Feature-selective Attention in Frontoparietal Cortex: Multivoxel Codes Adjust to Prioritize Task-relevant...

Area of Science:

• Neuroscience
• Cognitive Neuroscience
• Cognitive Psychology

Background:

• Human cognition exhibits remarkable flexibility, adapting to task objectives by selecting relevant information.
• The frontoparietal attention network, also known as multiple-demand (MD) regions, is crucial for cognitive flexibility.
• The adaptive coding hypothesis suggests MD regions dynamically adjust responses to prioritize behaviorally relevant information.

Purpose of the Study:

• To investigate if the representation of visual object features in MD regions flexibly adjusts based on task relevance.
• To explore the neural mechanisms underlying feature-selective attention within the MD network.

Main Methods:

• Utilized multivariate pattern analysis (MVPA) of fMRI data during a challenging visual object categorization task.
• Trained participants to categorize novel objects based on orthogonal dimensions (length and orientation).
• Employed alternating task blocks where either length or orientation was the relevant feature.

Main Results:

• Multivoxel activation patterns in MD regions encoded task-relevant distinctions more strongly than irrelevant ones.
• MD regions showed differential activation for length when length was relevant and for orientation when orientation was relevant.
• Demonstrated that the brain prioritizes and represents currently relevant stimulus features.

Conclusions:

• The findings support a flexible neural system within MD regions that adapts to behavioral demands.
• This adaptive representation mechanism provides a neural basis for feature-selective attention.
• The study elucidates how the brain dynamically prioritizes information processing based on current goals.