Differentiating Real-World Autobiographical Experiences without Recourse to Behaviour.

Area of Science:

• Cognitive Neuroscience
• Neuroimaging
• Machine Learning

Background:

• Investigating consciousness via brain activity is challenging.
• Autobiographical memory relies on frontal, parietal, and medial temporal lobe networks.
• Previous studies used static images; this study uses dynamic video stimuli.

Purpose of the Study:

• To decode brain activity differentiating first-person autobiographical experiences from similar third-person experiences.
• To identify neural mechanisms underlying autobiographical memory recall.
• To develop a novel paradigm for studying autobiographical memory in patient populations.

Main Methods:

• Participants recorded first-person videos of a mall visit.
• One week later, participants underwent fMRI scanning while viewing autobiographical and non-autobiographical videos.
• A machine-learning model classified video categories based on brain activity.

Main Results:

• The machine-learning model successfully classified autobiographical vs. non-autobiographical videos above chance, both within and across participants.
• Brain maps highlighted the fronto-parietal network, mid-temporal regions, and extrastriate cortex as critical for differentiation.
• This suggests a shared neural mechanism for distinguishing personal experiences.

Conclusions:

• A novel paradigm using first-person videos effectively captures autobiographical experiences.
• Shared neural mechanisms differentiate autobiographical from non-autobiographical memories.
• This method is suitable for patients unable to respond to traditional stimuli, like those with brain injuries.