Most of us have a vague intuition that something in the way an older person writes might carry information about how their mind is doing. The handwriting gets shaky, the letters drift, the pen lifts too often. But intuition is not measurement, and what a new study published in Frontiers in Human Neuroscience actually found is more specific, and more methodologically interesting, than the general idea that handwriting changes with age.
The headline finding is this: when researchers gave older adults with and without cognitive impairment two different writing tasks, one was not enough to tell the groups apart. The other was.
What the study did
Researchers at the University of Évora in Portugal recruited 58 older adults between the ages of 62 and 92, all living in care homes. Twenty had no diagnosis of cognitive impairment. Thirty-eight did. They performed writing exercises on a digital tablet using an active inking pen.
The tasks were structured in layers of increasing cognitive demand. First, participants were prompted to draw 10 horizontal lines within 20 seconds and make at least 10 dots on the paper during the same time frame. These were control tasks, designed to measure basic motor function rather than cognition. Then came the handwriting tasks: participants were asked to copy a sentence shown on a card, and to write a different sentence from dictation.
The digital tablet recorded not just the output, the words on the page, but the process: timing, pen pressure, number of strokes, pauses, vertical letter size, how long participants took before putting pen to paper. What the researchers were measuring was not legibility but the mechanics of how the brain organises a physical act in real time.
Why copying wasn’t enough
Results showed that neither of the pen control tasks could distinguish cognitive status between groups. As simple tasks, they mainly rely on basic motor control and may not be enough to reveal subtle differences that more cognitively taxing tasks can show. Copying tasks, which are more mentally demanding than pen control but less demanding than dictation, also didn’t show group differences but demonstrated a trend towards significance.
This is the finding that deserves more attention than it has received. Copying a sentence from a card is not a trivial act, but a supported one. The sentence is right there. The writer does not have to hold it in memory, process it from sound into symbol, or manage competing cognitive demands simultaneously. The motor system can carry out the task without the executive system being heavily involved. And so, in that condition, the two groups looked essentially the same.
What this means practically is that if someone is given only a copying task to assess their cognitive-motor function, they may appear to be doing fine when they are not. The deficit exists. The task just doesn’t reach it.
What dictation exposed
The results of the dictation tasks showed clear differences between the two participant groups. This could be due to the higher cognitive demand such tasks place on working memory and executive functions.
“Dictation tasks are more sensitive because they require the brain to do multiple things at once: listen, process language, convert sounds into written form, and coordinate movement,” said senior author Dr Ana Rita Matias, an assistant professor at the Department of Sport and Health at the University of Évora.
When a person writes from dictation, especially from a sentence they have not seen before, the brain cannot offload the cognitive work onto the visual environment. It has to hold the sentence in working memory while decoding the sounds, while planning the motor sequence, while executing it. Timing and stroke organization are closely linked to how the brain plans and executes actions, which depends on working memory and executive control. As these cognitive systems decline, writing becomes slower, more fragmented, and less coordinated.
The predictive variables that emerged from the data were precise. In the group with cognitive impairment, two predictors, start time and number of strokes, emerged as significant for the shorter sentence of the dictation task. For the more complex sentence three predictors, vertical size, start time, and total duration, were significant. Start time, in particular, reflects how long it takes the brain to initiate action after receiving the instruction. That delay is not random hesitation. It is the cost of cognitive load.
The sentence complexity effect
One detail in the findings points to something researchers will likely want to build on. Not all dictation sentences produced equal results. As the sentence became longer and linguistically more complex, more predictive variables came into view. A shorter, more predictable sentence placed less strain on working memory. A longer, less predictable one pushed harder on the executive system.
“Even within dictation tasks, differences can emerge,” Matias noted. “A longer, less predictable, or linguistically demanding sentence places greater strain on cognitive resources.”
This suggests that the sensitivity of the tool is not fixed. It scales with the cognitive demand of the task. Which in turn means that the design of the writing prompt matters as much as the measurement of the response. A well-designed task is not just harder; it is targeted at the specific cognitive resources most likely to be affected by early decline.
Why this matters for early detection
Cognitive decline is notoriously difficult to detect in its early stages. The tools clinicians currently rely on, standardised questionnaires, memory tests, verbal assessments, require either specialised administration or laboratory settings, and they often register decline only after it has already been progressing for some time.
The team said their approach, relying only on simple writing tasks and accessible digital tools, could serve as a practical way to monitor cognitive decline in a variety of settings, for example in doctors’ offices. Because it’s a non-invasive and relatively low-cost method, it could easily be integrated into routine clinical practice.
The appeal is the accessibility. A digital tablet and a pen. A sentence read aloud. A short observation that could be embedded in a routine appointment without requiring specialist referral or dedicated infrastructure. The measurement is not in what the person writes but in how they write it: the timing, the hesitation, the fragmentation of strokes, the delay before starting.
What the study cannot yet claim
The researchers are careful about the limits of what they have shown. The approach remains an emerging methodology, and future research will have to confirm the effects, also in the long term, in larger and more diverse populations. The results of the current study may therefore not be readily transferable. It also didn’t consider the use of medications and their possible influence.
Fifty-eight people, all living in care homes in Portugal, is a starting point rather than a conclusion. Handwriting is also shaped by education, dominant language, how much someone wrote by hand throughout their life, and individual motor variation that has nothing to do with cognition. A tool that works diagnostically will need to account for personal baselines rather than population averages.
But the core methodological point stands independently of those caveats. The task design is the instrument. Copying a familiar sentence does not ask enough of the brain to reveal what the brain can no longer do quietly. Dictation does.
A window, not a verdict
“Writing is not just a motor activity, it’s a window into the brain,” said Dr Matias. “We found that older adults with cognitive impairment displayed distinct patterns in the timing and organization of their handwriting movements. Tasks involving higher cognitive demands showed that cognitive decline is reflected in how efficiently and coherently handwriting movements are organized over time.”
The study’s contribution is not proving that cognitive decline affects handwriting. That has been observed for decades. The contribution is showing that the way researchers design the task determines what they can see, and that what looks like a simple writing exercise can, under the right conditions, function as a diagnostic instrument. The sentence on the card conceals. The sentence read aloud reveals.
Whether that distinction eventually makes its way into clinical practice will depend on replication, scale, and longitudinal validation. But the question the study raises is a precise one, and precision is where this kind of research has to start.
