| Home > Publications database > Load-dependent processing of prediction violations in task-irrelevant space |
| Journal Article | FZJ-2025-05716 |
; ; ; ; ;
2025
ARVO
Rockville, Md.
This record in other databases:
Please use a persistent id in citations: doi:10.1167/jov.25.14.6 doi:10.34734/FZJ-2025-05716
Abstract: Attentive and predictive mechanisms crucially shape perception, but the interplay between these fundamental processes remains poorly understood. Studies on interactions between attention and prediction have yielded discrepant results, potentially because of differences in task demands. The present study examined whether the perceptual load (i.e., task difficulty) affects predictive processing in task-relevant and task-irrelevant hemifields. To this end, we developed a novel delayed match-to-reference task that orthogonally manipulated task-relevance, prediction, and perceptual load. We hypothesized that a low-load condition should facilitate the processing of prediction violations (oddball effects) in task-irrelevant space because of the availability of spare processing resources. We analyzed accuracy and response time (RT) data from 28 healthy young participants with separate repeated measures analyses of variance. The results confirmed the effectiveness of the load manipulation because a high perceptual load significantly increased RTs and decreased accuracy. Notably, the accuracy analysis yielded a significant three-way interaction between task-relevance, prediction, and load. Post-hoc tests revealed that load modulated the processing of prediction violations in the task-irrelevant hemifield. Importantly, the prediction violation, induced by a low-frequency and task-irrelevant feature (orientation), reduced accuracy in the low-load but not in the high-load condition. This finding suggests that predictive processing in task-irrelevant space is contingent on the availability of processing resources, with high perceptual load inhibiting the processing of unexpected events in task-irrelevant regions. The present study shows that load is a crucial factor in the interaction between task-relevance and prediction.
|
The record appears in these collections: |