Ervin R. Hafter
Dept. of Psych., Univ. of California, Berkeley, CA 94720
Brief, incremental [S(+)] or decremental [S(-)] signals were added to an ongoing pedestal. For detection, correct responses were either ``yes'' to both S(+) and S(-) or ``no'' to nonsignal trials. For identification, a signal was presented on every trial and responses were either ``S(+)'' or ``S(-).'' Subjects monitored either a single channel or, in the ``dual task,'' two simultaneous channels with independent stimuli. Expanding on previous work with two visual stimuli, here there were one visual stimulus (on a CRT) and one auditory (pure tone). As before, dual-task performance in detection was better than in identification. In addition, detection was as good on each of two channels as on one alone, indicating parallel processing, but identification declined, suggesting shared attention. Testing the idea that pedestal-to-signal transients account for these differences, pedestals were turned off before and after the signal intervals, removing transients as sources of information. This did not affect identification, but detection was now worse than identification and declined in the dual task, suggestive of shared attention.