This post was written collaboratively with the authors.
Figure 1: Memory is malleable and vulnerable to
disruption for a time after learning.
People with amnesia in movies typically do not remember anything prior to the memory-erasing incident (usually a whack on the head). People may forget what caused the incident, what they had for lunch prior to the incident, or even their family members’ names. In real life, retrograde amnesia (i.e., loss of information acquired before the onset of memory impairment) is usually not so severe as to cause a loss of identity but it is temporally graded. Specifically, older memories are less vulnerable to disruption than recent memories. The process whereby memories become less vulnerable to disruption is called memory consolidation. Some structures (e.g., the hippocampus) become less critical, while cortical structures take on increasing importance as time passes after learning. Because we cannot whack people’s heads on purpose just to study memory, it is useful to study brain function in healthy individuals using non-invasive functional neuroimaging. Yet most of this work has used retrospective designs where participants are asked to remember information learned in the years to decades prior to the experiment. However, naturally occurring memory can be influenced by many factors, for example, by how much the material has been rehearsed or by how personally significant it may be. Prospective designs, where material is learned and then tested over a shorter time-period, provide a way to study memory consolidation where these variables may be more easily controlled. Using a prospective design, Christine Smith, Robert Clark, and Larry Squire looked at brain activity and connectivity changes in healthy individuals over a much shorter time scale (one hour to one month).Dr. Smith, Dr. Clark, and Dr. Squire work at the Veterans Affairs Healthcare System in San Diego and in the Department of Psychiatry at the University of California San Diego. They designed a prospective study to examine brain activity during short and longer-term memory recall. Examining brain activity at several time points is one way to determine which brain areas are involved in memory consolidation. In order to do this, they used functional magnetic resonance imaging (fMRI) to measure brain activity during a memory task. fMRI uses blood oxygen use as a measure of brain activity.
Fifteen participants studied 4 sets of 80 photos at different time points prior to the fMRI session. They studied one set of photos 30 days prior to the scan, a second set 7 days before the scan, a third set 1 day before the scan, and the fourth set 1 hour prior to the scan. During the fMRI session, participants viewed photos they had seen at the previous time-points mixed with ones they had never seen before. They stated whether the photo was old or new and rated the confidence of their memory on a scale of 1 to 6, with 1 being “Definitely sure the image is new” and 6 being “Definitely sure the image is old”.
|Figure 2: Experimental timeline.|
As expected, the participants remembered the more recently studied photos with greater accuracy and confidence than photos studied at the more remote time-periods. The response time was faster for the more recently viewed photos as well. In the fMRI data, activity increased in three cortical regions and decreased in one. The signal increased as a function of the age of the memory. These increases in cortical brain activity are consistent with the idea that memories consolidate as time passes after learning. Furthermore, detecting changes in the brain over the course of one month is consistent with similar changes observed in animal memory research over the same time scale. But were the changes in the brain related to memory consolidation or were they simply related to memory accuracy, memory confidence, and response time that also changed over the same time-period? To account for these behavioral changes, they used only the data for which participants had high accuracy and high confidence in their memory of the photos, thus equating performance for recent and remote memories. After equating for these parameters, there were still changes in activity in three brain regions, two increased and one decreased. Thus, it appears that most of the brain regions that were sensitive to the age of memory, were not responding to the changes in accuracy, confidence, or response times that accompanied those memories.
In the second part of the study, they assessed changes in brain connectivity between the hippocampus, a region involved in learning and memory, and other brain regions. After equating for behavioral changes, two areas had decreased connectivity with the hippocampus as a function of memory age. This is neat because it suggests that “the hippocampus plays a decreasing role in memory retrieval as memories age”, says Christine, a finding consistent with animal work and studies with memory-impaired patients. In the future, they will examine connectivity changes associated with other areas of the brain, such as the regions involved in visual processing and visual imagery.