A quick nap with deep N2 sleep could unlock hidden insights, as scientists find brainwave patterns during sleep best predict those breakthrough moments.
Study: N2 sleep promotes the occurrence of ‘aha’ moments in a perceptual insight task. Image Credit: Hlib Shabashnyi / Shutterstock
In a recent study published in the journal PLOS Biology, researchers found a role for sleep stage 2 (N2) and aperiodic neural activity in insight.
An insight, or “aha moment,” is a unique learning phenomenon that attracts significant research interest. The neural and cognitive mechanisms underlying insight have been described as restructuring existing task representations. Insight is often characterized by three key features: 1) an abrupt, nonlinear increase in task performance, 2) a variable, spontaneous delay before insight occurs, and 3) a selective occurrence in some individuals.
Sleep is a potential factor that facilitates the occurrence of insight. One study suggested that a full night’s sleep is beneficial for insight and found that more than twice as many participants had insight into a hidden rule of a task after sleep. However, other studies found no beneficial effect of sleep on insight.
These divergent findings may be due to specific sleep stages differentially influencing insight. For example, an influential 2021 study by Lacaux and colleagues suggested that N1 sleep, the lightest stage of sleep, was a "creative sweet spot" for gaining insight.
Moreover, distinct cognitive tasks benefit differently from sleep. The current study specifically employed a perceptual insight task (the Perceptual Spontaneous Strategy Switch Task, PSSST), in contrast to the mathematical tasks used in some previous research, which may influence how sleep stages affect insight.
The study and findings
In the present study, researchers investigated the links between insight and different sleep stages in a preregistered conceptual replication of the work by Lacaux et al. They recruited individuals aged 18-35 with a normal sleep-wake cycle and without color blindness, learning difficulty, or sleep disorders. Participants performed a perceptual insight task before and after a brief nap.
Participants were presented with a stimulus of purple or orange dots that moved in one of the four orthogonal directions. Dot motion exhibited varying noise levels across trials, making motion judgment relatively easier or more complicated in different trials. The primary task consisted of nine blocks, each comprising 100 trials, during which participants were required to press a button in response to the displayed stimulus and observe the resulting feedback.
Only stimulus motion correlated with correct response in the first three blocks. After 3.5 blocks, stimulus color also began to predict the correct response. Participants were allowed a 20-minute nap after four blocks, during which sleep and brain activity were monitored using electroencephalography (EEG). Subsequently, participants completed the remaining five blocks.
The subtle change in task structure from the middle of the fourth block provided a hidden opportunity to improve strategy through insight. Further, insight was tracked by monitoring rapid increases in performance on high-noise trials. Performance on high-noise trials was stable before the task structure was modified. A sudden shift to high accuracy on high-noise trials could indicate (gaining) insight about the color-based strategy.
Notably, 15 subjects gained insight before the nap and were excluded from analysis. EEG data quality precluded sleep classification for seven cases. As such, 68 subjects were included in post-nap data analysis. Of these, 48 participants (70.6%) had an abrupt, nonlinear improvement in performance post-nap and were stratified as insight participants.
Insight subjects had significantly higher mean accuracy across trial types and lower reaction times by the first half of block 8. The authors suggest that this indicates the nap period improved insight, as the 70.6% insight rate is substantially higher than the 49.5% rate observed in a previous study by the same group, which used a similar task but had no delay period. However, the authors also note the important caveat that the study did "not present a randomized manipulation of sleep, awake rest, and no rest."
Further, participants were stratified into three groups according to their vigilance state during rest: no sleep, sleep stage 1 (N1), or N2 sleep. Eighteen individuals were awake during rest, 22 reached N1 sleep, and 28 reached N2 sleep.
About 85.7% of the N2 sleep group gained insight, while 63.6% of the N1 sleep group gained insight. In contrast, only 55.5% of those who were awake gained insight. Contrary to the previous findings, the study aimed to replicate the results; however, it found no evidence that N1 sleep promotes insight. Instead, the researchers explored whether N2 sleep was the primary driver of insight. Fisher’s exact test indicated significantly higher insight prevalence following N2 sleep than without sleep. Results did not change when subjects with short N2 sleep episodes were excluded.
Notably, while the prevalence of insight was increased with N2 sleep, there was no impact on insight features (abruptness, delay, and selectivity). This means that N2 sleep increased the frequency of insight but did not change how abruptly, selectively, or with what delay insight occurred in those who experienced it. While the primary hypotheses about sleep stages were preregistered, further exploratory analyses examined the associations between aperiodic neural activity and insight, as prior studies suggested that regularization and noise facilitate abrupt, sudden performance changes (characterizing insight) and have been associated with aperiodic activity.
As such, aperiodic neural activity during the entire nap period was quantified by the spectral slope of the EEG power spectrum. The spectral slope was associated with stages of sleep and was the flattest in the no-sleep group and steepest in the N2 sleep group. Importantly, statistical model comparisons revealed that the spectral slope alone was the best standalone predictor of insight, subsuming the effect of sleep stage itself.
Finally, the team found that the spectral slope alone best predicts insight, rather than sleep stage or the combination of spectral slope and sleep stage. However, analyses on oscillatory activity showed no correlations with insight.
Conclusions
In summary, the findings suggest a beneficial effect of N2 sleep on the likelihood of insight. The spectral slope of the EEG power spectrum explained additional variance in the likelihood of insight beyond sleep stages and was the strongest individual predictor in statistical models. Thus, aperiodic neural activity but not oscillatory activity was the additional factor promoting insight. Since the spectral slope becomes steeper with sleep and predicts insight beyond sleep stages alone, deeper sleep may be necessary for insight.
The authors speculate that processes occurring in deeper sleep, such as synaptic downscaling (a form of neural regularization), may help create a "clean slate" that allows participants to more easily discover the hidden, optimal strategy after the nap. Future studies may evaluate the effect of a full night of sleep and different non-perceptual or non-mathematical tasks.