Researchers from the University of Toronto have confirmed the critical role played by the brain’s hippocampus in updating our memories when those recollections are shown to be inaccurate.
The hippocampus is part of the brain’s limbic system – the amalgam of neurological components that, among other functions, creates and retrieves our memories.
Previous research demonstrated that the hippocampus was involved in signaling when reality contradicts our expectations, which researchers refer to as a prediction error. When new information doesn’t match what we remember, we sense the signal as surprise – for example, when we hear a friend describe an event we both experienced and discover that some of the details they recall are different from what we remember. But no studies demonstrated conclusively that the hippocampus was involved in subsequently correcting that flawed memory, nor how the revision was made.
Now, in a study published this week in the Proceedings of the National Academy of Sciences (PNAS), researchers in the department of psychology in the Faculty of Arts & Science are among the first to show that the hippocampus is indeed involved in updating memories and that it does this by switching modes.
“What we found is that prediction errors change the role of the hippocampus,” explains lead author Alyssa Sinclair, who conducted the research as an undergraduate student in the department of psychology and is now at Duke University. “Surprise disrupts the stability of hippocampal patterns. Metaphorically, this is like losing your train of thought.
“Surprise captures attention and makes the hippocampus switch from ‘memory preservation mode’ to ‘memory updating mode.’ Prediction errors create opportunities to revise memories by correcting mistakes or adding relevant new information.”
The study’s co-authors include Professor R. Alison Adcock from Duke University and U of T’s Professor Morgan Barense and researchers Iva Brunec and Grace Manalili.
“Our result demonstrates the flexibility of the hippocampus in switching between different modes of processing,” says Barense.
“And the fact that it has underlying mechanisms that allow it to create and reinforce a memory, but also break down a memory and update it – I think that’s amazing. It shows that memory isn’t a video recording.”
The researchers made their discovery by showing study participants video clips featuring actions or sequences of events with notable endings – for example, a baseball player hitting a home run.
The next day, they showed half of the participants the same videos again. But for the other half, they interrupted the videos immediately before the critical moment in the action – in this case, in mid-swing before connecting with the baseball. To participants with a memory of the batter hitting the home run, the interruptions in the videos were prediction errors that prompted surprise.
A day later, the researchers tested all participants’ memories of the videos and found that memories for interrupted videos had been updated. Participants recalled more correct details about the interrupted videos, showing that surprise can strengthen memories. But they also had more false memories for other details, showing that memories were revised.
This demonstrated that prediction errors caused the hippocampus to switch to a “memory updating mode” that revises memories. The study participants likely updated memories by integrating details across different videos, creating false memories.
According to Sinclair and Barense, the insight has applications in a number of areas where memory and updating memories is important.
“The study tells us about how memories can be changed, for better or for worse,” says Sinclair. “We can potentially harness that knowledge in various interventions.”
When study participants viewed interrupted videos, many of the new memories they created were false; for example, they remembered the colour of a person’s hair incorrectly. This could point the way to refining law enforcement interview practices to minimize surprises that might lead an eye-witness to form a false memory – such as showing a witness previously unseen crime scene photos.
Also, the result suggests that, in the classroom, students might learn better from their mistakes if feedback is surprising, thereby putting the hippocampus in editing mode, and then by providing timely, accurate information to incorporate into memory.
And, in the case of individuals experiencing post-traumatic stress disorder (PTSD), the research could help identify opportunities to weaken or change harmful memories, reducing flashback and distress.
The research is just one piece of a larger puzzle Barense is trying to piece together through her research into the biological mechanisms behind memory and cognition.
“Memory is one of the most complex, if not the most complex process in the brain,” she says. “It is fundamental to our existence, our identity, our social interactions. So understanding how our memories are preserved, how they're updated, how this allows us to adaptively function in the world – I think that's a fascinating question.”