Sleep loss leads to the withdrawal of human helping across individuals, groups, and large-scale societies

Abstract

Humans help each other. This fundamental feature of homo sapiens has been one of the most powerful forces sculpting the advent of modern civilizations. But what determines whether humans choose to help one another? Across 3 replicating studies, here, we demonstrate that sleep loss represents one previously unrecognized factor dictating whether humans choose to help each other, observed at 3 different scales (within individuals, across individuals, and across societies). First, at an individual level, 1 night of sleep loss triggers the withdrawal of help from one individual to another. Moreover, fMRI findings revealed that the withdrawal of human helping is associated with deactivation of key nodes within the social cognition brain network that facilitates prosociality. Second, at a group level, ecological night-to-night reductions in sleep across several nights predict corresponding next-day reductions in the choice to help others during day-to-day interactions. Third, at a large-scale national level, we demonstrate that 1 h of lost sleep opportunity, inflicted by the transition to Daylight Saving Time, reduces real-world altruistic helping through the act of donation giving, established through the analysis of over 3 million charitable donations. Therefore, inadequate sleep represents a significant influential force determining whether humans choose to help one another, observable across micro- and macroscopic levels of civilized interaction. The implications of this effect may be non-trivial when considering the essentiality of human helping in the maintenance of cooperative, civil society, combined with the reported decline in sufficient sleep in many first-world nations.

Fig 1. In-laboratory Study 1.

(A) One night of sleep deprivation was associated with a significant decrease in helping desire, relative to the sleep-rested condition, for both circumstances involving strangers (left) and familiar others (right). (B) Activity in the social cognition network (left, meta-analysis-based activation mask, corrected for multiple comparisons, P_FDR < 0.01) was significantly reduced following sleep deprivation, relative to the sleep-rested condition (right). (C) The relative reduction in social cognition brain network activity under conditions of sleep loss was significantly associated with lower helping behavior across individuals. *P < 0.05; error bars reflect standard error of the mean. Individual data presented in this figure can be found in S1 Data. IFG, inferior frontal gyrus; mPFC, medial prefrontal cortex; MTG, middle temporal gyrus; TPJ, temporal-parietal junction.

Fig 2. Micro-longitudinal Study 2.

(A) Study design. Participants were asked to complete daily sleep logs and helping behavior questionnaires across 4 days, allowing for the assessment of free living, natural variations in both sleep, and helping choices across the micro-longitudinal assessment period. (B) Reduced sleep quality from 1 night to the next was associated with a significant decrease in helping choices from 1 corresponding day to the next and vice versa (β = 0.02 ± 0.01, P < 0.05, see Results for the full model). Violin plots depict the change in next-day helping behavior between the maximal and minimal sleep quality nights for each participant across the study period. *P < 0.05. Individual data presented in this figure can be found in S2 Data.

Fig 3. Online donation behavior—Study 3.

(A) Overall distribution of donation amounts obtained from US states that observe DST, from 2001 to 2016. Light green inset highlights DST-observing states (i.e., excluding Arizona and Hawaii). (B) Donation amount was significantly lower in the week of DST transition (light blue) relevant to other weeks in the surrounding months (β_{DST week} = −0.11 ± 0.04, P < 0.005, adjusted for donation day, month, and year, see Methods). (C) The reduction in donation amount observed in the weeks around DST (top panel, centered around the third week of March) was not evident in the transition to ST (bottom panel, centered around the second week of November), suggesting that insufficient sleep triggered by the transition to DST uniquely impacts donation behavior. *P < 0.05, ***P < 0.005; error bars reflect standard error of the mean. US base layer map was plotted using the free and open-source Plotly library for python (https://plotly.com/python/maps/). Individual data presented in this figure can be found in S3 Data. DST, daylight saving time; ST, standard time; US, United States.