Tania Singer, a celebrated neuroscientist and director at the Max Planck Institute for Human Cognitive and Brain Sciences in Leipzig, Germany, is known as one of the world’s foremost experts on empathy. In her research, she has sought to demonstrate that meditation can make people more kind and caring. The title of a profile of Singer written by this reporter in 2013 summed up her public image: Concentrating on Kindness.

But inside her lab, it was a very different story, eight former and current colleagues say in interviews with Science. The researchers, all but one of whom insisted on remaining anonymous because they feared for their careers, describe a group gripped by fear of their boss. “Whenever anyone had a meeting with her there was at least an even chance they would come out in tears,” one colleague says.

Singer, one of the most high-profile female researchers in the Max Planck Society (MPG), sometimes made harsh comments to women who became pregnant, multiple lab members told Science. “People were terrified. They were really, really afraid of telling her about their pregnancies,” one former colleague says. “For her, having a baby was basically you being irresponsible and letting down the team,” says another, who became a mother while working in Singer’s department.

Singer, who declined to answer questions for this story, is on a 1-year sabbatical, but her colleagues say they are speaking out now because MPG plans to eventually allow her to return to her lab.

Singer has acknowledged making mistakes in the past. “Problems associated to my exhaustion due to having to carry and be responsible for [a] huge and complex study,” were partly to blame, she wrote in a 12 February 2017 email to representatives of the department who had complained about working conditions to the institute’s scientific advisory board. In a 7 August letter to Science, Singer’s lawyer denied allegations of bullying, however. The letter said Singer had “apologized deeply” during a 2017 mediation process and had taken responsibility for the problems, for instance by asking for the sabbatical and the appointment of a temporary replacement. It also suggested that the allegations against Singer came from a “subgroup with its own strong interests and group dynamics.”

In a statement also issued on 7 August, MPG acknowledged it had learned of the allegations against Singer last year and said the society’s vice president, Bill Hansson, had investigated them, but that details are confidential. MPG says that “to calm the situation down,” it agreed to Singer’s sabbatical, which took effect in January. In a plan presented to the researchers on 25 July, MPG said it would separate Singer from her current colleagues and allow her to set up a new, smaller research group in Berlin for 2 to 3 years while the postdocs and Ph.D. students in Leipzig finish their projects and move on. (The Leipzig group, which once numbered more than 20 scientists, has dwindled to just five.) She would then return to her lab.

“It appears the Max Planck Society decided it would rather sacrifice another generation of students than risk a scandal,” says one former colleague. Asked how MPG would ensure that future students are treated better, a spokesperson says details of the plan are still being discussed.

Singer, the daughter of celebrated neuroscientist Wolf Singer, helped found a new field called social neuroscience; she rose to prominence with her work on empathy, including a landmark study published in Science in 2004 that showed watching a loved one experience pain activates the same brain areas as feeling physical pain directly. In 2013, she started a hugely ambitious study, The ReSource Project, in which 160 participants were trained for 9 months to demonstrate the power of meditation.

The study highlighted some of Singer’s strengths, colleagues say. “She’s creative, she can be charming, she knows how to make contacts and get resources. It’s a gift and it was necessary to make a project like this happen,” one colleague says. “Her superpower is vision,” another adds. “That original team of people that she put together was totally incredible.” Several people in her lab recall an impassioned speech Singer gave after the 2015 terrorist attacks in Paris, in which she argued that only a study like ReSource could prevent such acts in the future.

But colleagues say working with Singer was always difficult. She wanted to be in control of even the most minute research details but was often not available to discuss them. In-person meetings could quickly turn into a nightmare, one colleague says: “She gets extremely emotional and when that turns dark it is terrifying.” Another co-worker describes what happened after he told Singer some people in her group were unhappy: “She was very hurt by this and started crying and screaming,” he says. “It escalated to the extent that she left the room and went door to door in the institute in our department, crying, yelling to the people in the room ‘Are you happy here?’ When she came back, she said: ‘I just asked and everyone said they’re happy so it’s obviously you that’s the problem.’” (A colleague who says he was present corroborates the story.)

Almost every current or former lab member brought up Singer’s treatment of pregnant women; the issue was also on a list of grievances, shared with Science, that lab members say they drew up after a meeting with the scientific advisory board in February 2017 to record what was said. “Pregnancy and parental leave are received badly and denied/turned into accusations,” the notes say.

Bethany Kok, a former lab member who agreed to speak on the record because she is no longer working in neuroscience, says Singer reacted kindly when she first told her she was pregnant with twins. But the next day, Kok says, “She started screaming at me how she wasn’t running a charity, how I was a slacker and that I was going to work twice as hard for the time I would be gone.” A few weeks later, Kok says, she miscarried one of the twins and missed a lab meeting for an urgent medical appointment. “I got an email from Tania telling me that she wasn’t paying me to go to the doctor, that clearly I wasn’t using good judgment, and I was no longer allowed to go to the doctor during work hours.” (Kok says she no longer has access to the email.)

Singer’s lawyer says Singer never discriminated against pregnant women or any other group, and that events described by others “either did not happen or they happened very differently than described.”

Scientific discussions could also get overheated, lab members say. “It was very difficult to tell her if the data did not support her hypothesis,” says one researcher who worked with Singer. Singer’s lawyer stresses that no rules of scientific conduct were violated and that the institute’s scientific advisory board had rated the work of Singer’s department as “excellent.”

Contemporaneous notes from the researchers who took part in the 2017 meeting with the scientific advisory board—held after years of informal discussions and talks with an ombudsman that had led nowhere—include numerous other complaints, including “emotional abuse, threats, devaluation of work, and personal abilities.”

“Working with Tania is becoming increasingly difficult,” the lab members wrote after the meeting. “We represent the entire department. We would like some change but are unsure of how to approach the issue due to fear of retaliation.”

“One problem surely is that I have clearly underestimated the challenges associated with our ReSource study,” Singer wrote in the email she sent lab members afterward. She noted that “the conflict between the project’s need for long-term continuity and loyalty on the one hand and the researchers’ own divergent needs to move on with their own careers and lives.”

Six meetings with a mediator in the first half of 2017 did little to improve the situation, several people who attended the sessions say. Afterward, Hansson launched his investigation, but in December 2017, scientists at the institute were informed that MPG President Martin Stratmann had taken the matter into his own hands. On 20 December, Singer wrote to the group that she would be on a sabbatical for all of 2018, with former MPG Director Wolfgang Prinz taking over some of her duties.

The separation plan presented 2 weeks ago, which would take effect in January 2019, has two goals, according to official minutes from the 25 July meeting, obtained by Science: “1. Unencumbered continuation of everyone’s work. 2. An opportunity for Tania Singer to have an unencumbered new start.” Singer “has learned the lesson that groups that are too big carry the risk of losing contact with co-workers,” her lawyer writes. “A good work environment and dealing with each other respectfully are important to her.”

Several researchers say they are disappointed in MPG. “I had hoped that they take problems at their institutes seriously and act not only on the behalf of their directors but equally their employees. However, every decision was always dragged out, communication was nontransparent and top-down, and then finally a solution was presented to the employees that is really mostly a solution for Tania,” one says.

Several other scandals have recently rocked the prestigious MPG, which has an €1.8 billion annual budget and runs 84 independent institutes. Nikos Logothetis, a researcher at the Max Planck Institute for Biological Cybernetics in Tübingen, Germany, was indicted in February after accusations from animal welfare activists; the U.S. Society for Neuroscience and the Federation of European Neuroscience Societies sharply criticized MPG last weekfor not defending him and his colleagues adequately. Also this year, allegations of bullying and sexual harassment at the Max Planck Institute for Astrophysics in Garching, Germany, emerged. Guinevere Kauffmann, the director accused of bullying, received coaching and daily monitoring and now leads a drastically reduced group. In an interview with German newspaper Frankfurter Allgemeine Zeitung, Stratmann said the incident had shown that the society’s procedures for dealing with complaints did not work well. “I have to concede that, and for this reason we will improve it.”

Many say the years spent in Singer’s lab have left them disillusioned about science. “If the Max Planck Society has as its objective to create scientists, then this Ph.D. experience is not the way to do it,” says one.