Imagine being able to control machines by thinking. 

This communication link is known as a brain-machine interface and a new algorithm developed in Professor Brokoslaw Laschowski’s Computational Neuroscience Lab could soon make these interfaces more accurate and efficient. 

For brain-machine interfaces to work, an algorithm is needed to predict or “decode” human behaviour — such as speech or movement — from patterns of neural activity in the brain. This brain activity can be measured using functional MRI, electroencephalograms, or implanted electrodes, such as those developed by Neuralink. 

Today, brain-decoding algorithms exist, but they have significant limitations.

“Brain activity is highly subject-specific,” says Laschowski, a research scientist at the University Health Network, University of Toronto Robotics, and assistant professor (status) in the Faculty of Applied Science & Engineering. 

“Neural population activity in the brain varies considerably between and within subjects. That’s why building a universal brain-decoding algorithm is so challenging.” 

Most brain-decoding algorithms are optimized for individual subjects and tasks, requiring additional data collection and model retraining for each scenario, which is time-consuming and impedes clinical translation. Researchers in the Computational Neuroscience Lab are exploring ways to improve generalization.

“There’s an interesting phenomenon known as negative transfer,” says Laschowski. 

“In machine learning, the standard practice to improve model performance is to increase the size and diversity of the training dataset. However, due to negative transfer, increasing dataset diversity can sometimes degrade performance, leading to counterintuitive results where models trained on smaller datasets outperform those trained on larger ones,” he says. 

“This is why source selection for multi-subject brain decoding is important.” 

In a new study published on bioRxiv, Laschowski and Aidan Dempster (EngSci 2T5), now a PhD student in robotics at the University of Michigan, developed a new computational framework to minimize negative transfer in brain decoding by reframing source selection as a mixture model parameter estimation problem. This allows each source subject to contribute through a continuous mixture weight rather than being outright included or excluded. 

To calculate these weights, they developed a novel convex optimization algorithm based on the Generalized Method of Moments. By using model performance metrics as the generalized moment functions, their algorithm also more closely aligns with the mathematical foundations of domain adaptation theory, enhancing optimality guarantees. 

When tested on a brain-decoding dataset of more than 105 subjects, their algorithm achieved state-of-the-art performance while using 62% less training data, suggesting that performance gains stem from reduced negative transfer. 

“These findings challenge the dominant practice in machine learning, which focuses on developing and using large-scale datasets for training,” says Laschowski. 

“Our study shows that quality, not just quantity, is important when selecting source subjects to train a machine learning model for brain decoding.” 

A preliminary version of their algorithm was awarded Best Poster Award at the 2024 Toronto Robotics Conference.

These brain decoding algorithms can be used for a variety of applications.

One such example is a new interdisciplinary collaboration between Laschowski and Professor Hugh Liu (UTIAS), exploring how brain-decoding algorithms can be used to control and interact with autonomous drones. 

“My lab specializes in computational neuroscience, and his lab specializes in autonomous flight systems,” says Laschowski. “Together, we’re exploring how to combine our expertise to build something novel and advance our understanding of brains and machines.” 

In addition to brain-machine interfaces, his algorithms are also being used to support research in computational neuroscience, such as studying the underlying mechanisms and computations in the brain that give rise to the mind.

“What is the mind? What is thinking? Can we build an artificial brain? These are the sorts of grand questions that drive my research program. Our long-term mission is to reverse-engineer the human brain and discover fundamental principles of learning and intelligence,” says Laschowski.

“Understanding how the brain works is perhaps the greatest scientific question of all time.”

When it comes to mental wellness, support from family is often vital. Now, one family in particular is working to extend that support across the entire U of T Engineering community. 

Loui Pappas (CivE 8T8, MASc 8T9) and Sandra Gionas are the parents of Elena Pappas (ChemE 2T2). Inspired by her father’s example and her experiences at outreach programs such as the DEEP Summer Academy, Elena set her sights early on a career in engineering. 

“I had great summers at those programs, and I really got hooked on the idea of solving complex problems through math and science,” she recalls. 

“I was always dead set on being an engineer.” 

After successfully achieving admission to U of T Engineering, Elena quickly made lots of friends among her fellow first year students. She enjoyed being part of the community but she soon found that the actual classwork was presenting more of a challenge than she had expected. 

“I was really struggling to keep up, and during my final exam in first year, I had a full-blown panic attack,” she says. 

“I failed that course, and others. I wasn’t able to meet the goals I set out for myself.” 

Despite these challenges, her father Loui says that withdrawing was never an option for her. 

“It was difficult seeing your child struggle that first year, but Elena was adamant that she was going to get through it somehow, and with help and by her own resilience, that’s exactly what she did,” he says.  

Thanks to a referral by a friend, Elena ended up receiving a diagnosis of attention-deficit/hyperactivity disorder (ADHD). The discovery enabled her to better understand where her challenges were coming from and to develop new learning and coping strategies. 

She began second year as a part-time student, and soon was back on track, hunkering down the next few years and ultimately completing her degree in 2024. Today, she works in Calgary, Alberta as a Junior Project Coordinator at Lanmark Engineering Inc. 

After graduating from U of T Engineering, Loui Pappas had spent more than three decades as a business leader at Morrison Hershfield, a Canadian-based multidisciplinary engineering firm originally founded in 1946 by two U of T Engineering professors. The firm was acquired by Stantec in early 2024 — and as a longstanding shareholder, Pappas now had the means to make a new kind of legacy contribution to U of T Engineering. 

“Throughout my career, I have always enjoyed maintaining a relationship with and volunteering at the Faculty,” he says. 

“I’ve spoken to students at career fairs, participated in the Faculty’s broad-based assessment program, contributed to scholarships and many other fundraising initiatives . I have also seen the impact to students that alumni have in giving back, and I always knew I too wanted to create a legacy at U of T Engineering. Elena’s experience again became top of mind.” 

Sandra, a Trinity College graduate and former TV Ontario journalist, had covered mental health extensively in her research and reporting. 

“It’s one of the most overlooked areas of health,” she says. “Cancer and heart disease raise a lot of money, but mental health doesn’t always inspire the same level of generosity. Elena’s generation has done tremendous work in breaking the stigma, and we wanted to contribute to that momentum, and recognize her efforts specifically.” 

The family found the perfect fit in Skule™ Mental Wellness a student-run initiative within the Engineering Society that Elena herself once helped to establish. 

Together, they are strengthening the Skule™ Mental Wellness Bursary, which was created by students, for students, back in 2021. The bursary provides financial aid to students facing challenges related to mental health. The award has helped dozens of students so far, but demand for the bursary far outpaces the current funding. 

“The more we got into this, and had the chance to discuss the issue with Dean Chris Yip, the more we realized that immediate need is there in supporting mental health,” says Gionas.  

“We are also strongly supportive of the fact that, as a student-run initiative, the students themselves can rally behind and socialize it to further grow the bursary. 

“We’ve provided $20,000 initial funding for operational costs for two years, so that more students have immediate access to much needed counseling and support services. On top of that, we’ve also created a matching campaign to build up the bursary for the future, so that no students have to be turned away in the long term”.  Donations to the bursary are open now, and will be matched by the Pappas family up to a maximum of $50,000 over the course of the next year.  

“We know that the best and brightest kids from Canada and around the world are coming to U of T Engineering, the #1 program in Canada.” says Loui. 

“We want to help keep them in the faculty, and not have them leave the profession because of challenges that can be overcome with support and perseverance. I personally want to know that I did whatever I could to leave my engineering profession in good hands.” 

For Elena, the initiative closes the circle on her own experience. “At the end of the day, what students are dealing with when studying Engineering isn’t just school, it’s their lives,” she says. 

“This project is about making sure that they can get where they want to. That’s been a huge theme in my personal story. It just makes me feel incredible that future students will be able to get that help as well.” 

Join the Pappas family in supporting engineering student mental health. Thanks to the generosity of the Pappas family, all donations will be matched dollar-for-dollar up to a campaign total of $50,000, and new monthly gifts will be matched for one year.  

LaShawn Murray (MIE PhD student) is exploring how digital tools can improve health care delivery in First Nations communities, with a focus on reducing administrative burdens for nurses.

She is collaborating with OKAKI, an Alberta-based public health informatics social enterprise that delivers health services, professional consulting in privacy, analytics and program management, and develops custom software — with a focus on improving Indigenous health systems.

Murray is assessing the usability of the company’s novel AI scribe and its potential impact on clinical documentation workflows. AI scribes capture conversations between providers and patients to generate structured, regulatory-compliant clinical notes.

“Nurses have a high workload, especially on reserve, and documentation takes up a large part of their time,” says Murray, who is also a 2023 recipient of the IBET Momentum Fellowship in the Faculty of Applied Science & Engineering.

“They’re responsible for recording patient encounters and completing various forms to access services and resources.”

In remote or underserved Indigenous communities, nurses are often the first — and consistent — point of contact for primary care. With less frequent physician visits due to demand or shortages, nurses play a critical role in delivering day-to-day health services, including chronic disease management, homecare and emergency response.

“What’s different in OKAKI’s case is the focus on nurses and homecare professionals, rather than physicians working in hospitals or specialty settings,” says Murray. “There’s almost no data on how these tools impact nurses.”

As part of her dissertation, Murray recently began visits to four pilot sites in First Nations communities near Calgary and Edmonton. She’ll evaluate how the AI scribe is integrated into clinical workflows, assess the quality of the generated notes, and explore its effect on provider–patient interactions.

Her assessment will also look at technical performance, such as how long the scribe takes to launch, its ability to distinguish between voices, how much of the clinical encounter it captures, whether it saves time, and how well it documents care services, including medication names and acronyms.

“There’s also the question of how we communicate using accessible language with patients, while still needing to document using medical terminology, and if there is a way for the scribe to support that differentiation,” says Murray.

The work builds on earlier research conducted in collaboration with Women’s College Hospital in Toronto, where she evaluated the effectiveness of AI scribes in primary care as an IBET Fellow.

In a recent article published in JMIR Human Factors and co-authored with her supervisor, Professor Enid Montague (MIE), the team studied six AI scribes used at the hospital. The study offered one of the first systematic evaluations of AI scribes developed for — or currently in use in — Canadian primary care. Some tools assessed are also used in the United States.

Findings showed that AI scribes can help reduce administrative burdens by making them accessible across multiple platforms — such as mobile devices, tablets and desktop computers — which better align with the diverse workflows and preferences. Researchers also found limitations when it came to accuracy, such as multiple scribes’ ability to translate unrelated conversations, multiple speakers, and the complexity of conditions discussed.

The study also showed that AI scribes could generate high-quality medical notes even from transcripts that were not perfect, while in some cases, excellent transcripts did not lead to equally strong medical notes.

The study has given Murray valuable insights into how to apply her findings in the OKAKI project, and will continue with data collection for OKAKI over the next five months.

She emphasizes that the work is rooted in community collaboration.

“It’s a highly collaborative process; research done with communities, not for them,” says Murray.

“We’re co-designing the research, looking at what informed consent should look like, what language to use, and how to ensure it’s accessible to the people we’re working with.”

For the first time, U of T Engineering undergraduates in electronics courses will have access to automated pick and place machines to assemble printed circuit boards (PCBs) just as they do in industry.

The new laboratory space in the Myhal Centre for Engineering Innovation & Entrepreneurship will primarily serve students from the Department of Mechanical & Industrial Engineering and The Edward S. Rogers Sr. Department of Electrical & Computer Engineering.

PCBs with surface-mounted components are found in almost all electronic devices. They are essential components in a wide range of electronic devices and systems, including consumer electronics, such as smart phones and televisions. They are also used in medical devices, such as neurostimulators and magnetic resonance imaging (MRI) machines, and automotive systems, such as engine control units and advanced driver-assistance systems. The proliferation of PCBs with surface-mounted components has been made possible by their low cost of production with equipment showcased in this lab.

“Part of our goal is to introduce our students to how these components are placed onto PCBs, and that’s what these pick and place machines do,” says Tomas Bernreiter, Laboratory Engineer & Manager, MIE.

“Surface mount components are very small, often only a few millimeters in size, and therefore require specialized handling and placement equipment.”

The new PCBP Laboratory expands the soldering and electronic testing capabilities of the Myhal Fabrication Facility. Jointly funded by MIE, ECE and the faculty through the Dean’s Strategic Fund grants, it has 35 dedicated seats for electronics assembly and testing.

It will support hands-on lab components in MIE and ECE courses, including MIE366 and ECE295, where students build and populate circuit boards. The lab will also be available to student design teams for this purpose.

One of the first courses running design assignments in the new space is MIE346: Analog and Digital Electronics for Mechatronics, in which students design and build a variable power supply.

“Students design the electronic circuits for their project and then use the new lab to build and test the physical circuit board. The automated pick and place machines will place the small components onto a printed circuit board, after which the boards will be ready for solder reflowing,” says Bernreiter.

“They will then use special ovens in the new lab to reflow the solder to bond the surface mount components to the PCBs.”

The faculty aims to centralize all electronics and soldering work in this new space, freeing up the Myhal Light Fabrication Facility for more mechanical-based work.  Future MIE electric vehicle-based labs in power electronics will also utilize the pick and place machines.

“This is going to be the first time that undergraduate students are going to interact with pick and place machines,” says Bernreiter. “That’s something novel for the university.”

As the newest robotics researcher at the University of Toronto Institute for Aerospace Studies (UTIAS), Professor Nick Rhinehart and his team are developing autonomous systems that can operate safely and effectively in complex, unpredictable environments, such as roads and public spaces.  

Rhinehart joined the university in 2024 from Waymo, one of the leading autonomous ride-hailing services, where he was a member of the research team working on data-driven simulation and optimal driving. 

With the start of a new academic year, we sat down with Rhinehart to learn more about his research, what he’s most excited about and advice for students considering  graduate studies or a career in robotics. 

Please tell us a little bit about yourself.  

I’m from a small town in Pennsylvania. I first became interested in computer science as an undergrad at Swarthmore College, because it felt like the study of problem-solving itself. Over time, that curiosity evolved into a deeper interest in machine learning and robotics: how to build systems that perceive, learn and make decisions in the real world. 

Since then, I’ve worked as a researcher in both academic and industry settings, including as a grad student at Carnegie Mellon University, a postdoctoral researcher at UC Berkeley’s AI Research Lab and a senior research scientist at Waymo Research. Now I’m delighted to be at the University of Toronto as an assistant professor, where I lead the LEAF Lab (Learning, Embodied Autonomy, and Forecasting Lab). 

Why did you choose U of T? 

U of T is world-class when it comes to robotics and AI, and I was excited by its strong interdisciplinary culture. The chance to collaborate with brilliant researchers across engineering and computer science through the University of Toronto Robotics Institute was a big draw. Toronto itself was another draw because it’s vibrant, multicultural and full of the energy of people pursuing many different lifestyles and careers. Plus, I can easily visit my family back in Pennsylvania, and I genuinely enjoy the weather here (usually). 

What is the main research goal of your lab?  

At the LEAF Lab, we aim to develop broadly useful autonomous systems that efficiently and safely operate in complex environments by advancing the algorithmic foundations of robot learning. Our work combines learning, forecasting and control by teaching systems to anticipate changes in the world, adapt their behaviour and act intelligently. 

One of our specific interests is on learned model-based control methods, which learn to forecast the future from examples and can be used to plan robot behaviour. Some of my prior work on model-based methods is in the context of autonomous driving, where explicitly forecasting what might happen next is crucial to being safe and effective; it’d be very difficult to drive if people on the road weren’t at least somewhat predictable. 

Even if robots were perfect at forecasting and planning, it’s still a challenge to communicate to robots precisely how we want them to behave. This is another of my group’s interests — reward learning — which is about using information from people to model the essence of what we want robots to do. Broadly, we want robots to combine forecasting and reward learning so they can plan ahead, do what people actually want and improve with experience. 

What excites you most about being a faculty member at U of T? 

I’m most excited about building a thoughtful and collaborative lab of excellent researchers who want to solve tough problems and become leaders in robotics and machine learning. Being able to do that at a place like U of T, surrounded by talented students and colleagues, is an amazing opportunity. 

What advice do you have for students who are interested in pursuing a career or graduate studies in robotics? 

Read widely, build things, break them, and figure out why they broke. Get into research early if you can. Familiarize yourself both with recent research trends and foundations from the past. Always be on the lookout for hidden assumptions and insights. Ask a lot of “why?” questions (and try to answer them!) Develop opinions that you can back up with evidence, but keep an open mind. Dream big! 

What is something most people might not know about you? 

One of my favorite Wikipedia articles is “Timeline of the Far Future.” I like tracing the predictions back to their scientific roots, as well as thinking about which are inevitable, which seem unlikely, and which might mainly depend on choices people make.