MIT Researchers Map Brain Pathways of Visual Images Being Recognized
Researchers from the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) have discovered how some images linger in our memories while others fade away. Their study, published in PLOS Biology, reveals new insights into the brain dynamics that govern visual memorability.
Utilizing a combination of magnetoencephalography (MEG) and functional magnetic resonance imaging (fMRI), the team has, for the first time, pinpointed both the timing and the locations within the brain where memorable images are processed.
This dual approach allows for an unprecedented glimpse into the spatial and temporal aspects of how our brains perceive and remember images. The study involved showing 15 subjects 78 pairs of images that depicted the same concept but had different memorability scores — one image from each pair was easy to remember, while the other was easily forgotten.
These ranged across various themes, from urban landscapes to natural scenes, and everyday objects to human expressions. Benjamin Lahner, an MIT PhD student in electrical engineering and computer science and the study’s first author, explains, “We discovered that highly memorable images evoke stronger and more sustained responses in the brain, particularly in areas like the early visual cortex, which we previously underestimated in memory processing.“.
This “brain signature” of memorability, as Lahner calls it, appears about 300 milliseconds after an image is seen. It involves a broader network of brain regions than previously known, including the ventral occipital and temporal cortices, which handle aspects like color perception and object recognition.
The implications of these findings are profound, especially for clinical applications. Aude Oliva, CSAIL Senior Research Scientist and one of the study’s leaders, highlights the potential benefits for diagnosing and treating memory-related disorders. “Identifying specific brain signatures linked to memorability could lead to early biomarkers for conditions such as Alzheimer’s disease,” she states.
The study’s MEG/fMRI fusion method, enhanced by machine learning techniques, overcomes the limitations of previous research that could only focus on either the spatial or temporal dynamics of brain activity. This method provides a detailed “representational matrix” that illustrates how different brain regions respond to various images.
Despite these advancements, the researchers acknowledge limitations in their study, particularly in defining the precise functions of the identified brain regions in memory processing. However, the research paves the way for future exploration and potential therapeutic interventions tailored to individual neural profiles.
Wilma Bainbridge, an assistant professor of psychology at the University of Chicago who was not involved in the study, commented on the significance of these findings: “These insights bridge the gap between viewing something and storing it in memory, highlighting the cortical signals that determine what we remember and what we forget.“
Originally posted on OpenDataScience.com
Read more data science articles on OpenDataScience.com, including tutorials and guides from beginner to advanced levels! Subscribe to our weekly newsletter here and receive the latest news every Thursday. You can also get data science training on-demand wherever you are with our Ai+ Training platform. Interested in attending an ODSC event? Learn more about our upcoming events here.
