Niels Birbaumer, once a prominent neuroscientist at the University of Tübingen in Germany, fell hard from his pedestal three years back. He was accused of scientific misconduct for his controversial “brain-reading” research, stripped of funding, and fired from his job. His work, which showed that people paralyzed with ALS could communicate with a brain-computer interface, was retracted by the journal that published it, PLOS Biology.
Now, Birbaumer, who has fiercely defended his previous research, hopes to see his reputation restored.
On Tuesday, he published a new study in Nature Communications that builds upon his prior, discredited work. And Birbaumer and co-author Ujwal Chaudhary say they have won lawsuits that support the integrity of their PLOS report.
“Our study has proven that a person with locked-in syndrome is not a soulless body, but a body with will and desire. And the objection raised with our PLOS Biology paper will be proven right very soon,” Chaudhary said. “This is finally our redemption.”
Birbaumer agreed the research is about more than his reputation. “We didn’t do this for fun, or to advance technology, or for publishing papers,” he said. “We do these things because we want these people to be alive, even if society doesn’t want them to be.”
The pair’s new paper shows how a device implanted in the brain of a 34-year-old man with locked-in syndrome — on a ventilator, paralyzed, and unable to even move his eyes — could help him communicate in full sentences. Although brain-computer interface, or BCI, technology isn’t new, the case marks the first time it is known to have been used in a patient without any voluntary muscle control.
“This finding is important because it demonstrates for the first time that people in complete locked-in syndrome can not only follow instructions, but also willfully manipulate their brain signals to select choices presented to them and ultimately, to communicate,” said Femke Nijboer, biomedical signals and systems researcher at the University of Twente in the Netherlands. “The fact that he is present and seemingly cognitively intact is the big news. We also get a glimpse of his emotional state. … This guy seems to be doing OK and knows what he wants.”
Birbaumer’s decades-long neuroscience career led him to using brain-computer interface technology to read the thoughts of people who cannot otherwise speak. It’s a burgeoning field, but also controversial, with a litany of financial, ethical, and legal considerations. Debates rage on about the quality of life afforded to people who are conscious but immobile and on life support. Critics think such work can give both patients and families false hope. Proponents like Birbaumer say that patients can lead contented lives, even if they’re locked in.
In 2017, Birbaumer published headline-grabbing news in PLOS Biology indicating that he and his colleagues had enabled four people with paralysis caused by ALS to answer “yes” or “no” questions by placing a cap made of electrodes on their scalps. The paper generated a great deal of interest, particularly as the subjects reported that they had a high quality of life.
But a whistleblower at the University of Tübingen claimed that the study was not replicable. The DFG, which is Germany’s top research funding institution, investigated Birbaumer’s work and found the data incomplete and the results flawed; it levied serious sanctions on both Birbaumer and Chaudhary. The agency also recommended that PLOS retract two papers on the topic — saying that the scientists had a “special responsibility” toward this vulnerable patient population and failed them, “in particular by failing to document exactly the entire research procedure.” The situation became so fraught that Birbaumer moved out of the country, relocating to Italy.
Notably, dozens of scientists publicly supported Birbaumer and Chaudhary as signatories on a website that details and refutes the allegations against the pair.
“The episode with Niels’ PLOS 2017 paper is a very unfortunate one, which I don’t consider questions Niels’ scientific integrity,” said José del R. Millán, a professor of computer engineering and neurology at the University of Texas at Austin, who was not involved in Birbaumer’s research.
Millán said the Nature paper’s results don’t directly back up the PLOS Biology findings, given that the types of brain signals used were completely different, and the fact that the case involved someone who was not tracked in the prior study.
“But this new paper demonstrates Niels’ determination and rigor to find ways for [locked-in] patients after ALS to communicate by volitional modulation of brain signals,” Millán said. “He and his team are the first to have proven it.”
The particular implant used in this study was actually first shown to work in a patient with tetraplegia in 2006. The electrodes allowed the patient to move a cursor on a screen using a neural signal. Several studies have shown that this technology can be safe, can decode brain signals, and can drive external machines like robotic limbs.
But patients using these implants in the past had at least some sort of voluntary muscle control. The patient in the new research did not.
The 34-year-old man’s family had initially designed a schema to help him communicate — allowing him to use his eye movements to select letters off of a chart, ultimately spelling out sentences. As the neurodegeneration progressed, however, the man lost his ability to move his eyes — becoming completely locked in. The family, anticipating this, contacted Birbaumer in 2017, with hopes to test out a brain-computer interface to allow communication to continue, Chaudhary said.
“Since he’s lying in bed, unable to move, most people will think that he might not have a desire to live,” Chaudhary said. “But he’s the kind of person who wants all those things. His desire and will to live is very strong.”
The patient, who was not identified in the study, was able to provide consent to brain implantation using his eye movements, but it took six months to get through the ethical review process with BfArM, Germany’s regulatory agency for drugs and medical devices. By that time, the patient had zero voluntary muscle control. The team implanted the BCI into a portion of the patient’s brain that regulates motor function in March 2019.
It took several months to figure out how to help the patient communicate via the BCI. At first, the researchers asked the patient to try moving one body part to indicate “yes,” and another body part to say “no,” to see if that would light up certain brain regions and allow them to interpret the brain signals. It didn’t work.
“Then Birbaumer said we have to stop instructing the patient — and just say ‘use your own strategy to say yes and no,’” Chaudhry said.
The researchers used an auditory neurofeedback program, meaning that the patient was instructed to change the frequency of a sound tone with his mind. It’s still unclear exactly what the patient did with his brain, but at day 98 of the study, he was ultimately able to increase or decrease the frequency of the sound waves — making the computer play either fast or slow beeping noises.
“We don’t know how we did it. Nobody on earth can answer this question, not even the patient,” Birbaumer said. “Think of it this way: How do you think? Do you know? Nobody really knows how we have thoughts in our head, but we have them. It’s the same thing here.”
Researchers were able to translate these “yes” or “no” tones into sentences by using the alphabet schema the patient had been using before and had already memorized. It took about a minute to formulate each letter. The first thing he communicated was thanks to Birbaumer for the ability to speak again.
“He also instructed his wife to buy a mixer so he could have some goulash,” Chaudhary said. “He said he wanted beer, and wanted to watch TV shows with his son, who at that time was 3.”
The researchers meticulously documented their findings, capturing each moment on video. The study took place over 462 days. The patient is still alive.
“The methodology seems legitimate, careful, and detailed,” said Jane Huggins, director of the University of Michigan Direct Brain Interface Laboratory. “The messages typed show not only knowledge of spelling and semantics, but also of surroundings and social interactions.”
This case study is just a preliminary step in using these devices on a more broad scale, Huggins said.
“It is impossible to know yet what it will take to make it work for others,” she said. “But, the fact that it has worked for one person at such an advanced stage of neurodegenerative disease is an encouraging outcome for other people with neurodegenerative disease.”
Seward Rutkove, chair of neurology at Beth Israel Deaconess Medical Center in Boston, has no doubt that the technology works just as Birbaumer and Chaudhary describe — there have been many advances in BCIs in recent years. But he has misgivings about its practicality. It’s expensive, and reserved for the small slice of the population who wants to stay alive on a ventilator after complete neurodegeneration.
“I just think, as a real therapy or treatment in ALS, it’ll have a small market,” Rutkove said. “And I think where some of these things can get tricky are with the consent process — and patients knowing what they’re getting themselves into. Their quality of life isn’t going to be great.”
Birbaumer disagrees: “In general, the quality of life in complete paralysis is very good if you’re taken care of in a normal family environment,” he said.
Still, the subject is politically fraught.
“This work has political implications against euthanasia laws,” Birbaumer said. “In Europe, there’s a strong political force behind the liberalization of euthanasia laws, and so this type of research is very controversial.”
Chaudhary claims that he and Birbaumer were discredited not because of egregious scientific lapses, but because of these sorts of ethical considerations, as well as interpersonal politics.
“They were nitpicking,” Chaudhary said “In the end, it was more of the politics and personality of Niels which brought everything down.”
It’s been a struggle to right the ship; the peer-review process at Nature for their manuscript took about two years, Chaudhary said. Birbaumer has stopped working entirely, but is eager to clear his name.
“We were accused of falsifying the data, so I sued those people, particularly the university,” Birbaumer said. “We won the case, and that means that the organization who was responsible for this judgment has to lift the retraction.”
PLOS Biology did not respond to requests for comment. Chaudhary said there’s already been an agreement between the pair’s lawyers and the DFG; there’s been a settlement, and final resolution of the case is expected at the end of March.