The task was almost absurdly simple. Lie still in a magnetic resonance scanner at the Tel Aviv Sourasky Medical Center, pick one of four mental strategies you chose yourself before the run (recalling a trip, say), and try to push a number on a screen higher. The number, from 1 to 10, with a cartoon smiley beside it, was a readout of your own brain activity in real time, fed back to you from the reward circuitry deep in the midbrain. Do that for five trials, several runs, several days. Then, at the end of the last session, roll up your sleeve: a real hepatitis B vaccine, administered in the hospital. The question the researchers were asking was one that medicine has circled for a century without ever pinning down in people. Can the machinery of expectation reach past how you feel and into what your immune system actually builds?
First, a correction to the way this result has travelled. There was no placebo injection in this trial. Everyone who was vaccinated got the genuine hepatitis B vaccine. The placebo-adjacent part was upstream, in the mind: participants were trained with functional MRI neurofeedback to turn up activity in the reward mesolimbic system, and the strategies that worked best were ones built on positive expectation. Nitzan Lubianiker, Tamar Koren, Talma Hendler and colleagues ran it as a preregistered, double-blind randomized controlled trial (ClinicalTrials.gov NCT03951870), published in Nature Medicine on 19 January 2026. Eighty-five healthy adults aged 18 to 45, mean age about 25, 51 of them women, were randomized three ways: 34 to reward mesolimbic upregulation, 34 to upregulation of a non-mesolimbic control region, and 17 to no neurofeedback at all. Blood was drawn before vaccination and twice after, and 34 participants came back at three months.
What they found is narrower and more interesting than the headline version. Both neurofeedback groups did manage to raise reward mesolimbic activity, so the training worked. And across participants, the more a person increased activity specifically in the ventral tegmental area across runs, the larger their rise in hepatitis B antibody after the shot: a correlation of r = 0.31, P = 0.018. The nucleus accumbens did not show it. The control regions did not show it. Sustained ventral tegmental upregulation was linked to mental strategies that involved positive expectations rather than, say, pure relaxation. But here is the sentence that keeps this honest, and the authors put it in their own abstract: post-vaccination antibody levels did not differ between the groups. No adverse effects occurred. Being assigned to the reward arm did not, by itself, give anybody more antibody. What tracked with antibody was how well an individual actually engaged the circuit.
The idea did not come from nowhere, and it did not come from people. It came from mice. As the authors set out in their earlier preprint, posted to medRxiv in September 2024, rodent experiments had already shown that switching on dopamine neurons in the ventral tegmental area changes immune activity, with reported effects on recovery from bacterial infection and on lung tumour growth. That is animal work, done with chemogenetic and optogenetic tools that no human trial can replicate, and it tells you about mice. The Nature Medicine trial is the first attempt to ask whether something analogous happens when a person, consciously and voluntarily, drives the same region. One correlation in 85 young, healthy volunteers is not a therapy. It is a signal that the question is answerable.
For families reading this newspaper, the relevance is not that anybody should expect to think their way to a stronger immune response. It is that expectation is itself a brain function, with an anatomy, and brain diseases can damage it. In 2006, Fabrizio Benedetti and colleagues reported in the journal Pain that the expectation-related mechanisms underlying drug effects are disrupted in Alzheimer's disease, so that the same analgesic does less work in people whose prefrontal connectivity has been eroded. That finding, reported in a small group and never a reason to change anyone's treatment, reframes what the Tel Aviv trial is probing: not a trick of belief, but a physiological contribution that some people may have less of, through no fault of their own. It is also worth noting, as the authors disclose, that Hendler holds a role with GrayMatters Health and that several investigators have filed a patent covering neurofeedback and induction of an immune response. That does not make the data wrong. It is simply part of the record a careful reader is entitled to.