Static pay, shrinking prospects fuel neuroscience postdoc decline

Gerry Rubin loved his time as a Stanford University postdoctoral researcher. The two years were intense, he admits—he worked up to 80 hours a week in the lab, for pay that ended up being less than minimum wage. But he was 24, staying out of debt and could afford a car and a one-bedroom apartment (with a pool) in Palo Alto, California. And anyway, he was at Stanford, working down the hall from Nobel Prize winners, creating one of the first clone libraries. It felt like a dream.

In 1976, fresh off that postdoc position and still only 26, Rubin won a faculty position at Harvard Medical School, joining the nearly 72 percent of recent biomedical postdocs in the United States who were working in tenured or tenure-track positions that year. “No one felt exploited,” Rubin says, because “we felt it was a trade-off. We were learning new stuff, and then we were out the door.”

Things have changed a lot in 50 years. Rubin is now senior group leader at the Howard Hughes Medical Institute’s Janelia Research Campus, where his group studies neurobiology, and today postdoctoral positions often last three times as long as Rubin’s did. Also, trainees are coming to them later than he did—a typical postdoc now joins a lab in their 30s, a time when many hope to be starting families. The annual stipend from the U.S. National Institutes of Health (NIH) for a new postdoc was $10,000 in Rubin’s time. That’s about $64,000 in today’s money, which compares with the current base wage of $61,008. But these days, only about 26 percent of academic postdocs in the life sciences land faculty jobs.

The effects of rising living costs amid plummeting job prospects and stagnant salaries are fueling what some are calling a talent leak in basic neuroscience, and beyond. No longer viewed as an intense but intellectually rewarding training opportunity for future principal investigators (PIs), postdoctoral fellowships are now seen by many as cheap, long-term labor for established researchers. And a growing number of young scientists want out: the National Science Foundation does not break down postdocs beyond the categories of science and health, but those ranks dwindled from 55,748 to 54,415 in the U.S. between 2012 and 2022—about a 2 percent drop—after decades of growth. In the same time frame, the number of applicants for the NIH’s Ruth L. Kirschstein National Research Service Award postdoctoral fellowship dropped from 2,284 to 1,438—down by 37 percent. In 2023, the number of applicants dropped by another 17 percent to just 1,188.

The decline feels outsized in disciplines such as computational neuroscience, where tech companies investing in artificial intelligence have lured Ph.D.s away from academia with six-figure salaries, flexible work arrangements and generous 401Ks. “Computational neuroscientists now have opportunities that they didn’t have 5 or 10 years ago,” says Andrew Lo, Charles E. and Susan T. Harris Professor of Finance and Director of the Laboratory for Financial Engineering at the Massachusetts Institute of Technology (MIT) Sloan School of Management. “There is a porous membrane between academia and industry, particularly in the area of computational neuroscience.”

Kaela S. Singleton, who earned a Ph.D. in the interdisciplinary program in neuroscience at Georgetown University and is now director of grants management at the nonprofit organization Cure Alzheimer’s Fund, agrees that neuroscience is particularly vulnerable. “You have this highly skilled group of thinkers, leaders and problem solvers working through a system that hasn’t been updated in eons, and they’re outgrowing it,” says Singleton, who is also president of Black in Neuro. “It feels like other industries are recognizing that talent, building new opportunities or making folks offers they can’t refuse.”

It has been bad enough that, in late 2023, the NIH Advisory Committee to the Director Working Group on Re-Envisioning NIH-Supported Postdoctoral Training produced a report with recommendations for “improving the postdoctoral experience for both the postdoctoral scholar and the broader biomedical ecosystem.” In neuroscience circles, it seems nearly everyone agrees there is a problem; the question is whether academia is listening.

N

euroscience’s supply-and-demand problem was hinted at before even Rubin joined the postdoc ranks. In 1969, a study conducted by the National Research Council nicknamed postdoctoral fellowships the “invisible university”—an underrecognized and transient workforce whose numbers were growing. In 1976, MIT physicist Lee Grodzins predicted that the relative dearth of faculty positions would lead some strong Ph.D.s to ditch the postdoc path in favor of industry jobs.

Between 1979 and 2022, the number of graduate students in science and health fields in the U.S. more than doubled, from 285,770 to 622,534, and the number of postdocs nearly tripled, from 17,034 to 54,415. The number of junior faculty jobs in the life sciences also grew in the time period, but only from 11,300 in 1981 to 22,400 in 2021. This has left many postdocs in a kind of limbo, languishing for an extra year (or two, or three, or four), hoping to publish the paper that makes them a more competitive applicant for the few available faculty positions.

In this system, PIs get to retain, for years, an experienced researcher who can run experiments and publish but also write grants and mentor graduate students—all for about three-quarters of the salary of a staff scientist. This has been advantageous as science has become more labor intensive: Papers published before the mid-1970s had two authors, on average. Today, the average is six.

The tension between these forces is spilling out into the open. Social media is rife with postdocs describing their “exploitation,” from expectations of long hours to threats of lapsed visas for international postdocs who refuse to take on extra work. Postdocs have also become more forthcoming with their PIs. “If you contrast how people talked about this when I was starting my postdoc to now, it’s crazy,” says Jakob Voigts, a group leader at Janelia studying how animals learn from limited data. When Voigts was in training, senior peers told him to avoid mentioning industry, conceal if he had a romantic partner and otherwise hide aspects of his life that might make him appear “not serious about science,” thereby hurting his chances at a PI position. Now it’s “normal” for his postdocs to tell him “they might want to abandon the whole thing and go to industry,” he says. This meshes with findings of a study published last week, which found that about 41 percent of postdocs leave academia.

The COVID-19 pandemic seems to have made things worse. Social distancing made it harder to run experiments, share ideas with colleagues and get guidance from PIs. Half of postdocs surveyed in July 2020 said their level of job satisfaction had worsened that year, and nearly two-thirds said their career prospects had worsened, too. On top of that, the expectations around the postdoc experience had previously been passed down through a sort of cultural osmosis in the lab. But the pandemic and social distancing blocked that transmission, Voigts says, and “the whole expectation got completely reset.”

With in-person lab meetings and journal clubs on hiatus, discussions moved to social media, where postdocs used hashtags and threads on X (then Twitter) to commiserate about long hours and weekends without overtime pay. The pandemic “reframed a lot of people’s priorities,” says Ubadah Sabbagh, chief of staff at Arcadia Science and a former postdoc at MIT’s McGovern Institute for Brain Research. It forced people to pause and reflect on “how you’re investing your energy, your emotions, your thoughts in the world.”

The unhappiness has other causes. A 2021 survey by Columbia University’s postdoctoral workers union found that more than two-thirds of its postdocs said they had experienced power-based harassment, such as belittling, been given unreasonable workloads or had not received proper credit for their work.

Adding to this dissatisfaction is a feeling that some elements of the postdoc system affect certain groups more than others. The length of time that trainees spend in a postdoctoral position today makes it difficult to start a family, for instance, and this disproportionately affects women. Women also earn less than men, on average, and, according to the Columbia report, are more likely than men to experience harassment. Rates of reported harassment are also higher among other underrepresented groups. And a 2024 study found that international postdocs work more, earn less and have fewer benefits than their domestic peers. For underrepresented groups, “the system works in a way that they are being pushed out,” says Mayank Chugh, visiting assistant professor of biology at the College of William & Mary.

A

t the annual Society for Neuroscience meeting in October, Chugh gave a keynote address titled “Re-envisioning Postdoctoral Experience: A Game of Difficult Choices?” during the ALBA-International Brain Research Organization social event. Chugh asked postdocs in the audience, which also comprised graduate students and PIs, to describe challenges in their academic life on Post-it notes and stick them to poster boards around the room. Some focused on consistent expectations of overtime, poor job prospects and power-based harassment, but most of the Post-it notes dealt with inadequate compensation, Chugh says.

The minimum wage for postdocs has failed to keep up with inflation, and it also varies little between expensive cities and more affordable ones. Nanthia Suthana is associate professor of psychiatry, neurosurgery, bioengineering and psychology at the University of California, Los Angeles. She and her team use brain implants to stimulate and record brain activity as people engage in learning and memory tasks. She needs postdocs with basic neuroscience knowledge and engineering chops—people who would be strong candidates for industry jobs. Knowing that, and recognizing the high cost of living in Los Angeles, she pays her postdocs between $70,000 and $85,000 a year, depending on their years of experience, by dipping into her operating grants to top up their salaries. “How can I pay the minimum to a postdoc in Los Angeles? That’s just wrong,” she says.

In its 2023 report, the NIH advisory committee recommended boosting the base salary for postdocs to $70,000—an increase of about 15 percent—with an annual adjustment for inflation. But the agency responded by increasing the base wage by just 8 percent, stating that it would strive to reach the $70,000 target within five years (about $400 shy of what the current salary would be with inflation).

Singleton says she feels funding agencies have been “passing the buck” when it comes to paying postdocs a living wage. “They always put the burden on the next generation of scientists to push for it. And that ‘pushing’ takes effort and time away from the work that they do, and it comes unpaid,” she says.

Efforts by postdocs to unionize have yielded salary gains. For instance, postdocs at the University of Washington earn a minimum of $68,460 a year, and those at the University of California make upward of $66,737. In December 2024, a union representing postdocs at NIH facilities secured a new contract with a base salary of $68,544. The contract also includes the right to adhere to a 40-hour workweek, with some flexibility for projects occasionally requiring more work.

“We didn’t win everything that we wanted,” says Marjorie Levinstein, a neuropharmacology postdoc at the National Institute on Drug Abuse and a member of the bargaining committee, noting that they had hoped for higher pay, better support for international researchers and a child-care subsidy. But “we are happy with what we ended up with.” The contract includes paid leave for caregivers and stipulates that at least 10 percent of work time can be allocated to professional development and training.

Being transparent with trainees about job prospects in academia and alternatives to scarce faculty positions is key, Lo says. “We need to give students and postdocs a clearly defined career path that will allow them to not only further their own research and have impact but also raise a family if they choose to and be able to move on with other parts of their lives.”

V

oigts opened his lab at Janelia in 2022. He uses cameras and implanted brain electrodes to monitor mice and their brain activity around the clock. He needs computational neuroscientists in his lab, but most of his early postdoc applicants had a pure biology background. People with “a quantitative math background were conspicuously absent from the applicant pool,” he says.

Voigts suspects those missing computational applicants took up industry jobs. There were a record 751 rounds of venture capital funding into biopharmaceutical therapeutic and platform companies globally in 2021, and the year also saw a record amount of venture funding ($72 billion) into artificial-intelligence companies. In 2022, more than half of new Ph.D.s  in the U.S. in neuroscience-related fields such as biomedical and computational science accepted jobs in industry. Those workers earned a median salary of at least $110,000, roughly double the median salary of a postdoc at the time. With that kind of wage difference, it’s not surprising some new neuroscience Ph.D.s look to industry. “I think the workers are voting with their feet,” says Rubin, whose five most recent Ph.D. graduates went straight into industry. “And I would say they are wealthier and happier than my postdocs who went off and got good jobs as assistant professors at good research institutes.”

But the amount of available financing for startups is receding. The number of biopharma venture capital rounds was estimated to have fallen below 250 for last year, and by 2023, global AI funding had dipped to $40 billion. Voigts says that as the tech market has softened, he has seen an increase in postdoc applicants with a computational background, whom he needs to mine huge amounts of brain activity data. “It’s very clear when you look around now that people aren’t as optimistic about industry jobs anymore,” he says.

Funding at universities can go through similar cycles, and academia carries its own perks. Industry jobs can lack the intellectual freedom of academia—the flexibility to follow leads, says Letitia Weigand, director of the National Institute of Neurological Disorders and Stroke (NINDS) Office of Training and Workforce Development. Startups don’t offer great job security, either. Weigand would know: She spent three years at a biotech startup, after a postdoctoral fellowship and a longer stint in government, and she saw three corporate reorganizations. “You know what happens in a reorg in corporate America, right? You get on a call Monday morning at 9:30, and they say, ‘Everyone on this call has a role in this new structure.’ And you look at the Zoom screen, and your friends are gone,” she says. “I think sometimes people aren’t aware of those trade-offs.”

Changing career paths

A growing percentage of Ph.D.s are opting for jobs in industry instead of academia.

Creating neuroscience postdoc positions that bridge academia and industry could give trainees experience in both sectors, and the NIH advisory committee did recommend stronger ties to industry. Martin Reese, CEO of the biotech company Fabric Genomics, is all for it. “Then [postdocs] can decide what they want to do,” he says. Reese specializes in neural networks and designed gene-finding algorithms for the Human Genome Project in the 1990s. He started his first biotech company from his garage during his Ph.D. and sold it for about $70 million three months after graduating. Reese says he would happily match funding from the NIH to bring skilled industry postdocs on board at Fabric, his fourth company, which specializes in genomic medicine. At the very least, he says, the NIH should value industry experience and consulting work among grant applicants. Right now, they’re “purely looking at your academic career, and if you say that you would advise a company, it doesn’t count at all, which is stupid.”

Weigand acknowledges that funding structures for postdocs “were developed many, many years ago, and the guidelines around how to use them were for a different time, when people were being trained in different ways and had different aspirations.”

One solution might be more staff scientists. The NIH advisory committee suggested capping postdoc positions at a total of five years and increasing funding and support for staff scientists—researchers who aren’t interested in being PIs, but who, the advisory committee argued, “improve lab sustainability and efficiency by maintaining institutional knowledge and providing mentorship and support to others in the lab, thus reducing burden on principal investigators.” Staff scientists make more money than postdocs—typically upward of $80,000—and have better benefits and job security, though they have fewer opportunities for advancement.

There is already growth in this area. Although the number of postdocs in science and health in the U.S. has declined by about 2 percent over the past decade or so, the number of staff scientist jobs has grown from 19,411 to 27,924—up by about 43 percent.

With all this change, “maybe it’s time to take a step back and look at what the world needs,” Weigand says. “What the community needs and what science needs.”

W

hen Rubin was in his postdoc position, he had the luxury of supervisors who offered feedback and advice, and who willingly made calls and created opportunities for him. PIs often play a pivotal role in nurturing the strengths and aspirations of their trainees, but their mentoring skills can run the gamut, and regardless, Rubin says, some people in neuroscience “don’t have mentors who can do that for them.” That seemed particularly true during the pandemic, when more than a quarter of postdocs surveyed reported feeling unsupported by their PIs.

Toward that end, in 2017 NINDS launched the Landis Mentor Award, a $100,000 prize for NINDS-funded PIs who excel in training and mentorship. “We value people that really care about setting people up to be successful and aren’t just viewing them as a cog in the wheel,” Weigand says. She and her team at NINDS plan to collect data to assess the award’s impact on mentors and mentees. Meanwhile, recipients are using the money to support existing trainees and welcome new ones.

“We’re not only training the next generation of scientists; we’re sort of building a network of colleagues,” says Hongjun Song, Perelman Professor of Neuroscience at the University of Pennsylvania Perelman School of Medicine. Song received the Landis Award in 2022 and has used the money to support career development for a postdoc about to enter the job market as well as a new postdoc from an underrepresented group. He has also funded summer research stints for undergraduate and high school students. Song’s son worked in the lab when he was in high school and now works as a computer scientist for a pharmaceutical company, where he has monthly meetings with a mentor. They’re even more “goal-driven” in industry, Song says. “They want their new people to be successful so they can retain them.”

Among its other recommendations, the NIH advisory committee proposed taking steps to ensure that postdocs who experience harassment can change labs “without retaliation and with support to mitigate damage and delay to their well-being and careers.” Chugh says it’s not clear how effective a policy change might be, given that many decide to “just suck it up for two or three more years” and keep silent, but the NIH has asked for that kind of community feedback on the committee’s recommendations. In a statement, the agency “wholeheartedly” agreed with the need for change and said it’s particularly interested in how postdoctoral fellowships can be restructured to focus “more on ideas and creativity, and less on productivity.”

Yet Sabbagh says it’s paradoxical that the NIH advisory committee’s charge was to stem the decline in postdocs, yet one of its main recommendations was to increase pay. Because if pay raises are implemented with a flat NIH budget, that would actually decrease the number of postdocs. For instance, the NIH postdoc budget currently supports about 30,000 postdocs; if the budget stays the same, the pay hike would eliminate just shy of 4,000 NIH-funded postdocs.

“It’s a limited pot of money,” says Sabbagh, a member of the NIH advisory committee’s working group on postdocs. “If we have to increase compensation, something else has to give.”

Yet a smaller postdoc workforce may actually solve multiple problems. Large research groups are more expensive to run than smaller ones, but not necessarily better. Labs with direct costs exceeding $700,000 had a similar number of publications and a similar impact factor across their publications as labs with lower budgets, a 2010 analysis found. And the number and impact of publications from a lab does not track with the lab’s size, according to a 2015 study. There are diminishing returns, so to speak. Smaller labs with strong mentors at the helm seem to be the best bet. Janelia, for instance, has relatively small labs (two to six lab members plus a PI) but strong core facilities with skilled staff scientists. This structure has been great for postdocs, staff scientists and lab productivity, Voigts says. It’s also good for PIs: Howard Hughes Medical Investigator PIs, including all lab heads at Janelia, are assessed for reappointment every five to seven years, based not only on how many papers they publish but also the trajectories of their trainees. The quality of a PI’s mentorship, Rubin says, is weighted significantly.

Having smaller labs would also help lessen the number of researchers bottlenecked in postdoc positions, hoping for an academic position to open up. And it should help overworked PIs spend more time with the postdocs they do have.

Sabbagh would welcome that change. He thinks having fewer postdocs overall, while better compensating them for their expertise, would “absolutely” yield better science, and even improve productivity. Still, if he is wrong about that—if having fewer postdocs decreases productivity—then the poorer scientific output will be deserved, he adds.

“Then that’s the price we pay for underfunding science,” he says.