Computer program 'reads' minds
U.K. researchers have used a computer program to analyze human brain activity and "read" a person's memory of watching a short film.
"We've been able to look at brain activity for a specific episodic memory — to look at actual memory traces," Prof. Eleanor Maguire said in a statement.
The 10 volunteers in the study were each shown three seven-second movies, each showing a different woman in a different everyday situation.
In one of the movies, a woman looks through her purse for an envelope and then drops it in a mailbox. In a second film, another actress finishes her cup of coffee and drops the empty cup in a trash can.
The volunteers were then asked to recall each one of the films while they were inside a functional MRI scanner, a medical device that records brain activity by monitoring blood flow inside the brain.
Data from the MRI scans were then fed into a computer, where an algorithm studied the patterns and determined which of the movies the volunteers was thinking about at the time.
"The algorithm was able to predict correctly which of the three films the volunteer was recalling significantly above what would be expected by chance," said Martin Chadwick, lead author of the study, appearing Thursday in the journal Current Biology.
"This suggests that our memories are recorded in a regular pattern," he said.
The researchers focused on the medial temporal lobe of the brain, which includes the hippocampus, although areas throughout the brain are known to be active in storing and recalling memories.
"We found that our memories are definitely represented in the hippocampus. Now that we've seen where they are, we have an opportunity to understand how memories are stored and how they may change through time," said Maguire.
The researchers said the study provides more information about how memories are recorded in the brain.
They found that the hippocampus and surrounding areas of the brain were the key areas involved in recording memory. The study determined that the computer program performed best when it was analyzing activity in the hippocampus itself, suggesting it is particularly important in memory.