Integral membrane proteins (MPs), such as transporters, receptors, and ion channels, are of great interest because of their participation in various vital cellular functions including cell-cell interactions, ion transport, and signal transduction. However, studies of MPs are complicated because of their hydrophobic nature, heterogeneity, and low abundance. Cloud-point extraction (CPE) with the non-ionic surfactant Triton X-114 was performed to simultaneously extract and phase separate hydrophobic and hydrophilic proteins from Alzheimer's disease (AD) and unaffected control brain tissue. Quantitative proteomics analysis of temporal neocortex samples of AD patients and controls was performed using a shotgun approach based on stable isotope dimethyl labeling (DML) quantification technique followed by nanoLC-MS/MS analysis. A total of 1096 unique proteins were identified and quantified, with 40.3 % (211/524) predicted as integral MPs with at least one transmembrane domain (TMD) found in the detergent phase, and 10 % (80/798) in the detergent-depleted phase. Among these, 62 proteins were shown to be significantly altered (p-value <0.05), in AD versus control samples. In the detergent fraction, we found 10 hydrophobic transmembrane proteins containing up to 14 putative TMDs that were significantly up- or down-regulated in AD compared with control brains. Changes in four of these proteins, alpha-enolase (ENOA), lysosome-associated membrane glycoprotein 1 (LAMP1), 14-3-3 protein gamma (1433G), and sarcoplasmic/endoplasmic reticulum calcium ATPase2 (AT2A2) were validated by immunoblotting. Our results emphasize that separating hydrophobic MPs in CPE contributes to an increased understanding of the underlying molecular mechanisms in AD. Such knowledge can become useful for the development of novel disease biomarkers.