Neurocognitive processes that may underlie the age-related decline in memory performance, supplemented by relevant experimental results, will be discussed to summarize current knowledge and highlight unsolved questions. Thereafter, the focus will be on the results of an intentional old/new recognition study designed to test the hypothesis that impaired supervisory processes may cause age-related differences during encoding and subsequent episodic memory. In the study phase, young and elderly participants selected specific semantic information in a HIGH selection (HS) condition, or gave a general judgment about the meaning of a word in a LOW selection (LS) condition. Preliminary results indicate a late (starting around 800ms) left frontal negativity for HS words for both age groups during encoding, possibly reflecting activation of left inferior pre-frontal cortex. In the first test phase, the elderly showed lower hit rates than the young. This was especially pronounced in the HS condition. For the young, more positive going waveforms for hits relative to correct rejections were recorded at parietal locations (parietal Episodic Memory or EM effect). However, these waveforms were small or absent for the elderly. A repetition of the study-test cycle elicited similar ERP encoding effects. However, recognition performance for the elderly improved greatly to almost the level of the young and was accompanied by diminished ERP differences between the age groups. Demanding supervisory processes (HS) seem to complicate the first encoding phase, resulting in poor old/new recognition performance and small EM effects in the elderly. However, a second exposure to the items almost equalized behavioral and ERP recognition effects between the age groups.