Processing statistics: an examination of focused and distributed attention using event related potentials

Area of Science:

• Cognitive Psychology
• Neuroscience
• Visual Perception

Background:

• Human observers efficiently estimate statistical properties of visual scenes, like mean object size, even with numerous items.
• This capability is often attributed to distributed attention, contrasting with focused attention.
• The nature of working memory (WM) representations under distributed attention is debated: compressed or less structured?

Purpose of the Study:

• To investigate working memory (WM) loading during distributed attention tasks compared to focused attention tasks.
• To test the hypothesis that distributed attention leads to compressed WM representations.
• To explore the structure of representations in distributed versus focused attention using event-related potentials.

Main Methods:

• Used event-related potentials (ERPs) to measure contralateral delay activity (CDA), an index of WM loading.
• Compared CDA amplitude in a mean size estimation task (distributed attention) versus an object recognition task (focused attention).
• Participants performed tasks with varying set sizes (2, 4, or 8 target disks).

Main Results:

• CDA amplitude increased with set size in the object recognition task (focused attention).
• In the mean size estimation task (distributed attention), CDA amplitude remained high across all set sizes.
• CDA amplitude was significantly higher for set size 2 in the mean estimation task compared to the recognition task.

Conclusions:

• Distributed attention tasks, like mean size estimation, result in full working memory (WM) loading, even with small set sizes.
• WM representations under distributed attention are likely less structured, rather than compressed, compared to focused attention.
• Findings challenge the notion of compressed representations in distributed attention and highlight its unique WM demands.