Brain Boost: AI Personalizes Stimuli And Increases Attention And Focus At Home

This study developed a technology that uses artificial intelligence to personalize electrical neurostimulation in the brain, aiming to improve attention for long periods. The novelty is that the system adapts to each person (taking into account, for example, head size) and can be used at home. Tests have shown that it is especially effective in people with lower cognitive performance, without causing side effects. This approach can make cognitive enhancement more accessible, safe, and efficient in everyday life.

Sustained attention is the ability to maintain focus for long periods, such as when driving a car, studying, or working. This skill is essential for good performance in various daily tasks. When it fails, it can lead to accidents, errors at work, and even learning problems.

Furthermore, difficulties with sustained attention are common in several health conditions, such as ADHD, depression, schizophrenia, Alzheimer's, and even in people recovering from COVID-19 for a long time.

To improve this cognitive function, scientists have been testing different strategies, from meditation techniques and physical exercises to medications and mental training. One promising technique is neurostimulation, which uses very weak electrical currents in the brain to improve neural function.

A specific type, called tRNS (transcranial random noise stimulation), sends electrical signals at varying frequencies to gently stimulate the brain. The idea is that this improves communication between neurons, facilitating focus and attention. Unlike more invasive methods, tRNS is painless, generally safe, and can be used at home.

Despite its benefits, neurostimulation studies don't always yield consistent results. One problem is that many studies use a one-size-fits-all protocol, the so-called "one size fits all," ignoring the fact that each person has different characteristics, such as head shape, brain type, and initial attention span.

Another challenge is that most tests take place in laboratories, controlled environments that don't reflect the distractions and demands of the real world. This makes it difficult to apply the results outside the laboratory, such as at home or at work.

To overcome these limitations, researchers have created a system that uses artificial intelligence (AI) to personalize neurostimulation. This system adapts the electrical current parameters based on two main factors: the person's initial cognitive performance and their head size (measured by head circumference). This helps better approximate the brain's actual conditions and optimizes the effects of the stimulation.

This technique is called Personalized Bayesian Optimization (pBO). Over time, the system learns from data collected from multiple users, continually refining its parameters to achieve better results for people with similar profiles in the future.

Furthermore, this system allows neurostimulation to be performed at home, which brings many advantages. It eliminates the need for laboratory visits, reduces costs, and increases comfort. It also improves the validity of studies because the tests are conducted in a real environment, where the person actually lives, works, and plays.

In this study, healthy volunteers used the system at home while performing a specific sustained attention task. During the task, they received transcranial high-frequency random noise stimulation, a form of electrical stimulation that activates sodium channels in neurons, which helps transmit signals between them.

Previous studies have shown that this technique can improve attention in controlled environments. An important concept here is stochastic resonance, which suggests that a certain amount of "noise" (in this case, electrical noise) can, paradoxically, help the brain function better, as if it were a gentle nudge to break out of inertia.

People with lower cognitive performance tend to benefit more from this stimulation, as their brains are in less optimal condition.

The study had three main stages:

- Development of the pBO algorithm, which is the brain of the system, to discover which transcranial random noise stimulation parameters improve attention.

- Computer testing (in silico modeling) to compare this algorithm with other strategies, such as trial and error or non-personalized optimizations.

An experiment with volunteers, comparing three groups: one with personalized transcranial random noise stimulation (pBO-tRNS), one with standard tRNS (1.5 mA for all), and one with sham stimulation (placebo).

Participants completed a 20-minute air traffic control task designed by the U.S. Air Force Research Laboratory37 without stimulation, followed by 20 minutes of personalized neurostimulation.

The results showed that the pBO algorithm found an inverted-U relationship between current intensity and performance. This means that both very weak and very strong currents are less effective; there is a sweet spot of intensity that yields the best results.

And this sweet spot varies depending on the person's initial performance. People with larger heads required stronger currents, which confirms previous findings that show the current needs to travel through more tissue to reach the brain.

In computer tests, pBO performed better than the other techniques, even when the system had to deal with noisy (less accurate) data. When there is a lot of noise in the data, the AI needs to test more combinations to be sure of the best parameters, as if it needs more experimentation before it can be confident.

In the experiment with participants, the results showed that personalized transcranial random noise stimulation (pBO) was significantly more effective for people with low attention performance at baseline.

For these people, sustained attention improved more than with standard transcranial random noise stimulation or placebo. However, in people who already performed well, the stimulation made no difference.

These findings support the idea that neurostimulation is more beneficial for those in suboptimal cognitive condition, and reinforce that the effects are not random, but rather align with scientific theories such as stochastic resonance and the balance between neuronal excitation and inhibition.

This figure shows how a study tested the use of artificial intelligence (AI) to improve people's attention at home using electrical stimulation of the brain. First, participants received a brain stimulation device and a tablet with an attention test at home. They measured their head circumference and took a test to see how their initial focus was. This data was sent to an online AI platform, which calculated the best way to apply the electrical stimulation for each person. They then repeated the attention test, this time with personalized stimulation for 20 minutes. Graphs (b and c) show that the improvement in performance depends on the intensity of the stimulus, the person's initial performance, and head size. Graphs (d to g) compare different optimization methods and show that personalized AI (in red) performed better than random or generic methods, especially in people who had low attention performance to begin with. This suggests that tailoring stimulation with the help of AI can actually help those who most need to improve their focus.

It's important to highlight that the study found no significant side effects, and the current intensity did not influence the occurrence of discomfort. It was also observed that, even when performance improved, the speed and accuracy of responses were not negatively affected.

Another interesting point was that the pBO algorithm identified that head size directly influences the required current intensity. This confirms studies that have already shown that larger brains may require more current for the effect to reach the desired brain area.

Currently, this is typically done with expensive MRI scans. pBO offers a simpler and more affordable alternative, using only basic measurements and performance observations, without the need for expensive tests.

This type of technology has important implications for society. It aligns with UN goals to improve health and reduce inequalities by making cognitive enhancement accessible to more people, especially those who cannot access specialized clinics or laboratories.

To achieve this, it is essential that future applications consider ethical issues, such as protecting user data, ensuring the safe use of the technology, and ensuring fair provision for all.

Despite some limitations, such as the fact that the study was conducted online and that it could not use performance-based rewards (for ethical reasons), the results show that the combination of artificial intelligence and personalized neurostimulation works, is safe, and can transform the future of cognitive enhancement, both for healthy people and for the treatment of neurological diseases.

With further development, this approach could help people focus on their real lives, at home, work, or school, without relying on invasive or inaccessible solutions.

READ MORE:

Personalized home based neurostimulation via AI optimization augments sustained attention

Roi Cohen Kadosh, Delia Ciobotaru, Malin I. Karstens, and Vu Nguyen 

npj Digital Medicine, volume 8, Article number: 463 (2025)

https://doi.org/10.1038/s41746-025-01744-6

Abstract: 

Brain-based technologies for human augmentation face challenges in personalization and real-world translation. We present an AI-driven personalized Bayesian optimization algorithm that remotely adjusts neurostimulation parameters based on baseline ability and head anatomy to enhance sustained attention at home. Validated through in silico modeling and a double-blind, sham-controlled study, our approach aligns with MRI-based models and neurobiological theories, maximizing efficacy and enabling scalable, personalized cognitive enhancement and therapy in real-world settings.