In January 2025, a team led by Chun Shen at the University of Cambridge published a study that would have seemed implausible a generation ago. Using blood samples from 42,062 participants in the UK Biobank, they identified 175 distinct proteins whose levels were altered in socially isolated individuals. Twenty-six of those proteins were specifically associated with loneliness. More than half were prospectively linked to cardiovascular disease, type 2 diabetes, stroke, and early death over a 14-year follow-up period. The study, published in Nature Human Behaviour, was not measuring feelings. It was measuring molecules.
The idea that loneliness is bad for health is no longer controversial. Julianne Holt-Lunstad's landmark 2015 meta-analysis of 70 studies encompassing 3.4 million participants, published in Perspectives on Psychological Science, established that loneliness increases the risk of premature death by 26%, social isolation by 29%, and living alone by 32%. These effect sizes rival those of well-established risk factors like obesity and physical inactivity. But knowing that loneliness can kill and understanding how it does so are very different things.
Over the past two decades, researchers in neuroscience, genomics, endocrinology, and immunology have been tracing the biological pathways through which perceived social isolation gets under the skin. What they have found is that your body treats loneliness not as a mood but as a physical emergency — activating ancient survival systems that, when sustained over months or years, slowly damage nearly every organ system.
The Brain Registers Rejection as Physical Pain
The most striking finding in the neuroscience of loneliness came from Naomi Eisenberger's laboratory at UCLA. Using functional magnetic resonance imaging (fMRI), Eisenberger and colleagues had participants play a simple computer game called Cyberball while lying in a brain scanner. In the game, a virtual ball is tossed between players. At a certain point, the other "players" (actually a computer program) stop throwing the ball to the participant, effectively excluding them.
The brain scans revealed something remarkable: social exclusion activated the dorsal anterior cingulate cortex (dACC) and the anterior insula — the same neural regions that process the distressing, unpleasant component of physical pain. As Eisenberger described in a 2012 review in Psychosomatic Medicine, "the neural bases of social pain overlap with the neural bases of physical pain." This was not metaphor. The brain was processing social rejection through the same circuitry it uses for a broken bone or a burn.
Subsequent work has added nuance to this finding. The overlap is not total — social and physical pain activate overlapping but not identical networks, and some researchers have debated the degree of shared versus distinct processing. But the core insight has been replicated across multiple studies and paradigms: the brain does not treat social disconnection as merely unpleasant. It treats it as threatening.
Loneliness Triggers the Same Craving Circuitry as Hunger
If social pain shares neural territory with physical pain, what about social need? In 2020, Livia Tomova and colleagues at MIT published a study in Nature Neuroscience that addressed this question directly. They recruited 40 participants and subjected each to two separate 10-hour deprivation conditions: one day of complete social isolation (no contact with any other person) and one day of food fasting.
After each deprivation, participants were placed in an fMRI scanner and shown images of social interactions or food. The results were striking: the substantia nigra pars compacta and ventral tegmental area (SN/VTA) — a midbrain region rich in dopaminergic neurons — responded selectively to food cues after fasting and to social cues after isolation. The magnitude of the craving response correlated with participants' self-reported longing. In other words, the brain's dopamine-driven reward system treated social contact as a need comparable to food.
This study was small (n=40) and examined acute isolation, not the chronic loneliness that drives health outcomes. But it offered biological evidence for what the late John Cacioppo, a pioneer of loneliness research at the University of Chicago, had long argued: that humans have a fundamental need for social connection that is rooted in evolutionary biology, and that the brain monitors this need the way it monitors other survival requirements.
The Hypervigilance Trap
Cacioppo's most influential contribution to the field may be his Evolutionary Theory of Loneliness, which explains not just why loneliness is painful but why it tends to perpetuate itself. The theory proposes that when an individual perceives themselves as socially isolated, their brain automatically shifts into a hypervigilant state — scanning the environment for social threats rather than social opportunities.
The evidence for this comes from both behavioral and neuroimaging studies. In electroencephalography (EEG) research, Cacioppo and colleagues found that lonely individuals' brains differentiated social threat images from non-social threats in approximately 116 milliseconds — less than half the time it took non-lonely individuals (~252 ms). This is faster than conscious awareness. The lonely brain is not choosing to be suspicious; it is wired for it.
The practical consequence is a self-reinforcing cycle. Heightened threat detection leads to withdrawal from social situations, which deepens isolation, which intensifies hypervigilance. Cacioppo described this as a "self-fulfilling prophecy" of loneliness — and it helps explain why simply telling lonely people to "put themselves out there" is neurobiologically naive. Their brains are literally processing social encounters differently.
It is worth noting that the hypervigilance response may be adaptive in the short term. For an early human separated from their group, heightened alertness to both social cues and threats could facilitate reconnection or protect against danger. The problem is that in modern life, perceived social isolation can persist for months or years — and a stress response designed for temporary emergencies becomes a chronic burden on the body.
How Loneliness Rewrites Your Immune System
Perhaps the most consequential biological discovery in loneliness research comes from the laboratory of Steve Cole at UCLA. Cole, a genomics researcher, has spent more than a decade studying what happens to gene expression in the white blood cells of lonely people.
What he and his colleagues found is a pattern they call the Conserved Transcriptional Response to Adversity (CTRA). In lonely individuals, the genes responsible for inflammation are upregulated — switched on more aggressively — while genes that mount antiviral responses are downregulated. The result is an immune system tilted toward chronic inflammation and away from the targeted defense against viruses.
A 2015 study by Cole and Cacioppo, published in the Proceedings of the National Academy of Sciences, examined this pattern in both humans and rhesus macaques — a primate species known for complex social behavior. In both species, loneliness was associated with higher levels of immature monocytes (a type of white blood cell that promotes inflammation) in the bloodstream. The CTRA pattern was not just a statistical association; it was bidirectional. Lonely gene expression predicted future loneliness measured more than a year later, and loneliness predicted future gene expression. The biological and psychological states were reinforcing each other.
The implications are significant. Chronic low-grade inflammation is now understood to be a driver of cardiovascular disease, type 2 diabetes, neurodegenerative conditions, and certain cancers. If loneliness systematically upregulates inflammatory gene expression, it provides a concrete molecular pathway from social disconnection to physical disease.
It is important to be precise about the evidence here. Cole's studies have typically used relatively small human samples, and the CTRA pattern has been studied more extensively in the context of general adversity (poverty, bereavement, trauma) than loneliness specifically. The macaque findings provide important replication across species, but the causal chain from gene expression to disease in lonely humans has not been demonstrated in a single longitudinal study. What exists is a biologically plausible mechanism, supported by converging evidence, rather than definitive proof.
From Proteins to Disease: The Cambridge Proteomics Study
This is where the 2025 Cambridge study becomes particularly important. While Cole's work examined gene expression in white blood cells, Shen and colleagues measured actual protein levels circulating in the bloodstream of 42,062 people. They used a proteome-wide association approach across 2,920 plasma proteins, making it the largest study of its kind.
Among the proteins linked to social isolation and loneliness, many were implicated in inflammation, antiviral responses, and complement systems — closely matching the CTRA pattern that Cole had identified at the gene expression level a decade earlier. Approximately 85% of the proteins associated with loneliness overlapped with those associated with social isolation, suggesting shared biological pathways.
Critically, the study went beyond cross-sectional association. Using Mendelian randomization — a statistical technique that uses genetic variants as natural experiments to test causal direction — the researchers found evidence suggesting causal relationships from loneliness to five specific proteins. Two of these, ADM (adrenomedullin) and ASGR1 (asialoglycoprotein receptor 1), are already known to be involved in cardiovascular regulation and liver metabolism.
More than half of the loneliness-associated proteins were prospectively linked to cardiovascular disease, type 2 diabetes, stroke, and mortality during the 14-year follow-up period. This begins to close the gap between "loneliness changes your biology" and "those biological changes lead to disease."
Stress Hormones and Sleep
The biological story extends beyond the immune system. Research on the hypothalamic-pituitary-adrenal (HPA) axis — the body's central stress-response system — has shown that loneliness is associated with dysregulated cortisol patterns. In a study published in Psychoneuroendocrinology, Cacioppo and colleagues found that lonely individuals showed flattened diurnal cortisol rhythms: less of the normal morning spike and less of the evening decline. A flattened cortisol slope has been associated in other research with poorer cognitive function, increased inflammation, and higher mortality risk.
The cortisol findings are not entirely consistent across studies. Some research in older adults has found lower overall cortisol in lonely individuals, while studies in younger adults have found higher levels. This may reflect different stages of HPA axis dysfunction — initial overactivation followed by eventual exhaustion — or it may reflect different measurement methods and populations. The field has not fully resolved this question.
Sleep provides another biological channel. A 2025 study published in Cogent Psychology using a nationally representative U.S. sample found that loneliness was significantly associated with insomnia severity. Notably, the relationship appears to work not primarily through total sleep duration but through sleep fragmentation — lonely individuals wake more frequently during the night. Research using ecological momentary assessment (real-time daily tracking) published in Sleep Health in 2024 confirmed a bidirectional relationship: lonelier days predicted worse sleep that night, and worse sleep predicted lonelier days the following day.
The sleep connection matters because sleep disruption is itself a driver of inflammation, impaired immune function, and cardiovascular risk. It may represent one of several amplifying loops through which loneliness compounds its own biological effects.
What This Evidence Does and Does Not Show
Taken together, the biological research on loneliness tells a coherent story: perceived social isolation activates neural alarm systems (social pain, craving, hypervigilance), which trigger stress-response pathways (HPA axis, sympathetic nervous system), which alter gene expression and protein levels in immune cells (CTRA pattern), which create a chronic inflammatory state that increases vulnerability to cardiovascular disease, metabolic disorders, and possibly neurodegeneration. A 2026 meta-analysis by Zheng and colleagues, published in the British Journal of Health Psychology and synthesizing 167 studies with 303,643 participants across 36 countries, confirmed that loneliness is associated with worse health even in healthy populations (r = −.22) — suggesting these biological pathways operate before clinical disease appears.
But intellectual honesty requires noting what remains uncertain:
- Most evidence is correlational. The Cambridge proteomics study's Mendelian randomization analysis is an exception, but the full causal chain from loneliness to brain activation to gene expression to disease has not been demonstrated in a single longitudinal study.
- Effect sizes are moderate. The Zheng meta-analysis found an overall effect of r = −.22, and over 92% of the variability in effect sizes reflected true differences across studies. Loneliness is one risk factor among many, not a deterministic sentence.
- Individual variation is large. Not everyone who is lonely develops inflammation or disease. Genetic predisposition, other health behaviors, and the duration and intensity of loneliness all likely moderate the relationship.
- The direction of causation may be partly reversed. People with chronic illness may become socially isolated because of their condition, creating a chicken-and-egg problem that even sophisticated statistical methods cannot fully resolve.
What the evidence does establish, with reasonable confidence, is that loneliness is not merely a subjective feeling with no physical consequences. It produces measurable changes in brain function, hormone regulation, gene expression, protein levels, and immune function — changes that map onto known pathways for disease.
Why This Matters Beyond the Laboratory
The biological research on loneliness carries an important implication: interventions that reduce loneliness may need to be understood as health interventions, not merely social programs. If loneliness dysregulates cortisol, inflames the immune system, and fragments sleep, then addressing it is not a luxury but a public health strategy — much as reducing smoking or improving diet is.
It also suggests that the quality of social contact matters as much as the quantity. The biological responses described here are triggered by perceived social isolation — the subjective feeling of disconnection, not simply being alone. A person can have a large social network and still feel lonely, triggering the same stress cascades. Conversely, even a small number of genuinely supportive relationships may be protective. As the research on interventions suggests, the most effective approaches address the cognitive distortions that sustain loneliness — the hypervigilance and social threat perception that Cacioppo identified — rather than simply increasing social contact.
The body, it turns out, is paying close attention to whether we feel connected. The science of the last two decades has made that attention visible — in brain scans, blood proteins, and gene transcripts. Whether that knowledge translates into better health depends on whether we take it as seriously as we take cholesterol levels and blood pressure readings.
See the Research in Action
YaraCircle is built on the friendship science we cover here.