Coding of episodic memory in the human hippocampus.

Area of Science:

• Neuroscience
• Cognitive Neuroscience
• Computational Neuroscience

Background:

• Neurocomputational models propose that episodic memories are encoded by sparse, stimulus-specific neural codes in the human hippocampus.
• A related theory suggests that active neural assemblies suppress competing neurons, a process termed neural sharpening.

Purpose of the Study:

• To investigate episodic memory coding in the human hippocampus and amygdala using single-neuron recordings.
• To test the hypothesis of sparse distributed coding and neural sharpening in human episodic memory.

Main Methods:

• Recorded single-neuron activity from 20 epilepsy patients during a continuous recognition memory task.
• Utilized intracranial monitoring to capture neural responses in the hippocampus and amygdala.

Main Results:

• Observed sparse distributed coding in the left hippocampus, with a small fraction of neurons strongly responding to specific repeated words.
• Found that each repeated word activated a distinct small neuronal fraction, while most neurons showed reduced firing rates.
• This pattern supports the neural sharpening mechanism during memory retrieval.

Conclusions:

• The findings provide strong single-cell evidence for sparse distributed coding in human episodic memory within the hippocampus.
• The results align with and offer support for longstanding neurocomputational predictions previously lacking direct human hippocampal data.