Form and Function of Sleep Spindles across the Lifespan

Abstract

Since the advent of EEG recordings, sleep spindles have been identified as hallmarks of non-REM sleep. Despite a broad general understanding of mechanisms of spindle generation gleaned from animal studies, the mechanisms underlying certain features of spindles in the human brain, such as "global" versus "local" spindles, are largely unknown. Neither the topography nor the morphology of sleep spindles remains constant throughout the lifespan. It is likely that changes in spindle phenomenology during development and aging are the result of dramatic changes in brain structure and function. Across various developmental windows, spindle activity is correlated with general cognitive aptitude, learning, and memory; however, these correlations vary in strength, and even direction, depending on age and metrics used. Understanding these differences across the lifespan should further clarify how these oscillations are generated and their function under a variety of circumstances. We discuss these issues, and their translational implications for human cognitive function. Because sleep spindles are similarly affected in disorders of neurodevelopment (such as schizophrenia) and during aging (such as neurodegenerative conditions), both types of disorders may benefit from therapies based on a better understanding of spindle function.

Figure 1

Spindles generation within the CT-TRN-TC circuit. (a) Sleep spindles are generated through reciprocal interactions of the TRN (green) and TC (red) neurons. Spindles can be initiated by excitatory input from CT (blue) neurons, which also synchronize spindle oscillations across the thalamocortical network. Representative neural firing patterns (right) show the order and organization of firing for CT, TRN, and TC neurons, respectively (adapted from [9]). As TC neurons fire, they provide input to both cortical and TRN neurons, and these populations phase-lock their firing. As cortical, TRN, and TC neurons desynchronize their firing, the spindle oscillation wanes. (b) The CT-TRN-TC circuit may also generate local spindles following waking experience. For example, training on a motor task leads to increased sleep spindles over the contralateral motor cortex during subsequent sleep (left; adapted from [10]). There are several possible mechanisms by which changes to CT-TRN-TC circuitry during waking could cause subsequent local spindle increases (right). TC projections may show changes to dynamics (e.g., firing rate or bursting) based on prior waking experience (1). Experience may alter TRN excitability or bursting during subsequent sleep (2). CT feedback may be altered by experience to increase synchronization or amplification of spindles (3). Finally, intracortical plasticity could amplify spindles locally (4).

Figure 2

Changes in spindle form across the lifespan. (a) Heat maps depict topographical spindle density during early development, adolescence, and aging. Spindles are initially seen over central areas of the brain and gradually develop over frontotemporal areas during the first year of life [11]. During adolescence, density reaches a relative maximum with equal distribution across frontal, central, and parietal leads. In aging, there is a return to the same pattern seen earlier in development, with highest density at central leads [12]. Below the heat maps are representations of spindle morphology at these ages. (b) Sleep spindle density (blue trace) increases throughout early development, peaking during puberty and steadily declining from adolescence to old age [13, 14]. Duration of sleep spindles (orange trace), on the other hand, peaks early in life, and then generally declines over the lifespan. Spindle amplitude (purple trace) is relatively small early in development, increasing to maximum values over the first year of life, and then steadily declines until old age [15]. (c) Neurodevelopmental milestones.