Weiss, Kirsten2016-12-192016-12-192016-09https://hdl.handle.net/11299/183304University of Minnesota M.S. thesis. September 2016. Major: Food Science. Advisor: Zata Vickers. 1 computer file (PDF); vii, 105 pages.Because all volatile compounds in food products do not contribute to the perceived aroma, a reliable procedure for determining which compounds do contribute is important. The typical procedure used involves instrumental and sensory methods to isolate, identify, and quantify odorous volatiles followed by recombination and omission testing. The principle idea behind omission testing is quite simple; if a volatile compound contributes to the perceived aroma of the model, people will notice if it is removed. If a compound is removed from a full model and no one can detect that it has been removed, it must not be important. When compounds are combined at concentrations with similar perceived intensities, they tend to fuse or blend creating an aroma with unique characteristics unlike those of the individual compounds. In this situation is omission testing valid? The objective of this study was to determine whether people could learn to discriminate between a five-compound odor mixture with all compounds at the same perceived intensity and the same mixture with any one or two of the five compounds removed. We selected panelists (n = 18) based on their ability to correctly order the perceived intensities of five concentrations of each of 4 compounds (butyric acid, furaneol, methional, and δ-decalactone). During preliminary test sessions we selected, for each panelist separately, concentrations of each of the four compounds that were equivalent in intensity to 5 ppm acetylpropionyl. We then constructed, for each participant, a mixture of the five compounds that were matched in intensity. Panelists then participated in 20 sessions consisting of a series of A-not-A Tests with corrective feedback. During each of these sessions panelists were presented with ten, complete 5-component mixtures, five, 4-component mixtures (each missing a different component), and five, 3-component mixtures (each missing 2 different components). We used signal detection theory and the discriminability index (d’) to evaluate the omission testing data by sessions, panelists, and omitted compounds. The panelists, as a group, were able to discriminate between the full five-compound mixture and mixtures with any one of the five compounds removed after 60 trials. However, low d’ values indicated that the mixtures were extremely difficult to discriminate. We screened out 13 of 31 initial panelists because we could not establish intensity matches between the standard and each of the 4 other compounds. Seven of the 18 who qualified were not able to successfully discriminate between the complete mixture and the n-1 mixture even after 200 trials.enThesis or Dissertation