Journal of Comparative and Physiological Psychology 1976, Vol. 90, No. 1, 67-77
Some Effects of Conditioned Aversion on Food Intake and Olfactory Bulb Electrical Responses in the Rat Jeanne Pager and Jean-Pierre Royet Claude Bernard University, Lyon-Villeurbanne, France Rats maintained on an unadulterated synthetic food, available from 8 a.m. to 10 a.m. everyday, were submitted to an aversive conditioning schedule on which a first ingestion of eucalyptol-flavored food (EF) was followed by an apomorphine injection (20 mg/kg, ip). In the first experiment the daily food intake was measured from Day 1 to 17, during the first and second hours of the meal. The EF was offered on Days 8, 9, and 17 during the first or the second hour of the meal (Series B or A). On Day 8, the meal was followed in a group of rats by the apomorphine injection. As compared with the intake of Day 8, the mean EF intake of Day 9 was significantly decreased in Series A and B, and of Day 17 in Series A only. No significant EF-intake modification could be observed in a saline-injected group or in an untreated control group. In the second experiment, rats bearing bulbar electrodes for the chronic recording of multiunit mitral cell responses received a 2-hr EF meal before the apomorphine injection. They were stimulated by puffs of odors of pure eucalyptol, unadulterated food, and EF and recorded in hungry and satiated states. Before the aversive conditioning, a significantly greater occurrence of positive responses to the odors of unadulterated food and EF was observed in hungry rats compared with satiated rats. The eucalyptol odor yielded equivalent patterns of responses in hungry and satiated rats before and after conditioning. Conditioning did not alter the modulated responses to unadulterated food odor (a greater occurrence of positive responses was still observed in hungry rats) but modified the responses to the odor of EF (the same high rate of positive responses was then observed in satiated and hungry rats). Electrophysiological data are discussed in terms of palatability changes and food-odor meaning.
The mammalian olfactory bulb appears to be a complex multistage modulated system for information processing, and extensive studies have shown correlations between the olfactory sense and the regulation of various behavior patterns necessary for the survival and perpetuation of the species. Data are still lacking, however, concerning the way in which the anatomical and electrophysiological features of the olfactory system could account for these correlations, although the general theoretical schema established by Le Magnen (1971, 1974) has proved to be a
clarifying tool for the study of the olfactory regulation of behavior. Without preconceived ideas concerning the structures involved, the animal is considered as a cybernetic system endowed with an internal input collecting neural and humoral signals characterizing a given internal state and its corresponding specific arousal state, while an external input transduces and processes sensory cues from the surrounding milieu, External and internal information is integrated so that adapted behavior is released and adequately sustained at the output of
~The^data were presented at the First Congress of the European Chemoreception Research Organi-
*he *??*% has
PrOvided thf the ext?''nal ^ admitted significant cues relative to a
zation, Campus of Orsay, France, July 1974. Spe- specific behavioral arousal state, cial thanks are due A. Holley for his helpful critiIn the particular case of nutritional intercism during preparation of the manuscript, and
Qal gtate and
olfactory input, the set of cen-
i^Tr^.^tS^'rE acknowledged for the pictures. Requests for reprints should be sent to Jeanne
trifugal pathways reaching the olfactory bulbs was proposed as a neural support for the specific internal state that might control
Pager, Laboratoire d'Electrophysiologie, Uni- external ;nmlt it wag verified (Pacer pxtprna i m versite- Claude Bernard, 43 Boulevard du 11 Put' "> was verinea J^ager, novembre 1918, 69621 Lyon-Villeurbanne, France. Giachetti, Holley, & Le Magnen, 1972) that 67
JEANNE PAGER AND JEAN-PIERRE ROYET
the response patterns of mitral cells in the olfactory bulbs of chronically implanted rats were modified by internal state in such a way that there was a significantly greater occurrence of positive responses when food was delivered to hungry as opposed to satiated animals; and on the contrary, amyl acetate stimuli elicited the same slightly positive response pattern in both alimentary states. This selective and specific activation disappeared in the mitral cell layer in the same side as a unilateral transverse section of the anterior limb of anterior commissure (Pager, 1974a). It was thus established that the nutritional state was able to modulate the olfactory input through a selective centrifugal mechanism. Among the many points that remained to be investigated to build up a more precisely detailed model of behavior was that of definition of the proper external cues corresponding to the considered internal state. In terms of food intake behavior, the problem was to determine which criteria conferred nutritional activating properties to an odor. An electrophysiological investigation of palatability learning brought a few data on this point. It was observed (Pager, 1974b) that with nonpalatable eucalyptol stimuli, adult rats gave no significantly different mitral cell response patterns in hunger and satiety even after having eaten an eucalyptol-odorized diet exclusively for 1 mo. However, if consumption of the eucalyptol food began at birth (from the mother's milk) and was maintamed to adulthood, eucalyptol odor, when tested in hungry and satiated rats, elicited modulated response patterns, as would be expected of all food odors; this modulation was shown as a significant decrease of positive responses occurring during satiety. The alimentary character of an odor did not seem therefore to be strictly predetermined but resulted from the animal's previous experience; it was expressed at the mitral level by an electrophysiological equivalent of palatability, which we termed "admissibility," computed from the recorded response patterns. The next stage was to examine whether aversive conditioning could provide symmetrical results on the modification of the alimentary character of an odor. Relative to gustatory aversive condition-
ing, data concerning olfactory learned aversions are more scarce, possibly because gustation is the privileged sensory modality in this kind of experimental association (Domjan & Wilson, 1972; Garcia & Koelling, 1967; Rozin, 1969). Predetermined pathways could facilitate the convergence of visceral and gustatory afferences, and the odor alone should become aversive only when followed immediately and repetitively with toxicosis (Hankins, Garcia, & Rusiniak, 1973). Conditioned olfactory aversion was obtained nevertheless for several days for new odors paired with toxicosis (Lorden, Kenfield, & Braun, 1970) or with metabolic malaise (Simson & Booth, 1973). In fact, as is probably the case in natural food intake behavior during which animals smell and taste the ingested substances simultaneously, both gustatory and olfactory cues, which had been implicated in the establishment of palatability learning, interfered equally in aversive learning as it was performed in the present experiment. The purpose of the experiment was not to achieve a strictly olfactory aversion but to study the effects of a learned nutritional aversion on the olfactory input. The experimental procedure for this purpose was to accustom the animals to a poorly odorized synthetic diet and then on a given day to deliver a meal of the same diet to which was added an unknown flavor (eucalyptol), followed by the injection of a gastrointestinal sickening agent (apomorphine). The toxic effects on the intake of the diet with the new smell and taste were further measured. For electrophysiological recording, hungry and satiated rats were used; electrodes were chronically implanted in the mitral layer of each olfactory bulb, and the rats were submitted to the aversion-conditioning schedule. EXPERIMENT 1 This experiment was designed to determine a proper schedule for an olfacto-gustatory learned aversion to be used for a further electrophysiological study of selective mitral modulation. Method Subjects. The subjects were two series of male Wistar rats, each split into three groups: in Series A, Group 1, n = 9, mean weight = 195 g; Groups
CONDITIONED AVERSIONS AND OLFACTORY BULB RESPONSES 2 and 3, n = 6 each, mean weight = 246 g each and in Series B, Group 1, n = 8, mean weight = 192 g; Groups 2 and 3, n = 6 each, mean weight = 236 g and 225 g, respectively. The rats were housed in individual cages, with water freely available. Their food, a synthetic unadulterated diet mainly constituted of casein, starch, saccharose, and peanut oil with mineral salts and added vitamins, was delivered during a single daily meal between 8 a.m. and 10 a.m. As needed in the experiment, the diet was in some cases flavored with .5% eucalyptol. Experimental procedure. The rats had access to their food between the bars of a grid in such a way that they could gnaw the pellets but not carry them away in the cage. The daily intake was weighed from Day 1 to 17 for the first and the second hour separately. On Day 8 the rats were offered a 1-hr meal of unadulterated basic diet and the eucalyptol-flavored diet during the other hour of the meal. The animals in each group were randomly assigned to Series A or B, which differed according to whether the unadultered diet was offered during the first or the second hour. The two-dish meal was followed by an apomorphine injection in Group 1, an equivalent isotonic saline injection in Group 2, and no particular treatment in Group 3. Two-dish meals similar to those of Day 8 were offered again to all the rats on Days 9 and 17. The apomorphine dose of 20 mg/kg appeared to be necessary for detectable behavioral effects to occur (stereotyped motor patterns). Results
The food intake data, corresponding to the mean daily intake of both diets for each
experimental group, are presented in Figure 1 (Series A) and Figure 2 (Series B). The mean values of eucalyptol-flavored diet intake on Days 8, 9, and 17 were submitted to variance analyses, considering both the day of measurement and the treatment (Group 1, 2, or 3) independently for Series A and B (F and t values corresponding to each series have been noted FA and FB, tA, and ts, respectively). In either series, significant effects appeared to be related to the treatment, FA(2, 55) = 9.3, p < .001; Fs(2, 58) = 4.08, p < .05. On the first eucalyptol-flavored diet presentation, no significant difference appeared among the three groups of rats, FA(2, 17) = .6; FB(2, 18) = 1.9. On the day following injection (Day 9), the eucalyptol-flavored diet intakes differed significantly, FA(2, 17) = 6.11, p< .01; FB(2, 18) = 14.7, p < .001. In both series, however, the control Groups 2 and 3 gave statistically equivalent results, while the apomorphine groups (1A and IB) displayed significantly decreased intakes: Group 1 vs. 2, tA(l2) = 2.233, p < .05, and *B(13) = 2.798, p < .02; Group 1 vs. 3,
Some Effects of Conditioned Aversion on Food Intake and Olfactory Bulb Electrical Responses in the Rat Jeanne Pager and Jean-Pierre Royet Claude Bernard University, Lyon-Villeurbanne, France Rats maintained on an unadulterated synthetic food, available from 8 a.m. to 10 a.m. everyday, were submitted to an aversive conditioning schedule on which a first ingestion of eucalyptol-flavored food (EF) was followed by an apomorphine injection (20 mg/kg, ip). In the first experiment the daily food intake was measured from Day 1 to 17, during the first and second hours of the meal. The EF was offered on Days 8, 9, and 17 during the first or the second hour of the meal (Series B or A). On Day 8, the meal was followed in a group of rats by the apomorphine injection. As compared with the intake of Day 8, the mean EF intake of Day 9 was significantly decreased in Series A and B, and of Day 17 in Series A only. No significant EF-intake modification could be observed in a saline-injected group or in an untreated control group. In the second experiment, rats bearing bulbar electrodes for the chronic recording of multiunit mitral cell responses received a 2-hr EF meal before the apomorphine injection. They were stimulated by puffs of odors of pure eucalyptol, unadulterated food, and EF and recorded in hungry and satiated states. Before the aversive conditioning, a significantly greater occurrence of positive responses to the odors of unadulterated food and EF was observed in hungry rats compared with satiated rats. The eucalyptol odor yielded equivalent patterns of responses in hungry and satiated rats before and after conditioning. Conditioning did not alter the modulated responses to unadulterated food odor (a greater occurrence of positive responses was still observed in hungry rats) but modified the responses to the odor of EF (the same high rate of positive responses was then observed in satiated and hungry rats). Electrophysiological data are discussed in terms of palatability changes and food-odor meaning.
The mammalian olfactory bulb appears to be a complex multistage modulated system for information processing, and extensive studies have shown correlations between the olfactory sense and the regulation of various behavior patterns necessary for the survival and perpetuation of the species. Data are still lacking, however, concerning the way in which the anatomical and electrophysiological features of the olfactory system could account for these correlations, although the general theoretical schema established by Le Magnen (1971, 1974) has proved to be a
clarifying tool for the study of the olfactory regulation of behavior. Without preconceived ideas concerning the structures involved, the animal is considered as a cybernetic system endowed with an internal input collecting neural and humoral signals characterizing a given internal state and its corresponding specific arousal state, while an external input transduces and processes sensory cues from the surrounding milieu, External and internal information is integrated so that adapted behavior is released and adequately sustained at the output of
~The^data were presented at the First Congress of the European Chemoreception Research Organi-
*he *??*% has
PrOvided thf the ext?''nal ^ admitted significant cues relative to a
zation, Campus of Orsay, France, July 1974. Spe- specific behavioral arousal state, cial thanks are due A. Holley for his helpful critiIn the particular case of nutritional intercism during preparation of the manuscript, and
Qal gtate and
olfactory input, the set of cen-
i^Tr^.^tS^'rE acknowledged for the pictures. Requests for reprints should be sent to Jeanne
trifugal pathways reaching the olfactory bulbs was proposed as a neural support for the specific internal state that might control
Pager, Laboratoire d'Electrophysiologie, Uni- external ;nmlt it wag verified (Pacer pxtprna i m versite- Claude Bernard, 43 Boulevard du 11 Put' "> was verinea J^ager, novembre 1918, 69621 Lyon-Villeurbanne, France. Giachetti, Holley, & Le Magnen, 1972) that 67
JEANNE PAGER AND JEAN-PIERRE ROYET
the response patterns of mitral cells in the olfactory bulbs of chronically implanted rats were modified by internal state in such a way that there was a significantly greater occurrence of positive responses when food was delivered to hungry as opposed to satiated animals; and on the contrary, amyl acetate stimuli elicited the same slightly positive response pattern in both alimentary states. This selective and specific activation disappeared in the mitral cell layer in the same side as a unilateral transverse section of the anterior limb of anterior commissure (Pager, 1974a). It was thus established that the nutritional state was able to modulate the olfactory input through a selective centrifugal mechanism. Among the many points that remained to be investigated to build up a more precisely detailed model of behavior was that of definition of the proper external cues corresponding to the considered internal state. In terms of food intake behavior, the problem was to determine which criteria conferred nutritional activating properties to an odor. An electrophysiological investigation of palatability learning brought a few data on this point. It was observed (Pager, 1974b) that with nonpalatable eucalyptol stimuli, adult rats gave no significantly different mitral cell response patterns in hunger and satiety even after having eaten an eucalyptol-odorized diet exclusively for 1 mo. However, if consumption of the eucalyptol food began at birth (from the mother's milk) and was maintamed to adulthood, eucalyptol odor, when tested in hungry and satiated rats, elicited modulated response patterns, as would be expected of all food odors; this modulation was shown as a significant decrease of positive responses occurring during satiety. The alimentary character of an odor did not seem therefore to be strictly predetermined but resulted from the animal's previous experience; it was expressed at the mitral level by an electrophysiological equivalent of palatability, which we termed "admissibility," computed from the recorded response patterns. The next stage was to examine whether aversive conditioning could provide symmetrical results on the modification of the alimentary character of an odor. Relative to gustatory aversive condition-
ing, data concerning olfactory learned aversions are more scarce, possibly because gustation is the privileged sensory modality in this kind of experimental association (Domjan & Wilson, 1972; Garcia & Koelling, 1967; Rozin, 1969). Predetermined pathways could facilitate the convergence of visceral and gustatory afferences, and the odor alone should become aversive only when followed immediately and repetitively with toxicosis (Hankins, Garcia, & Rusiniak, 1973). Conditioned olfactory aversion was obtained nevertheless for several days for new odors paired with toxicosis (Lorden, Kenfield, & Braun, 1970) or with metabolic malaise (Simson & Booth, 1973). In fact, as is probably the case in natural food intake behavior during which animals smell and taste the ingested substances simultaneously, both gustatory and olfactory cues, which had been implicated in the establishment of palatability learning, interfered equally in aversive learning as it was performed in the present experiment. The purpose of the experiment was not to achieve a strictly olfactory aversion but to study the effects of a learned nutritional aversion on the olfactory input. The experimental procedure for this purpose was to accustom the animals to a poorly odorized synthetic diet and then on a given day to deliver a meal of the same diet to which was added an unknown flavor (eucalyptol), followed by the injection of a gastrointestinal sickening agent (apomorphine). The toxic effects on the intake of the diet with the new smell and taste were further measured. For electrophysiological recording, hungry and satiated rats were used; electrodes were chronically implanted in the mitral layer of each olfactory bulb, and the rats were submitted to the aversion-conditioning schedule. EXPERIMENT 1 This experiment was designed to determine a proper schedule for an olfacto-gustatory learned aversion to be used for a further electrophysiological study of selective mitral modulation. Method Subjects. The subjects were two series of male Wistar rats, each split into three groups: in Series A, Group 1, n = 9, mean weight = 195 g; Groups
CONDITIONED AVERSIONS AND OLFACTORY BULB RESPONSES 2 and 3, n = 6 each, mean weight = 246 g each and in Series B, Group 1, n = 8, mean weight = 192 g; Groups 2 and 3, n = 6 each, mean weight = 236 g and 225 g, respectively. The rats were housed in individual cages, with water freely available. Their food, a synthetic unadulterated diet mainly constituted of casein, starch, saccharose, and peanut oil with mineral salts and added vitamins, was delivered during a single daily meal between 8 a.m. and 10 a.m. As needed in the experiment, the diet was in some cases flavored with .5% eucalyptol. Experimental procedure. The rats had access to their food between the bars of a grid in such a way that they could gnaw the pellets but not carry them away in the cage. The daily intake was weighed from Day 1 to 17 for the first and the second hour separately. On Day 8 the rats were offered a 1-hr meal of unadulterated basic diet and the eucalyptol-flavored diet during the other hour of the meal. The animals in each group were randomly assigned to Series A or B, which differed according to whether the unadultered diet was offered during the first or the second hour. The two-dish meal was followed by an apomorphine injection in Group 1, an equivalent isotonic saline injection in Group 2, and no particular treatment in Group 3. Two-dish meals similar to those of Day 8 were offered again to all the rats on Days 9 and 17. The apomorphine dose of 20 mg/kg appeared to be necessary for detectable behavioral effects to occur (stereotyped motor patterns). Results
The food intake data, corresponding to the mean daily intake of both diets for each
experimental group, are presented in Figure 1 (Series A) and Figure 2 (Series B). The mean values of eucalyptol-flavored diet intake on Days 8, 9, and 17 were submitted to variance analyses, considering both the day of measurement and the treatment (Group 1, 2, or 3) independently for Series A and B (F and t values corresponding to each series have been noted FA and FB, tA, and ts, respectively). In either series, significant effects appeared to be related to the treatment, FA(2, 55) = 9.3, p < .001; Fs(2, 58) = 4.08, p < .05. On the first eucalyptol-flavored diet presentation, no significant difference appeared among the three groups of rats, FA(2, 17) = .6; FB(2, 18) = 1.9. On the day following injection (Day 9), the eucalyptol-flavored diet intakes differed significantly, FA(2, 17) = 6.11, p< .01; FB(2, 18) = 14.7, p < .001. In both series, however, the control Groups 2 and 3 gave statistically equivalent results, while the apomorphine groups (1A and IB) displayed significantly decreased intakes: Group 1 vs. 2, tA(l2) = 2.233, p < .05, and *B(13) = 2.798, p < .02; Group 1 vs. 3,
