The impact of different bitter taste compounds on the retronasal perception of coffee aroma was investigated. A sorted napping experiment was carried out on thirteen compounds at iso-intense bitter concentrations. Differences in perceptual bitter sub-qualities among the compounds were reported by Multidimensional Scaling (MDS) and Multiple Factor Analysis (MFA) analyses. Seven exemplar compounds were further selected to investigate the impact of taste sub-qualities on cross-modal flavor interactions. In general, the different bitter compounds, when paired with a coffee aroma isolate, significantly modified the perception of the retronasal coffee aroma profile. Interestingly, the three bitter compounds endogenous to coffee had the most similar impact on the coffee aroma profile. Further sensory analysis of these sample sets indicated no significant effect of the bitter compounds on the orthonasal perception. Gas Chromatography/Mass Spectrometry (GC/MS) analysis of the volatile composition of the samples headspace also indicated negligible impact of the bitter compounds on aroma release. Altogether evidence of cross-modal interactions occurring at a higher cognitive level were demonstrated in a complex food sample, supporting the importance of multi-modal sensory integration on flavor perception.

Little information exists regarding ethanol thresholds or how thresholds change depending upon the matrix of evaluation. The objective of this experiment was to determine the detection threshold of ethanol in water and the just-noticeable difference (JND) thresholds for ethanol in water and beer. We also sought to determine whether ethanol thresholds correlate with frequency of ethanol consumption. The best estimate of threshold (BET) was determined for 75 subjects in water and/or beer using the ascending method of limits. In water, the group detection BET was 0.66 ± 0.08% with no significant difference between males and females. In water and beer containing 5% ethanol, the group JND BETs were 7.9 ± 1.1% and 11.0 ± 1.1%, respectively. The significant increase of ethanol thresholds in beer compared to water likely reflects the impact of the matrix on the sensory properties evoked by ethanol. No significant differences between males and females were found for ethanol JND thresholds in water, however, in beer, JND thresholds were significantly lower for males compared to females.Ethanol JND thresholds obtained in water and beer were not associated with self-reported frequency of alcohol consumption. Practical ApplicationsThe present study provides relevant details regarding the levels of ethanol that can be discriminated in water and beer. The sensory qualities of ethanol have been proposed to be the most consistent and relevant stimuli to immediately predict alcohol's postabsorptive effects. This implies that humans can accurately detect and discriminate between varying levels of ethanol. However, the present results suggest that the presence of other flavor compounds in a complex matrix such as beer makes discriminating ethanol levels difficult. K E Y W O R D Sbeer, ethyl alcohol, matrix effect, threshold

Utilizing immersive technologies to reintroduce the environmental context (i.e., visual, auditory, and olfactory cues) in sensory testing has been one area of research for improving panelist engagement. The current study sought to understand whether pairing smart-speaker questionnaires in immersive spaces could positively affect the panelist experience through enhanced ecological validity. To this end, subjects performed an immersive consumer test in which responses were collected using a traditional computer-based survey, a smart-speaker approach incorporating a direct translation of the computer questionnaire into a verbal survey requiring numeric responses, and an optimized smart-speaker survey with alternative question formatting requiring spoken word-based responses. After testing, participants answered the Engagement Questionnaire (EQ) to assess participant engagement during the test, and the System Usability Scale (SUS) survey to understand the ease, and potential adoption, of using the various survey technologies in the study. Results indicated that the traditional computer-based survey was the most engaging (p < 0.001) and usable (p < 0.001), with no differences found between the two smart-speaker surveys (p = 0.803 and p = 0.577, respectively). This suggests that the proposed optimizations for the smart-speaker surveys were not robust enough to influence engagement and usability, and further research is needed to enhance their conversational capabilities.