A new computational technique could make it easier to engineer useful proteins

To engineer proteins with useful functions, researchers usually begin with a natural protein that has a desirable function, such as emitting fluorescent light, and put it through many rounds of random mutation that eventually generate an optimized version of the protein.

This process has yielded optimized versions of many important proteins, including green fluorescent protein (GFP). However, for other proteins, it has proven difficult to generate an optimized version. MIT researchers have now developed a computational approach that makes it easier to predict mutations that will lead to better proteins, based on a relatively small amount of data.

Using this model, the researchers generated proteins with mutations that were predicted to lead to improved versions of GFP and a protein from adeno-associated virus (AAV), which is used to deliver DNA for gene therapy. They hope it could also be used to develop additional tools for neuroscience research and medical applications.

Woman gestures with her hand in front of a glass wall with equations written on it.
MIT Professor of Brain and Cognitive Sciences Ila Fiete in her lab at the McGovern Institute. Photo: Steph Stevens

“Protein design is a hard problem because the mapping from DNA sequence to protein structure and function is really complex. There might be a great protein 10 changes away in the sequence, but each intermediate change might correspond to a totally nonfunctional protein. It’s like trying to find your way to the river basin in a mountain range, when there are craggy peaks along the way that block your view. The current work tries to make the riverbed easier to find,” says Ila Fiete, a professor of brain and cognitive sciences at MIT, a member of MIT’s McGovern Institute for Brain Research, director of the K. Lisa Yang Integrative Computational Neuroscience Center, and one of the senior authors of the study.

Regina Barzilay, the School of Engineering Distinguished Professor for AI and Health at MIT, and Tommi Jaakkola, the Thomas Siebel Professor of Electrical Engineering and Computer Science at MIT, are also senior authors of an open-access paper on the work, which will be presented at the International Conference on Learning Representations in May. MIT graduate students Andrew Kirjner and Jason Yim are the lead authors of the study. Other authors include Shahar Bracha, an MIT postdoc, and Raman Samusevich, a graduate student at Czech Technical University.

Optimizing proteins

Many naturally occurring proteins have functions that could make them useful for research or medical applications, but they need a little extra engineering to optimize them. In this study, the researchers were originally interested in developing proteins that could be used in living cells as voltage indicators. These proteins, produced by some bacteria and algae, emit fluorescent light when an electric potential is detected. If engineered for use in mammalian cells, such proteins could allow researchers to measure neuron activity without using electrodes.

While decades of research have gone into engineering these proteins to produce a stronger fluorescent signal, on a faster timescale, they haven’t become effective enough for widespread use. Bracha, who works in Edward Boyden’s lab at the McGovern Institute, reached out to Fiete’s lab to see if they could work together on a computational approach that might help speed up the process of optimizing the proteins.

“This work exemplifies the human serendipity that characterizes so much science discovery,” Fiete says.

“This work grew out of the Yang Tan Collective retreat, a scientific meeting of researchers from multiple centers at MIT with distinct missions unified by the shared support of K. Lisa Yang. We learned that some of our interests and tools in modeling how brains learn and optimize could be applied in the totally different domain of protein design, as being practiced in the Boyden lab.”

For any given protein that researchers might want to optimize, there is a nearly infinite number of possible sequences that could generated by swapping in different amino acids at each point within the sequence. With so many possible variants, it is impossible to test all of them experimentally, so researchers have turned to computational modeling to try to predict which ones will work best.

In this study, the researchers set out to overcome those challenges, using data from GFP to develop and test a computational model that could predict better versions of the protein.

They began by training a type of model known as a convolutional neural network (CNN) on experimental data consisting of GFP sequences and their brightness — the feature that they wanted to optimize.

The model was able to create a “fitness landscape” — a three-dimensional map that depicts the fitness of a given protein and how much it differs from the original sequence — based on a relatively small amount of experimental data (from about 1,000 variants of GFP).

These landscapes contain peaks that represent fitter proteins and valleys that represent less fit proteins. Predicting the path that a protein needs to follow to reach the peaks of fitness can be difficult, because often a protein will need to undergo a mutation that makes it less fit before it reaches a nearby peak of higher fitness. To overcome this problem, the researchers used an existing computational technique to “smooth” the fitness landscape.

Once these small bumps in the landscape were smoothed, the researchers retrained the CNN model and found that it was able to reach greater fitness peaks more easily. The model was able to predict optimized GFP sequences that had as many as seven different amino acids from the protein sequence they started with, and the best of these proteins were estimated to be about 2.5 times fitter than the original.

“Once we have this landscape that represents what the model thinks is nearby, we smooth it out and then we retrain the model on the smoother version of the landscape,” Kirjner says. “Now there is a smooth path from your starting point to the top, which the model is now able to reach by iteratively making small improvements. The same is often impossible for unsmoothed landscapes.”

Proof-of-concept

The researchers also showed that this approach worked well in identifying new sequences for the viral capsid of adeno-associated virus (AAV), a viral vector that is commonly used to deliver DNA. In that case, they optimized the capsid for its ability to package a DNA payload.

“We used GFP and AAV as a proof-of-concept to show that this is a method that works on data sets that are very well-characterized, and because of that, it should be applicable to other protein engineering problems,” Bracha says.

The researchers now plan to use this computational technique on data that Bracha has been generating on voltage indicator proteins.

“Dozens of labs having been working on that for two decades, and still there isn’t anything better,” she says. “The hope is that now with generation of a smaller data set, we could train a model in silico and make predictions that could be better than the past two decades of manual testing.”

The research was funded, in part, by the U.S. National Science Foundation, the Machine Learning for Pharmaceutical Discovery and Synthesis consortium, the Abdul Latif Jameel Clinic for Machine Learning in Health, the DTRA Discovery of Medical Countermeasures Against New and Emerging threats program, the DARPA Accelerated Molecular Discovery program, the Sanofi Computational Antibody Design grant, the U.S. Office of Naval Research, the Howard Hughes Medical Institute, the National Institutes of Health, the K. Lisa Yang ICoN Center, and the K. Lisa Yang and Hock E. Tan Center for Molecular Therapeutics at MIT.

Researchers reveal roadmap for AI innovation in brain and language learning

One of the hallmarks of humanity is language, but now, powerful new artificial intelligence tools also compose poetry, write songs, and have extensive conversations with human users. Tools like ChatGPT and Gemini are widely available at the tap of a button — but just how smart are these AIs?

A new multidisciplinary research effort co-led by Anna (Anya) Ivanova, assistant professor in the School of Psychology at Georgia Tech, alongside Kyle Mahowald, an assistant professor in the Department of Linguistics at the University of Texas at Austin, is working to uncover just that.

Their results could lead to innovative AIs that are more similar to the human brain than ever before — and also help neuroscientists and psychologists who are unearthing the secrets of our own minds.

The study, “Dissociating Language and Thought in Large Language Models,” is published this week in the scientific journal Trends in Cognitive Sciences. The work is already making waves in the scientific community: an earlier preprint of the paper, released in January 2023, has already been cited more than 150 times by fellow researchers. The research team has continued to refine the research for this final journal publication.

“ChatGPT became available while we were finalizing the preprint,” explains Ivanova, who conducted the research while a postdoctoral researcher at MIT’s McGovern Institute. “Over the past year, we’ve had an opportunity to update our arguments in light of this newer generation of models, now including ChatGPT.”

Form versus function

The study focuses on large language models (LLMs), which include AIs like ChatGPT. LLMs are text prediction models, and create writing by predicting which word comes next in a sentence — just like how a cell phone or email service like Gmail might suggest what next word you might want to write. However, while this type of language learning is extremely effective at creating coherent sentences, that doesn’t necessarily signify intelligence.

Ivanova’s team argues that formal competence — creating a well-structured, grammatically correct sentence — should be differentiated from functional competence — answering the right question, communicating the correct information, or appropriately communicating. They also found that while LLMs trained on text prediction are often very good at formal skills, they still struggle with functional skills.

“We humans have the tendency to conflate language and thought,” Ivanova says. “I think that’s an important thing to keep in mind as we’re trying to figure out what these models are capable of, because using that ability to be good at language, to be good at formal competence, leads many people to assume that AIs are also good at thinking — even when that’s not the case.

It’s a heuristic that we developed when interacting with other humans over thousands of years of evolution, but now in some respects, that heuristic is broken,” Ivanova explains.

The distinction between formal and functional competence is also vital in rigorously testing an AI’s capabilities, Ivanova adds. Evaluations often don’t distinguish formal and functional competence, making it difficult to assess what factors are determining a model’s success or failure. The need to develop distinct tests is one of the team’s more widely accepted findings, and one that some researchers in the field have already begun to implement.

Creating a modular system

While the human tendency to conflate functional and formal competence may have hindered understanding of LLMs in the past, our human brains could also be the key to unlocking more powerful AIs.

Leveraging the tools of cognitive neuroscience while a postdoctoral associate at Massachusetts Institute of Technology (MIT), Ivanova and her team studied brain activity in neurotypical individuals via fMRI, and used behavioral assessments of individuals with brain damage to test the causal role of brain regions in language and cognition — both conducting new research and drawing on previous studies. The team’s results showed that human brains use different regions for functional and formal competence, further supporting this distinction in AIs.

“Our research shows that in the brain, there is a language processing module and separate modules for reasoning,” Ivanova says. This modularity could also serve as a blueprint for how to develop future AIs.

“Building on insights from human brains — where the language processing system is sharply distinct from the systems that support our ability to think — we argue that the language-thought distinction is conceptually important for thinking about, evaluating, and improving large language models, especially given recent efforts to imbue these models with human-like intelligence,” says Ivanova’s former advisor and study co-author Evelina Fedorenko, a professor of brain and cognitive sciences at MIT and a member of the McGovern Institute for Brain Research.

Developing AIs in the pattern of the human brain could help create more powerful systems — while also helping them dovetail more naturally with human users. “Generally, differences in a mechanism’s internal structure affect behavior,” Ivanova says. “Building a system that has a broad macroscopic organization similar to that of the human brain could help ensure that it might be more aligned with humans down the road.”

In the rapidly developing world of AI, these systems are ripe for experimentation. After the team’s preprint was published, OpenAI announced their intention to add plug-ins to their GPT models.

“That plug-in system is actually very similar to what we suggest,” Ivanova adds. “It takes a modularity approach where the language model can be an interface to another specialized module within a system.”

While the OpenAI plug-in system will include features like booking flights and ordering food, rather than cognitively inspired features, it demonstrates that “the approach has a lot of potential,” Ivanova says.

The future of AI — and what it can tell us about ourselves

While our own brains might be the key to unlocking better, more powerful AIs, these AIs might also help us better understand ourselves. “When researchers try to study the brain and cognition, it’s often useful to have some smaller system where you can actually go in and poke around and see what’s going on before you get to the immense complexity,” Ivanova explains.

However, since human language is unique, model or animal systems are more difficult to relate. That’s where LLMs come in.

“There are lots of surprising similarities between how one would approach the study of the brain and the study of an artificial neural network” like a large language model, she adds. “They are both information processing systems that have biological or artificial neurons to perform computations.”

In many ways, the human brain is still a black box, but openly available AIs offer a unique opportunity to see the synthetic system’s inner workings and modify variables, and explore these corresponding systems like never before.

“It’s a really wonderful model that we have a lot of control over,” Ivanova says. “Neural networks — they are amazing.”

Along with Anna (Anya) Ivanova, Kyle Mahowald, and Evelina Fedorenko, the research team also includes Idan Blank (University of California, Los Angeles), as well as Nancy Kanwisher and Joshua Tenenbaum (Massachusetts Institute of Technology).

For people who speak many languages, there’s something special about their native tongue

A new study of people who speak many languages has found that there is something special about how the brain processes their native language.

In the brains of these polyglots — people who speak five or more languages — the same language regions light up when they listen to any of the languages that they speak. In general, this network responds more strongly to languages in which the speaker is more proficient, with one notable exception: the speaker’s native language. When listening to one’s native language, language network activity drops off significantly.

The findings suggest there is something unique about the first language one acquires, which allows the brain to process it with minimal effort, the researchers say.

“Something makes it a little bit easier to process — maybe it’s that you’ve spent more time using that language — and you get a dip in activity for the native language compared to other languages that you speak proficiently,” says Evelina Fedorenko, an associate professor of neuroscience at MIT, a member of MIT’s McGovern Institute for Brain Research, and the senior author of the study.

Saima Malik-Moraleda, a graduate student in the Speech and Hearing Bioscience and Technology Program at Harvard University, and Olessia Jouravlev, a former MIT postdoc who is now an associate professor at Carleton University, are the lead authors of the paper, which appears today in the journal Cerebral Cortex.

Many languages, one network

McGovern Investivator Ev Fedorenko in the Martinos Imaging Center at MIT. Photo: Caitlin Cunningham

The brain’s language processing network, located primarily in the left hemisphere, includes regions in the frontal and temporal lobes. In a 2021 study, Fedorenko’s lab found that in the brains of polyglots, the language network was less active when listening to their native language than the language networks of people who speak only one language.

In the new study, the researchers wanted to expand on that finding and explore what happens in the brains of polyglots as they listen to languages in which they have varying levels of proficiency. Studying polyglots can help researchers learn more about the functions of the language network, and how languages learned later in life might be represented differently than a native language or languages.

“With polyglots, you can do all of the comparisons within one person. You have languages that vary along a continuum, and you can try to see how the brain modulates responses as a function of proficiency,” Fedorenko says.

For the study, the researchers recruited 34 polyglots, each of whom had at least some degree of proficiency in five or more languages but were not bilingual or multilingual from infancy. Sixteen of the participants spoke 10 or more languages, including one who spoke 54 languages with at least some proficiency.

Each participant was scanned with functional magnetic resonance imaging (fMRI) as they listened to passages read in eight different languages. These included their native language, a language they were highly proficient in, a language they were moderately proficient in, and a language in which they described themselves as having low proficiency.

They were also scanned while listening to four languages they didn’t speak at all. Two of these were languages from the same family (such as Romance languages) as a language they could speak, and two were languages completely unrelated to any languages they spoke.

The passages used for the study came from two different sources, which the researchers had previously developed for other language studies. One was a set of Bible stories recorded in many different languages, and the other consisted of passages from “Alice in Wonderland” translated into many languages.

Brain scans revealed that the language network lit up the most when participants listened to languages in which they were the most proficient. However, that did not hold true for the participants’ native languages, which activated the language network much less than non-native languages in which they had similar proficiency. This suggests that people are so proficient in their native language that the language network doesn’t need to work very hard to interpret it.

“As you increase proficiency, you can engage linguistic computations to a greater extent, so you get these progressively stronger responses. But then if you compare a really high-proficiency language and a native language, it may be that the native language is just a little bit easier, possibly because you’ve had more experience with it,” Fedorenko says.

Brain engagement

The researchers saw a similar phenomenon when polyglots listened to languages that they don’t speak: Their language network was more engaged when listening to languages related to a language that they could understand, than compared to listening to completely unfamiliar languages.

“Here we’re getting a hint that the response in the language network scales up with how much you understand from the input,” Malik-Moraleda says. “We didn’t quantify the level of understanding here, but in the future we’re planning to evaluate how much people are truly understanding the passages that they’re listening to, and then see how that relates to the activation.”

The researchers also found that a brain network known as the multiple demand network, which turns on whenever the brain is performing a cognitively demanding task, also becomes activated when listening to languages other than one’s native language.

“What we’re seeing here is that the language regions are engaged when we process all these languages, and then there’s this other network that comes in for non-native languages to help you out because it’s a harder task,” Malik-Moraleda says.

In this study, most of the polyglots began studying their non-native languages as teenagers or adults, but in future work, the researchers hope to study people who learned multiple languages from a very young age. They also plan to study people who learned one language from infancy but moved to the United States at a very young age and began speaking English as their dominant language, while becoming less proficient in their native language, to help disentangle the effects of proficiency versus age of acquisition on brain responses.

The research was funded by the McGovern Institute for Brain Research, MIT’s Department of Brain and Cognitive Sciences, and the Simons Center for the Social Brain.

School of Science announces 2024 Infinite Expansion Awards

The MIT School of Science has announced nine postdocs and research scientists as recipients of the 2024 Infinite Expansion Award, which highlights extraordinary members of the MIT community.

The following are the 2024 School of Science Infinite Expansion winners:

  • Sarthak Chandra, a research scientist in the Department of Brain and Cognitive Sciences, was nominated by Professor Ila Fiete, who wrote, “He has expanded the research abilities of my group by being a versatile and brilliant scientist, by drawing connections with a different area that he was an expert in from his PhD training, and by being a highly involved and caring mentor.”
  • Michal Fux, a research scientist in the Department of Brain and Cognitive Sciences, was nominated by Professor Pawan Sinha, who wrote, “She is one of those figurative beams of light that not only brilliantly illuminate scientific questions, but also enliven a research team.”
  • Andrew Savinov, a postdoc in the Department of Biology, was nominated by Associate Professor Gene-Wei Li, who wrote, “Andrew is an extraordinarily creative and accomplished biophysicist, as well as an outstanding contributor to the broader MIT community.”
  • Ho Fung Cheng, a postdoc in the Department of Chemistry, was nominated by Professor Jeremiah Johnson, who wrote, “His impact on research and our departmental community during his time at MIT has been outstanding, and I believe that he will be a worldclass teacher and research group leader in his independent career next year.”
  • Gabi Wenzel, a postdoc in the Department of Chemistry, was nominated by Assistant Professor Brett McGuire, who wrote, “In the one year since Gabi joined our team, she has become an indispensable leader, demonstrating exceptional skill, innovation, and dedication in our challenging research environment.”
  • Yu-An Zhang, a postdoc in the Department of Chemistry, was nominated by Professor Alison Wendlandt, who wrote, “He is a creative, deep-thinking scientist and a superb organic chemist. But above all, he is an off-scale mentor and a cherished coworker.”
  • Wouter Van de Pontseele, a senior postdoc in the Laboratory for Nuclear Science, was nominated by Professor Joseph Formaggio, who wrote, “He is a talented scientist with an intense creativity, scholarship, and student mentorship record. In the time he has been with my group, he has led multiple facets of my experimental program and has been a wonderful citizen of the MIT community.”
  • Alexander Shvonski, a lecturer in the Department of Physics, was nominated by Assistant Professor Andrew Vanderburg, who wrote, “… I have been blown away by Alex’s knowledge of education research and best practices, his skills as a teacher and course content designer, and I have been extremely grateful for his assistance.”
  • David Stoppel, a research scientist in The Picower Institute for Learning and Memory, was nominated by Professor Mark Bear and his research group, who wrote, “As impressive as his research achievements might be, David’s most genuine qualification for this award is his incredible commitment to mentorship and the dissemination of knowledge.”

Winners are honored with a monetary award and will be celebrated with family, friends, and nominators at a later date, along with recipients of the Infinite Mile Award.

Exposure to different kinds of music influences how the brain interprets rhythm

When listening to music, the human brain appears to be biased toward hearing and producing rhythms composed of simple integer ratios — for example, a series of four beats separated by equal time intervals (forming a 1:1:1 ratio).

However, the favored ratios can vary greatly between different societies, according to a large-scale study led by researchers at MIT and the Max Planck Institute for Empirical Aesthetics and carried out in 15 countries. The study included 39 groups of participants, many of whom came from societies whose traditional music contains distinctive patterns of rhythm not found in Western music.

“Our study provides the clearest evidence yet for some degree of universality in music perception and cognition, in the sense that every single group of participants that was tested exhibits biases for integer ratios. It also provides a glimpse of the variation that can occur across cultures, which can be quite substantial,” says Nori Jacoby, the study’s lead author and a former MIT postdoc, who is now a research group leader at the Max Planck Institute for Empirical Aesthetics in Frankfurt, Germany.

The brain’s bias toward simple integer ratios may have evolved as a natural error-correction system that makes it easier to maintain a consistent body of music, which human societies often use to transmit information.

“When people produce music, they often make small mistakes. Our results are consistent with the idea that our mental representation is somewhat robust to those mistakes, but it is robust in a way that pushes us toward our preexisting ideas of the structures that should be found in music,” says Josh McDermott, an associate professor of brain and cognitive sciences at MIT and a member of MIT’s McGovern Institute for Brain Research and Center for Brains, Minds, and Machines.

McDermott is the senior author of the study, which appears today in Nature Human Behaviour. The research team also included scientists from more than two dozen institutions around the world.

A global approach

The new study grew out of a smaller analysis that Jacoby and McDermott published in 2017. In that paper, the researchers compared rhythm perception in groups of listeners from the United States and the Tsimane’, an Indigenous society located in the Bolivian Amazon rainforest.

pitch perception study
Nori Jacoby, a former MIT postdoc now at the Max Planck Institute for Empirical Aesthetics, runs an experiment with a member of the Tsimane’ tribe, who have had little exposure to Western music. Photo: Josh McDermott

To measure how people perceive rhythm, the researchers devised a task in which they play a randomly generated series of four beats and then ask the listener to tap back what they heard. The rhythm produced by the listener is then played back to the listener, and they tap it back again. Over several iterations, the tapped sequences became dominated by the listener’s internal biases, also known as priors.

“The initial stimulus pattern is random, but at each iteration the pattern is pushed by the listener’s biases, such that it tends to converge to a particular point in the space of possible rhythms,” McDermott says. “That can give you a picture of what we call the prior, which is the set of internal implicit expectations for rhythms that people have in their heads.”

When the researchers first did this experiment, with American college students as the test subjects, they found that people tended to produce time intervals that are related by simple integer ratios. Furthermore, most of the rhythms they produced, such as those with ratios of 1:1:2 and 2:3:3, are commonly found in Western music.

The researchers then went to Bolivia and asked members of the Tsimane’ society to perform the same task. They found that Tsimane’ also produced rhythms with simple integer ratios, but their preferred ratios were different and appeared to be consistent with those that have been documented in the few existing records of Tsimane’ music.

“At that point, it provided some evidence that there might be very widespread tendencies to favor these small integer ratios, and that there might be some degree of cross-cultural variation. But because we had just looked at this one other culture, it really wasn’t clear how this was going to look at a broader scale,” Jacoby says.

To try to get that broader picture, the MIT team began seeking collaborators around the world who could help them gather data on a more diverse set of populations. They ended up studying listeners from 39 groups, representing 15 countries on five continents — North America, South America, Europe, Africa, and Asia.

“This is really the first study of its kind in the sense that we did the same experiment in all these different places, with people who are on the ground in those locations,” McDermott says. “That hasn’t really been done before at anything close to this scale, and it gave us an opportunity to see the degree of variation that might exist around the world.”

A grid of nine different photos showing a researcher working with an individual at a table. The individuals are wearing headphones.
Example testing sites. a, Yaranda, Bolivia. b, Montevideo, Uruguay. c, Sagele, Mali. d, Spitzkoppe, Namibia. e, Pleven, Bulgaria. f, Bamako, Mali. g, D’Kar, Botswana. h, Stockholm, Sweden. i, Guizhou, China. j, Mumbai, India. Verbal informed consent was obtained from the individuals in each photo.

Cultural comparisons

Just as they had in their original 2017 study, the researchers found that in every group they tested, people tended to be biased toward simple integer ratios of rhythm. However, not every group showed the same biases. People from North America and Western Europe, who have likely been exposed to the same kinds of music, were more likely to generate rhythms with the same ratios. However, many groups, for example those in Turkey, Mali, Bulgaria, and Botswana showed a bias for other rhythms.

“There are certain cultures where there are particular rhythms that are prominent in their music, and those end up showing up in the mental representation of rhythm,” Jacoby says.

The researchers believe their findings reveal a mechanism that the brain uses to aid in the perception and production of music.

“When you hear somebody playing something and they have errors in their performance, you’re going to mentally correct for those by mapping them onto where you implicitly think they ought to be,” McDermott says. “If you didn’t have something like this, and you just faithfully represented what you heard, these errors might propagate and make it much harder to maintain a musical system.”

Among the groups that they studied, the researchers took care to include not only college students, who are easy to study in large numbers, but also people living in traditional societies, who are more difficult to reach. Participants from those more traditional groups showed significant differences from college students living in the same countries, and from people who live in those countries but performed the test online.

“What’s very clear from the paper is that if you just look at the results from undergraduate students around the world, you vastly underestimate the diversity that you see otherwise,” Jacoby says. “And the same was true of experiments where we tested groups of people online in Brazil and India, because you’re dealing with people who have internet access and presumably have more exposure to Western music.”

The researchers now hope to run additional studies of different aspects of music perception, taking this global approach.

“If you’re just testing college students around the world or people online, things look a lot more homogenous. I think it’s very important for the field to realize that you actually need to go out into communities and run experiments there, as opposed to taking the low-hanging fruit of running studies with people in a university or on the internet,” McDermott says.

The research was funded by the James S. McDonnell Foundation, the Canadian National Science and Engineering Research Council, the South African National Research Foundation, the United States National Science Foundation, the Chilean National Research and Development Agency, the Austrian Academy of Sciences, the Japan Society for the Promotion of Science, the Keio Global Research Institute, the United Kingdom Arts and Humanities Research Council, the Swedish Research Council, and the John Fell Fund.

The brain runs an internal simulation to keep track of time

Clocks, computers, and metronomes can keep time with exquisite precision. But even in the absence of an external time keeper, we can track time on our own. We know when minutes or hours have elapsed, and we can maintain a rhythm when we dance, sing, or play music. Now, neuroscientists at the National Autonomous University of Mexico and MIT’s McGovern Institute and have discovered one way the brain keeps a beat: It runs an internal simulation, mentally recreating the perception of an external rhythm and preparing an appropriately timed response.

The discovery, reported January 10, 2024, in the journal Science Advances, illustrates how animals can think about imaginary events and use an internal model to guide their interactions with the world. “It’s a real indication of mental states as an independent driver of behavior,” says neuroscientist Mehrdad Jazayeri, an investigator at the McGovern Institute and an associate professor of brain and cognitive sciences at MIT.

Predicting the future

Jazayeri teamed up with Victor de Lafuente, a neuroscientist at the National Autonomous University of Mexico, to investigate the brain’s time-keeping ability. De Lafuente, who led the study, says they were motivated by curiosity about how the brain makes predictions and prepares for future states of the world.

De Lafuente and his team used a visual metronome to teach monkeys a simple rhythm, showing them a circle that moved between two positions on a screen to set a steady tempo. Then the metronome stopped. After a variable and unpredictable pause, the monkeys were asked to indicate where the dot would be if the metronome had carried on.

Monkeys do well at this task, successfully keeping time after the metronome stops. After the waiting period, they are usually able to identify the expected position of the circle, which they communicate by reaching towards a touchscreen.

To find out how the animals were keeping track of the metronome’s rhythm, de Lafuente’s group monitored their brain activity. In several key brain regions, they found rhythmic patterns of activity that oscillated at the same frequency as the metronome. This occurred while the monkeys watched the metronome. More remarkably, it continued after the metronome had stopped.

“The animal is seeing things going and then things stop. What we find in the brain is the continuation of that process in the animal’s mind,” Jazayeri says. “An entire network is replicating what it was doing.”

That was true in the visual cortex, where clusters of neurons respond to stimuli in specific spots within the eyes’ field of view. One set of cells in the visual cortex fired when the metronome’s circle was on the left of the screen; another set fired when the dot was on the right. As a monkey followed the visual metronome, the researchers could see these cells’ activity alternating rhythmically, tracking the movement. When the metronome stopped, the back-and-forth neural activity continued, maintaining the rhythm. “Once the stimulus was no longer visible, they were seeing the stimulus within their minds,” de Lafuente says.

They found something similar in the brain’s motor cortex, where movements are prepared and executed. De Lafuente explains that the monkeys are motionless for most of their time-keeping task; only when they are asked to indicate where the metronome’s circle should be do they move a hand to touch the screen. But the motor cortex was engaged even before it was time to move. “Within their brains there is a signal that is switching from the left to the right,” he says. “So the monkeys are thinking ‘left, right, left, right’—even when they are not moving and the world is constant.”

While some scientists have proposed that the brain may have a central time-keeping mechanism, the team’s findings indicate that entire networks can be called on to track the passage of time. The monkeys’ model of the future was surprisingly explicit, de Lafuente says, representing specific sensory stimuli and plans for movement. “This offers a potential solution to mentally tracking the dynamics in the world, which is to basically think about them in terms of how they actually would have happened,” Jazayeri says.

 

Complex, unfamiliar sentences make the brain’s language network work harder

With help from an artificial language network, MIT neuroscientists have discovered what kind of sentences are most likely to fire up the brain’s key language processing centers.

The new study reveals that sentences that are more complex, either because of unusual grammar or unexpected meaning, generate stronger responses in these language processing centers. Sentences that are very straightforward barely engage these regions, and nonsensical sequences of words don’t do much for them either.

For example, the researchers found this brain network was most active when reading unusual sentences such as “Buy sell signals remains a particular,” taken from a publicly available language dataset called C4. However, it went quiet when reading something very straightforward, such as “We were sitting on the couch.”

“The input has to be language-like enough to engage the system,” says Evelina Fedorenko, Associate Professor of Neuroscience at MIT and a member of MIT’s McGovern Institute for Brain Research. “And then within that space, if things are really easy to process, then you don’t have much of a response. But if things get difficult, or surprising, if there’s an unusual construction or an unusual set of words that you’re maybe not very familiar with, then the network has to work harder.”

Fedorenko is the senior author of the study, which appears today in Nature Human Behavior. MIT graduate student Greta Tuckute is the lead author of the paper.

Processing language

In this study, the researchers focused on language-processing regions found in the left hemisphere of the brain, which includes Broca’s area as well as other parts of the left frontal and temporal lobes of the brain.

“This language network is highly selective to language, but it’s been harder to actually figure out what is going on in these language regions,” Tuckute says. “We wanted to discover what kinds of sentences, what kinds of linguistic input, drive the left hemisphere language network.”

The researchers began by compiling a set of 1,000 sentences taken from a wide variety of sources — fiction, transcriptions of spoken words, web text, and scientific articles, among many others.

Five human participants read each of the sentences while the researchers measured their language network activity using functional magnetic resonance imaging (fMRI). The researchers then fed those same 1,000 sentences into a large language model — a model similar to ChatGPT, which learns to generate and understand language from predicting the next word in huge amounts of text — and measured the activation patterns of the model in response to each sentence.

Once they had all of those data, the researchers trained a mapping model, known as an “encoding model,” which relates the activation patterns seen in the human brain with those observed in the artificial language model. Once trained, the model could predict how the human language network would respond to any new sentence based on how the artificial language network responded to these 1,000 sentences.

The researchers then used the encoding model to identify 500 new sentences that would generate maximal activity in the human brain (the “drive” sentences), as well as sentences that would elicit minimal activity in the brain’s language network (the “suppress” sentences).

In a group of three new human participants, the researchers found these new sentences did indeed drive and suppress brain activity as predicted.

“This ‘closed-loop’ modulation of brain activity during language processing is novel,” Tuckute says. “Our study shows that the model we’re using (that maps between language-model activations and brain responses) is accurate enough to do this. This is the first demonstration of this approach in brain areas implicated in higher-level cognition, such as the language network.”

Linguistic complexity

To figure out what made certain sentences drive activity more than others, the researchers analyzed the sentences based on 11 different linguistic properties, including grammaticality, plausibility, emotional valence (positive or negative), and how easy it is to visualize the sentence content.

For each of those properties, the researchers asked participants from crowd-sourcing platforms to rate the sentences. They also used a computational technique to quantify each sentence’s “surprisal,” or how uncommon it is compared to other sentences.

This analysis revealed that sentences with higher surprisal generate higher responses in the brain. This is consistent with previous studies showing people have more difficulty processing sentences with higher surprisal, the researchers say.

Another linguistic property that correlated with the language network’s responses was linguistic complexity, which is measured by how much a sentence adheres to the rules of English grammar and how plausible it is, meaning how much sense the content makes, apart from the grammar.

Sentences at either end of the spectrum — either extremely simple, or so complex that they make no sense at all — evoked very little activation in the language network. The largest responses came from sentences that make some sense but require work to figure them out, such as “Jiffy Lube of — of therapies, yes,” which came from the Corpus of Contemporary American English dataset.

“We found that the sentences that elicit the highest brain response have a weird grammatical thing and/or a weird meaning,” Fedorenko says. “There’s something slightly unusual about these sentences.”

The researchers now plan to see if they can extend these findings in speakers of languages other than English. They also hope to explore what type of stimuli may activate language processing regions in the brain’s right hemisphere.

The research was funded by an Amazon Fellowship from the Science Hub, an International Doctoral Fellowship from the American Association of University Women, the MIT-IBM Watson AI Lab, the National Institutes of Health, the McGovern Institute, the Simons Center for the Social Brain, and MIT’s Department of Brain and Cognitive Sciences.

Deep neural networks show promise as models of human hearing

Computational models that mimic the structure and function of the human auditory system could help researchers design better hearing aids, cochlear implants, and brain-machine interfaces. A new study from MIT has found that modern computational models derived from machine learning are moving closer to this goal.

In the largest study yet of deep neural networks that have been trained to perform auditory tasks, the MIT team showed that most of these models generate internal representations that share properties of representations seen in the human brain when people are listening to the same sounds.

The study also offers insight into how to best train this type of model: The researchers found that models trained on auditory input including background noise more closely mimic the activation patterns of the human auditory cortex.

“What sets this study apart is it is the most comprehensive comparison of these kinds of models to the auditory system so far. The study suggests that models that are derived from machine learning are a step in the right direction, and it gives us some clues as to what tends to make them better models of the brain,” says Josh McDermott, an associate professor of brain and cognitive sciences at MIT, a member of MIT’s McGovern Institute for Brain Research and Center for Brains, Minds, and Machines, and the senior author of the study.

MIT graduate student Greta Tuckute and Jenelle Feather PhD ’22 are the lead authors of the open-access paper, which appears today in PLOS Biology.

Models of hearing

Deep neural networks are computational models that consists of many layers of information-processing units that can be trained on huge volumes of data to perform specific tasks. This type of model has become widely used in many applications, and neuroscientists have begun to explore the possibility that these systems can also be used to describe how the human brain performs certain tasks.

“These models that are built with machine learning are able to mediate behaviors on a scale that really wasn’t possible with previous types of models, and that has led to interest in whether or not the representations in the models might capture things that are happening in the brain,” Tuckute says.

When a neural network is performing a task, its processing units generate activation patterns in response to each audio input it receives, such as a word or other type of sound. Those model representations of the input can be compared to the activation patterns seen in fMRI brain scans of people listening to the same input.

In 2018, McDermott and then-graduate student Alexander Kell reported that when they trained a neural network to perform auditory tasks (such as recognizing words from an audio signal), the internal representations generated by the model showed similarity to those seen in fMRI scans of people listening to the same sounds.

Since then, these types of models have become widely used, so McDermott’s research group set out to evaluate a larger set of models, to see if the ability to approximate the neural representations seen in the human brain is a general trait of these models.

For this study, the researchers analyzed nine publicly available deep neural network models that had been trained to perform auditory tasks, and they also created 14 models of their own, based on two different architectures. Most of these models were trained to perform a single task — recognizing words, identifying the speaker, recognizing environmental sounds, and identifying musical genre — while two of them were trained to perform multiple tasks.

When the researchers presented these models with natural sounds that had been used as stimuli in human fMRI experiments, they found that the internal model representations tended to exhibit similarity with those generated by the human brain. The models whose representations were most similar to those seen in the brain were models that had been trained on more than one task and had been trained on auditory input that included background noise.

“If you train models in noise, they give better brain predictions than if you don’t, which is intuitively reasonable because a lot of real-world hearing involves hearing in noise, and that’s plausibly something the auditory system is adapted to,” Feather says.

Hierarchical processing

The new study also supports the idea that the human auditory cortex has some degree of hierarchical organization, in which processing is divided into stages that support distinct computational functions. As in the 2018 study, the researchers found that representations generated in earlier stages of the model most closely resemble those seen in the primary auditory cortex, while representations generated in later model stages more closely resemble those generated in brain regions beyond the primary cortex.

Additionally, the researchers found that models that had been trained on different tasks were better at replicating different aspects of audition. For example, models trained on a speech-related task more closely resembled speech-selective areas.

“Even though the model has seen the exact same training data and the architecture is the same, when you optimize for one particular task, you can see that it selectively explains specific tuning properties in the brain,” Tuckute says.

McDermott’s lab now plans to make use of their findings to try to develop models that are even more successful at reproducing human brain responses. In addition to helping scientists learn more about how the brain may be organized, such models could also be used to help develop better hearing aids, cochlear implants, and brain-machine interfaces.

“A goal of our field is to end up with a computer model that can predict brain responses and behavior. We think that if we are successful in reaching that goal, it will open a lot of doors,” McDermott says.

The research was funded by the National Institutes of Health, an Amazon Fellowship from the Science Hub, an International Doctoral Fellowship from the American Association of University Women, an MIT Friends of McGovern Institute Fellowship, a fellowship from the K. Lisa Yang Integrative Computational Neuroscience (ICoN) Center at MIT, and a Department of Energy Computational Science Graduate Fellowship.

What does the future hold for generative AI?

Speaking at the “Generative AI: Shaping the Future” symposium on Nov. 28, the kickoff event of MIT’s Generative AI Week, keynote speaker and iRobot co-founder Rodney Brooks warned attendees against uncritically overestimating the capabilities of this emerging technology, which underpins increasingly powerful tools like OpenAI’s ChatGPT and Google’s Bard.

“Hype leads to hubris, and hubris leads to conceit, and conceit leads to failure,” cautioned Brooks, who is also a professor emeritus at MIT, a former director of the Computer Science and Artificial Intelligence Laboratory (CSAIL), and founder of Robust.AI.

“No one technology has ever surpassed everything else,” he added.

The symposium, which drew hundreds of attendees from academia and industry to the Institute’s Kresge Auditorium, was laced with messages of hope about the opportunities generative AI offers for making the world a better place, including through art and creativity, interspersed with cautionary tales about what could go wrong if these AI tools are not developed responsibly.

Generative AI is a term to describe machine-learning models that learn to generate new material that looks like the data they were trained on. These models have exhibited some incredible capabilities, such as the ability to produce human-like creative writing, translate languages, generate functional computer code, or craft realistic images from text prompts.

In her opening remarks to launch the symposium, MIT President Sally Kornbluth highlighted several projects faculty and students have undertaken to use generative AI to make a positive impact in the world. For example, the work of the Axim Collaborative, an online education initiative launched by MIT and Harvard, includes exploring the educational aspects of generative AI to help underserved students.

The Institute also recently announced seed grants for 27 interdisciplinary faculty research projects centered on how AI will transform people’s lives across society.

In hosting Generative AI Week, MIT hopes to not only showcase this type of innovation, but also generate “collaborative collisions” among attendees, Kornbluth said.

Collaboration involving academics, policymakers, and industry will be critical if we are to safely integrate a rapidly evolving technology like generative AI in ways that are humane and help humans solve problems, she told the audience.

“I honestly cannot think of a challenge more closely aligned with MIT’s mission. It is a profound responsibility, but I have every confidence that we can face it, if we face it head on and if we face it as a community,” she said.

While generative AI holds the potential to help solve some of the planet’s most pressing problems, the emergence of these powerful machine learning models has blurred the distinction between science fiction and reality, said CSAIL Director Daniela Rus in her opening remarks. It is no longer a question of whether we can make machines that produce new content, she said, but how we can use these tools to enhance businesses and ensure sustainability. 

“Today, we will discuss the possibility of a future where generative AI does not just exist as a technological marvel, but stands as a source of hope and a force for good,” said Rus, who is also the Andrew and Erna Viterbi Professor in the Department of Electrical Engineering and Computer Science.

But before the discussion dove deeply into the capabilities of generative AI, attendees were first asked to ponder their humanity, as MIT Professor Joshua Bennett read an original poem.

Bennett, a professor in the MIT Literature Section and Distinguished Chair of the Humanities, was asked to write a poem about what it means to be human, and drew inspiration from his daughter, who was born three weeks ago.

The poem told of his experiences as a boy watching Star Trek with his father and touched on the importance of passing traditions down to the next generation.

In his keynote remarks, Brooks set out to unpack some of the deep, scientific questions surrounding generative AI, as well as explore what the technology can tell us about ourselves.

To begin, he sought to dispel some of the mystery swirling around generative AI tools like ChatGPT by explaining the basics of how this large language model works. ChatGPT, for instance, generates text one word at a time by determining what the next word should be in the context of what it has already written. While a human might write a story by thinking about entire phrases, ChatGPT only focuses on the next word, Brooks explained.

ChatGPT 3.5 is built on a machine-learning model that has 175 billion parameters and has been exposed to billions of pages of text on the web during training. (The newest iteration, ChatGPT 4, is even larger.) It learns correlations between words in this massive corpus of text and uses this knowledge to propose what word might come next when given a prompt.

The model has demonstrated some incredible capabilities, such as the ability to write a sonnet about robots in the style of Shakespeare’s famous Sonnet 18. During his talk, Brooks showcased the sonnet he asked ChatGPT to write side-by-side with his own sonnet.

But while researchers still don’t fully understand exactly how these models work, Brooks assured the audience that generative AI’s seemingly incredible capabilities are not magic, and it doesn’t mean these models can do anything.

His biggest fears about generative AI don’t revolve around models that could someday surpass human intelligence. Rather, he is most worried about researchers who may throw away decades of excellent work that was nearing a breakthrough, just to jump on shiny new advancements in generative AI; venture capital firms that blindly swarm toward technologies that can yield the highest margins; or the possibility that a whole generation of engineers will forget about other forms of software and AI.

At the end of the day, those who believe generative AI can solve the world’s problems and those who believe it will only generate new problems have at least one thing in common: Both groups tend to overestimate the technology, he said.

“What is the conceit with generative AI? The conceit is that it is somehow going to lead to artificial general intelligence. By itself, it is not,” Brooks said.

Following Brooks’ presentation, a group of MIT faculty spoke about their work using generative AI and participated in a panel discussion about future advances, important but underexplored research topics, and the challenges of AI regulation and policy.

The panel consisted of Jacob Andreas, an associate professor in the MIT Department of Electrical Engineering and Computer Science (EECS) and a member of CSAIL; Antonio Torralba, the Delta Electronics Professor of EECS and a member of CSAIL; Ev Fedorenko, an associate professor of brain and cognitive sciences and an investigator at the McGovern Institute for Brain Research at MIT; and Armando Solar-Lezama, a Distinguished Professor of Computing and associate director of CSAIL. It was moderated by William T. Freeman, the Thomas and Gerd Perkins Professor of EECS and a member of CSAIL.

The panelists discussed several potential future research directions around generative AI, including the possibility of integrating perceptual systems, drawing on human senses like touch and smell, rather than focusing primarily on language and images. The researchers also spoke about the importance of engaging with policymakers and the public to ensure generative AI tools are produced and deployed responsibly.

“One of the big risks with generative AI today is the risk of digital snake oil. There is a big risk of a lot of products going out that claim to do miraculous things but in the long run could be very harmful,” Solar-Lezama said.

The morning session concluded with an excerpt from the 1925 science fiction novel “Metropolis,” read by senior Joy Ma, a physics and theater arts major, followed by a roundtable discussion on the future of generative AI. The discussion included Joshua Tenenbaum, a professor in the Department of Brain and Cognitive Sciences and a member of CSAIL; Dina Katabi, the Thuan and Nicole Pham Professor in EECS and a principal investigator in CSAIL and the MIT Jameel Clinic; and Max Tegmark, professor of physics; and was moderated by Daniela Rus.

One focus of the discussion was the possibility of developing generative AI models that can go beyond what we can do as humans, such as tools that can sense someone’s emotions by using electromagnetic signals to understand how a person’s breathing and heart rate are changing.

But one key to integrating AI like this into the real world safely is to ensure that we can trust it, Tegmark said. If we know an AI tool will meet the specifications we insist on, then “we no longer have to be afraid of building really powerful systems that go out and do things for us in the world,” he said.

The brain may learn about the world the same way some computational models do

To make our way through the world, our brain must develop an intuitive understanding of the physical world around us, which we then use to interpret sensory information coming into the brain.

How does the brain develop that intuitive understanding? Many scientists believe that it may use a process similar to what’s known as “self-supervised learning.” This type of machine learning, originally developed as a way to create more efficient models for computer vision, allows computational models to learn about visual scenes based solely on the similarities and differences between them, with no labels or other information.

A pair of studies from researchers at the K. Lisa Yang Integrative Computational Neuroscience (ICoN) Center at MIT offers new evidence supporting this hypothesis. The researchers found that when they trained models known as neural networks using a particular type of self-supervised learning, the resulting models generated activity patterns very similar to those seen in the brains of animals that were performing the same tasks as the models.

The findings suggest that these models are able to learn representations of the physical world that they can use to make accurate predictions about what will happen in that world, and that the mammalian brain may be using the same strategy, the researchers say.

“The theme of our work is that AI designed to help build better robots ends up also being a framework to better understand the brain more generally,” says Aran Nayebi, a postdoc in the ICoN Center. “We can’t say if it’s the whole brain yet, but across scales and disparate brain areas, our results seem to be suggestive of an organizing principle.”

Nayebi is the lead author of one of the studies, co-authored with Rishi Rajalingham, a former MIT postdoc now at Meta Reality Labs, and senior authors Mehrdad Jazayeri, an associate professor of brain and cognitive sciences and a member of the McGovern Institute for Brain Research; and Robert Yang, an assistant professor of brain and cognitive sciences and an associate member of the McGovern Institute. Ila Fiete, director of the ICoN Center, a professor of brain and cognitive sciences, and an associate member of the McGovern Institute, is the senior author of the other study, which was co-led by Mikail Khona, an MIT graduate student, and Rylan Schaeffer, a former senior research associate at MIT.

Both studies will be presented at the 2023 Conference on Neural Information Processing Systems (NeurIPS) in December.

Modeling the physical world

Early models of computer vision mainly relied on supervised learning. Using this approach, models are trained to classify images that are each labeled with a name — cat, car, etc. The resulting models work well, but this type of training requires a great deal of human-labeled data.

To create a more efficient alternative, in recent years researchers have turned to models built through a technique known as contrastive self-supervised learning. This type of learning allows an algorithm to learn to classify objects based on how similar they are to each other, with no external labels provided.

“This is a very powerful method because you can now leverage very large modern data sets, especially videos, and really unlock their potential,” Nayebi says. “A lot of the modern AI that you see now, especially in the last couple years with ChatGPT and GPT-4, is a result of training a self-supervised objective function on a large-scale dataset to obtain a very flexible representation.”

These types of models, also called neural networks, consist of thousands or millions of processing units connected to each other. Each node has connections of varying strengths to other nodes in the network. As the network analyzes huge amounts of data, the strengths of those connections change as the network learns to perform the desired task.

As the model performs a particular task, the activity patterns of different units within the network can be measured. Each unit’s activity can be represented as a firing pattern, similar to the firing patterns of neurons in the brain. Previous work from Nayebi and others has shown that self-supervised models of vision generate activity similar to that seen in the visual processing system of mammalian brains.

In both of the new NeurIPS studies, the researchers set out to explore whether self-supervised computational models of other cognitive functions might also show similarities to the mammalian brain. In the study led by Nayebi, the researchers trained self-supervised models to predict the future state of their environment across hundreds of thousands of naturalistic videos depicting everyday scenarios.

“For the last decade or so, the dominant method to build neural network models in cognitive neuroscience is to train these networks on individual cognitive tasks. But models trained this way rarely generalize to other tasks,” Yang says. “Here we test whether we can build models for some aspect of cognition by first training on naturalistic data using self-supervised learning, then evaluating in lab settings.”

Once the model was trained, the researchers had it generalize to a task they call “Mental-Pong.” This is similar to the video game Pong, where a player moves a paddle to hit a ball traveling across the screen. In the Mental-Pong version, the ball disappears shortly before hitting the paddle, so the player has to estimate its trajectory in order to hit the ball.

The researchers found that the model was able to track the hidden ball’s trajectory with accuracy similar to that of neurons in the mammalian brain, which had been shown in a previous study by Rajalingham and Jazayeri to simulate its trajectory — a cognitive phenomenon known as “mental simulation.” Furthermore, the neural activation patterns seen within the model were similar to those seen in the brains of animals as they played the game — specifically, in a part of the brain called the dorsomedial frontal cortex. No other class of computational model has been able to match the biological data as closely as this one, the researchers say.

“There are many efforts in the machine learning community to create artificial intelligence,” Jazayeri says. “The relevance of these models to neurobiology hinges on their ability to additionally capture the inner workings of the brain. The fact that Aran’s model predicts neural data is really important as it suggests that we may be getting closer to building artificial systems that emulate natural intelligence.”

Navigating the world

The study led by Khona, Schaeffer, and Fiete focused on a type of specialized neurons known as grid cells. These cells, located in the entorhinal cortex, help animals to navigate, working together with place cells located in the hippocampus.

While place cells fire whenever an animal is in a specific location, grid cells fire only when the animal is at one of the vertices of a triangular lattice. Groups of grid cells create overlapping lattices of different sizes, which allows them to encode a large number of positions using a relatively small number of cells.

In recent studies, researchers have trained supervised neural networks to mimic grid cell function by predicting an animal’s next location based on its starting point and velocity, a task known as path integration. However, these models hinged on access to privileged information about absolute space at all times — information that the animal does not have.

Inspired by the striking coding properties of the multiperiodic grid-cell code for space, the MIT team trained a contrastive self-supervised model to both perform this same path integration task and represent space efficiently while doing so. For the training data, they used sequences of velocity inputs. The model learned to distinguish positions based on whether they were similar or different — nearby positions generated similar codes, but further positions generated more different codes.

“It’s similar to training models on images, where if two images are both heads of cats, their codes should be similar, but if one is the head of a cat and one is a truck, then you want their codes to repel,” Khona says. “We’re taking that same idea but applying it to spatial trajectories.”

Once the model was trained, the researchers found that the activation patterns of the nodes within the model formed several lattice patterns with different periods, very similar to those formed by grid cells in the brain.

“What excites me about this work is that it makes connections between mathematical work on the striking information-theoretic properties of the grid cell code and the computation of path integration,” Fiete says. “While the mathematical work was analytic — what properties does the grid cell code possess? — the approach of optimizing coding efficiency through self-supervised learning and obtaining grid-like tuning is synthetic: It shows what properties might be necessary and sufficient to explain why the brain has grid cells.”

The research was funded by the K. Lisa Yang ICoN Center, the National Institutes of Health, the Simons Foundation, the McKnight Foundation, the McGovern Institute, and the Helen Hay Whitney Foundation.