Temporal dynamics of adapting to novel contexts during the generalization of learned spatial suppression

The spatial suppression of a high-probability distractor location (HPDL) acquired through statistical learning critically reduces its attentional priority. The present study conducted four experiments to systematically investigate the mechanisms underlying how this learned suppression generalizes to novel tasks. Specifically, we focused on how generalization is influenced by the priority-enhancing attentional capture of a salient target and the competing demands of a newly introduced suppression process. Following training, Experiments 1 and 2 employed a salient color singleton target to test attentional capture. Experiment 1 reused training stimuli, whereas Experiment 2 introduced novel shapes to determine if the capture effect of novel stimuli would completely override the learned suppression. Experiment 3 introduced a new suppression process to evaluate the impact of competing suppression demands. Finally, Experiment 4 utilized a feature search paradigm as a test task devoid of color singletons to eliminate these interfering factors. The results revealed that the intense capture effect of novel stimuli in Experiment 2 completely masked the generalization. Furthermore, sliding window analyses in Experiments 1 and 3 uncovered a dynamic process where generalization only manifested during the middle of the test phase, indicating a competition between these newly introduced factors and the previously acquired suppression. Conversely, eliminating all interference in Experiment 4 yielded a highly stable and persistent generalization of the HPDL suppression. These findings demonstrate that novel interfering factors drive a dynamic adjustment of the HPDL weight within the priority map, revealing the highly flexible and adaptive nature of human attentional control.