of all HT, WDR, and LT neurons. Note that during the dynamic phase of the heat stimuli, the discharge rate of LT-THT neurons increased in a near-linear fashion from 39 to 46°C followed by a decrease; suggesting inputs from warm receptors (see text).

View larger version (34K): [in this window] [in a new window]   Fig. 9. Static, dynamic, and mixed responses of 1 WDR and 2 LT THT neurons to innocuous and noxious cold stimuli. A-C: peristimulus time histograms illustrating responses to mechanical stimuli. D-F: peristimulus time histograms illustrating their responses to cold stimuli. Stimulus temperature and temperature tracing are shown above histograms. Note that the neuron on the left responded maximally during the steady-state phase of the stimulus (i.e., the static phase), the neuron in the middle during the heating ramp (i.e., the dynamic phase), and the neuron on the right exhibited a maximal peak during heating ramp and a slow adaptation during the steady-state (i.e., a mixed response). G: stimulus-response curves illustrating mean responses of individual HT, WDR, and LT neurons during the static phase of the stimulus. Mean responses during the dynamic phase of the cold stimulus are shown only for LT neurons because HT neurons did not respond during this phase and WDR neurons responded rarely. Thick lines show the overall mean responses of all HT, WDR, and LT neurons. Note that during the cooling phase the discharge rate of LT-THT neurons decreased in a near-linear fashion from 20 to 10°C; suggesting inputs from cold receptors (see text).

ORAL-SENSITIVE NEURONS. Twenty-one neurons responded to mechanical stimulation of the oral mucosa, tongue, or lips. They were classified as HT in 9 cases, WDR in 9 cases, and LT in 3 cases. The majority of oral-sensitive THT neurons were recorded in the dorsomedial third of laminae V (11 units) and III-IV (2 units) of C₁ (6 units), and Vc (7 units). The other seven oral-sensitive neurons were recorded in the LRF, and the recording site of one neuron was not found. The cutaneous receptive fields of the THT neurons varied; they extended over small mandibular or maxillary areas in three (23%) cases, mandibular and maxillary areas in 4 (31%) cases, and the entire face in six (46%) cases. Regardless of the cutaneous receptive field size, maximal neuronal responses were most commonly induced by stimulating intraoral structures. An example of an oral-sensitive HT-THT neuron is illustrated in Fig. 10 (left). This hypothalamic projecting neuron was recorded in the most dorsomedial portion of lamina V and exhibited an oral receptive field that included the tongue, hard palate, and the upper lip (A). It responded more vigorously to noxious mechanical stimuli of the hard palate than the tongue (B). The mechanical response profiles of 19 oral-sensitive neurons are illustrated in C. Their mean (± SE) firing rates to brush, pressure, pinch, and crush were 4.7 ± 1.6, 12.2 ± 3.0, 30.6 ± 4.6, and 37.2 ± 4.2 spikes/s, respectively. The heat response profiles of eight oral-sensitive neurons are illustrated in D. Their mean (± SE) firing rates to 39, 41, 46, 50, and 55°C were 0.1 ± 0.1, 0.3 ± 0.3, 4.1 ± 1.5, 17.9 ± 6.3, and 27.6 ± 8.6 spikes/s, respectively. About 15% of the oral-sensitive THT neurons also responded to innocuous cold stimuli (data not shown).

View larger version (34K): [in this window] [in a new window]   Fig. 10. Response properties of oral-sensitive (left), cornea-sensitive (middle), and dura-sensitive (right) THT neurons. A's: locations of low-threshold points for antidromic activation, recording sites, and receptive fields. B's: peristimulus time histograms illustrating responses to mechanical stimulation of the tongue and hard palate, mechanical and chemical stimulation of the cornea and periorbital skin, and mechanical stimulation of the dura and periorbital skin. Numbers in parentheses depict mean response (spikes/s) to each stimulus. C's and D's: stimulus response curves illustrating the response profiles of all oral-, cornea-, and dura-sensitive neurons to mechanical and thermal stimulation of the lips, periorbital skin, and skin innervated by the ophthalmic branch of the trigeminal nerve, respectively. Thick lines denote means.

CORNEA-SENSITIVE NEURONS. Thirteen neurons responded to mechanical stimulation of the cornea. They were classified as HT in four cases, WDR in seven cases, and LT in two cases. All THT neurons were recorded in lamina V at the level of caudal Vc (5 units) and rostral C₁ (4 units). Of the remaining four neurons, three were recorded in the LRF, and the recording location of one unit was not identified. Although no attempt was made to map the corneal receptive fields of these neurons, they seemed to respond to stimulation of all four corneal quadrants. In all but one case, the cutaneous receptive field extended over the periorbital skin; they included the ophthalmic and maxillary skin in all cases, and the mandibular skin in less than half of the cases. An example of a cornea-sensitive THT neuron is illustrated in Fig. 10 (middle). This hypothalamic projecting neuron was recorded in the most lateral region of lamina V, and its receptive field included the cornea and the periorbital skin (A). It responded preferentially to noxious mechanical stimuli of the ophthalmic skin and most vigorously to mechanical and chemical (nicotine) stimulation of the cornea (B). The mechanical response profiles of 10 cornea-sensitive neurons are illustrated in C. Their mean firing rates to brush, pressure, pinch, and crush were 10.8 ± 5.8, 31.0 ± 10.0, 41.7 ± 7.2, and 39.6 ± 6.5 spikes/s, respectively. The heat response profiles of nine cornea-sensitive neurons are illustrated in D. Their mean firing rates to 39, 41, 46, 50, and 55°C were 0.1 ± 0.1, 0.8 ± 0.5, 3.4 ± 1.5, 7.8 ± 1.6, and 15.1 ± 3.5 spikes/s, respectively. In five cases, 1 M nicotine was applied to the cornea for 30 s. The neuronal responses in all cases were vigorous.

DURA-SENSITIVE NEURONS. Ten THT neurons responded to electrical and mechanical stimulation of the dura mater overlying the transverse sinus. The majority (8 units) were recorded in the most lateral region of lamina V at the level of Vc, and only two neurons were recorded in laminae I-III or in C₁. The dural receptive fields of these neurons were usually small and restricted to the transverse or superior sagittal sinuses. All dura-sensitive neurons also exhibited cutaneous receptive fields (which varied in size). Regardless of the cutaneous receptive field size, maximal neuronal responses were most commonly elicited from the periorbital skin region. Eight of the 10 dura-sensitive neurons were physiologically characterized based on their responses to cutaneous stimulation; all were classified as WDR. An example of a dura-sensitive WDR-THT neuron is illustrated in Fig. 10 (right). This hypothalamic projecting neuron (A1) was recorded in the most lateral region of lamina V (A2). It exhibited ophthalmic receptive fields on the skin (A3) and the dura (A4). It responded preferentially to noxious mechanical stimulation of the ophthalmic skin and to dural brushing (B). The mechanical response profiles of eight dura-sensitive neurons are illustrated in C. Their mean firing rate to brush, pressure, pinch, and crush were 9.0 ± 2.4, 38.1 ± 7.3, 50.6 ± 7.4, and 43.1 ± 5.7 spikes/s, respectively. The heat response profiles of four dura-sensitive neurons are illustrated in D. Their mean firing rates to 39, 41, 46, 50, and 55°C were 4.0 ± 3.0, 5.8 ± 2.5, 16.0 ± 4.1, 33.8 ± 7.3, and 33.0 ± 1.0 spikes/s, respectively.

LRF-RHT NEURONS. Eighteen hypothalamic projecting neurons were recorded in the LRF. Their recording sites were in general more medial and slightly more ventral than the recording sites of the THT neurons in the ventrolateral area of lamina V (Fig. 3). Their most distinct property was their cutaneous receptive fields. They included orofacial and extracephalic skin areas in all cases (Fig. 4). In 11 cases, receptive fields extended over the entire body. An example of a HT LRF-RHT neuron is illustrated in Fig. 11. This hypothalamic projecting neuron exhibited a whole body receptive field. Like most LRF-RHT neurons in this study, it responded more vigorously to mechanical stimulation of ipsilateral orofacial organs (i.e., tongue, cornea, skin) compared with mechanical stimulation of contralateral orofacial organs or any of the extracephalic regions (i.e., paws and tail). Figure 12 illustrates the responses of all 18 LRF neurons to mechanical stimulation of their facial (i.e., trigeminal, Fig. 12A) and extracephalic (i.e., non-trigeminal, Fig. 12B) receptive fields. The means (±95% confidence interval) of the responses to brush, pressure, pinch, and crush are shown in Fig. 12C. Responses induced by trigeminal receptive fields stimuli were significantly greater than the responses to the respective stimuli of the non-trigeminal receptive fields (P values given in the right column of the table). These findings indicate that LRF neurons responded to brush and pressure of their trigeminal but not non-trigeminal receptive fields and that pinch and crush induced larger responses when applied to their trigeminal than to non-trigeminal receptive field.

View larger version (24K): [in this window] [in a new window]   Fig. 11. Response properties of an LRF-RHT neuron classified as HT. Because this neuron exhibited a whole-body receptive field, innocuous and noxious mechanical stimulation were applied to all indicated skin areas on both sides of the body. The peristimulus time histograms show the responses to brush, pressure, pinch, and crush. Horizontal lines depict stimulus duration and numbers in parentheses indicate means of responses. Recording site is shown in the circle within the rat drawing. Note that the neuron responded only to noxious stimuli and that the responses to stimulation of ipsilateral orofacial organs and skin areas were more vigorous than the responses to stimulation of extracephalic skin areas or the contralateral head.

View larger version (22K): [in this window] [in a new window]   Fig. 12. Comparisons between the response magnitudes of LRF-RHT neurons to mechanical stimulation of their trigeminal and non-trigeminal receptive fields. A and B: plots of stimulus-response curves illustrating the response profiles of LRF-RHT neurons to mechanical stimulation of their trigeminal and non-trigeminal receptive fields. C: mean ± 95% confidence interval of the responses to brush, pressure, pinch, and crush stimuli applied to trigeminal and non-trigeminal receptive fields. The P values show that LRF neurons respond to brush and pressure of their trigeminal but not non-trigeminal receptive fields and that pinch and crush induce larger responses when applied to trigeminal compared with non-trigeminal receptive fields. Trend analysis (Spearman rank correlation) showed a significant correlation between the stimulus intensity and the response magnitude.

COMPARISONS OF THE RESPONSE PROFILES OF C1-THT, VC-THT, AND LRF-RHT NEURONS. The response profiles of C₁-THT, Vc-THT, and LRF-RHT were compared. These data are presented in Fig. 13, where the response profiles of all HT, WDR, and LT neurons recorded in each of the indicated areas were grouped. Within all three areas, responses to pressure were significantly larger than responses to brush, and responses to pinch were significantly larger than responses to pressure. Among the three groups, the magnitude of the responses to brush (e.g., C₁-THT brush vs. Vc-THT brush vs. LRF-RHT brush), pressure, pinch, and crush were not different (Fig. 13A). The mean ± 95% confidence interval of the responses are shown in the tables on the right. The trends of increased response magnitudes with increased stimulus intensity were also significant for all three groups (P values shown in the row marked "correlation").

View larger version (48K): [in this window] [in a new window]   Fig. 13. Comparisons between the response magnitudes of hypothalamic-projecting C1, Vc, and LRF neurons to mechanical and thermal stimulation of their orofacial receptive fields. A: plots of stimulus response curves illustrating response profiles to mechanical stimuli. B: plots of stimulus response curves illustrating response profiles to heat stimuli. C: plots of stimulus response curves illustrating response profiles to cold stimuli. Right: tables the mean ± 95% confidence interval of the responses and the P values of the trend analysis (Spearman rank for the mechanical stimuli and regression analysis for the heat and cold stimuli). These values show that C1, Vc, and LRF neurons are all capable of encoding the intensity of the noxious mechanical and heat stimuli and that C1 and Vc but not LRF neurons can encode the intensity of noxious cold stimuli.

Axonal mapping

ANTIDROMIC MAPPING IN THE CONTRALATERAL MIDBRAIN AND DIENCEPHALON. Anatomical (Burstein et al. 1987; Cliffer et al. 1991) and physiological (Burstein et al. 1991; Dado et al. 1994a; Katter et al. 1996a) studies indicated that most spinohypothalamic tract axons reach the hypothalamus through the posterior thalamus and the supraoptic decussation. To determine whether THT and RHT neurons also share this anatomical approach, we attempted to map their axonal routes between the midbrain and the hypothalamus. In 72 cases (54 THT and 18 RHT), neurons that were antidromically activated from the contralateral hypothalamus were also antidromically activated with currents of 50 µA from the contralateral midbrain and posterior diencephalon. Figure 14 illustrates an experiment in which a neuron was antidromically activated from low-threshold points in the contralateral midbrain, contralateral caudal diencephalon, and contralateral hypothalamus. The neuron was initially activated antidromically from a lowest threshold point in the contralateral lateral hypothalamus (point a). The lesion marking the location of this point was found just medial to the optic tract, within the supraoptic decussation. The antidromic latency from this point was 2.7 ms (Fig. 14C), and the minimum current required to activate the neuron was 18 µA. When the stimulating electrode was moved in the medial, lateral, dorsal, or ventral directions, higher currents were required to activate the neuron antidromically. Prior to the removal of the first stimulating electrode from the lowest threshold point in the hypothalamus, a second stimulating electrode was used to search for a low-threshold point for antidromic activation of that same neuron from the midbrain (point d). This point was found between the periaqueductal gray and the superior colliculus (Fig. 14B). The antidromic latency from this point was 1.4 ms (Fig. 14C), and the minimum current required to activate the neuron was 11 µA. The first stimulating electrode was then moved 1.5 mm posteriorly, and 11 electrode penetrations were made across the mediolateral extent of the contralateral posterior hypothalamus and thalamus. At this level, the low-threshold point was also located between the internal capsule and the optic tract (point b). The antidromic latency from this point was 2.0 ms, and the minimum current required to activate the neuron was 14 µA. At the level of the posterior commissure (3.5 mm posterior to point a), eight electrode penetrations were made contralaterally. At this level, the lowest threshold point for antidromic activation was located in the substantia nigra pars compacta (point c; 1.6 ms, 9 µA). That the same neuron was antidromically activated from each of the lowest threshold points (a-c), was confirmed by colliding the antidromic spikes evoked at each of these points with the antidromic spikes evoked from the second stimulating electrode in the midbrain (Fig. 14C, d-a, d-b, and d-c). Based on the recorded latencies from each point and their distances from the recording site, we estimated that the conduction velocity of this axon was 6.2 m/s to point a, 7.7 ms to point b, and 8.3 ms to point c.

View larger version (36K): [in this window] [in a new window]