Reaction times predict dynamic brain representations measured with MEG for only some object categorisation tasks.

Area of Science:

• Cognitive Neuroscience
• Neuroimaging
• Computational Neuroscience

Background:

• Behavioral reaction times (RTs) link behavior to brain representations in neuroimaging.
• Object representations exist in a multidimensional space where distance indicates neural similarity.
• Faster RTs correlate with greater distance from the animacy decision boundary.

Purpose of the Study:

• Investigate if the RT-distance relationship is consistent across different object categories.
• Examine the temporal dynamics of the RT-distance relationship during object processing.
• Determine if behavioral RTs predict representational distance for categories beyond animacy.

Main Methods:

• Used magnetoencephalography (MEG) to record brain activity.
• Presented stimuli of animate, inanimate, and ambiguous objects (robots, toys).
• Conducted four semantic categorization tasks (animacy, living, moving, human-similarity) and linked RTs to MEG decoding.

Main Results:

• A sustained RT-distance relationship was observed for animacy and human-similarity categories over time.
• This sustained relationship was absent for living and moving categories.
• MEG data showed comparable classification accuracy across all four category organizations.

Conclusions:

• Behavioral RTs predict representational distance for object categorization based on human-similarity, not just animacy.
• The RT-distance relationship's temporal dynamics vary across different semantic categories.
• Further research is needed to understand why this relationship differs between categories.