Elon Musk recently that Blindsight, a cortical implant to restore vision, would have low resolution at first “but may ultimately exceed normal human vision.”

That pronouncement is unrealistic at best, according to new research from the 91̽��.

, lead author and 91̽��professor of psychology, said Musk’s projection for the latest project rests on the flawed premise that implanting millions of tiny electrodes into the visual cortex, the region of the brain that processes information received from the eye, will result in high-resolution vision.

For the study, , the researchers created a computational model that simulates the experience of a wide range of human cortical studies, including an extremely high-resolution implant like Blindsight. One simulation shows that a movie of a cat at a resolution of 45,000 pixels is crystal-clear, but a movie simulating the experience of a patient with 45,000 electrodes implanted in the visual cortex would perceive the cat as blurry and barely recognizable.

That’s because a single electrode doesn’t represent a pixel, Fine said, but instead stimulates, at best, a single neuron.

On a computer screen, pixels are tiny ‘dots.’ But that’s not the case in the visual cortex. Instead, each neuron tells the brain about images within a small region of space called the “receptive field,” and the receptive fields of neurons overlap. This means that a single spot of light stimulates a complex pool of neurons. Image sharpness is determined not by the size or number of individual electrodes, but the way information is represented by thousands of neurons in the brain.

“Engineers often think of electrodes as producing pixels,” Fine said, “but that is simply not how biology works. We hope that our simulations based on a simple model of the visual system can give insight into how these implants are going to perform. These simulations are very different from the intuition an engineer might have if they are thinking in terms of a pixels on a computer screen.”

The researchers’ approach was to use a wide range of animal and human data to generate computational “virtual patients” that show, for the first time, how human electrical stimulation in the visual cortex might be experienced. Even blurry vision would be a life-changing breakthrough for many people, Fine said, but these simulations — which represent the likely best-case scenario for visual implants — suggest that caution is appropriate.

While Fine said Musk is making important strides in the engineering challenge of visual implants, a big obstacle remains: Once the electrodes are implanted and stimulating single cells, you still need to recreate a neural code — a complex pattern of firing over many thousands of cells — that creates good vision.

“Even to get to typical human vision, you would not only have to align an electrode to each cell in the visual cortex, but you’d also have to stimulate it with the appropriate code,” Fine said. “That is incredibly complicated because each individual cell has its own code. You can’t stimulate 44,000 cells in a blind person and say, ‘Draw what you see when I stimulate this cell.’ It would literally take years to map out every single cell.”

So far, Fine said scientists have no idea of how to find the correct neural code in a blind individual.

“Somebody might one day have a conceptual breakthrough that gives us that Rosetta Stone,” Fine said. “It’s also possible that there can be some plasticity where people can learn to make better use of an incorrect code. But my own research and that of others shows that there’s currently no evidence that people have massive abilities to adapt to an incorrect code.”

Without that sort of development, the vision provided by Blindsight and similar projects will remain fuzzy and imperfect — no matter how sophisticated the electronic technology.

For now, the models developed in the study could be used by researchers and companies to aid in the placement of existing devices and the development of new technology, among other benefits. Entities like the Food and Drug Administration and Medicare could also gain insight into what sort of tests are important when evaluating devices. Further, the models provide realistic expectations for surgeons, patients and their families.

“Many people become blind late in life,” Fine said. “When you’re 70 years old, learning the new skills required to thrive as a blind individual is very difficult. There are high rates of depression. There can be desperation to regain sight. Blindness doesn’t make people vulnerable, but becoming blind late in life can make some people vulnerable. So, when Elon Musk says things like, ‘This is going to better than human vision,’ that is a dangerous thing to say.”

, 91̽��professor of psychology, was a co-author. The research was funded by the National Institutes of Health.

For more information, contact Fine at ionefine@uw.edu.

Research has shown that people who are born blind or become blind early in life often have a more nuanced sense of hearing, especially when it comes to musical abilities and tracking moving objects in space (imagine crossing a busy road using sound alone).  For decades scientists have wondered what changes in the brain might underlie these enhanced auditory abilities.

Now, a pair of research papers published the week of April 22 from the 91̽�� — in the , the other in the — use   to identify two differences in the brains of blind individuals that might be responsible for their abilities to make better use of auditory information.

“There’s this idea that blind people are good at auditory tasks, because they have to make their way in the world without visual information. We wanted to explore how this happens in the brain,” said , a 91̽��professor of psychology and the senior author on both studies.

Instead of simply looking to see which parts of the brain were most active while listening, both studies examined the sensitivity of the brain to subtle differences in auditory frequency.

“We weren’t measuring how rapidly neurons fire, but rather how accurately populations of neurons represent information about sound,” said , a graduate student in the 91̽��Department of Psychology and lead author on the Journal of Neuroscience paper.

That found that in the auditory cortex, individuals who are blind showed narrower neural “tuning” than sighted subjects in discerning small differences in sound frequency.

“This is the first study to show that blindness results in plasticity in the auditory cortex. This is important because this is an area of the brain that receives very similar auditory information in blind and sighted individuals,” Fine said. “But in blind individuals, more information needs to be extracted from sound — and this region seems to develop enhanced capacities as a result.

“This provides an elegant example of how the development of abilities within infant brains is influenced by the environment they grow up in.”

The second study examined how the brains of people who are born blind or become blind early in life — referred to as “early blind” individuals — represent moving objects in space. The research team showed that an area of the brain called the hMT+ — which in sighted individuals is responsible for tracking moving visual objects — shows neural responses that reflect both the motion and the frequency of auditory signals in blind individuals. This suggests that in blind people, area hMT+ is recruited to play an analogous role — tracking moving auditory objects, such as cars, or the footsteps of the people around them.

The paper in the Journal of Neuroscience involved two teams — one at the UW, the other at the University of Oxford in the United Kingdom. Both teams measured neural responses in study participants while participants listened to a sequence of Morse code-like tones that differed in frequency while the fMRI machine recorded brain activity.  The research teams found that in the blind participants, the auditory cortex more accurately represented the frequency of each sound.

“Our study shows that the brains of blind individuals are better able to represent frequencies,” Chang said. “For a sighted person, having an accurate representation of sound isn’t as important because they have sight to help them recognize objects, while blind individuals only have auditory information. This gives us an idea of what changes in the brain explain why blind people are better at picking out and identifying sounds in the environment.”

The Proceedings of the National Academy of Sciences study examined how the brain’s “recruitment” of the hMT+ region might help blind people track the motion of objects using sound. Participants once again listened to tones that differed in auditory frequency, but this time the tones sounded like they were moving. As has been found in previous studies, in blind individuals the neural responses in area hMT+ contained information about the direction of motion of the sounds, whereas in the sighted participants these sounds did not produce significant neural activity.

By using sounds that varied in frequency, the researchers could show that in blind individuals, the hMT+ region was selective for the frequency as well as the motion of sounds, supporting the idea that this region might help blind individuals track moving objects in space.

“These results suggest that early blindness results in visual areas being recruited to solve auditory tasks in a relatively sophisticated way,” Fine said.

This study also included two sight-recovery subjects — individuals who had been blind from infancy until adulthood, when sight was restored via surgery in adulthood. In these individuals, area hMT+ seemed to serve a dual purpose, capable of processing both auditory and visual motion. The inclusion of people who used to be visually impaired lends additional evidence to the idea that this plasticity in the brain happens early in development, Fine said, because the results show that their brains made the shift to auditory processing as a result of their early-life blindness, yet maintains these abilities even after sight was restored in adulthood.

According to Fine, this research extends current knowledge about how the brain develops because the team was not only looking at which regions of the brain are altered as a result of blindness, but also examining precisely what sort of changes — specifically, sensitivity to frequency —might explain how early blind people make sense of the world. As one of the study participants described it, “You see with your eyes, I see with my ears.”

Both studies were funded by the National Eye Institute and the National Institutes of Health. The Proceedings of the National Academy of Sciences study was co-authored by of the 91̽��and Fang Jiang of the University of Nevada, Reno. The Journal of Neuroscience study was co-authored by Chang and Huber, as well as Ivan Alvarez, Aaron Hundle and Holly Bridge of the University of Oxford.

###

For more information, contact Fine at ionefine@uw.edu or 206-685-6157, or Chang at kchang4@uw.edu.

Five years ago, Nature — one of the most prestigious research journals in science — published an pledging to improve on the low number of women editors and authors in its pages.

For many readers and scientists, that acknowledgement was a long time in coming. Yet with the hindsight of today’s re-examination of the treatment of women at all levels of society, the editorial could seem almost prescient.

In the time since that editorial, however, not much has changed, according to a new 91̽�� study published and cited in a printed March 7 in Nature. The preliminary study, by 91̽��psychology professor and doctoral student , finds that many high-profile neuroscience journals had a low representation of female authors. For example, fewer than 25 percent of Nature research articles listed women as the first author — usually the junior scientist who led the research. Among last authors — typically the senior laboratory leader — just over 15 percent were women. Nature’s top-tier competitor, Science, had similarly low numbers of women authors.

What most concerned the 91̽��team was that over a 12-year period ending in 2017, the percentage of female authors across these journals showed little improvement: less than 1 percent annually, with many journals showing no increase at all.

For the sake of comparison, the 91̽��team also looked at the number of women who received major National Institute of Health grants during the same time period. Those numbers were much higher, and increased slowly but steadily, with just under 30 percent of grants in 2017 awarded to women.

“These research grants are awarded based on significance, impact and productivity. We shouldn’t see this huge discrepancy between NIH funding and last authorship in high impact journals,” Fine said. It’s particularly troubling, the study’s authors say, given that publishing in high-profile journals is virtually imperative for winning academic awards or positions at top-ranked institutions.

Gender disparities in STEM fields has garnered more attention in recent years. While National Science Foundation-compiled data show that women make up a proportion of STEM faculty, their numbers remain significantly lower than those of men. A 2016 by the Society for Neuroscience showed that a little more than half of neuroscience doctorates are awarded to women, but women make up an average of only 30 percent of neuroscience faculty.

Other studies of gender and authorship have also pointed to the possible contribution of publication bias. A small-scale focusing on Nature Neuroscience, in 2016, showed similar results to the 91̽��findings. And in 2013, a led by the UW’s Jevin West and Carl Bergstrom, though an analysis of publications in the JSTOR digital library, found that women also are much less likely to be featured in prominent first- or last-author positions.

The issue extends beyond science: In spring 2017, an at the University of Liverpool found that papers written by female economists took an average of six months longer to get published than those written by men.

For this study, Shen, Fine, and their psychology co-authors research associate Jason Webster and professor, turned to the MEDLINE database of articles, which is hosted by the U.S. National Library of Medicine. They focused on 15 journals that publish neuroscience research, accounting for nearly 167,000 research articles from 2005 to 2017, and analyzed the author bylines using another database that predicts gender based on more than 216,000 distinct first names.

Some journals did have a proportionate number of female authors. The journals with the highest percentage of first authors were Neuropsychology Review (53 percent) and Brain (43 percent); among last authors, numbers were highest in Neuropsychology Review (39 percent) and Current Opinion in Neurobiology (27 percent).

“From our analysis, it is not that women are not conducting research and publishing, they are just much less likely to get their work into the really high-profile journals,” Shen said.

Fine and Shen suggest several solutions for all journals: to record and report article authorship by gender; to train reviewers to avoid bias, provide reviewers with more specific review criteria, akin to those required for grant awards; to adopt double-blind reviewing; or to establish byline quotas.

“It’s ridiculous to think bias isn’t at play in these very elite journals,” Fine says. “There are glass ceilings in technology, in politics, in business. It’s very hard not to believe that this is not just another glass ceiling.”

Increasing the number of women faculty in STEM fields is the goal of the . But if publication presents a barrier, then some universities may be challenged to hire and promote women, said Eve Riskin, 91̽��associate dean of engineering for diversity and access, professor of electrical engineering and faculty director of ADVANCE.

“Research shows that diverse teams lead to better solutions,” Riskin said. “Research also shows that female students in STEM do better when they have female faculty as instructors.  Holding women to higher standards for publication makes it harder for universities to increase their number of female faculty members in STEM and in leadership positions.”

The study’s authors have also made their code publicly available, with the hope that students or faculty in other fields will take on the same challenge, determine the gender breakdown of bylines in a given set of journals, and call for change.

“These journals make a lot of money and wield a huge amount of power. Finding a way to fix this problem is the least they can do,” Fine said. “They are under the same legal obligations to avoid discrimination as other businesses.”

###

For more information, contact Shen at psy.alicia@gmail.com or Fine at ionefine@uw.edu or 206-685-6157.

A new 91̽�� study seeks to answer that question and offers visual simulations of what someone with restored vision might see. The study concludes that while important advancements have been made in the field, the vision provided by sight recovery technologies may be very different from what scientists and patients had previously assumed.

In a  published Aug. 3 in the journal Philosophical Transactions B, 91̽��researchers used simulations to create that mimic what vision would be like after two different types of sight recovery therapies.

Lead author , a 91̽��associate professor of psychology, said the simulations are unprecedented.

“This is the first visual simulation of restored sight in any realistic form,” she said. “Now we can actually say, ‘This is what the world might look like if you had a retinal implant.’”

Fine said the paper aims to provide information about the quality of vision people can expect if they undergo sight restoration surgery, an invasive and costly procedure.

“This is a really difficult decision to make,” she said. “These devices involve long surgeries, and they don’t restore anything close to normal vision. The more information patients have, the better.”

More than 20 million Americans aged 18 and older have experienced vision loss, according to the , and rates of vision loss are expected to double by 2030 as the nation’s population ages.

For many of these patients, vision loss occurs after light enters the eye and lands on the retina, a thin layer at the back of the eye that contains millions of nerve cells. Among those are cells called rods and cones, which convert light into electrical impulses that are transmitted to vision centers in the brain. Loss of rods and cones is the primary cause of vision loss in diseases such as or .

But those diseases leave most remaining neurons within the retina relatively intact, and various technologies under development aim to restore vision by targeting the surviving cells.

This is a pivotal time for the industry, Fine said, with one company that has a device on the market and several others set to enter the market in the next five to 10 years.

Two of the most promising devices, she said, are electric prostheses, which enable vision by stimulating surviving cells with an array of electrodes placed on the retina, and optogenetics, which insert proteins into the surviving retinal cells to make them light-sensitive.

But the devices have a major shortcoming, co-author said, since stimulating the surviving cells in a retina is unlikely to produce vision that is close to normal.

“The retina contains a vast diversity of cells that carry distinct visual information and respond differently to visual input,” said Boynton, a 91̽��psychology professor.

“Electrically stimulating the retina excites all of these cells at the same time, which is very different from how these cells respond to real visual input.”

There are similar issues with optogenetics, Boynton said. “The optogenetic proteins that are currently available produce sluggish responses over time, and they are limited in the number of different cell types that they can separately target,” he said.

These limitations in both technologies mean that patients may see fuzzy, comet-like shapes or blurred outlines, or they may experience temporary visual disappearances if an object moves too fast.

Previous simulations of restored vision have used a “scoreboard model,” a grid of dots similar to the scoreboard at a football game, in which each electrode produces a visible dot in space. Together, that collection of dots is intended to demonstrate what someone with restored vision will see.

Fine said the new simulations show that the scoreboard model, which is sometimes used to test devices, doesn’t provide a good representation of the quality of vision sight restoration technologies are likely to produce. More realistic models are needed, she said, to give patients, clinicians and researchers a better idea of how those technologies will work in the real world.

Fine said better simulations can provide valuable information about how implants need to be improved to produce more natural vision.

“As these devices start being implanted in people, we can compare different types of devices and the different perceptual outcomes of each,” she said. “The path to fully restored eyesight is an elusive target. We need to start developing more sophisticated models of what people actually see.

“Until we do that, we’re just shooting in the dark in trying to improve these implants.”

But a study published three years after the operation found that the then-49-year-old could see colors, motion and some simple two-dimensional shapes, but was incapable of more complex visual processing.

Hoping May might eventually regain those visual skills, 91̽�� researchers and colleagues retested him a decade later. But in a now available online in Psychological Science, they report that May — referred to in the study as M.M. — continues to perform significantly worse than sighted control group participants.

The conclusion: May’s vision remains very limited 15 years after the surgeries. Though disappointing, the results provide valuable information that can help researchers better understand how vision develops and which visual processing tasks are most vulnerable to sight deprivation.

“With sight-restoration procedures becoming more developed, we’re going to see more and more cases where people are blind for long periods of time and then get their sight back,” said senior author , a 91̽��associate professor of psychology.

“But we know very little about what happens in their brains during that period. That is going to be one of the fundamental questions going forward — what happens when the lights are turned off, and what happens when you turn them back on?”

May went blind at age 3 when a jar of chemicals exploded in his face. He went on to work for the CIA and became a successful entrepreneur, founding the Sendero Group, a company that makes GPS and talking-map products for blind people. May is also a motivational speaker and holds the world downhill skiing speed record, 65 mph, for a completely blind person.

But fully restored sight has eluded May, and his unusual case has puzzled researchers. There were few previous cases of restored vision before his — the last well-documented one was in 1963 — and scientists knew little about whether people whose sight is restored as adults can regain functional vision, and if so, how long that might take.

In the recent tests, May was shown images of household objects and faces, and also video clips while his brain responses were measured with fMRI (functional magnetic resonance imaging). As with the tests a decade earlier, May did not have normal brain responses to three-dimensional objects or faces, consistent with his inability to make sense of these stimuli.

Researchers believe that’s because May’s brain, like those of other people who went blind at an early age, has adjusted to respond to other stimuli, such as sound or touch.

“We suspect that Mike lost vision at an age when these brain regions were able to take on new roles,” said joint first author Jason Webster. “It remains to be seen what these areas are doing now.”

May’s case is particularly interesting, Fine said, because his blindness started when the visual system is already developed, but the ability to perceive objects and faces is still evolving.

“He lost his vision at an age when vision is pretty good, but he was still young enough for it to deteriorate,” she said.

The findings, the researchers say, indicate that visual function for tasks such as object recognition and face processing continues to develop through childhood and early adolescence and remains sensitive to loss of sight for several years afterward.

The good news, said joint first author and 91̽��graduate student Elizabeth Huber, is that the findings imply that adults’ vision is relatively fixed, meaning that as visual losses increase in an aging population, the chances of restoring useful sight to older people are good.

“This study is encouraging because it suggests that if someone loses sight later in life, it may still be possible to restore relatively normal vision, even after many years of blindness,” she said.

May told the researchers he uses his other senses to compensate for his poor vision.

“I have learned what works with vision and what doesn’t, so I really don’t challenge my vision much anymore,” he said in the paper. “Where motion or colors might be clues, I use my vision. Where details might be required, like reading print or recognizing who someone is, I use tactile and auditory techniques.”

Other co-authors are 91̽��psychology professor ; Alyssa Brewer at the University of California, Irvine; Donald MacLeod at the University of California, San Diego; Brian Wandell at Stanford University; and Alex Wade at the University of York in the U.K.