Attention and reaction times to signals of uncertain modality.

Visual, auditory, and tactile reaction time (RT) signals were used in an a-reaction task. The main independent variable was the predictability of signal modality, which was varied by cuing the relevant modality or modalities before each trial. The response requirement was nondiscriminative with respect to modality. Three experiments showed that (a) RT's were longer when signal modality was uncertain, the more so with three possible modalities than with two; (b) this effect of uncertainty was approximately the same whether varied within subjects or between subjects; and (C) the effect of uncertainty was somewhat smaller on tactile RTs than on visual or auditory RTs. Experiment 4 examined change in this uncertainty effect with practice. The uncertainty effect declined over 11 daily sessions to the point of virtual absence from auditory and tactile RTs but was restored or increase will respect to all three signals following one session of discrimination RTs ("respond if visual, refrain if auditory or tactile"). The results are interpreted as showing that attention can be allocated to sensory modalities and that the implied selective process is concerned with modality "identification," though not in a way consistent with a channel-switching model thereof.