Attention to multiple cues by severely mentally retarded adults: effects of single-component pretraining.

The present investigation examined whether the amount of compatible single-element training affected response accuracy for conditional-discrimination tasks requiring attention to multiple cues. Four severely mentally retarded adults participated, and a multiple baseline across subjects and reversal design was employed. Following acquisition of conditional discriminations by individually training each stimulus element, transfer tests were provided. In the first transfer condition, two intermixed conditional-discrimination tasks were presented containing identical stimuli to those individually trained. The second transfer condition consisted of two intermixed conditional-discrimination tasks composed of novel S- stimuli and pretrained S+ stimuli. Finally, a control procedure was administered in which conditional-discrimination tasks contained all untrained cues. The severely mentally retarded subjects did learn to respond to multiple cues. Establishing separate control by each component produced attention to two aspects of compound stimuli with few errors occurring. In addition, these broadened attentional skills generalized to multistimulus tasks containing all or some of the original pretrained stimuli. If single-element training was omitted, performance on conditional-discrimination tasks was greatly impaired.