Allen Braun's federal lab spent decades studying brain activity during stuttering, reciting, or listening to poetry (above), and normal language processing but he was fired last year, and the use of his data has been prohibited, sparking protests from Braun's collaborators.

Firing of veteran NIH scientist prompts protests over publication ban

At least two dozen junior and senior researchers are stuck in scientific limbo after being barred from publishing data collected over a 25-year period at a National Institutes of Health (NIH) lab. The unusual ban follows the firing last summer of veteran neurologist Allen Braun by the National Institute on Deafness and Other Communication Disorders (NIDCD) for what many scientists have told Science are relatively minor, if widespread, violations of his lab’s experimental protocol.

Most of the violations, which were unearthed after Braun himself reported a problem, involve the prescreening or vetting of volunteers for brain imaging scans and other experiments on language processing. The fallout from the case was recently chronicled on a blog by one of Braun’s former postdocs, and it highlights a not-uncommon problem across science: the career harm to innocent junior investigators following lab misconduct or accidental violations on the part of senior scientists. But this case, say those familiar with it, is extreme.

“We’re truly collateral damage,” says Nan Bernstein Ratner of the University of Maryland in College Park, who researches stuttering. She spent 5 years collaborating with Braun. Now, two of her graduate students have had to shift their master’s theses topics, and an undergraduate she mentored cannot publish a planned paper. “The process has been—you can use this term—surreal.”

Braun, who had been at NIH 32 years, including serving as chief of NIDCD’s Language Section since 1994, has filed a legal claim appealing the termination of his employment and would not comment. Andrew Griffith, the scientific director of NIDCD in Bethesda, Maryland, also declined to comment. “Unfortunately, the matter you reference is currently in litigation and therefore NIH is not in a position to provide any information or comment,” he emailed.

We’re truly collateral damage.

Nan Bernstein Ratner, University of Maryland

No one disputes that Braun’s lab failed to properly follow its experimental plan. An audit commissioned by NIH found, among other issues, that more than 200 healthy volunteers had not had a medical history and a physical signed off on by Braun, as required by the lab’s protocol. Nonetheless, affected researchers, senior scientists who are not impacted, and a patient advocacy group all oppose NIH’s decision to halt data publication. There is no evidence, they have argued in letters to NIH officials, that the violations compromised the bulk of the data or the safety of study volunteers.

They also question the breadth of the ban. Although the audit examined only volunteers newly enrolled since September 2009, NIDCD determined that the violations it found were sufficiently severe that all data from Braun’s studies back to 1992 were unusable. These ranged from functional MRI (fMRI) studies on people who stutter to how stroke victims regain speech to the neurological basis of more complex language functions. (Science found no evidence that NIH officials are seeking retractions for the dozens of published papers that used those same data sets.)

Braun’s lab began unraveling in 2015, when it included two postdocs and four other scientists. According to several sources, Braun learned that a new study volunteer had been incorrectly “coded.” The individual, who shared the same name and age as another volunteer, was misidentified as having already been part of the study, and was not screened as a new volunteer should be. Braun reported the incident to his institutional review board (IRB) and the protocol was suspended. NIDCD then commissioned an external audit of the lab.

The audit, which is dated February 2016 and which Science obtained, noted that Braun had not signed off on histories and physicals for 206 of the 424 volunteers whose records the audit examined. But the audit also noted that of those 206, all but five had received a history and physical elsewhere at the agency, because they were participating in other NIH studies, too.

The audit cited other documents missing from Braun’s lab, including screening questionnaires and, in the case of two female volunteers, records of a pregnancy test before they’d entered the scanner. (Pregnant women are typically excluded from MRI studies without special approval.) 

A month after the audit was completed, the IRB wrote to Braun that it “classified the [violations] as serious deviations and serious unanticipated problems.” But its memo went on: “The Board had insufficient information to determine whether the non-compliance invalidated the study data. The IRB strongly supports the use and publication of data whenever possible so that the contribution of participants is not negated.”

The IRB requested that Braun’s work remain suspended until the protocol violations could be resolved. But “if there is appropriate remediation,” it wrote to Braun, the work could continue. This would include re-education of all investigators and staff on the study, as well as, at minimum, better monitoring of the research, medical record forms that support “proper and complete documentation” of various study features, and perhaps additional safeguards.

A patient advocacy group whose members volunteered for Braun’s research lobbied NIDCD to free up those data. “I wish to express my opinion that Dr. Braun’s co-researchers be allowed to submit research findings for publication in a timely manner,” wrote Gerald Maguire, a psychiatrist at the University of California, Riverside, School of Medicine, in his role as chair of the Board of Directors of the National Stuttering Association, in a 27 March 2016 letter to Michael Gottesman, NIH deputy director for intramural research. “Stuttering suffers from a dearth of adequate research.”

The University of Maryland's Nan Bernstein Ratner collaborated with fired investigator Allen Braun of the National Institute on Deafness and Other Communications Disorders, and has been advocating for publication of data from his lab on behalf of junior scientists inside and outside NIH.

University of Maryland Photo Service

In June 2016, in the same week the IRB planned to discuss a remediation plan, Braun was fired and his lab was shuttered. Because his protocol was still suspended, research from the lab couldn’t be published until NIH administrators gave the green light. Affected researchers pleaded their case. “We believe there are viable ways that data validity can be assured to everyone’s satisfaction,” a group of 15 impacted researchers, most of them junior, wrote to NIDCD administrators. “We sincerely ask you to formulate a plan for data access and publication for all former trainees in Dr. Braun’s lab and his outside collaborators.”

NIDCD held firm. “We regret NIDCD cannot approve further publications,” wrote Carter Van Waes, the institute’s clinical director, to one of the signatories, Jed Meltzer of the University of Toronto in Canada. Meltzer was a postdoctoral fellow in Braun’s lab from 2006 to 2010 and continued to collaborate with him. “I have a graduate student who’s got a paper he can’t publish,” Meltzer says. “It’s like one-quarter of his Ph.D.” Another former Braun postdoc is on the job market and has approximately 10 papers that cannot be published.

The toll on junior scientists disturbs David Wright of Michigan State University in East Lansing, who from 2012 to 2014 ran the Office of Research Integrity at the Department of Health and Human Services. Although there are times when data cannot be salvaged, when a principal investigator is fired “the institute really ought to extend itself significantly to protect” the careers of investigators who did nothing wrong, Wright says. He cautions that he’s unfamiliar with details of the Braun case, but notes that supporting junior scientists is not only “a matter of fairness and decency, but also in the public interest.”

Scientists familiar with fMRI scans are also puzzled by the publication ban. “There’s nothing here that would flaw this data in any way,” says Joshua Shimony, a neuroradiologist at Washington University in St. Louis in Missouri. And he emphasizes that Braun’s lapses would not have jeopardized safety, as fMRI scans are extremely low risk. His view echoes those of seven senior NIH scientists who specialize in MRI research; they wrote a letter on Braun’s behalf before he was fired.

Especially frustrating to Ratner was that NIDCD declined to share whether the volunteers in her group’s stuttering research were among those flagged as violations. On a Friday afternoon earlier this month, she pleaded her case yet again. One week later, a senior NIDCD official replied. “We have extensively reviewed the matter and stand by our decision,” he wrote. “We sincerely regret the impact this has on anyone associated with this protocol.”

*Correction, 27 February, 10:22 a.m.: This article mistakenly indicated that the Office of Research Integrity is within NIH. It is within the Department of Health and Human Services, of which NIH is a part.