Characterisation of fresh bread flavour: Relationships between sensory characteristics and volatile composition

The sensory properties and volatile composition of bread flavour were measured to allow improved understanding of perceived bread freshness. Twenty bread varieties consisting of specialty breads (n = 10) and commercial breads (n = 10) were evaluated by descriptive sensory analysis, and volatile composition of all breads was measured by proton transfer reaction mass spectrometry (PTR-MS). The specialty breads (n = 10) studied had been evaluated by consumers, and perceived freshness was known. All sensory attributes and 33 mass ions representative of the PTR-MS spectra significantly (p < 0.05) distinguished between the different breads. Partial least squares regression (PLSR) was used to model and predict sensory profiles as a function of volatile composition for all breads. In addition, a separate model that related volatile composition to known consumer freshness of the 10 specialty breads was created. For each model, accuracy was validated by comparing the differences between predicted and actual, sensory and freshness intensities.