For almost ten years, researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have conducted studies to understand why some images are more memorable than others. The team used magnetoencephalography (MEG), which records timing of brain activity, and functional magnetic resonance imaging (fMRI), which identifies active brain regions, to discern when and where in the brain processing of a memorable image occurs.
The team’s study, recently published in PLOS Biology, made use of 78 pairs of images that were similar conceptually but had different memorability scores. These images were shown to 15 participants, each scene varying from skateboarding to natural landscapes to urban scenes to faces displaying different expressions. Researchers found that there are various regions in the brain that are responsible for the encoding and retention processes involved in memorability.
CSAIL affiliate and MIT PhD student in electrical engineering and computer science, Benjamin Lahner, explains that memorable images show stronger and more sustained responses in the brain, particularly in the early visual cortex, which is crucial to memory processing. On the contrary, the response to less memorable images quickly subsides. These findings could redefine our understanding of how memories are formed and retained, potentially paving the way for future clinical applications, especially in the diagnosis and treatment of memory-related disorders.
Lahner’s team developed a fusion method combining MEG and fMRI techniques, which enables recording the brain’s spatial and temporal dynamics. With the aid of machine learning, they created a “representational matrix”, a detailed chart showing the neural responses in the brain. This technique allowed them to pinpoint when and where the brain processes visual images.
However, the researchers acknowledge limitations to their work; it identifies brain regions showcasing significant memorability effects, but it can’t reveal how these regions contribute to memory encoding/retrieval.
Aude Oliva, CSAIL Senior Research Scientist, highlights the clinical potential of understanding the neural bases of memorability, particularly early diagnosis and treatment of memory-related disorders like Alzheimer’s.
According to Wilma Bainbridge, assistant professor of psychology at the University of Chicago, the findings provide crucial insights into neural activities associated with memory processing, which are instrumental in determining which information will be stored or forgotten.
Researchers who also contributed to the study include Yalda Mohsenzadeh from Western University, and Caitlin Mullin from York University. Their research received support from the National Institutes of Health, Vannevar Bush Faculty Fellowship, the Office of Naval Research, National Science Foundation, Multidisciplinary University Research Initiative, Army Research Office, and EECS MathWorks Fellowship.
