Milk proteins are known to bind volatile flavour compounds to varying extents, depending on the nature of the protein and flavour compound. Processing conditions, such as temperature and pH, are also known to have an influence on the interactions between milk proteins and flavour compounds. These interactions cause a great challenge for flavour scientists because they influence the perceived aroma profile of food products significantly, in particular in low-fat food products. The objectives of this research were to develop a headspace solid-phase microextraction (SPME) method followed by gas chromatography with flame ionisation detection (GC-FID) for the investigation of protein-flavour interactions, and to determine binding parameters of the hydrophobic flavour compound, 2-nonanone, to individual milk proteins - namely, β-lactoglobulin (β-lg), α-lactalbumin (α-la), bovine serum albumin (BSA), αs1-casein, and β-casein -, whey protein isolate (WPI), and sodium caseinate. Secondly, it was the aim to compare the binding of the structurally similar flavour compounds - 2-nonanone, 1-nonanal, and trans-2-nonenal – to WPI in aqueous solution, and to investigate the effect of heat and high pressure treatment, and pH on the extent of protein-flavour binding. The final objective was to investigate the in vivo release of the reversibly bound flavour compound, 2-nonanone, from WPI and sodium caseinate using proton transfer-reaction mass spectrometry (PTR-MS), and to understand the effect of viscosity on flavour release in vivo. The binding of the model flavour compound 2-nonanone to individual milk proteins, WPI, and sodium caseinate in aqueous solutions was investigated, using headspace SPME followed by GC-FID. The 2-nonanone binding capacities decreased in the order: BSA > β-lg > α-la > αs1-casein > β-casein, and the binding to WPI was stronger than the binding to sodium caseinate. All proteins appeared to have one binding site for 2-nonanone, except for BSA which possessed two classes of binding sites. The influence of heat treatment, high pressure processing and pH of the protein solutions on the binding of 2-nonanone, 1-nonanal, and trans-2-nonenal to WPI was determined. The binding of these compounds to WPI decreased in the order: trans-2-nonenal > 1-nonanal > 2-nonanone. The binding of 2-nonanone appears to involve hydrophobic interactions only, whereas the aldehydes, in particular trans-2-nonenal, also react through covalent binding. Upon both heat and high pressure denaturation, the binding of 2-nonanone to WPI decreased, the binding of 1-nonanal remained unchanged, while the binding of trans-2-nonenal increased. The binding affinity of the flavour compounds and WPI increased with increasing pH, which is likely to result from pH dependent conformational changes of whey proteins. The in vivo flavour (2-nonanone) release from solutions of WPI and sodium caseinate was investigated using proton-transfer-reaction mass spectrometry. During consumption, 2-nonanone was partly released from WPI, whereas there was no significant release from sodium caseinate. Even after swallowing of the samples, a substantial amount of flavour was detected in the breath, suggesting that the milk proteins interact with the mucosa in the mouth and throat, resulting in a further release of flavour from mucosa-bound proteins. An increase in viscosity of the protein solutions by the addition of carboxymethylcellulose enhanced the release of 2-nonanone from WPI, and resulted in 2-nonanone release from sodium caseinate. This may be due to a thicker coating of the mucosa with the sample solution after swallowing due to the higher viscosity, resulting in additional release of protein-bound flavour. These findings contribute to the knowledge of the interactions that occur between flavour compounds and proteins, which is required to improve food flavouring and to make protein based foods, e.g., low-fat dairy products, sensorily more acceptable to the consumer. The results also emphasize a careful choice of food processing conditions, such as temperature, high pressure or pH to obtain a desirable flavour profile.