Dream machines: how IT is changing the world of neuroscience
We talk to computer scientist and entrepreneur Jamil El Imad about the cutting-edge intersection of neuroscience and IT
Arriving home at her poky London flat after a difficult day, Emma closes the door on the outside world, hangs up her coat and tries to forget the stress she is under. She changes into comfy pyjamas, settles in on the couch under a soft blanket and, plugging in her smartphone, places a small virtual reality (VR) headset over her head.
Floating before her eyes is a menu and the words “Choose your dream”. She sees a range of scenarios: a Buddhist monastery high in the Himalayas, the bright white sands of a deserted Hebridean beach, a steaming Icelandic hot spring, a fragrant Californian redwood grove. With a nod, Emma selects an Alpine meadow, and enters the dream scenario.
At first she sees nothing but a drifting white mist, but as she relaxes, feeling the tension draining from her neck and shoulders, her heart rate slows, her breathing becomes shallower, and the fog begins to part. She sees first a carpet of wildflowers spreading out before her. As she concentrates, the mist rolls back to reveal the full scene. She looks up at the clear sky and sees birds overhead. She hears the mountain breeze and cowbells in the distance. A valley somewhere in Austria is spread out before her. Emma sits back on her sofa, and feels herself like a feather on the wind, a thousand miles from her troubles.
It sounds like the opening to a Philip K Dick novel or a treatment for the next season of Black Mirror, but actually the technology Emma might one day use exists right now in prototype form.
It’s called the Dream Machine, it’s designed to improve mindfulness and concentration, and it’s the brainchild of computer scientist and serial entrepreneur Jamil El Imad. It is the result of his work at Switzerland’s École Polytechnique Fédérale de Lausanne (EPFL) on the Human Brain Project, a multi-year programme that is bringing together researchers from across Europe to advance the fields of neuroscience and computing.
So how did a computer scientist become interested in neuroscience? The answer goes back to the 1970s, when El Imad left Lebanon for the UK to pursue a career in IT, working at first with IBM punch cards. Stints in the IT departments of oil giant Halliburton and French pharmaceuticals company Roussel followed (a job he found in the back pages of Computer Weekly), before a return to the Middle East in an IT services role in the early 2000s. However, by his own admission, by 2005 he was starting to get bored.
“I believe intelligence is biological, not logical”
Jamil El Imad
“I returned to the UK and rejoined academia at Imperial [College London],” he tells Computer Weekly over coffee around the corner from his Mayfair pad. “I wanted to work on something disruptive and, at that time, I took a very early interest in VR, and cloud was just starting to be talked about back in 2008.”
At first, El Imad perceived the cloud merely as a godsend for the IT department. “Most developments in the 1980s and 1990s were bespoke,” he says. “Nobody was taking a one-size-fits-all approach, and it was a huge effort to implement, unlike today. Now we have applications that allow every company to customise and use technology however they want.
“In my view, cloud has come as a saviour from IT overspend, when I think about the inefficiencies of managing large IT platforms, when you’re able to access computing on demand and not have to budget for peak. In my capacity planning, I always used to have to budget for a peak of maybe two hours a day, and for the other 22 hours, my machine was running at 20% capacity.”
The novelty of the cloud led El Imad to look at other emerging areas, such as VR and artificial intelligence (AI). But he says the penny really dropped about the link between computer science and neuroscience when IBM Watson beat Jeopardy.
El Imad proclaims himself somewhat sceptical about some of the grandiose claims made about AI, saying: “I believe intelligence is biological, not logical.”
He adds: “I can get a computer to read 1,000 books on a subject and answer my questions with cognitive computing, but that’s not intelligence, that’s intellect. I want to build applications that capitalise on that.”
Understanding the brain
Reduced attention span, poor sleep quality, increased anxiety levels – much has been written about the negative effects that having such a glut of technology at our fingertips can have on our brains, and the evidence against smartphones, social media and so on continues to mount.
El Imad says people are now overusing their mental capacity and that the human brain is not built to handle such a deluge of information, hence the growing mental health crisis around the world.
“We are overusing our mental capacity,” he says. “We are not able to handle so much multitasking, and we are constantly trying to move at the speed of information, rather than at our own speed. I think this is why you see a lot of mental health problems today.
“The numbers speak for themselves when it comes to the rise in conditions such as ADD [attention deficit disorder]. The teams I have worked on, computer scientists and physicians, all draw this conclusion.”
El Imad adds: “I was under the impression that we really understand the brain, but the truth is that nobody really knows how it works. The truth is that in the past 30 years, neuroscience has not advanced nearly as much as it should have. These are not my words, but the words of those who work in this field. This is why there must be a renewed effort to bring it back to public attention.”
He believes technology can both address the shortfall in research and resolve the global mental health crisis, hence his involvement in the field and the development of the so-called Dream Machine described above.
About 15 years ago, together with some investor partners, El Imad was instrumental in the formation of the Brain Forum, a charity designed to promote neuroscience and technology, bringing together computer scientists, clinicians and investors to exchange ideas and increase the sum of knowledge in this field.
Biosensor data to fight illness
At the same time, El Imad came to the realisation that in terms of digitisation, the healthcare sector was some way behind verticals such as financial services. This was the impetus behind the founding of NeuroPro, which he set up to try to change this.
“We decided to focus on brain signals and brain data because this type of content seemed to be very popular, not only in the health sector but also in a lot of applications around the internet of things,” he says.
Ultimately, the ecosystem that NeuroPro created can be described as like iTunes, but for biosensor data generated by electroencephalography (EEG), says El Imad. The VMLPro (Virtual Mobile Laboratory) cloud sits on Amazon Web Services (AWS) infrastructure and provides an as-a-service environment in which to store, manage and process EEG data in such a way that the datasets can be used to run remote monitoring on patients, perform remote diagnostics and collaborate with other clinicians.
The NeuroPro cloud enables clinicians to process EEG data on any device using very complex algorithms that it would not be possible for individual hospitals to run on their own, says El Imad.
“To give you an example, at a hospital in Switzerland that we work with, they have an algorithm that they need to run prior to performing brain surgery. It reads very complex brain signals and processes terabytes of data to try to tell them where the incision should take place. But it takes them six days for the algorithm to give them the output they need, subject to which they proceed with the surgery. We brought that down to six hours,” he says.
“That can save lives, hospital bookings, and not only that, but we can potentially make it, if they wish, a revenue generator, because if it is good enough for them, it must be good enough for other hospitals. That is the power of the cloud, and I would say that any enterprise, any CIO who is looking to transform, must look seriously at cloud. If you can overcome the challenges of security, the efficiencies can be considerable.”
With the VMLPro platform built at NeuroPro, El Imad is working with a number of institutions specialising in neurological conditions.
Among them is the Swiss Epilepsy Centre, which has been operating since the days when epilepsy was treated by locking people away for life, but is now one of the most advanced institutions in its field anywhere in the world.
“They are experimenting with it because at the moment a lot of their specialists have to travel to other hospitals to work with local doctors to analyse data and diagnose,” says El Imad. “Our cloud allows collaboration on the fly. It doesn’t matter where the patient is being monitored – now two doctors from two different hospitals can watch the EEG data in real time.”
But the NeuroPro cloud is about more than just enabling collaboration. El Imad is also working with researchers to use the data he has collected from epilepsy patients to allow machine learning algorithms to mine it for new information and patterns that humans might not be able to spot, which may in future be able to improve the patients’ quality of life, maybe even predict and ward off seizures.
“A lot of research on seizure prediction that we’ve worked on is to do with actually picking up environmental changes,” he says. “Circumventing the problem is as good as curing the problem.”
The implications of this volumetric pattern matching can be extended way beyond epilepsy. Today, a skilled clinician can recognise tiny changes in scans over time to detect cancer growing in the lungs, for example. But imagine a world where a constant flow of data and a powerful algorithm can detect even tinier changes in the body months before a doctor, or before the patient even notices they are ill. Ultimately, says El Imad, would you prefer to cure your cancer, or to never have it in the first place?
Biosensor data for peace of mind
The cloud has myriad novel applications, but is now well-established technology. VR is arguably older, but despite having been around in some form or another for the best part of three decades, it is only now beginning to come into its own – and it is here where El Imad is doing work that could change society, not just healthcare.
Back at the EPFL and with assistance from the Human Brain Project, El Imad and his team have devised something they call reality substitution, which uses video, not computer-generated images (CGI), to generate a VR environment.
“Basically, we transport you live to a new environment that is being filmed on a rig that we built using 14 GoPros,” he says. “We took the rig to a mountaintop to test it for people with a fear of heights. Imagine – if you have a phobia, you can film it and you can practise exposing yourself to it at home.”
This brings us back to the hypothetical Emma and her personal Dream Machine, which marries EEG data and VR to promote mindfulness and concentration.
By transporting the user to the environment of their choice, the program helps them to relax and concentrate by rewarding them. It does this by reading their EEG data to detect activity in the relevant parts of the brain, slowing heart rate, and so on.
Read more about healthcare innovation
- Healthcare organisations should be prepared to see applications for blockchain and more uses for AI in 2018. Analytics provided by EHR suppliers will also emerge.
- As healthcare executives begin to build their technology roadmaps, AI and advanced analytics, as well as continuing to focus on security, are essential components to include.
- The FCC’s vote to repeal net neutrality could hurt telemedicine providers and consumers and limit the ability of some healthcare systems to exchange data-intensive medical images.
When the biological conditions are right, the user is portrayed within the environment as a feather on the breeze, and the fog described earlier begins to part, revealing the scene to the user bit by bit. At the end of the session, they receive a personalised score, which they can then work to improve.
“You have no idea how relaxing it is,” says El Imad. “You don’t want to take it off.”
He envisages that one day the Dream Machine might be used to help employees de-stress in a busy office, or even in the gym. But wherever it is used, he is clear that it will be a product that is available for consumers to buy, and not a medical service.
“We are going down the non-clinical route and doing non-clinical trials because these tools [smartphones and headsets] are available in the home,” he says. “You don’t need specialist equipment to enjoy it.”
It seems ironic that in dealing with the problems that technology has caused for our mental wellbeing, it is technology that may ride to the rescue, and El Imad is all too aware of this. But, he says, the way the world and society now work lend themselves to a technological solution.
“What is the alternative?” he says. “It would be nice if we all had access to a nice park, but things are not like that – people live in high-rise blocks with no outdoor space. So why should they not have that sort of experience and that moment of pleasure?”