These results demonstrate that the parabrachial nucleus is required for acquisition of a learned aversion to a taste in sham feeding rats. The same animals, however, were able to learn to avoid 100% corn oil when it was paired with injections of LiCl. The results are consistent with previous observations that PBN lesions disrupted CTA in real feeding rats when the CS was a taste, but not when it was a trigeminal cue: e.g. capsaicin or 100% corn oil [11
]. In contrast, lesions encompassing the thalamic taste and oral trigeminal relays failed to block learned aversions to either sucrose or corn oil. This is consistent with prior studies showing that acquisition of a gustatory CTA is intact following lesions centered on the thalamic taste area. As such, the data did not support our hypothesis that thalamic lesions centered in the trigeminal relay would interfere with a learned aversion to 100% corn oil. The data are, however, in keeping with the findings in the companion manuscript [14
] showing that lesions of the TOA also failed to reduce operant responding for either a sucrose or a corn oil reward.
Although a range of stimulus properties may potentially contribute to the development of a CTA, these data make it clear that the caloric value of the CS is not necessary for conditioned aversion learning. This experiment was conducted entirely using a sham feeding procedure that prevents substantial metabolic feedback. Nevertheless, all the controls and all the TOAx rats acquired conditioned taste aversions in a manner similar to real feeding animals using the same protocol. These results strengthen other experiments using non-nutritive oral stimuli such as saccharin [12
] or capsaicin [11
] or brief intraoral infusions of the CS [6
] and suggest that taste factors, alone, can support the establishment of a LiCl-induced conditioned aversion.
Lesions of the PBN but not the TOA increased the intake of both the sucrose and corn oil CS upon first exposure. Because the PBNx rats did not completely avoid corn oil until after two pairings with LiCl, they appeared to be slow to acquire the aversion. The initial intake of the corn oil, however, was more than three times greater in the PBNx rats than in the controls. Even so, the slope of decreasing intake in the two groups was near parallel indicating that learning speed was the same. When the CS was 0.3M sucrose, the same PBNx rats again sham drank more than the controls upon the first exposure but the difference was not significant. After three acquisition trials, these rats still did not avoid sucrose. Although sucrose intake decreased from its first trial maximum, it did not drop significantly below baseline water. Further, during the 2-bottle test, PBNx rats still tended to prefer sucrose.
Neophobia is defined by animals consuming only small amounts on first exposure to a novel food. Therefore, the increase in CS intake upon the first exposure by the PBNx rats could be interpreted as a lack of neophobia. Regardless, this increase in intake did not contribute to the failure of PBNx rats to acquire a CTA. Although the rats exhibited higher intake of both oil and sucrose, they learned an aversion to oil but not to sucrose. In fact, rats with PBN lesions often fail to show neophobia because their initial CS intakes are more than the intact rats [10
]. Furthermore, in studies when the initial intake did not differ between groups, the PBNx rats still failed to acquire a CTA while the control rats ceased intake after a single pairing with LiCl [8
]. As such, a general failure to exhibit taste neophobia cannot account for the failure of PBNx rats to acquire a CTA to a gustatory stimulus.
Unlike the PBN lesions, the TOA lesions had no effect on the acquisition of conditioned aversion to either sucrose or corn oil. Although the lesions included both the oral trigeminal and taste area of the thalamus, they failed to disrupt CTA. As mentioned, this result is consistent with several similar experiments in rats with lesions centered on the gustatory thalamus [17
]. These findings can be contrasted with other reports that apparently similar thalamic damage did impair acquisition of a CTA [22
]. These studies, however, employed only a single taste-LiCl pairing. Our experiments and Lasiter's (1985), on the other hand, used 3 CS-US pairings. After a single pairing, some rats with thalamic lesions exhibit somewhat less reduction in CS (sucrose or oil) intake than the controls, but they catch up after the second pairing. In the present case, this second trial difference was not statistically significant; in others it was [19
]. Thus, when only a single CS-US pairing is employed, impaired acquisition could be reported [21
The present thalamic lesions, however, extended into the gustatory area, but were centered on the oral trigeminal relay just lateral to taste. Thus, the behavioral data failed to support the hypothesis that the thalamic trigeminal area is necessary for learning an aversion to corn oil. This appears to contradict our assumption that the trigeminal system processes the sensory properties of corn oil needed for a CTA. It remains possible that the spinal or the principal sensory trigeminal nuclei are necessary for learned aversions to corn oil but this leaves questions as to how the sensory activity reaches the reward system from the brainstem.
The available anatomical data indicate that the trigeminal system reaches the forebrain via the thalamus and large lesions there have little if any effect on CTA regardless of the CS moiety [19
]. Several scenarios could explain this apparent conundrum but none is entirely satisfactory. Because the hypothesis requires that the sensory properties of oil are tactile, an obvious explanation could be that this assumption is incorrect. Olfactory sensibility is not required for oils to serve as a CS in a CTA [25
]. Nevertheless, the odor of oil might substitute if other sensory information was absent. Although we did not test this possibility directly, in the first extinction trial, we did use mineral rather than corn oil. The PBNx rats avoided it, i.e. they generalized to an oil without a vapor phase and thus without an olfactory component.
If the important sensory information mediating oil recognition is neither tactile nor olfactory, then it must be gustatory. If taste is important, however, then PBN damage would have interfered with the formation of an oil CTA. In rats, the PBN is an obligate synapse for gustatory afferent activity destined for the forebrain [26
]. Parabrachial damage blocks CTA when the CS is a taste stimulus, but not if it is primarily trigeminal, i.e. capsaicin or corn oil [11
One possibility is that the pertinent sensory activity produced by oil is trigeminal but not lemniscal. First, primary afferent axons from the lingual branch of V terminate in the rostral nucleus of the solitary tract [29
]. In the vicinity of this terminal field, NST neurons respond to tactile stimulation of the intraoral cavity [30
]. Few of these cells project to the PBN [32
] but, as with some NST taste neurons, they may terminate locally in the subjacent reticular formation [33
]. Although it is not known if these neurons respond to oral oil stimulation, they represent a possible non-lemniscal route for trigeminal afferent activity to reach the forebrain.
Another possible route is the dorsal ascending secondary trigeminal tract [34
]. In primates and some other species, this projection arises from the dorsomedial corner of the principal sensory trigeminal nucleus (dmPV) and ascends ipsilaterally in the central tegmental tract to the ventroposteromedial nucleus of the thalamus [35
]; see Norgren & Leonard, 1973 for further references). Although this area contains neurons that respond to intraoral tactile stimulation, in rodents, this ipsilateral non-lemniscal trigeminal projection is sparse at best or absent altogether [24
]. Regardless, this pathway terminates in the VPM along with the trigeminal lemniscus. Thus it also would be interrupted by thalamic lesions, leaving us back where we started without extra-thalamic oral trigeminal projections to the forebrain.
Finally, several other possible relays exist through which trigeminal afferent activity could reach the forebrain such as the posterior thalamic nuclei, the central gray, and even the hypothalamus [24
]. Neurons in the spinal trigeminal and paratrigeminal nuclei that project directly to the central gray or hypothalamus are normally associated with pain [24
]. Trigeminal research focuses primarily on vibrissae and pain; little effort goes into the intraoral realm. The paucity of intraoral functional data frustrates evaluating the role of these extra-thalamic projections in sensing oil.
The current data eliminated the parsimonious hypothesis that prompted the experiments in the first place. We assumed that the rewarding aspects of oral oil sensibility are trigeminal and that the principal and spinal trigeminal nuclei do not have substantial direct projections to the ventral forebrain as is the case for taste. As a result, we predicted that thalamic VPM lesions would block an oil CTA but damage to the PBN would not. Our data supported the latter but not the former. This leaves open how oil becomes rewarding and underscores the need for electrophysiological data from trigeminal sensory neurons during ingestion.
The results summarized in the first paper in this series demonstrate that PBN damage reduces spontaneous intake of sucrose and interferes with operant responding for it as well. This second paper documents that the same lesions also eliminate acquiring a learned aversion to sucrose. Based on prior research, we can extend these deficits to gustatory stimuli in general. The PBN is less important for any of these tasks when corn oil, rather than sucrose, serves as the oral stimulus. Regardless of the stimulus modality, TOA damage exerted little or no influence on the same tasks. In a subsequent experiment, however, the same animals failed to demonstrate avoidance of a saccharin cue when it was paired with a drug of abuse (Nyland, Li, Liang, & Grigson, in preparation). This demonstrates that the thalamic lesions did affect at least one taste-guided behavior. In prior experiments using similar procedures, somewhat more medial thalamic damage as well as lesions of gustatory cortex disrupted suppression of a taste CS when it was paired with either a highly preferred drug or sucrose [38
]. The third paper in this series, then, uses sham feeding of sucrose and corn oil to test whether the PBN or TOA lesion will disrupt comparison of the relative value of different stimulus concentrations in an anticipatory contrast paradigm.