The purpose of the present study was to investigate whether depressed and anxious people showed implicit memory bias for negative information. Such bias is indicated when performance on an implicit memory task is better for negative rather than positive information. Previous studies in the area have had two methodological problems: (a) Subjects' performance on implicit memory tasks was possibly contaminated with their explicit memory, because the tasks were performed right after the learning session; (b) level of processing in the learning session were deep. Therefore, in the present study, performance on an implicit memory task was measured twice: immediately and two months after the learning session. In addition, lexical and self-reference judgment tasks were used as non-deep and deep tasks, respectively. Results were as follows: 1) Depressed people did not show implicit memory bias for negative information; 2) anxious people showed it two months after the learning session. Finally, maintenance processes of depression and anxiety disorders were discussed.

Key words:implicit memory bias, depression, anxiety, level of processing