Expecting Someone Else's Pain Reshapes Empathy Itself, Brain Study Finds

An event-related-potential study of 26 healthy young adults finds that expecting another person's pain does not simply blunt or heighten one empathic response; it splits it into an early self-protective reflex and a later, smaller but still measurable empathic signal.

A study using scalp electroencephalography has found that expecting to see another person in pain changes the brain's response to that pain in two separate stages, not one, according to a paper published February 15, 2026 in the peer-reviewed journal Behavioral Sciences. Researchers led by Bingni Huang recorded event-related potentials, voltage changes tied to specific moments of processing, while 26 healthy university students (12 men, ages 18 to 27) viewed images of hands in painful or non-painful situations, each preceded by a cue signaling whether pain was coming. Behaviorally, participants responded faster and more accurately when a painful image had been signaled in advance, the paper reports.

At the neural level, the earlier signal, a component called N2 that peaks roughly a fifth of a second after a cue, was larger when pain was expected, a pattern the authors interpret as an early avoidance reaction tied to self-protection. Later, in a component called P3, two things happened at once: overall P3 amplitude was smaller under pain expectation, consistent with the brain allocating fewer cognitive resources once forewarned, but painful images still produced a bigger P3 than neutral ones even inside that expected condition. The authors call this an avoidance-approach dual-drive model: expectation appears to trigger a fast, protective pullback while a slower, weaker empathic response persists underneath it.

The design was a small, tightly controlled 2x2 within-subject experiment: 28 students were originally recruited and two were dropped for excessive signal artifact, leaving the 26 analyzed. That is standard practice for this kind of ERP work but it means the finding rests on a single sample of young, healthy adults performing a laboratory task with photographs, not a clinical population and not real-time caregiving. The paper does not test people caring for someone with dementia, older adults, or repeated real-world exposure to a person's distress, all of which are different from a one-time image-viewing task in a lab.

The result matters to this section because so much of what is written here about expectation and the brain concerns what a person predicts will happen to their own body: placebo pain relief, anticipated symptoms, expected decline. This paper applies the same predictive framework to watching someone else, and its central claim, an early avoidance signature layered under a later, reduced but real empathic one, offers language for something families and clinicians who witness distress day after day may recognize: bracing for another person's pain does not erase the empathic response, it appears to reshape its timing.

The authors describe their model as a first pass; whether the same two-stage pattern holds outside a 26-person student sample, in older adults, in people who witness pain repeatedly over months or years, or in the context of a specific dementia diagnosis, is a separate question the paper does not address and does not claim to answer.