Modulation of receptive field properties of thalamic somatosensory neurons by the depth of anesthesia. The dominant frequency of electrocorticographic (ECoG) recordings was used to determine the depth of halothane or urethan anesthesia while recording extracellular single-unit responses from thalamic ventral posterior medial (VPM) neurons. A piezoelectric stimulator was used to deflect individual whiskers to assess the peak onset latency, magnitude, probability of response, and receptive field (RF) size. There was a predictable increase in the dominant ECoG frequency from deep stage IV to light stage III-1 anesthetic levels. There was no detectable frequency at stage IV, a 1- to 2-Hz dominant frequency at stage III-4, 3–4 Hz at stage III-3, 5–7 Hz at stage III-2, and a dual 6- and 10- to 13-Hz pattern at stage III-1. Reflexes and other physical signs showed a correlation with depth of anesthesia but exhibited too much overlap between stages to be used as a criterion for any single stage. RF size and peak onset latency of VPM neurons to whisker stimulations increased between stage III-4 and III-1. A dramatic increase in RF size and response latency occurred at the transition from stage III-3 (RF size ∼2 whiskers, latency ∼7 ms) to stage III-2 (RF size ∼6 whiskers, latency ∼11 ms). Response probability and magnitude decreased from stage III-4 to stage III-3 and III-2. No responses were ever evoked in VPM cells by vibrissa movement at stage IV. These changes in VPM responses as a function of anesthetic depth were seen only when the nucleus principalis (PrV) and nucleus interpolaris (SpVi) trigeminothalamic pathways were both intact. Eliminating SpVi inputs to VPM, either by cutting the primary trigeminal afferent fibers to SpVi or cutting axons projecting from SpVi to VPM, immediately reduced the RF size to fewer than three whiskers. In addition, the predictable changes in VPM response probability, response magnitude, and peak onset latency at different anesthetic depths were all absent after SpVi pathway interruption. We conclude that 1) the PrV input mediates the near “one-to-one” correspondence between a neuronal response in VPM and a single mystacial whisker, 2) in contrast, the SpVi input to VPM is primarily responsible for the RF properties of VPM neurons at light levels of anesthesia and presumably in the awake animal, and 3) alterations in VPM responses produced by changing the depth of anesthesia are due to its selective influence on the properties mediated by SpVi inputs at the level of the thalamus.
Somatosensory information from the vibrissae of the rat is carried by the trigeminal nerve to the trigeminal complex in the brain stem (Cajal 1911; Hayashi 1980;Vincent 1913). The trigeminal complex projects mainly to the ventral posterior medial (VPM) nucleus of the contralateral thalamus with less dense inputs to other regions of the diencephalon (Bruce et al. 1987; Erzurumlu and Killackey 1981; Lund and Webster 1967; Peschanski 1984; Rhoades et al. 1987; Smith 1973). VPM, in turn, projects to layer IV and adjacent layer III of primary somatosensory cortex (Jensen and Killackey 1987; Killackey 1973; Killackey and Leshin 1975). The majority of the trigeminal neurons that send projections to VPM are located in nucleus principalis (PrV) and nucleus interpolaris (SpVi) (Bruce et al. 1987; Erzurumlu and Killackey 1981; Peschanski 1984). Previous studies have shown that the response properties of VPM neurons are very similar to those of PrV projection neurons. The majority of thalamic projection neurons in PrV respond only to the movement of a single vibrissa (Jacquin et al. 1988; Nord 1968;Shipley 1974; Waite 1984), a receptive field (RF) size comparable with VPM cells (Brown and Waite 1974; Ito 1988; Rhoades et al. 1987; Sugitani et al. 1990; Waite 1973a,b). Thus there appears to be a “one-to-one” correspondence among a single mystacial whisker, a single cluster of neurons in PrV, and a group of neurons in VPM. In contrast, the average RF size of thalamic projection neurons in SpVi is six to eight vibrissae (Jacquin et al. 1986a, 1989a,b;Woolston et al. 1982), suggesting that this anatomic input is ineffective in driving VPM neurons.
Rhoades et al. (1987) showed that a lesion of SpVi had no detectable effect on VPM response properties, indicating that this input was ineffective under their recording conditions. However, when PrV was destroyed, the average RF size of VPM neurons was enlarged from one to six vibrissae. They interpreted the RF enlargement as an “unmasking” of the SpVi influence on VPM neurons after the excitotoxic destruction of PrV.
A recent study in intact rats under light fentanyl anesthesia indicated that VPM cells can have large RF sizes (Simons and Carvell 1989) as well as single whisker RFs, suggesting that RF size may be sensitive to the anesthetic or arousal state of the animal. Two explanations are plausible for the variation in thalamic RF size;1) SpVi projects to a separate population of VPM cells that are difficult to record from, or 2) SpVi inputs on VPM cells are actively suppressed under some conditions. The recent report of Simons and Carvell of VPM neurons with multiwhisker RFs in normal animals, in combination with the lesion studies of Rhoades et al. (1987), led us to favor the idea that deeper levels of anesthesia may selectively gate the influence of SpVi inputs on VPM cell responses. To test this hypothesis, we assessed VPM RF properties at different stages of anesthesia, before and after SpVi pathway interruption. We compared VPM responses under halothane (an inhalation anesthesia) with those under urethan anesthesia, administered by intraperitoneal injection. By using halothane anesthesia we were able to record from single cells continuously at different stages of anesthesia. These results were compared with recordings collected from animals at each single stage of urethan anesthesia. ECoG and physical signs were used to evaluate the animal’s anesthetic stage while recording RF properties of VPM cells.
Forty adult, male Long-Evans (hooded) rats ranging in weight from 250 to 300 g obtained from the Charles River Laboratory were used in these experiments. Thirteen animals were used for RF mapping under halothane anesthesia, and 27 were used for recording under urethan, 15 of which were used for SpVi pathway cut, and the remaining 12 were used for trigeminal tract transection.
Preparation and single-unit recording
Animals were maintained under anesthesia with either halothane, via a tracheostomy, or urethan, by an intraperitoneal injection (1.5 g/kg of body weight). The animal’s body temperature was monitored with a rectal thermister and maintained at 37–38°C on a thermostatically controlled heating pad. Two small openings were made in the skull with a dental drill; one opening was made over the right parietal cortex for recording the EcoG, and the other opening was made over the left VPM nucleus for advancing a carbon fiber microelectrode. A small incision was made in the dura with a 26-gauge needle before applying 4% agar in 0.9% saline at 37.0°C to the surface of the skull over both openings. A cavity was sculpted in the agar over the left side opening and filled with warm saline. The carbon fiber microelectrode (0.5–2.0 MΩ at 1 KHz) Armstrong-James and Millar 1979) was attached to a Kopf stepping microdrive and advanced into VPM. Single units were located extracellularly and isolated with an amplitude-time window discriminator (BAK Instruments). The initial waveform was compared with subsequent waveforms on a digital storage oscilloscope to confirm that the same unit was still isolated. Recording sites were marked by making electrolytic microlesions (10 μA for 5 s) that were located after histological processing of the brain.
Stimulation and data acquisition
Recording methods used in this study were similar to those reported previously (Lee et al. 1994a). In brief, a piezoelectric mechanical stimulator was used to deflect individual whiskers for a duration of 10 ms. The stimulator consisted of a glass capillary glued to one end of the piezoelectric element. The stimulator was calibrated so that the tip of the capillary moved 0.3–0.5 mm in a single direction at the onset of a square-wave pulse applied across the piezoelectric element. The tip of the capillary was positioned just under but not touching the vibrissa to produce a transient mechanical deflection of the vibrissa in the upward direction followed by a passive return to its resting position. The upward deflection of the whiskers was chosen because a previous study from this laboratory indicated that the majority of neurons in VPM are most sensitive to the movement of whiskers in the upward direction (Lee et al. 1994a). The stimulator and data acquisition were controlled by a computer (IBM-AT and “MI2” program, Modular Instruments) that was interfaced through a series of high-speed clocks and memory modules. The action potentials collected during the first 200 ms after the whisker stimulus were collected by the computer interface module and displayed as a peristimulus time histogram (PSTH). The stimulus-dependent times and values were stored on a hard disk for further “off-line” analysis.
Single units were included into the analysis with the following procedure. Unitary responses were initially identified qualitatively by hand-held stimulators. Once a responsive neuron was isolated from background spikes, the piezoelectric stimulator was used to quantitatively measure responses to all whiskers in the whisker pad. All units isolated in this way had a center RF with a response probability of 1.0 and RF size that varied as a function of anesthetic stage.
To specify the level of anesthesia at the time the data for a single cell were collected, we used a classification of the clinical stages of anesthesia developed in humans by Guedel (1920). Guedel divided the depth of anesthesia into four stages from light to deep as I, II, III, and IV, with stage III subdivided into III-1, III-2, III-3, and III-4. The stages of anesthesia can be estimated under a single anesthetic fairly accurately with various physiological signs. However, the additional use of electroencephalographic (EEG) or ECoG data adds a more precise characterization of the level of anesthesia (Pichlmayr et al. 1984). To achieve this we usedKubicki’s (1968) EEG correlates for Guedel’s stages of anesthesia. The induction stages I and II show dominant alpha (8–12 Hz) followed by β frequencies (13–30 Hz); anesthetic stage III-1 has δ (0.5–3 Hz), θ (4–7 Hz), and β frequencies (13–30 Hz); stage III-2 has δ–θ (0.5–8 Hz) frequencies; stage III-3 has δ (0.5–3 Hz) frequencies; stage III-4 has δ (0.5–2 Hz) frequencies plus burst suppression; and stage IV has total suppression (flat line) of the EEG. Halothane- and urethan-induced anesthetic stages were determined with ECoG to define the stage of anesthesia.
Low impedance (<500 KΩ at 100 Hz) tungsten microelectrodes (A-M Systems) placed in the parietal cortex were used in our experiments to obtain continuous ECoG information. The total bandwidth was analogue filtered from 0.1 to 60 Hz and digitally filtered between 0.1 and 32 Hz. The raw ECoG trace was displayed “on-line” on one computer screen as a digitized signal taken at a sampling frequency of 200 Hz and on another screen as a color-coded power band frequency distribution that was obtained by computer analyzed spectral analysis (fast Fourier transform, FFT). The most prevalent frequency (highest power) within a 2-s epoch was defined as the dominant frequency. The dominant ECoG frequency was correlated with physical signs, i.e., with respiratory rate, heart rate, coordinated movement of the entire vibrissa pad (vibrissae movement), corneal reflex, light reflex, and pinch withdrawal reflex.
Transection of primary trigeminal afferents to SpVi or its trigeminothalamic fibers
To transect the primary sensory afferents to SpVi, a 3-mm cut was made in the medial–lateral direction at the level of the subnucleus oralis. Specifically, a #10 scalpel blade was mounted in a stereotaxic holder and advanced into the brain stem at the coordinates AP = −11.0 mm, D = 10.0 mm, and ML = 1.0 to 4.0 mm (Paxinos and Watson 1982). For each case, the extent of the cut through the spinal trigeminal tract was analyzed histologically. Recording unitary activity in VPM at the time of the cut allowed us to compare the RF properties from at least one neuron in each animal before and after the sectioning procedure. In each case, a transient (1–2 s) increase in the firing of VPM neurons indicated that the spinal trigeminal tract was being cut. Only those cases with histologically verified complete cuts were included in this study.
By using similar recording and sectioning techniques, the secondary sensory projection fibers from SpVi to VPM were transected with the coordinates ML = 1.0 mm, D = 10.0 mm, and AP = from −11.0 to −14.0 mm (Paxinos and Watson 1982) to make a 3-mm cut in the anterior–posterior (parasagittal) direction. Therefore the SpVi projection fibers to the thalamus were cut at the level of the SpVi nucleus. Similar to transections of the spinal trigeminal tract, a cut through the SpVi projections fibers inevitably led to a brief, marked increase in the spontaneous firing of VPM neurons. Only those cases with histologically verified transections through the entire extent of the SpVi trigeminothalamic pathway were included in this study. In both tractotomy paradigms the anesthetic stage was maintained at the presectioning stage, either stage III-3 or stage III-2.
In animals in which horseradish peroxidase (HRP; Sigma Type VI; 20% in saline) was used to verify the disruption of SpVi pathway by the transection procedure (n = 15), the recording electrode in VPM was replaced with a 10 μl Hamilton syringe, which injected 0.1 μl HRP almost immediately after the SpVi pathway cut. In three HRP cases, injections were made bilaterally in both VPM nuclei, which facilitated the comparison of back-filled axons and cells in SpVi for the intact and disrupted SpVi–VPM pathways. After the HRP injections, animals survived for 7 days, after which time they were perfused and processed for HRP histochemistry,
At the end of each recording experiment, rats were overdosed with urethan. Animals were then perfused transcardially with a prerinse of 200 ml heparinized PBS followed by 4% paraformaldehyde in PBS solution. HRP-reacted cases were fixed with 1.25% glutaldehyde–1% paraformaldehyde solution in PBS. The neuraxis was transected at the midbrain, and sections were cut at 75-μm thickness through the brain stem in the horizontal plane and through the forebrain in the coronal plane. Individual sections were collected and mounted in serial order and stained with cresyl violet. If HRP injections were made in VPM, the sections were reacted with the method of Itoh et al. (1979). Histological localization of recording sites was documented by locating microlesions made at the time of recording. At the current levels used a blanched circular electrolytic microlesion with a diameter of 50–120 μm could be found in every case. The locations of the single units within VPM were determined with the microdrive reading and angle of penetration of the recording and by accounting for tissue shrinkage during histological processing. Only units histologically verified to be in VPM were included in this study.
The data files for each unit were cast “on-line” as PSTHs and were quantitatively analyzed “off-line” after the recording session. The four standard parameters were response magnitude, response probability, peak onset latency, and RF size. The peak onset latency was calculated by collecting spikes in 1-ms bins after the onset of whisker movement and determining the most frequent time at which the first response occurred within 2–50 ms for each of the 50 trials. The response magnitude was defined as the spike count between 2 and 50 ms after the stimulus divided by the number of trials (50 trials) after subtracting equivalent prestimulus spontaneous activity. Response probability was calculated by correlating the number of stimuli that elicited one or more spikes within the 2- to 50-ms poststimulus time window, with a probability of 1.0 representing a one-to-one correlation. A response to the 50 stimuli was defined as greater than or equal to five spikes 2–50 ms after stimulus onset, at least three of which must be in the same 1-ms bin (for statistical significance seeArmstrong-James and Fox 1987). These response criteria minimize the risk of including random events and stimulus artifacts as responses. All statistical P values presented in this study were derived by paired two-tailed t-test analysis. All error bars presented in this paper represent SE. The number of vibrissae eliciting a response in an individual cell defined the cell’s RF size.
A composite of the physiological changes at various anesthetic stages were derived from 13 animals under the conditions of halothane anesthesia and 27 under urethan anesthesia. Under halothane anesthesia, 13 VPM units were assessed for their RF characteristics at all 5 anesthetic stages, i.e., stage IV to stage III-1. Under urethan anesthesia, 155 VPM units were tested for their RF properties at 1 of 4 anesthetic stages, stages IV, III-4, III-3, or III-2. Stage III-1, the lightest anesthetic stage (Fig.1), could not be obtained under urethan anesthesia. RF data from trigeminal tract cut and SpVi pathway cut cases were averaged from 115 and 144 VPM units, respectively, under urethan anesthesia.
Relationship between stages of anesthesia and the dominant ECoG frequency
Figure 1 lists the Guedel clinical stages of anesthesia and our corresponding ECoG and physical sign data. There were striking similarities between halothane and urethan anesthesia, although halothane levels could be adjusted more quickly by changing the concentration of the inhaled gas, whereas urethan was metabolized slowly and stages changed over periods measured in hours. Because the ECoG frequencies and physical sign data were indistinguishable with these two anesthetics, the data collected under the two anesthetics were combined. Although there was a consistent gradual change in the physical signs present from light stage III-1 to deep stage IV of anesthesia, the physical signs alone did not distinguish among the different anesthetic stages. For example, heart rate and respiratory rate were both more rapid at stage III-1 than at stage IV, but the absolute rate could not be used to determine the specific stage of anesthesia because the range of values overlapped with those in more than one adjacent stage. The same overlap appeared with the testing of reflexes such as withdrawal reflex, eyelid reflex, and corneal reflex. The presence or absence of a given reflex could be used only as a general indicator of the level of anesthesia. In contrast, the dominant ECoG frequency could be used to distinguish between adjacent anesthetic stages. A dominant frequency of 6 Hz linked with 10–13 Hz was seen at stage III-1, pure 5–7 Hz at stage III-2, 3–4 Hz at stage III-3, 1–2 Hz at stage III-4, and suppression of all frequencies at stage IV.
The results in Fig. 1 show that the dominant ECoG frequencies for halothane and urethan were identical at all stages. ECoG tracings under halothane (Fig. 2 A) are representative of both anesthetics. Deeper than light stage III-2, the transitions between stages were clear. The stage III-2 to III-3 transition was characterized by an increase in amplitude and a decrease in frequency in the ECoG. The change from stage III-3 to III-4 showed a decrease in amplitude accompanied by burst suppression and the appearance of spindles. Stage III-4 changed to complete suppression (flat line) in stage IV.
The raw ECoG was impossible to use on-line as a rapid indicator of state while collecting single unit response data. FFT analysis of the raw ECoG data depicted as a color-coded power band spectrum facilitated rapid discrimination between stages. In Fig. 2 B a dominant 3- to 4-Hz frequency was present at stage III-3. As the level of anesthesia was lightened, the dominant frequency shifted to 5 Hz, consistent with stage III-2, and even lighter levels of anesthesia caused a shift to 6 + 10 Hz, indicative of stage III-1. When moving in the opposite direction, from lighter to deeper levels of anesthesia, the same stage transitions were observed. By using the FFT on-line, we were able to determine the anesthetic stage of the animal and maintain the desired stage under either halothane or urethan anesthesia for extended periods of time.
Effect of anesthesia on the RF size of VPM neurons
After developing an effective way to determine the anesthetic stage of an animal, VPM RF properties could be collected at defined stages, III-1 through IV for halothane and III-2 through IV for urethan. RF properties of individual neurons could be collected at each stage with halothane because of the rapid response to different inhaled concentrations. Under halothane, response properties of cells were collected only after an animal was maintained at a defined stage for ≥10 min. The data presented in Fig. 3are the average RF size for four cells that were followed through all stages and nine cells that were characterized between stages III-4 and III-2. This represents the recordings from 13 different animals because we recorded from only one cell per animal to avoid cumulative effects of repeated stage changes.
Under urethan anesthesia we recorded from different populations of cells at stages III-2, III-3, and III-4 (Fig.4, data labeled Normal). In several experiments cells recorded at III-4 were held while moving to a deeper level of anesthesia (stage IV). The data generated from recording under urethan and halothane were very similar. No responses could be obtained at stage IV with either anesthetic. At stages III-4 and III-3 the RF size was on average between one and three vibrissae. A dramatic change in RF size occurred reliably between stages III-3 and III-2. At this transition, the average RF size changed from approximately two vibrissae to approximately six to eight vibrissae. Similarly, at III-1 a large RF size in the range of six vibrissae was seen under halothane anesthesia.
Effect of anesthesia on SpVi pathway
To determine whether the large RF sizes found in VPM cells at III-2 were dependent on SpVi, we selectively disconnected the SpVi pathway to VPM in two ways that are shown schematically in Fig.5 A. The first method was to cut the trigeminal ganglion cell axons after they enter the brain stem at the pons and course in an ascending limb, which innervates PrV, and a descending limb, which forms the spinal trigeminal tract (Cajal 1911; Hayashi 1980).
The trigeminal nerve axons were cut after they enter the pons and descend and before they terminate in SpVi (labeled SpVt in Fig.5 A). Figure 4 shows that after tractotomy the RF of VPM cells under urethan anesthesia ranged from one to two vibrissae in stage III-4, III-3, and even at stage III-2. As expected at stage IV, no responses were elicited. The small RF size at all stages leaves the RFs in VPM indistinguishable from those of cells in PrV. Cutting the descending collateral branches of the primary afferent fibers to PrV and/or disconnecting the extensive intersubnuclear projections between SpVi and PrV may alter the response properties of PrV neurons (Jacquin et al. 1986b, 1990). Therefore a second strategy was devised to disconnect SpVi from the thalamic circuit without directly damaging the primary sensory and intersubnuclear axons to PrV. This was achieved by cutting the SpVi trigeminothalamic fibers before they decussate and ascend in the trigeminal lemniscus with the PrV projections to the contralateral thalamus (labeled SpVp in Fig.5 A). The separate trajectory and level of decussation of the lemniscal fibers from SpVi and PrV are shown diagrammatically in Fig.5 A. The anatomic separation of the two groups of trigeminothalamic fibers enabled us to selectively cut the SpVi fibers while leaving the PrV axons intact. Figure 5 B demonstrates the normal distribution of retrogradely labeled spinothalamic neurons after an HRP injection into the contralateral VPM. After the tractotomy procedure, the same HRP injection showed normal labeling of projection neurons in PrV but absence of HRP-positive neurons in SpVi (Fig.5 C).
Figure 4 shows that cutting the projection fibers from SpVi has an identical effect to cutting the tract fibers; the RF of VPM cells was between one and two vibrissae from stages III-2 to III-4 (urethan). Again, no responses were present at stage IV (n = 10). Therefore removal of the SpVi input to VPM resulted in RF sizes in VPM cells indistinguishable from those reported for PrV cells (Shipley 1974). These results show that either eliminating the SpVi influence on VPM or changing the depth of anesthesia will alter the RF properties of VPM cells. These two sets of data indicate a strong effect of anesthesia, specifically on the SpVi contribution to VPM cell responses.
Convergence of SpVi and PrV pathways onto single VPM neurons
Previous studies provided evidence that the terminal fields of PrV and SpVi projections overlap in VPM (Ma et al. 1987;Peschanski 1984). Further, the PrV fibers terminate on large (proximal) dendrites and somas, and the SpVi axons target small (distal) dendrites (Williams et al. 1991). To date there was no clear demonstration that the two inputs in fact converge on single VPM neurons. To provide such evidence, the first cell from each SpVi deletion case (n = 15) was recorded from before and immediately after pathway section to show that both PrV and SpVi contributed to the response properties of the same neuron (Fig.6). The change in RF size after disruption of the SpVi inputs occurred as quickly as the RF size could be assessed (measured in seconds). Figure 6 shows the combined results from individual VPM cells recorded before and after tract or projection cut. In all cases the RF size of a cell at either stage III-2 (P < 0.0005) or III-3 (P < 0.005) decreased immediately in size, providing physiological evidence for both PrV and SpVi influences on an individual VPM cell.
Effect of anesthesia on VPM RF properties
In addition to RF size, we were interested in determining how anesthesia affected other response properties, such as probability, magnitude, and latency. Response probability and response magnitude of VPM cells are affected similarly by the level of anesthesia and by SpVi projection fiber section (Fig. 7). There is a significant drop in response probability and magnitude elicited by vibrissae comprised in the RF of VPM cells when moving from stage III-4 to lighter levels III-3 and III-2. In contrast to the magnitude and probability of response, the peak latency of response to deflection of vibrissae making up a VPM cell’s RF increased significantly at lighter levels of anesthesia (III-3 and III-2) and decreased after SpVi fiber section. The increase in average latency represents the recruitment of whiskers that produce responses at longer latency via the SpVi pathway. The effect of removing the SpVi thalamic projection fibers is to eliminate these differences between stage III-4 and lighter anesthetic stages. When comparing VPM responses at each anesthetic level, with and without SpVi deafferentation, significant differences are found only between stages III-3 and III-2. No differences in peak latency, probability, or magnitude occur at stage IV whether the SpVi projection is intact or sectioned.
In summary, when the depth of anesthesia permits the influence of SpVi to be detected in VPM cell responses (stages III-3 and III-2), the net effect of SpVi is to increase the number of whiskers that contribute to total RF size, each of which elicits a weaker, longer latency response. The response latency of the principal whisker alone after SpVi deafferentation did not change as a function of anesthetic depth.
These results show a selective effect of urethan and halothane anesthesia on the contribution of the SpVi input to VPM cell responses (see Figs. 3, 4, and 7). Large RF sizes averaging between five and eight vibrissae were never observed after removal of the SpVi contribution, e.g., by deafferenting SpVi, by cutting the SpVi projection to the thalamus, or by inducing anesthetic stage III-4. The lesion-induced deletion of SpVi inputs was followed by an immediate reduction in RF size, thus providing the first physiological demonstration of convergent PrV and SpVi inputs on single VPM neurons.
These results support the finding that many anesthetics used alone produce equivalent EEG changes (Kubicki et al. 1979) and that the dominant EEG or ECoG frequency can be used to distinguish among anesthetic stages in mammals. We determined the anesthetic stage in a rodent and then maintained the animal at that stage under halothane or urethan anesthesia for an extended period of time. By using these techniques, we were able to accurately determine the anesthetic stage of the animal while quantitatively measuring the response properties of single units in VPM. This degree of control permitted us to assess the effect of anesthesia on VPM response properties and to make comparisons among experimental paradigms. The correlation between depth of anesthesia and RF size was noted in previous studies of VPM response properties (Armstrong-James and Callahan 1991; Waite 1973a,b). Determining the size of the RF that corresponds to a specific anesthetic stage permits us to compare the present results with previously reported RF sizes, which vary from one to many vibrissae. The traditional view that VPM cells respond to a single whisker in rodents was derived from studies that used deep barbiturate or urethan anesthesia. Our findings suggest that the data in these studies accurately describe the response of thalamic VPM neurons at anesthetic stage III-4. More recently, an alternative anesthetic procedure, fentanyl (a potent analgesic) in conjunction with Flaxedil (a muscle relaxant), resulted in a high percentage of VPM units that responded to multiple whiskers (Simons and Carvell 1989). Our data suggest that fentanyl produces anesthetic conditions in which the SpVi pathway is actively contributing to and responsible for the multiwhisker RF size of VPM neurons.
Under urethan anesthesia, at a level where VPM cells showed an average RF size of 1.2 whiskers (our stage III-4), Rhoades et al. (1987)reported that acute and chronic kainic acid lesions of SpVi produced no detectable changes in RF properties. Our data indicate that they were recording at an anesthetic stage where SpVi was not expressed in the VPM RFs because SpVi contributes to RF size only at stage III-3 and lighter. Further, there is no significant difference between the RF size of VPM cells at III-4 when the SpVi projection is intact or disconnected from VPM. These results are consistent with the interpretation of Rhoades and coworkers that SpVi does not appear to contribute to VPM RF properties under their conditions. However, after kainic acid destruction of PrV at the same anesthetic level (III-4), Rhoades and colleagues made the important observation that vibrissa-responsive cells in VPM could show an average RF size of 6.3 whiskers. We repeated these experiments and confirmed their results (Friedberg et al. 1989). Thus the destruction of PrV in some way appears to result in the attenuation of the effect of anesthesia on the SpVi pathway activation of VPM.
The detailed anatomy of the trigeminothalamic projections is centrally important to understanding the response properties of VPM cells. Our finding that the PrV and SpVi fibers in their initial trajectory are completely separated before forming the trigeminal lemniscus confirms the previously described location of each of these projections (Jacquin et al. 1986a; Smith 1973). When the trigeminal projection fibers reach the thalamus, they overlap in their distribution in the VPM nucleus (Peschanski 1984). Our physiological data support the idea of overlapping terminal fields of PrV and SpVi fibers and show that an individual VPM cell receives convergent PrV and SpVi inputs. Although the number of cells whose responses were characterized before and after SpVi pathway disruption is small (n = 15), it is important to note that in every case of SpVi pathway disruption there was an immediate decrease in RF size (P < 0.0005 at III-2 and P< 0.005 at III-3). The entire RF of two cells was lost after sectioning the SpVi axons (results from pilot study not shown), which suggests that a few cells in VPM may be totally dependent on SpVi (i.e., do not receive PrV input), but a firm conclusion would need to be supported by a much larger sample.
The separate contribution of SpVi and of PrV to VPM cell responses can be distinguished by analyzing response probability, response magnitude, and response latency in addition to RF size. The PrV contribution can be characterized as a shorter onset latency, higher probability, and higher magnitude response. This response is consistent with the morphology of PrV axon terminals; they are very large (∼10-μm diam), round vesicle-containing profiles that form multiple asymmetrical contacts close to the cell body of VPM neurons (Lee 1981; McAllister and Wells 1981;Peschanski et al. 1984, 1985; Tripp and Wells 1978; Williams et al. 1991). In contrast, the SpVi contribution can be characterized as a longer-onset latency, smaller probability, and smaller magnitude response. SpVi axons are believed to form smaller, round vesicle, asymmetrical contacts on the primary dendrite, distal to the PrV contacts of VPM cells (Williams et al. 1991). Thus PrV and SpVi inputs related to whisker sensation may relay two fundamentally different types of information. The PrV pathway is dominated by a short-latency, sensory-mediated activity related to a single whisker. The SpVi pathway, on the other hand, relays information from a number of different whiskers at a significantly longer latency than the PrV input. The convergence of this complex whisker-related information is regulated at the thalamic level by inhibition from the thalamic reticular nucleus (TRN) (Lee et al. 1994a,b). The feedback inhibition from TRN is extremely effective in suppressing evoked responses in VPM for ∼30–40 ms after a VPM neuron fires an action potential. Consequently, the precise meaning and texture of the sensory information that is relayed to SI cortex appears to be critically dependent on the integration of PrV and SpVi inputs and the state of the reticular activating system.
Identifying the relative contributions of SpVi and PrV leads to the conclusion that there is an interaction between the influence of these two pathways at their site of convergence on an individual VPM neuron. We have shown the response properties of VPM cells when both PrV and SpVi are intact and when the PrV input is isolated by disconnecting the SpVi pathway. To clearly determine the extent of interaction between the PrV and SpVi inputs on individual VPM cells, the response properties of VPM neurons mediated by SpVi in isolation need to be assessed. In this way the individual contribution of PrV and SpVi to VPM cell responses can be compared with the response properties of VPM neurons when both the PrV and SpVi input are functional. In view of our findings on anesthetic effects, we would predict interactions between these two inputs on VPM neurons at anesthetic stages lighter than III-4 as well as in the awake animal.
The most important conclusion that can be drawn from this study is that the level of arousal can dramatically influence sensory processing of sensory stimuli. Therefore either precise control or knowledge of the state of the animal must be maintained to accurately interpret RF data. This is especially true for any plasticity study that proposes to demonstrate the reorganization of sensory representations caused by manipulation of the central or peripheral nervous system.
We thank Drs. Irving T. Diamond, Michael Armstrong-James, and Karen L. Furie for critical evaluation of the manuscript and helpful suggestions and P. Krueger for technical assistance. We are grateful to F. Dorman for help in preparation of this manuscript.
This study was supported by National Institute of Neurological Disorders and Stroke Grants NS-13031 and NS-25907.
Address for reprint requests: S. M. Lee, Division of Neurosurgery, 18-228 NPI, Box 957039, UCLA School of Medicine, Los Angeles, CA 90095.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 1999 The American Physiological Society