Targeted Memory Reactivation During Slow-Wave Sleep May Facilitate Facial Recognition

Recent research suggests targeted memory reactivation might help with facial recognition and name recall, but quality and quantity of sleep may be important factors in effectiveness.

During targeted memory reactivation(TMR), information is associated with a sound or smell as someone learns it. Then, the same sensory cue is presented as the subject sleeps, and without waking them. When subjects wake up, they are likely to remember the information associated with the cue better than other information that was equally learned, according to a new study.

The study, published in NPJ Science of Learning, found that TMR may help improve face memory.

Researchers recruited 24 participants (8 male, 16 female) with a mean age of 23 (range, 18-31) years from the Northwestern University population and surrounding area. Participants were tasked with memorizing the faces and names of 2 hypothetical classes: a Latin American history class of 40 students and a Japanese history class of 40 students.

First, participants were taught the names and faces in each class, and each class was associated with a distinct instrumental music track—either traditional Latin American music or traditional Japanese music.

Each student was presented to participants twice as an initial learning exercise, with their faces shown at quarter-right and quarter-left orientation. After each set of 5 students, participants did a recall activity in which they tried to type the corresponding first name as each 5 face appeared on a screen. They received feedback in the form of faces paired visually and audibly with the correct name.

After initial learning, participants were shown the same faces one at a time, and entering a correct name dropped that face from the training list. Finally, participants completed a visualization exercise before repeating the entire learning process for the next class.

Before and after napping, participants were tested on their recall. They were shown 230 faces sequentially, deciding whether each face belonged to Class 1, Class 2, or neither. The images included the 80 learned faces, 70 new faces, and the 80 learned faces rotated at a different angle. To test name recall, participants were tasked with typing a person’s first name when shown a picture of the corresponding face. Cued recall tests were completed for both classes.

Between waking up and recall testing, participants were given an interference task to learn 20 new faces and names with no recall step.

While napping, half of the participants heard cues for Class 1 and the other half heard cues for Class 2. Background music associated with the cued class was played at a low volume continuously as names were spoken as loudly as possible without causing participants to awaken.

Electroencephalogram (EEG), electrooculogram, and electromyogram signals also were recorded during the napping phase. Mean napping time was 59 (range, 32-92) minutes, and TMR commenced a mean 7.3 (range, range 1.6-21) minutes into the naps. TMR was started manually when signs of N3 sleep (low EEG slow waves) appeared and was paused when participants moved out of N3 sleep or woke up; it restarted once a participant entered N3 sleep again.

Overall, TMR had a varied effect depending on a participant’s N3 sleep duration. Before and after sleep, memory testing showed similar recognition scores. Participants recognized a mean 97% of old faces both pre- and postsleep testing. Old faces rotated from their original view were recognized 92% of the time before sleep and 91% after sleep. In addition, pre-sleep, 74% of new faces were correctly identified as new vs 85% post sleep.

The effect of cueing on recall was analyzed based on total, N2, N3, and REM sleep durations. However, the effect of cueing was only found to be significantly correlated with the N3 sleep stage—and was even more strongly associated with the relative duration of N3 sleep. The results for face recognition paralleled these findings.

In line with previous research, TMR’s effect on recall depended on undisturbed sleep. Sleep disruption was negatively correlated with cueing’s effect on recall, but not with face recognition.

“We already know that some sleep disorders like apnea can impair memory,” lead author Nathan Whitmore, a PhD candidate in the Interdepartmental Neuroscience Program at Northwestern, said in a statement. “Our research suggests a potential explanation for this: Frequent sleep interruptions at night might be degrading memory.”

Overall, the findings suggest presenting cues during sleep influenced facial recognition and name recall, reactivating recent memories when sleep was ample and undisturbed. The authors note that quality and quantity of sleep are factors of interest that are sometimes overlooked in literature on TMR.

This new line of research will let us address many interesting questions—like whether sleep disruption is always harmful or whether it could be used to weaken unwanted memories,” senior author Ken Paller, PhD, professor of psychology and director of the Cognitive Neuroscience Program at Weinberg College of Arts and Sciences at Northwestern, said. “At any rate, we are increasingly finding good reasons to value high-quality sleep.”

Reference

Whitmore NW, Bassard AM, Paller KA. Targeted memory reactivation of face-name learning depends on ample and undisturbed slow-wave sleep. NPJ Sci Learn. Published online January 12, 2022. doi:10.1038/s41539-021-00119-2