Neurodegenerative diseases, such as Parkinson's and Alzheimer's, constitute a growingthreat with ever increasing prevalence. These diseases are characterized by the presence ofprotein deposits in the patient's brain that are called amyloids. Several proteins wereidentified in these deposits as being the molecular hallmark of the disorder, among whichwe can cite alpha-synuclein for Parkinson’s disease and tau for Alzheimers’s disease.Protein amyloid aggregation is central to neurodegenerative diseases and hence constitutesa target of choice for diagnostic and therapeutic attempts. Itis characterized by the formation of a structural cross-beta pattern, which is a stack ofbeta-sheets, usually forming long fibrils. Under specific conditions, larger aggregates can beobtained, such as micrometer-sized particles, including so-called particulates andspherulites. Several pieces of evidence suggest that the formation of such aggregates, and especially at early-stages, can be involved in protein toxicity. Yet, the reasons for theaggregation to occur are not well understood. In this work, we aimed at deciphering thefundamental principles underlying protein amyloid aggregation by studying the changesin protein ad hydration water dynamics, the understanding of which might help in the development ofwater-dynamics based diagnostic methods.We employed mainly incoherent neutron scattering (on SPHERES at the MLZ and IN16B at the ILL)and molecular dynamics simulations. Theformer provides ensemble averaged information on hydrogen motions in the system, and thelatter provides a fully atomistic picture from which dynamical and structural aspects canbe investigated.Studying alpha-synuclein, we could show that protein backbone and side-chain motions - that is,internal dynamics - is barely affected by aggregation. However, hydration water motions areincreased around amyloid fibrils. The increased dynamics originates from a fraction ofwater molecules being displaced from the protein hydrophobic core to the hydrophilictermini regions when fibrils are formed. Hence, it results in a higher water entropy in fibrils,where the central cross-beta pattern appears highly efficient in protecting itself frominteracting with the solvent.For gammaS-crystallin, comparison of the internal protein dynamics of the wild-type proteinwith a G18V mutant revealed that the mutant is less dynamic, whatever itsaggregation state. This observation, along with the comparison of protein dynamics withtheir relative hydropathy index, indicates that the internal dynamics depends strongly onthe amino acid composition, but not on the aggregation state. In addition, other factorscan affect protein dynamics, such as the presence of metal ions.The measurements carriedout on insulin, in the presence or absence of zinc show that the metal promotes proteinhydration at pH 1.8, where it interacts loosely with the protein. The zincaffects also aggregate-aggregate interaction, probably by electrostatic screening as theformation of spherulites is facilitated in the absence of the metal.Eventually, the possibility to unambiguously and simultaneously access internal dynamicsand center-of-mass diffusion was demonstrated by carrying out so-called fixed- window scanson the IN16B instrument at the ILL. This novel technique applied to lysozyme showed thatparticulate formation occurs in a one-step process, and the internal dynamics remainsconstant all along. This pilot experiment opens up the possibility to study fibrilformation of pathologically relevant proteins.Taken together, the aforementioned results demonstrate that we can now study the amyloidaggregation process with great detail, and there is a great opportunity to extend this workwithin a biological context, in order to link the biophysical properties of protein amyloidaggregation with its effects and toxicity in-vivo