学習と記憶: 覚えて知っているタスク

Our experience of memory is varied and complex. Sometimes we can remember events in vivid detail, while other times we may only have a vague sense of familiarity.

The first type, a recollected memory, is one that is remembered with strong details about the time at which it was learned—such as a dining experience the previous evening, where not only was the lobster dinner recalled, but also were the paintings on the wall and the restaurant staff who served you.

On the other hand, a familiar memory is similar to a recollected one in that it is known, but differs in that it is recalled without any explicit details surrounding the event. That is, a familiar memory lacks specifics about the setting, like the waiter who served dinner or what the décor was.

This video demonstrates how to combine functional magnetic resonance imaging—fMRI—with a task called Remember-Know to investigate how the brain—especially the hippocampus—responds to judgments made towards repeated or novel images based on previous work performed by Gimbel and Brewer.

In this experiment, participants are asked to complete two phases: initial encoding and fMRI testing. In part one, encoding, they are exposed to colored pictures of nameable objects, such as an apple, which they must remember.

Following each item’s presentation, a question is asked, promoting participants’ attention during this process.

Afterwards, in the second phase—fMRI testing—participants are placed inside a scanner and, via a projection system, are shown images: those previously observed along with brand new ones.

A fixation cross precedes each picture to optimize the separation of the brain’s hemodynamic responses across the different presentations.

Upon seeing each image, participants are asked to respond in one of three ways: ‘remember’ if the item can be recalled along with specific details about its presentation; ‘know’, if it’s familiar but they cannot recall specific details about seeing it before; or ‘new’, if the object was not seen at all.

In this case, the dependent variable is the intensity of the hemodynamic signal measured after each response type. The extent of activation can then be visualized into clusters of voxels on an anatomical brain scan.

The hippocampus—a region in the medial temporal lobe notably studied in learning and memory studies—is expected to show greater activation during the ‘remember’ trials than during the ‘know’ and ‘new’ trials.

These findings would support a dual-process theory of memory recall, where the hippocampus supports recollection and a different neural region—one outside of the hippocampus—generates familiarity.

For experimental control and safety concerns, recruit participants who are right-handed, with normal or corrected-to-normal vision, no history of psychological disorders or suffering from claustrophobia, and without any metal in their body.

Have them fill out a magnetic resonance screening form, with additional questions related to their health and safety encompassing the scanning session.

Before sending the participant into the scanner, sit in front of a laptop and expose them to objects that they need to remember for the next session. Explain that they will now view 256 color images, each for 3 s. To ensure that they are paying attention, instruct them to press the ‘F’ key to indicate that an object is living or ‘J’ if the item is non-living.

After the participant views all of the images, further explain that those pictures, along with an additional 256 novel items, will be shown inside the scanner. Also introduce them to the MR-safe button-box that they will use to classify items—as ‘remember’, ‘know’, or ‘novel’—when they appear onscreen.

In preparation to enter the scanning room, ask the participant to remove all metal objects from their body, including cell phones, watches or jewelry, wallets, keys, belts, and coins, due to the strong magnetic field. Use a metal detector to verify that no metal items remain.

Next, escort the participant near the scanner. Provide earplugs to protect their ears from loud noises and earphones so that they can hear you during the scan. Have them lie down on the bed with their head in the coil, and secure it with foam pads to avoid excessive movement and blurring during the scan.

Place a mirror above the participant’s eyes to reflect a screen at the back of the scanner bore. Make sure that they are equipped with a squeeze ball in case of an emergency during the scan and the button response box. Also, remind them that it’s very important to keep their head as still as possible throughout the experiment.

After raising the scanner bed, align the participant and send them into the bore. In the adjacent room, collect high-resolution anatomical images before starting the event-related, functional phase. Synchronize the start of the stimulus presentation with the start of the functional scan, and allow the participant to complete 512 trials.

To conclude the session, bring them out of the scanning room. Debrief them by providing an explanation of the study and compensation for their participation.

To begin the analysis, first pre-process the data by performing correction to reduce motion artifacts, temporal filtering to remove signal drifts, and spatial smoothing to increase the signal-to-noise ratio.

Then, create a model of the expected hemodynamic response for each task condition. Fit the data to this model, resulting in a statistical map for each subject, where the value at each voxel represents the extent to which that voxel was involved in the task condition.

Register the participant’s brain to a standard atlas to combine data across subjects. To perform a group-level analysis, threshold the statistical maps, taking into account correction for multiple comparisons. Only accept significant voxels if they also occur within a cluster of a given size to minimize false-positive results.

Using these extracted clusters, overlay them on an average anatomical brain. Note that the activation measured during the ‘know’ trials was subtracted from that in the ‘remember’ trials. The hippocampus, outlined here in yellow, showed significantly more activation for ‘remember’ trials compared to ‘know’ trials.

To examine hippocampal activation in more detail, plot the percentage of signal change across time after the onset of the stimulus.

Inspection of this time-course of activity revealed that the hippocampus responded positively when participants explicitly reported remembering the stimuli and when identifying new stimuli—noted here with a positive deflection.

In contrast, it responded negatively or very little when participants reported feelings of familiarity or did not remember images at all.

These results support a dual process theory of memory recall, where the hippocampus is involved with memory recollection but not familiarity.

Now that you are familiar with designing an fMRI experiment to understand brain activation during judgments of recollection and familiarity in typical adults, let’s look at additional studies that apply the Remember-Know paradigm.

If the hippocampus plays a central role in recollection, its absence might reveal dissociations in memory retrieval. This scenario can be addressed by comparing patients with bilateral hippocampal damage versus controls—individuals without any such damage.

Interestingly, patients with damage showed impaired memory recollection compared to controls, whereas both groups performed equally well during familiarity trials. Taken together, these results support a specific role of the hippocampus in recollection processes.

On the contrary, if individuals showed increased hippocampal volumes, we’d predict that they’d also display enhanced recollection.

One such example exists and involves London taxicab drivers, who were shown to augment their hippocampal gray matter after years of memorizing extensive and complex routes around the city. With their larger hippocampi and superb memory, they transport passengers to their correct destination in a timely manner.

Researchers are also interested in gaining further insight into the mechanisms responsible for memory retrieval in order to enhance it in other ways. Take for instance, a college psychology lecture, where large amounts of information are presented. Knowing that material is familiar is not helpful for an exam.

Instead, a student needs something else—beyond having that cup of coffee—to aid in remembering. Perhaps, taking a memory-enhancing compound would allow improved recall of the entire discussion to ace that important test.

You’ve just watched JoVE’s introduction to Remember-Know task. Now you should have a good understanding of how to design and conduct the memory recall experiment in conjunction with functional neuroimaging, how to analyze and interpret differential brain activation results, and finally how to apply the paradigm to real-life scenarios.

Thanks for watching!