Neural decoding of autobiographical mental image features with a general semantic model

Area of Science:

• Cognitive Neuroscience
• Neuroimaging

Background:

• Language comprehension and autobiographical imagination were traditionally thought to involve separate neurocognitive systems.
• The semantic system (lateral temporal, parietal, frontal lobes) and the autobiographic/episodic system (midline cortical areas, hippocampus) were considered distinct.

Purpose of the Study:

• To investigate the extent to which language comprehension engages the same brain areas as autobiographical experience imagination.
• To identify shared neural representations between sentence comprehension and self-generated autobiographical mental imagery.

Main Methods:

• Functional Magnetic Resonance Imaging (fMRI) was used on 50 participants imagining unique experiences.
• A pre-trained semantic decoder, derived from fMRI data during sentence reading, was applied to reconstruct self-reported experiential features from mental imagery fMRI data.
• Zero-shot decoding was performed across participants and cognitive tasks.

Main Results:

• Evidence of shared representations was found in both midline and lateral cortical areas during sentence comprehension and autobiographical mental imagery.
• These shared representations were most prominent in midline regions.
• Participant-specific autobiographical feature ratings were successfully decoded from fMRI data using the semantic decoder.

Conclusions:

• Neural encoding of sentence meaning shares cortical areas and representational codes with self-generated autobiographical mental images.
• This challenges the traditional view of separate neurocognitive systems for language and autobiographical memory.