Nouvelles
Portiloop: a potential brain-wave trainer to boost memory
It’s a device that could one day improve our concentration, or even slow cognitive decline. In an article published in the scientific journal PLOS ONE, a team from Polytechnique Montréal has just unveiled Portiloop, a tool that scans brain waves for specific signatures in real time and then stimulates the brain in order to “train” it.

With the Portiloop, we will soon be able to know if the amplification of fast brain waves helps to improve memory processes during sleep. (Photo: MIST Laboratory)
While we sleep, our brain goes through different phases, each with its own function. During light sleep, for example, the brain stores in its long-term memory what it has learned throughout the day. Deep sleep, meanwhile, enables the brain to rid itself of cellular debris and carry out other memory activities.
During each of these phases, the brain produces specific brain waves that act as actual signatures of what is happening. Scientists study them through electroencephalography (EEG), by attaching electrodes to the head of a guinea pig. Much more than just a way to recognize the phases of sleep, these signatures indicate which precise processes are taking place between neurons.
Slow waves of about 1 hertz (Hz), for example, signal that a memory consolidation process is taking place. They can even be interacted with by playing short pink noises (similar to white noise) at the exact moment they are produced in order to increase the amplitude of these waves in the brain. This process, known as “closed-loop deep-brain stimulation,” improves participants’ memory capacity.
To gain precision and improve brain feedback approaches, researchers in this field are now looking to interact with faster waves, those of a brief event called a “sleep spindle,” which lasts only one to two seconds and occurs suddenly, without warning. Experts believe that this could further improve the memory process, but this remains to be tested.
The problem is that the tools available to researchers are not fast enough to identify these signals and then react by sending a signal. What’s more, they are expensive (more than $30,000) and complex to use.
But this situation is changing.

Professor Giovanni Beltrame (Photo: Caroline Perron)
A neural network that could train another
In collaboration with a research group in the Department of Psychology at Concordia University, the team led by Giovanni Beltrame, a full professor in Polytechnique Montréal’s Department of Computer Engineering and Software Engineering, has designed a device that rivals and even surpasses the performance of the tools currently available to neurophysiology researchers. All this for only about $500 in parts and materials!
The group unveiled its device in this article published in the scientific journal PLOS One on August 22, 2022.
The Portiloop, as it is called, fits on the corner of a table. Thanks to its integrated electroencephalogram, it recognizes the early signs of a “sleep spindle” and reacts in about 0.25 seconds by sending a signal, even before the end of the brief event.
It does this by using an artificial neural network—artificial intelligence—trained on data from 180 individuals. It now detects sleep spindle events within the first milliseconds of their occurrence, giving it enough time to react and then send a sound signal to the brain to potentially increase the amplitude of brain waves.
“This is the first time that a research group has been able to stimulate sleep spindles in this way,” says Professor Beltrame, who wrote the paper, together with the team led by his spouse at Concordia University, Professor Emily B.J. Coffey. “We collect data, analyze it and stimulate in a quarter of a second.”
According to Professor Beltrame, this innovation opens the door to new research possibilities, including those related to the cognitive decline that comes with old age. “Experts have noticed that the amplitude and frequency of sleep spindles decrease with age,” Professor Beltrame explains. “With our tool, they will be able to see if it is possible to compensate for this by stimulating the sleep spindles to improve memory.”
Eventually, such a tool could find its way into bedrooms to improve the brain’s memory capacity during sleep, but we are not there yet. The Portiloop should find its place in neurophysiology laboratories. Of the five prototypes assembled so far, two are already in laboratories at Concordia and McGill universities, where they are being tested.
The group of researchers from Polytechnique and Concordia is working to improve its artificial intelligence tool so that it can adapt to each new patient.
“EEG signatures change from one individual to another,” Professor Beltrame explains. “We’d like to make it even more efficient.”
In the meantime, his team is making the entire Portiloop assembly process freely available to provide laboratories around the world with a low-cost solution to boost neurophysiology research.
Learn more
The article: “The Portiloop: A Deep-Learning-Based Open Science Tool for Closed-Loop Brain Stimulation” published in PLOS One
MIST Laboratory website
Department of Computer Engineering and Software Engineering website (in French)
Professor Giovanni Beltrame’s expertise
Professor Emily Coffey’s expertise