In a longitudinal design, 16 inpatients with bipolar mood disorder and 16 normal control participants were administered measures of backward masking. Bipolar inpatients were assessed while actively manic and again following manic episode. Clinical state was determined from ratings on an expanded version of the Brief Psychiatric Rating Scale. Two backward masking paradigms were used: (a) a staircase method, which yielded a critical interstimulus interval, and (b) set interstimulus intervals, which provided a masking function. Bipolar patients performed significantly worse than the normal controls at both sessions, but the Group X Session interaction was nonsignificant with both masking procedures. The masking performance deficit for the manic patients was not related to the presence of psychotic symptoms but seemed to be partially associated with lithium treatment. The results indicate that the impaired masking performance of manic patients is not strictly limited to the period of the manic episode.