“We know that every patient is different. No two patients with dementia are the same,” says Toronto neurologist and entrepreneur Nardin Samuel. And yet, she adds, they may end up with similar treatment plans — with varying results. By harnessing data and AI, Samuel is hoping to decode not only why conditions vary so widely between individuals but also why certain therapeutics might work for some people and not others.
At Cove Neurosciences, Samuel and co-founder Irene Harmsen have developed an advanced platform that analyzes brain imaging data from electroencephalography (EEG), magnetic resonance imaging (MRI) and magnetoencephalography (MEGs) from patients with various motor, mood and memory disorders. The aim, says Samuel, is to help researchers identify biomarkers associated with neurologic and psychiatric disorders, speed up clinical trials and pave the way for personalized treatment.
Mental health is a growing concern — it’s estimated one in five Canadians experience mental illness every year. How can data unlock our understanding of different disorders and potentially improve treatment?
If you think about something like depression, you give a patient a medication, and you say come back in four to six weeks and we’ll reassess. There are no really good standard objective ways of measuring that. But if the patient has some sort of EEG metric or a proprietary biomarker that’s measured, however, you can try to track their response to therapy. We do this in every other area of medicine. When a patient takes medication for diabetes or cholesterol you can measure the biomarkers. We started Cove to figure out how we can actually harness brain data to figure out ways to personalize treatment for neurologic and psychiatric disorders.
What disorders are you looking to tackle first?
Currently, we’re focusing on Alzheimer’s disease; we hope to implement these biomarkers in a clinical trial setting. One of the things that I’m so excited about clinically and academically is that we’re working to understand the functional connectivity in the brain as it relates to disease. This could have major market implications for therapeutic development and therapeutic targeting.
What sort of brain data are you using?
We pride ourselves in being multimodal — combining data from different sources. We’re applying for a pretty substantial grant to help with the integration of multimodal imaging with EEG and MRI. You really need to learn how to separate signal from noise — what is actually real biologic data. That’s not easy and it’s something that we’re working on.
Cove’s work is mainly focused on supporting clinical research and development. Could the data be integrated with wearable technology for personalized health care and monitoring? What are some of the challenges with wearables?
That’s something that we’re investigating. We have been working with groups in the U.S. on how we can utilize remote devices so that it decentralizes the data collection process. So you have patients with epilepsy or stroke or something along those lines and you can actually collect data at home.
With any wearables for personalized health care and monitoring, there are a couple of challenges. The first being the hardware itself. There’s variability in the quality of the devices and the quality of data these devices are gathering. If you think of a smartwatch, you can get an Apple Watch or you can get a really cheap watch that doesn’t have the same level of functionality. Then there’s a software component and how that data is processed and transmitted from the wearable. Is it in real time? And what are the different algorithms to analyze that data? The third aspect is the user. In the case of EEG, if there’s someone putting on an EEG cap, there’s operator variability in terms of accuracy. If it’s something that’s wearable at home, then there could also be variability in how someone is putting it on. These things are really important — you need to ensure there’s consistency in the data. So you want quality hardware devices that are generating solid data and also good software to process that data.
Is there a particular neurological or neuropsychiatric issue you wish we had more control over?
Mental health and physical health are so important to living a full and meaningful life. If you had some mechanism or a tool to modify multiple different diseases at once rather than one drug for this and one drug for that — that could be very powerful. I think that neuromodulation is a very exciting area of neuroscience. It’s heading in the direction where you could actually treat psychiatric disorders, neurological disorders with the same tool. So can you eradicate multiple birds with one stone? Maybe.
You are working toward better treatments. But what do you suggest people do to maintain their own brain health?
Some of the best things you can do for yourself are spending time outdoors, getting vitamin D, getting sunshine and moving your body. Humans are social beings, so interacting with others, having a community, and having meaningful human connections with family and friends is so important. One thing that is very underrated — especially in the medical community — is sleep. So there are many neurologic and other physiologic disorders that are associated with lack of sleep. It’s very important to get a good night’s rest as often as possible. That really can’t be emphasized enough.
Ana Fonseca writes about technology for MaRS. Torstar, the parent company of the Toronto Star, has partnered with MaRS to highlight innovation in Canadian companies.
