Even when two odors are difficult to tell apart, can repeatedly receiving chocolate after one of them lead to different responses? In a small 2014 experiment, there was no clear improvement in choosing the correct odor, but differences appeared in inhalation, reaction time, and skin conductance responses to the odor associated with a reward.
The researchers wanted to find out whether learning through a pleasant experience changes odor discrimination, and whether any change is specific to one odor or spreads to similar ones.
A small addition to lemon and strawberry odors
The researchers prepared lemon fragrance and a version blended with 2% Hedione, as well as strawberry fragrance and a version blended with 3% Hedione. Hedione is a fragrance material reminiscent of jasmine. In a preliminary test, performance in distinguishing each original odor from its lightly blended version was not significantly above chance.
Eighteen people who liked chocolate and were not dieting took part in the main study. Two were excluded because of recording problems, leaving 16 in the analysis. They were asked not to eat for four hours beforehand to increase their motivation to eat. A device delivered the odors to the nose while inhalation and skin conductance responses were recorded.
In a control task comparing Hedione alone with odorless air, participants discriminated above chance. The problem was not that the material itself could not be perceived, but that a small amount mixed into lemon or strawberry was difficult to distinguish.
Only one odor signals chocolate
Just one of the four odors was made a signal for receiving 0.5 g of chocolate. No chocolate followed the other three. The reward-predicting odor was counterbalanced across participants, and milk or dark chocolate was used according to individual preference. Each odor was presented 18 times, for 72 presentations in total.
After smelling an odor, participants pressed a key to remove a gray patch covering the screen, revealing either a chocolate image or a symbol indicating no reward. Key-press speed was recorded, but they were told that how they pressed did not determine the reward. They were also asked to work out which odor predicted chocolate.
Before and after this learning phase, participants completed a task in which they selected the one different odor from three presentations. The lemon pair and strawberry pair were tested separately, with accuracy and inhalation measured.
No clear gain in accuracy, but a difference in inhalation
Accuracy in the three-choice task was not significantly above the chance level of approximately one-third, either before or after learning. The results did not show that the pair containing the rewarded odor had become clearly distinguishable.
After learning, however, inhalation volume was greater for the chocolate-associated odor than for its similar counterpart that did not predict chocolate. This difference was absent before learning. No comparable significant change appeared for the other pair, which was not associated with reward.
This was a difference between two odors after learning. To avoid effects of the passage of time, such as habituation, the researchers emphasized this relative difference. They did not directly test whether inhalation volume for the rewarded odor alone increased from before to after learning.
Specific responses, but liking spreads to a similar odor
Among the 13 participants with key-press records, mean reaction time after the chocolate-predicting odor was approximately 384 milliseconds, shorter than the approximately 406-millisecond average for the other three odors combined. The difference was absent in the first half of learning and significant in the second half.
Skin conductance responses during odor ratings after learning were also greater for the rewarded odor than for the other three. This analysis included 13 participants, excluding those without responses recorded at the appropriate time. During the three-choice discrimination task, however, there was no significant difference in skin conductance responses. The same physiological response did not occur in every setting.
Self-reported liking showed a somewhat different pattern. The rewarded odor alone was not significantly better liked than the other three. Instead, the pair consisting of the rewarded odor and its similar counterpart was better liked overall than the unrewarded pair. The authors interpreted this as liking generalizing to similar odors, while responses such as inhalation appeared more selectively for the odor predicting reward.
Looking at answers and bodily responses together
What makes the result interesting is that differences not apparent in discrimination accuracy emerged in other measures. The authors regard this as evidence of implicit discrimination arising through reward learning.
That does not mean bodily responses reliably reveal a person’s “true preferences.” Participants consciously searched for the relationship between odors and chocolate, and the odors themselves were perceptible stimuli. This was a short experiment with few participants and a limited set of fragrance materials. A lack of difference in discrimination performance does not establish that the odors could not be distinguished by any method. Brain activity and the long-term persistence of learning were not measured.
Summary
After learning, the chocolate-associated odor elicited greater inhalation volume than its similar unrewarded counterpart, and faster key responses in comparison with the other odors combined. Skin conductance responses during subsequent ratings were also greater, while a clear improvement in three-choice discrimination was not established.
Liking also spread beyond the rewarded odor to its similar counterpart. The study shows that, when learning about odors, being able to report a distinction, finding an odor pleasant, and responding physiologically do not necessarily change in the same way.
The research received support through a Swiss National Science Foundation NCCR and a research grant from Firmenich, which also supplied fragrance materials and technical assistance. Two coauthors were affiliated with Firmenich at the time. The original conflict-of-interest statement declares no commercial or financial relationships that could be construed as a potential conflict of interest.
Reference
Pool E, Delplanque S, Porcherot C, Jenkins T, Cayeux I, Sander D. Sweet reward increases implicit discrimination of similar odors. Frontiers in Behavioral Neuroscience. 2014;8:158. doi:10.3389/fnbeh.2014.00158
