Loading wiki pages...

Wiki Version:
<p>Introduction: A strong body of evidence suggests newly formed memories become reactivated during non-rapid eye movement (NREM), leading to their stabilization. In the EEG, reactivation seems to positively relate to slow oscillations (&lt; 1 Hz) and sleep spindles [short bursts (0.5-2 s) between 11-16 Hz]. Interestingly, numerous studies demonstrate reactivation can be elicited by presenting learning-related stimuli during post-learning sleep, a process termed targeted memory reactivation (TMR). In two previous studies, we found sleep spindles occurring after TMR cues positively predict memory. Surprisingly, we also found spindles occurring within short periods before TMR cues negatively predict memory, perhaps because of the refractory period present between successive spindles. Here, we will attempt to manipulate the time between spindle onsets and TMR cues by tracking spindles in real-time. Half the cues will be presented 0.25 s after spindle offset (early condition), whereas the other half will be presented 3.5 s after spindle onset (late condition). We predict better memory in the late condition relative to the early condition. Finally, note that in a previous real-time study, we found strong indications for a refractory period effect on memory, but behavioral and neural measures indicate we waited too long in the late condition (6 s), suggesting we may have fallen out of the optimal spindle-TMR cue delay. </p> <p>Methods: Task - Subjects first over-learn arbitrary associations between environmental sounds and picture items of famous faces and places. Next, they learn the locations of those items against a background spatial grid, wherein they learn to a criterion of placing each object within a short distance of its correct location (150 pixels) once. They take a test on each item once before an afternoon nap. During the nap, upon online indications of slow wave sleep (SWS), we will begin the real-time algorithm to play the early and late sounds in a random order at appopriate times with respect to spindles. After the nap, subjects will return after a 2.5 hr delay to take a final spatial test without sounds and a sound-picture test. The primary task of interest is the spatial test, whereas we expect fairly high (near ceiling) performance of the sound-picture test. </p> <p>Delivering TMR cues - We aim to cue within SWS due to its well-documented benefits for declarative memories and its high threshold for awakening due to stimuli. However, as stage-2 may confer similar benefits, if over 45 mins have passed and we have not reached our minimum threshold of cues, we will deliver cues in stage-2 sleep. We followed a similar strategy in previous studies. </p> <p>Inclusion criterion - In most previous studies, we require subjects to hear at least one round of cues. However, waiting for spindles to introduce cues makes it far more difficult to get through this minimum number of sounds. Therefore, in this study, we will exclude subjects only if they have not heard at least half of the sounds. For those subjects hearing fewer than one whole round of sounds, we will run analyses on only the cued sounds in the early and late condition. We aim to collect N=20 subjects meeting this criterion.</p> <p>Oscillatory power - As in previous studies, we will use the root-mean square (RMS) of the artifact-rejected EEG signal filtered within specific oscillatory bands. For this calculation, we calculate RMS power at each time point with a +/- 200 ms moving window. Next, we average across all cues within a particular condition for each subject and calculate statistics using within-subject contrasts. </p> <p>Planned analyses and hypotheses: Behavioral - our primary contrast involves the amount of forgetting across the nap (post-nap - pre-nap memory) for information paired with early and late sounds. In previous studies, we found pre-nap memory significantly negatively predicts forgetting, presumably due to ceiling and regression to the mean effects. Therefore, here we will adjust the raw number of pixels to account for this relationship by calculating the residual between the actual and expected amount of forgetting for each item. We predict better memory retention using this measure in the late condition than the early condition. </p> <p>Neural - In previous studies, we found sigma power (11-16 Hz) at the CPz electrode approximately 1000-1500 ms post-cue positively predicted memory and approximately -1500-1000 ms pre-cue negatively predicted memory. Therefore, we expect these measures to again have these predictive relationships over memory, although it is an open question whether they will be as predictive when spindle power variability is more controlled with our real-time method. </p> <p>Notes about timing of this registration and previous versions: This posting is placed after 4 subjects have been collected, which we wanted to use to ensure the validity of the current real-time procedure after a few changes to the previous approach. However, nothing has changed in our hypotheses about this project since its inception during the late spring of 2016 (see the poster from the Cognitive Neuroscience Society conference in April, 2017 and Margaret Wang's undergraduate thesis published in May, 2017).</p>
OSF does not support the use of Internet Explorer. For optimal performance, please switch to another browser.
This website relies on cookies to help provide a better user experience. By clicking Accept or continuing to use the site, you agree. For more information, see our Privacy Policy and information on cookie use.

Start managing your projects on the OSF today.

Free and easy to use, the Open Science Framework supports the entire research lifecycle: planning, execution, reporting, archiving, and discovery.