New Insights into How the Human Brain Categorizes and Stores Visual Memories through Neural Timing

Researchers at the University of Southern California have unveiled groundbreaking findings on how the human brain organizes visual memories. In a recent study published in Advanced Science, scientists combined human brain recordings with advanced machine learning techniques to decode the internal neural processes responsible for categorizing visual images. The study sheds light on the brain’s 'filing cabinet' mechanism, demonstrating that the hippocampus sorts visual information into categories such as animals, plants, buildings, vehicles, and tools.

Using intracranial recordings from 24 epilepsy patients, the team observed that the brain's neurons encode visual memories through the precise timing of their electrical spikes, rather than just the firing rate. This temporal coding allows for a high level of specificity, enabling researchers to accurately determine which category of image a person is recalling, solely based on neural activity data.

The innovation in this research lies in the development of an interpretable decoding model that analyzes the spatio-temporal spike patterns of ensembles of neurons. This approach reveals that the hippocampus employs a distributed and millisecond-scale timing code to represent visual memory categories, with a majority of neurons contributing briefly to each memory encoding. Such efficiency minimizes energy use while maximizing memory storage capacity.

This discovery advances our understanding of how the brain manages complex, high-dimensional information without storing every object individually. It underscores the potential for future brain-computer interfaces, including memory prostheses aimed at restoring cognitive function in patients with neurological conditions like dementia and hippocampal dysfunction.

Led by Dr. Dong Song and Dr. Charles Liu, the study utilized recordings from epilepsy patients performing a visual memory task, where they viewed images from different categories. The findings demonstrate that the brain sorts memories into categories in a way that is both decipherable and precise, thanks to the timing of neuronal spikes. This research not only answers fundamental questions about memory storage but also paves the way for innovative neurorestorative technologies.

For more details, see the original publication: Xiwei She et al, "Distributed Temporal Coding of Visual Memory Categories in Human Hippocampal Neurons Revealed by an Interpretable Decoding Model," Advanced Science (2025). Source: https://medicalxpress.com/news/2025-07-human-brain-visual-memories-precise.html