Neuroinflammation is believed to play a key role in Alzheimer’s disease, as evidenced by recent genome wide association studies highlighting several risk genes associated with the immune response. Misfolded and aggregated proteins bind to pattern recognising receptors on glial cells in the brain, causing the production and release of pro-inflammatory messengers, which contribute to disease progression. By studying the relationship between protein aggregation and neuroinflammation at early stages, we can begin to uncover the mechanisms underlying the initiation of the disease. Chapter 3 describes the optimisation of a neuroinflammation assay that uses a BV2 microglial cell line to test the inflammatory potential of synthetic amyloid beta 42 aggregates. In conjunction with a liposome-permeability assay and an immunoprecipitation assay, this assay has been used to validate a proprietary amyloid beta 42 antibody from MedImmune (now AstraZeneca). Most neuroinflammation studies have focused extensively on synthetic aggregates and to a much lesser extent on endogenous soluble aggregates from human tissue. The aggregates present in humans are still poorly characterised due to a lack of suitable methods required for characterising the low concentration of heterogeneous aggregates present. In Chapter 4, a variety of biophysical methods have been employed to characterise the soluble aggregates present in human Alzheimer’s disease brains at Braak stage III. We have identified the similarities and differences between the soluble aggregates in eight different regions by providing a detailed characterisation of their size, morphology, structure, neurotoxicity, inflammatory potential, and capability to permeabilise a lipid membrane. This data shows that soluble aggregates of a range of sizes and morphologies, capable of causing inflammation, are already present in all brain regions at Braak stage III and that aggregation is occurring by the same processes all over the brain to a greater or lesser extent.