Learning and language in the unconscious human hippocampus

Area of Science:

• Neuroscience
• Cognitive Science

Background:

• Consciousness is key to cognition, but its necessity for higher-order perception is debated.
• Understanding unconscious processing is crucial for defining consciousness.

Purpose of the Study:

• To investigate if learning, semantic processing, and prediction occur in the human brain during unconsciousness.
• To explore the hippocampus's role in processing sensory stimuli under general anesthesia.

Main Methods:

• Recorded neural activity in the human hippocampus using high-density Neuropixels microelectrodes in anesthetized patients.
• Presented auditory stimuli (tones and language) to assess neural responses.
• Utilized a recurrent neural network model to analyze learning and prediction mechanisms.

Main Results:

• Hippocampal neurons detected oddball tones, with learning effects increasing over time.
• A neural network model confirmed emergent learning and tone discrimination.
• Neural ensembles processed semantic and grammatical features of speech, predicting upcoming word information.

Conclusions:

• Complex sensory processing, including learning and prediction, occurs in the hippocampus during unconsciousness.
• The hippocampus demonstrates significant cognitive function independent of conscious awareness.
• Findings challenge the necessity of consciousness for certain higher-order perceptual tasks.