MERIT: A No-Code Platform for Unifying High-Precision Cognitive Assessment Across Laboratory and Real-World Settings
This paper introduces MERIT, a no-code platform that unifies high-precision cognitive assessment across laboratory and real-world settings by demonstrating that browser-based timing jitter can significantly distort neural and behavioral measures while providing a native mobile solution with sub-millisecond stimulus accuracy that ensures task consistency across environments.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
For decades, the study of how the human mind works has relied on a strict trade-off. To measure the speed of a thought or the precise moment a brain reacts to a sight, researchers needed specialized, expensive equipment locked inside a quiet laboratory. These machines could record events with millisecond accuracy, capturing the split-second electrical sparks of the brain as they happened. But this precision came at a cost: the experiments were rigid, difficult to set up, and impossible to take outside the lab. Conversely, the modern world offers a way to reach people anywhere, anytime, using the smartphones and tablets they carry in their pockets. Yet, these everyday devices have historically been too imprecise for serious science. The software that runs them, particularly web browsers, often introduces tiny, unpredictable delays—sometimes tens of milliseconds—that blur the very timing researchers need to measure. This gap has forced scientists to choose between high-quality data from a few people in a lab or lower-quality data from many people in the real world, but never both at once.
A team of researchers has now built a new system designed to close this gap, allowing the same precise experiment to run on a participant's own phone just as accurately as it would on a laboratory computer. They call this system MERIT. It is a tool that lets researchers design complex mental tasks without writing any computer code, using a visual builder similar to arranging blocks. Once a task is created, it runs on a powerful engine embedded directly into mobile apps, bypassing the slow and variable web browsers that have plagued previous attempts. The system is so precise that it can measure the time between visual cues with an error margin of less than one millisecond on modern devices, though slightly higher on older models. This level of accuracy is critical because the brain's electrical signals, which researchers often record alongside these tasks, happen in the blink of a second. If the timing of the task is off by even a fraction of a second, the connection between the behavior and the brain activity becomes impossible to interpret.
To prove their system works, the researchers first had to answer a fundamental question: how much timing error can a brain measurement actually tolerate before the data becomes useless? They took a large, public dataset of brain recordings and artificially injected random delays into the timing of the visual cues, simulating the kind of errors that happen on ordinary computers. They then measured how these delays affected three specific brain signals that appear when people see or hear something. They found that for a fast brain signal known as the N170, which occurs roughly 170 milliseconds after a face is seen, the timing error could not exceed about 15.8 milliseconds if the researchers wanted to keep their measurements reliable. Slower brain signals could tolerate a bit more error, up to 20 milliseconds or slightly more. Crucially, they discovered that the type of error mattered just as much as the amount. A system that makes rare, large mistakes is actually better for science than one that makes consistent, small mistakes, because the rare mistakes cost less to the analysis than the consistent errors do.
With these limits established, the team tested their new platform, MERIT, on a variety of real-world devices. They measured the timing precision of the system on several current smartphones and tablets, as well as an older budget phone. On the newest devices, the system was incredibly steady, with timing errors ranging from 0.71 to 0.96 milliseconds. Even on the older, slower phone, the error was only 3.73 milliseconds. Every single measurement was well below the 15.8-millisecond limit required for the most sensitive brain signals. This means that for the first time, a researcher can ask a participant to perform a cognitive task on their own phone at home, and be confident that the timing is precise enough to link that behavior to brain activity recorded in a lab. The system achieves this by running the task directly on the phone's operating system rather than through a web browser, and by using a specialized connection that sends timing signals to laboratory equipment in real time.
The platform is not just about timing; it is about flexibility and accessibility. Researchers can now build tasks that involve drawing, continuous movement, or complex sequences of choices, all without needing to hire a programmer. These tasks can be set up to run once, or to appear at specific times throughout the day and week, allowing scientists to study how attention and mood change over time in a person's natural environment. The system supports four main types of tasks: quick reaction tests, continuous control games where a person must steady a moving object, drawing exercises that capture the speed and pressure of a pen stroke, and sequential puzzles. Because the same task definition is used whether the participant is in a lab or at home, the data collected in the field can be directly compared to the brain and body data collected in the lab. This allows scientists to understand what a slow reaction time means in terms of brain activity, even when that reaction time is measured in a living room rather than a laboratory.
The implications of this work extend beyond just better data; they change who can do the science. By removing the need for coding skills and expensive hardware, the platform opens the door for a wider range of researchers to design and run high-quality studies. It also allows for new types of research that were previously impossible, such as tracking cognitive changes in patients over months or years, or studying how the brain works in the middle of a busy day. The system is designed to be sustainable, with the developers planning to keep it available to the scientific community through a model that supports long-term maintenance and updates. While no system is perfect, and the researchers acknowledge that older devices or specific operating system settings can still introduce small variations, the results show that the gap between laboratory precision and real-world convenience has been bridged. The same instrument that measures the brain in a controlled setting can now measure the mind in the wild, with a clarity that was once thought to be out of reach.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.