Oxytocin in the Spinal Cord Explains Why Massage Reduces Pain and Improves Wellbeing

Massage has long been used to reduce pain and improve well-being, but these two outcomes are often treated as separate: one “physical,” one “psychological.”

Results from a new study argues they may be tightly linked through a single mechanism—oxytocin acting on spinal cord circuits. In other words, part of massage’s therapeutic effect may begin at the first central processing station for touch and pain: the spinal dorsal horn, not only in the brain.

The study was conducted by a large, highly interdisciplinary international team, bringing together expertise that rarely sits in a single laboratory. The authors span neuroscience, pain medicine, social neuroscience, spinal cord physiology, genetics, and clinical research.

The authors start from a simple observation across social mammals: pleasant social touch reliably improves mood and can reduce pain. Oxytocin is a strong candidate mediator because it rises during bonding and supportive touch, and it also increases during distress and pain. That dual activation has created confusion—why would a “bonding hormone” also turn on during aversive states? Their proposal is that oxytocin is a state-dependent modulator: whether the trigger is comforting touch or pain, oxytocin shifts sensory processing toward outcomes that promote recovery and social support.

In the human arm of the study, participants received a controlled, light-pressure, skin-to-skin back massage versus a control condition. After massage, participants reported better mood, showed higher salivary oxytocin, and had physiological signs consistent with reduced stress (including lower cortisol and heart rate). Importantly, everyone rated the massage as pleasant—so the intervention reliably produced a positive affective state.

The study then tests whether this oxytocin-linked state change alters how people experience an unpleasant stimulus. Participants received mild electrical stimulation (uncomfortable but not painful). After massage, people rated the same stimulation as less unpleasant. The key clinical implication is that massage may not only change “pain intensity,” but also reduce the aversive emotional tone attached to an unpleasant bodily signal.

To understand where this shift might be generated, the team used mouse experiments that allow direct circuit mapping. They show that gentle, social-touch–like stimulation activates oxytocin neurons in the hypothalamus. Interestingly, noxious stimuli (like painful heat) and neuropathic pain states also activate these oxytocin neurons. That supports the idea that oxytocin is recruited by “salient body states” across the valence spectrum—pleasant and unpleasant.

Crucially, they identify a descending pathway: hypothalamic oxytocin neurons project to the superficial dorsal horn of the spinal cord, where oxytocin receptors are highly expressed. In both mice and humans, oxytocin receptor–expressing neurons are abundant in dorsal horn regions that integrate pain and touch inputs. This positioning makes the spinal cord a plausible site where internal state (via oxytocin) can reshape incoming sensory information before it ever reaches higher brain centres.

What does oxytocin do to spinal processing? The paper’s central claim is that it rebalances excitation and inhibition in dorsal horn circuits. Oxytocin receptors are found on both excitatory and inhibitory interneurons, but oxytocin tends to boost inhibitory responses and dampen certain excitatory outputs—especially those driven by C-fiber inputs that can carry nociceptive and affective touch signals. Functionally, that pattern looks like a “gain control” system that can quiet pain-linked pathways while allowing—or even amplifying—touch-reward pathways.

To connect this circuitry to meaningful outcomes, the authors test pain relief in neuropathic models. When they experimentally activate spinal oxytocin receptor circuits, mice show reductions in both sensory hypersensitivity (how strongly they react) and affective-motivational pain (how aversive the state is), assessed with behavioural paradigms that separate “pain unpleasantness” from reflexive withdrawal. This matters for therapists because it aligns with real-world patient reports: “It still hurts, but it bothers me less,” or “I can cope with it better.”

One of the most striking findings is that spinal oxytocin signalling appears necessary for the rewarding properties of gentle touch. When oxytocin receptors were selectively reduced in the relevant spinal segments, gentle brushing no longer produced a preference for the “touch-paired” environment—yet basic touch detection remained intact. That suggests the spinal cord is not just a passive cable transmitting touch; it may contribute to whether touch feels comforting and socially rewarding, which could be especially relevant in persistent pain.

The authors also propose that oxytocin doesn’t act uniformly; it biases specific output pathways. In simplified terms, spinal pathways associated with rewarding touch are facilitated, while pathways associated with avoidance and aversion are suppressed. This bidirectional tuning could explain the “two-for-one” clinical effect: massage can simultaneously make social touch feel better and make pain feel less threatening.

Finally, they bridge back to humans using spinal and cortical somatosensory evoked potentials. After massage, the timing of spinal sensory signals changes, and the magnitude of oxytocin increase correlates with this spinal-level modulation. Combined with reduced unpleasantness ratings, this supports the translational idea that oxytocin-linked changes in early spinal processing are measurable in people—not just inferred from animal work.

For therapists, the practical takeaway is not “oxytocin explains everything,” but that it offers a plausible mechanism for why touch-based interventions can shift both pain experience and emotional state. It also reinforces the idea that massage may work best when it provides not only mechanical input, but a salient, safe, socially meaningful sensory context—because the nervous system may be using that context to decide how to weight incoming signals.