Where do I go? Decoding temporal neural dynamics of scene processing and visuospatial memory interactions using convolutional neural networks

Comprehensive Summary

This study explored how bottom-up scene features and top-down contextual knowledge integrate during visual perception and processing. EEG recordings were obtained from 30 participants while they performed scene memory and visuospatial tasks in an alternating manner. During this investigative procedure, the researchers manipulated the number of open doors as well as applied CNNs with Grad-CAM to decode neural dynamics. The researchers identified an early occipitoparietal integration window in which visual features and spatial memory can converge. To continue, it was found that newly acquired spatial knowledge altered neural responses. These findings demonstrate that spatial knowledge exerts a rapid top-down influence on scene perception.

Outcomes and Implications

This study highlights how spatial memory can shape scene perception, showing that top-down knowledge influences neural processing. For the medical community, this suggests that disruptions in perception-memory integrations could be detected and monitored using EEG.