Decreasing sleep requirement with increasing numbers of neurons as a driver for bigger brains and bodies in mammalian evolution

Abstract

Mammals sleep between 3 and 20 h d(-1), but what regulates daily sleep requirement is unknown. While mammalian evolution has been characterized by a tendency towards larger bodies and brains, sustaining larger bodies and brains requires increasing hours of feeding per day, which is incompatible with a large sleep requirement. Mammalian evolution, therefore, must involve mechanisms that tie increasing body and brain size to decreasing sleep requirements. Here I show that daily sleep requirement decreases across mammalian species and in rat postnatal development with a decreasing ratio between cortical neuronal density and surface area, which presumably causes sleep-inducing metabolites to accumulate more slowly in the parenchyma. Because addition of neurons to the non-primate cortex in mammalian evolution decreases this ratio, I propose that increasing numbers of cortical neurons led to decreased sleep requirement in evolution that allowed for more hours of feeding and increased body mass, which would then facilitate further increases in numbers of brain neurons through a larger caloric intake per hour. Coupling of increasing numbers of neurons to decreasing sleep requirement and increasing hours of feeding thus may have not only allowed but also driven the trend of increasing brain and body mass in mammalian evolution.

Keywords: brain size; evolution; metabolite clearance; number of neurons; sleep.

Figure 1.

Daily sleep requirement is predicted to correlate with the ratio between neuronal density and surface area (D/A) in the cerebral cortex. The three images illustrate blocks of cortical tissue of similar pial surface area but different neuronal densities (either in different species or in different developmental stages of a same species). Because the production of metabolites per neuron per unit time is presumed to be similar across them (see text), the concentration of sleep-inducing metabolites produced per time and unit area (shading in each block) should be proportional to the density of neurons. Thus, when both the average neuronal density in the tissue and the total pial surface area vary across species (not shown), the rate of increase of the concentration of sleep-inducing metabolites should accompany D/A, the ratio between neuronal density and surface area in the cerebral cortex. The bottom part of the figure shows how a smaller D/A would then allow for longer times spent in waking until a critical concentration of sleep-inducing metabolites is reached, and thus lead to a smaller daily sleep requirement.

Figure 2.

Neuronal density mm⁻² (D/A) is the parameter that best correlates with total sleep hours per day across mammalian species. Each data point corresponds to one species, as in table 1. Values shown indicate the Spearman correlation coefficient for each graph. All values of p < 0.005 except for the two correlations in the bottom row, where p > 0.1. Red, primates; orange, eulipotyphlans; green, glires; blue, afrotherians; pink, artiodactyls; grey, scandentia. (Online version in colour.)

Figure 3.

Total daily sleep requirement decreases with increasing number of neurons in the cerebral cortex across non-primate species through a decrease in D/A (neuronal density mm⁻²). (a) Total daily sleep requirement scales as a power function of neuronal density mm⁻² with an exponent of 0.133 ± 0.023 across all 24 mammalian species in the dataset (r² = 0.601, p < 0.0001). (b) Neuronal density mm⁻² decreases steeply with increasing number of cortical neurons as a power function of exponent −1.694 ± 0.080 across non-primates (r² = 0.971, p < 0.0001) and less steeply across primates (exponent −1.233 ± 0.144, r² = 0.924, p < 0.0001). (c) Total daily sleep requirement decreases as a power function of the number of cortical neurons across non-primates (exponent, −0.266 ± 0.034, r² = 0.809, p < 0.0001), but not across primates (p = 0.2597). Red, primates; orange, eulipotyphlans; green, glires; blue, afrotherians; pink, artiodactyls; grey, scandentia. (Online version in colour.)

Figure 4.

Decrease in total sleep time in early postnatal development in the rat is associated with a decrease in neuronal density mm⁻². Graphs show the ontogenetic variation in different parameters pertaining to cortical morphology and cellular composition from birth (day 0) onwards. The shaded bar indicates the period of rapid decrease in % time spent asleep and in neuronal density mm⁻² (D/A). O/N, other cells/neurons ratio, which approximates the maximal glia/neuron ratio. Data available in table 2.