Attention deficits are encountered very early in the development of Alzheimer's disease. While these deficits may be detected using classic clinical tests or even from simple observation, experimental tools enable a more precise evaluation of these deficits, typically by subtracting between conditions which vary the quantity of necessary attention, and by recording response times which allow a more precise and modulated measurement. The sensitivity of these tools can be further increased by analyzing the intra-individual variability of performance in these experiments, this variability being particularly important during the earliest stages of the disease. We present the cognitive aspects of these deficits by regrouping them according to the type of experiment used to highlight each deficit. Thus, we distinguish deficits of spatial attention orientation, of executive control and of sustained attention. In each section we present several of the more often employed experiments (spatial cueing, spatial search, dual-tasks, conflict tasks, vigilance tests), and draw a parallel between these experiments and everyday life.