Frequent lucid dreaming associated with increased functional connectivity between frontopolar cortex and temporoparietal association areas

Abstract

Humans typically lack awareness that they are dreaming while dreaming. However, at times a remarkable exception occurs and reflective consciousness can be regained while dreaming, referred to as lucid dreaming. While most individuals experience lucid dreams rarely there is substantial variance in lucid dream frequency. The neurobiological basis of lucid dreaming is unknown, but evidence points to involvement of anterior prefrontal cortex (aPFC) and parietal cortex. This study evaluated the neuroanatomical/neurofunctional correlates of frequent lucid dreams and specifically whether functional connectivity of aPFC is associated with frequent lucid dreams. We analyzed structural and functional magnetic resonance imaging from an exceptional sample of fourteen individuals who reported ≥3 lucid dreams/week and a control group matched on age, gender and dream recall that reported ≤1 lucid dream/year. Compared to controls, the frequent lucid dream group showed significantly increased resting-state functional connectivity between left aPFC and bilateral angular gyrus, bilateral middle temporal gyrus and right inferior frontal gyrus, and higher node degree and strength in left aPFC. In contrast, no significant differences in brain structure were observed. Our results suggest that frequent lucid dreaming is associated with increased functional connectivity between aPFC and temporoparietal association areas, regions normally deactivated during sleep.

Figure 1

Seed-based resting-state functional connectivity differences between frequent lucid dream and control groups. Top panel: (a) Seed region of left aPFC with significant differences between groups. To estimate connectivity, spherical ROIs of 6 mm radius were defined in aPFC based on the peak voxel reported in Dresler et al. which had increased fMRI BOLD signal response during signal-verified lucid REM sleep dreaming. (b) The frequent lucid dream group showed increased resting-state functional connectivity between left aPFC and the bilateral angular gyrus (AG), bilateral middle temporal gyrus (MTG) and right inferior frontal gyrus (IFG). All clusters are significant at p < 0.05, corrected for multiple comparisons at the cluster level. Middle panel: Volume slices illustrating bilateral MTG and IFG results. Bottom panel: Volume slices illustrating bilateral AG results.

Figure 2

Clusters in lateral parietal cortex showing increased resting-state functional connectivity with aPFC in the frequent lucid dream group overlaid with cytoarchitectonic subdivisions of IPL/IPS. The angular gyrus can be subdivided into anterior (PGa; blue outline) and posterior (PGp; white outline) subdivisions based on cytoarchitecture. IPS can be divided into three subdivisions (hlP1 on the posterior lateral bank- yellow outline, hlP2 which is anterior to hIP1- purple outline, and hlP3 which is posterior and medial to both subdivisions- green outline). The cluster peak as well as maximal cluster extent localized bilaterally to a dorsal segment of the anterior angular gyrus (PGa). Region-of-interest (ROI) analysis revealed increased connectivity between left aPFC and bilateral PGa (blue outline) [all p < 0.05].

Figure 3

Whole-brain graph-theoretic network differences between frequent lucid dream and control groups. (a) aPFC node (red sphere) with significantly higher degree (k) and strength (s) in the frequent lucid dream group from axial (top panel) and left sagittal (bottom panel) views. (b) Left panel: Mean node degree (top row) and strength (bottom row) over density (cost factor) thresholds 0.05 ≤ δ ≤ 0.35 (step size 0.01) for frequent lucid dream (blue triangles) and control groups (red circles) for significant node shown in panel a. Shaded regions show 95% confidence intervals for each δ. Right panel: boxplots of area under the curve (AUC) for frequent lucid dream and control groups. The bottoms and tops of the boxes show the 25th and 75th percentiles (the lower and upper quartiles), respectively; the inner white band shows the median; and the whiskers show the most extreme data points not considered outliers (outliers are plotted separately with red squares). Asterisks indicate significant differences (p < 0.05) between conditions with a nonparametric bootstrap test after correcting for multiple comparisons against a surrogate max t distribution across all nodes.