Dynamics of flavor perception: combining sensory methods and direct injection mass spectrometry

In this study a combination of sensory and analytical techniques was used to study the effect of roasting degree and sugar addition to coffee on flavor perception and volatile release. The evolution of in-mouth perception was measured with a descriptive dynamic method called “Temporal Dominance of Sensations (TDS)” performed by 18 trained panelists. Panelists were asked to evaluate two espresso coffees (light and dark roasted) in presence and absence of sugar. The panelists used selected flavor attributes (sweet, sour, bitter, astringent, roasted, burnt, caramel, nutty, vegetal) to describe and evaluate the products during 1 min. The volatile release was measured by nosespace (NS) analysis performed with Proton Transfer Reaction-Time of Flight-Mass Spectrometry simultaneously to sensory evaluation for 3 min TDS profiles showed marked differences. Roasting increased dominance of astringent, bitter, burnt and roasted attributes. The presence of sugar in coffee resulted in a total change in the “palette” of attributes that the panelists used to describe dominance. The NS data showed a marked impact of roasting with an increase in overall release for at least 21 mass peaks that could be assigned to key aroma compounds of coffee. NS analysis also showed that the release curves fell into two distinct groups characterized by different patterns. Compounds belonging to cluster 1 were associated to release curves that increased steeply, reached maximum intensity after ~10 s and tailed down relatively fast, almost reaching zero by the end of the TDS/NS session. The release curves of cluster 2 mass peaks increased more slowly, reached maximum intensity at 20 s or later and 60 s after sample introduction, still retained around 20% of maximum intensity. This observation led to detection of candidate mass peaks tentatively assigned to various N-heterocyclic compounds, to be used to explain some of the changes in temporal dominance for burnt and roasted attributes in different time phases. For sugar addition, this factor had almost no measurable effect on NS composition indicating that the effect observed in TDS could rather be imputed to perceptual/cognitive phenomena probably related to cross-modal interactions between taste and olfactory stimuli. The present work represents the first example of simultaneous application of TDS and NS to a food matrix. Quite interestingly, most mass peaks of coffee nosespace fell into two distinct classes according to their release profiles. Some of these peaks were designated as candidate “temporal dominance markers” to be used to describe dominance mechanisms