How ‘neural fingerprinting’ could analyse our minds
New brain-scanning technology is both tantalising and disturbing
John Thornhill
Everyone reading this article can congratulate themselves on their amazing abilities and good fortune.
You are the lucky owner of the most remarkable processing machine in the known universe: your brain.
The 3lbs of wetware inside your skull is the original neural network and vastly more efficient than the computational kind.
Whereas today’s monster data centres consume huge amounts of electricity, your brain operates on just 20 watts a day, which you can generate by eating a cheeseburger.
Even a junk food-fuelled brain is intuitively capable of doing things that the best AI model finds difficult: dreaming up transformative creative projects; distinguishing between a plastic bag and a rock on a motorway; and detecting irony, for example.
But in many ways neuroscientists remain mystified by how this network of 86bn neurons works.
Recent advances in quantum sensing and AI are opening up promising new avenues of research.
Yet, as is so often the case, the positive possibilities of this technology are accompanied by fears of potentially negative applications.
One of the most intriguing brain-scanning techniques is magnetoencephalography (MEG), a non-invasive way of mapping the electronic activity of the brain.
This methodology is tricky, though, given the need to chill superconducting sensors to -269C.
But the recent development of optically pumped magnetometers (OPMs), which do not require cryogenic cooling, has enabled researchers to gather equivalent data in 30 minutes from a wearable helmet containing 64 sensors.
Using these scanners, neuroscientists can create unique “neural fingerprints” of an individual’s brain.
Researchers contrast the difference between traditional brain scanners and OPM-MEG devices as that between looking at the stars through binoculars and the James Webb Space Telescope.
Matt Brookes, physics professor at the University of Nottingham and chair of Cerca Magnetics, a start-up that is pioneering OPM-MEG helmets, says that today’s commonly used magnetic resonance imaging (MRI) devices are good at finding structural abnormalities in the brain, such as “a hole or a lump or a bump” that may be a tumour.
But OPM-MEG devices can detect functional abnormalities, such as schizophrenia, epilepsy and dementia.
Neural fingerprints will give us a “huge amount of data that will be very, very useful for us to know”, Brookes tells me.
More than 10 research universities in Europe and North America are already using Cerca’s technology.
The company is also seeking regulatory approval from the US Food and Drug Administration to register its helmets as medical devices, an approval process that can take several years.
The aim is to create vast data sets of neural activity to deepen our understanding of brain disorders and lead to better medical care.
In future, Brookes suggests, patients may be routinely screened for early signs of dementia in the same way that women are regularly scanned for breast cancer.
But Brookes is also concerned by how neural fingerprinting could potentially lead to negative science fiction-type scenarios.
He speculates that we may soon be able to tell what kind of an upbringing a child has had from examining their brain function.
The brains of children from privileged families who benefited from experiential stimuli will develop in different ways from those from more deprived backgrounds.
“You could imagine how that could be incredibly useful.
You could also imagine how it could be completely misused,” he says.
Separate research from scientists at University of California, Los Angeles has demonstrated how a wearable, non-invasive brain computer interface can help paralysed patients by reading brain signals to move a robotic arm or a computer cursor.
Using a different technique, known as electroencephalography (EEG), the UCLA team enabled patients to translate thought into movement, assisted by an AI co-pilot.
Several private companies, including Elon Musk’s Neuralink, are developing invasive brain-computer interface (BCI) technology by implanting electrodes in the skull.
Last month, Musk’s rival Sam Altman, chief executive of OpenAI, launched a new venture called Merge Labs to compete with Neuralink.
And China has signalled its intent to become a world leader in BCI technology within five years.
Mind-reading may be the stuff of science fiction.
But a very primitive form of it may be approaching faster than we think.
Brain scans are rightly treated as confidential medical data subject to high standards of data protection.
Even so, expect the budding international movement to preserve our cognitive liberty to build momentum.
0 comments:
Publicar un comentario