How the Way We Recall Information Changes Our Memories

How Memory Retrieval Alters Neural Representation

We typically think of memory as a fixed resource stored away in the brain, ready to be retrieved unchanged whenever we need it. But psychologists have found that the very act of accessing a memory updates it for the future. Our memory is far from simply a fixed resource that we consult as needed, says Marc Coutanche, an associate professor of psychological sciences at Rice University. It’s dynamic and adaptive. Memory retrieval actively updates our memories rather than just ‘playing them back.’ Coutanche and Amy Qi, a fourth-year doctoral student at Rice University, set out to determine whether the specific manner in which people access information leaves a lasting trace detectable during subsequent retrievals. Imagine thinking about a pet dog: on one occasion you might focus on the unique pattern of its fur, while on another you might categorize it as a dog or consider its place in your family. The researchers wanted to discover whether these different pathways of recall shape subsequent brain activity.

Brain Imaging and Computational Learning Methods

To investigate this dynamic, the researchers asked study participants to learn pairs of images and unfamiliar Dutch words while undergoing functional magnetic resonance imaging, or fMRI. Later in the study, researchers prompted the participants to recall these objects in distinct ways—either by focusing on a specific visual feature, placing them in a broader category, or thinking about a theme. When tested on their memory later, participants performed equally well on tests regardless of how they had previously thought about the information. However, underlying neural activity revealed a completely different story about how the brain stores and handles recalled data. What I find most interesting is that we showed that a memory seems to carry a trace of how it was used before, Qi explains. The neural activity related to the memory not only reflected what participants were doing in the moment but also something about how that memory had been accessed before.

Visual Processing and Computational Insights

Further analysis showed that thinking about objects by category made the neural patterns associated with different objects look more alike within areas of the brain involved in visual processing. Conversely, focusing intensely on an individual object made its neural pattern more distinct in a brain region responsible for processing words. These findings reinforce the view that human memory changes with use rather than remaining a static record. Furthermore, the study highlights the power of advanced computational tools to uncover subtle patterns in the brain that basic behavioral tests might miss. By combining fMRI with machine learning techniques, the researchers detected how information had previously been accessed within the hippocampus and visual cortex. Without looking at the subtle patterns of activity with techniques like this, we might have missed the findings entirely, Coutanche notes. Computational methods like this allow us to detect and understand subtle patterns of activity in a way that gives real insights into how the brain remembers.