Brain levels of amyloid-β (Aβ) are frequently assessed in transgenic mice models of Alzheimer's disease. The procedure involves tissue sample homogenization using different buffers in a sequential process. No attempt has been made to assess if these procedures are able to extract the total amount of Aβ present in the samples. Here we present data suggesting that standard protocols can lead to a dramatic underestimation of the Aβ content. Results show that higher extraction buffer volumes and at least two repetitions of the soluble and membrane-bound extraction steps are necessary for a more accurate estimation of the Aβ content in brain tissues.