Relationship between volatile compounds and sensory attributes of olive oils by the sensory wheel

Sixty-five volatile compounds and 103 sensory attributes were evaluated in 32 virgin olive oil samples from three different Mediterranean countries. Volatile compounds were analyzed with a dynamic headspace gas-chromatographic technique by using a thermal desorption cold-trap injector. The sensory analysis was conducted by six panels composed of assessors from the United Kingdom, Spain, the Netherlands, Greece and Italy. principal-components analysis of sensory attributes was used to construct a statistical sensory wheel that represents the whole virgin olive oil flavor matrix. This wheel is composed of seven sectors that show the basic perceptions produced by the oil: green, bitter-pungent, undesirable, ripe olives, ripe fruit, fruity and sweet. The boundaries of each sector were calculated from the circular standard deviation of its sensory attributes. The relationship between sensory and instrumental analysis was determined by projecting volatiles onto the sensory wheel. Correlations of each volatile with the first two components of the principal-components analysis were taken as its coordinates (x,y) in the sensory wheel. Volatiles took up the most appropriate place within the sector that corresponded with their perception, and often close to the sensory attributes that explained their sensory properties. A gas-chromatographic/sniffing method was applied to virgin olive oil samples to assess the aroma notes that corresponded to olive oil volatile compounds and to verify the relationships found by the sensory wheel procedure. Most (89%) of the volatiles were well classified. Use of the statistical sensory wheel as an appropriate method to relate volatile and sensory data was clearly demonstrated.