In search of flavour-nutrient learning. A study of the Samburu pastoralists of North-Central Kenya

Much of our dietary behaviour is learned. In particular, one suggestion is that 'flavour-nutrient learning' (F-NL) influences both choice and intake of food. F-NL occurs when an association forms between the orosensory properties of a food and its postingestive effects. Unfortunately, this process has been difficult to evaluate because F-NL is rarely observed in controlled studies of adult humans. One possibility is that we are disposed to F-NL. However, learning is compromised by exposure to a complex Western diet that includes a wide range of energy-dense foods. To test this idea we explored evidence for F-NL in a sample of semi-nomadic pastoralists who eat a very limited diet, and who are lean and food stressed. Our Samburu participants (N = 68) consumed a sensory-matched portion (400 g) of either a novel low (0.72 kcal/g) or higher (1.57 kcal/g) energy-dense semi-solid food on two training days, and an intermediate version on day 3. Before and after each meal we measured appetite and assessed expected satiation and liking for the test food. We found no evidence of F-NL. Nevertheless, self-reported measures were very consistent and, as anticipated, expected satiation increased as the test food became familiar (expected-satiation drift). Surprisingly, we observed insensitivity to the effects of test-meal energy density on measures of post-meal appetite. To explore this further we repeated a single training day using participants (N = 52) from the UK. Unlike in the Samburu, the higher energy-dense meal caused greater suppression of appetite. These observations expose interesting cross-cultural differences in sensitivity to the energy content of food. More generally, our work illustrates how measures can be translated to assess different populations, highlighting the potential for further comparisons of this kind.