Neurobiology of Loneliness, Isolation, and Loss: Integrating Human and Animal Perspectives

Abstract

In social species such as humans, non-human primates, and even many rodent species, social interaction and the maintenance of social bonds are necessary for mental and physical health and wellbeing. In humans, perceived isolation, or loneliness, is not only characterized by physical isolation from peers or loved ones, but also involves negative perceptions about social interactions and connectedness that reinforce the feelings of isolation and anxiety. As a complex behavioral state, it is no surprise that loneliness and isolation are associated with dysfunction within the ventral striatum and the limbic system - brain regions that regulate motivation and stress responsiveness, respectively. Accompanying these neural changes are physiological symptoms such as increased plasma and urinary cortisol levels and an increase in stress responsivity. Although studies using animal models are not perfectly analogous to the uniquely human state of loneliness, studies on the effects of social isolation in animals have observed similar physiological symptoms such as increased corticosterone, the rodent analog to human cortisol, and also display altered motivation, increased stress responsiveness, and dysregulation of the mesocortical dopamine and limbic systems. This review will discuss behavioral and neuropsychological components of loneliness in humans, social isolation in rodent models, and the neurochemical regulators of these behavioral phenotypes with a neuroanatomical focus on the corticostriatal and limbic systems. We will also discuss social loss as a unique form of social isolation, and the consequences of bond disruption on stress-related behavior and neurophysiology.

Keywords: CRH; dopamine; isolation; loneliness; loss; opioids; oxytocin.

FIGURE 1

Schematic of proposed insular cortex activity in lonely individuals. Schematic of proposed insular cortex activity in lonely individuals (modeled from Barrett and Simmons, 2015). Prediction neurons (green) are “inflexible” or sending strong prediction signals to dysgranular and granular layers. Simultaneously, social cues/stimuli with negative valence (that match the prediction) are attended to more strongly than those that would generate a prediction error. This would lead to insufficient activation of prediction-error neurons, and therefore no error signal being relayed to prediction neurons to update the simulation. Oxytocin is particularly well situated to modulate the activity of precision neurons and has been hypothesized to do so in a similar model for autism spectrum disorders (reviewed in Quattrocki and Friston, 2014). Adapted from Barrett and Simmons (2015), with permission from Springer Nature.

FIGURE 2

Human brain networks and brain regions within networks that have been implicated in loneliness. Colors represent the network that each region or group of regions belongs to (Note: this is not an extensive list of all known neural networks, only those that are most implicated in human loneliness). Dotted lines represent reduced resting state activity (for individual brain regions) or reduced functional connectivity (arrows between regions/networks) in lonely individuals compared to non-lonely individuals. Bolded lines represent increased resting state activity (for individual brain regions) or increased functional connectivity (arrows between regions/networks) in lonely individuals. Arrowheads touching network boxes denote altered connectivity between networks, while arrowheads touching brain regions denote altered connectivity between regions. Gray boxes represent regions receiving direct stimuli about the environment (i.e., exteroception, interoception, proprioception) and relay such information to other regions defined in these networks; however, information about their resting state activity or connectivity in lonely individuals is unknown at this time. For a recent review of the structural and functional alterations in brain regions and networks associated with loneliness, see Lam et al. (2021).

FIGURE 3

Rodent model depicting some of the known neurochemical circuits that regulate hypervigilance (green), social motivation (orange), passive coping (blue), and pain (purple), with particular focus on the four main neurochemical systems discussed in this review: oxytocin (OT/OTR), dopamine (DA/D1/D2), endogenous opioids (DOR/KOR/MOR), and corticotrophic releasing hormone (CRH/CRH-R1/2). 5-HT, serotonin; ACC, anterior cingulate cortex; ACTH, adrenocorticotrophic hormone; BLA, basolateral amygdala; BNST, bed nucleus of the stria terminalis; CeA, central amygdala; DRN, dorsal raphe nucleus; GR, glucocorticoid receptor; Hp, hippocampus; IC, insular cortex; MR, mineralocorticoid receptor; NAc, nucleus accumbens; PFC, prefrontal cortex; PVN, paraventricular nucleus of the hypothalamus; VTA, ventral tegmental area.