Why Arithmetic Is Often the First Thing to Go: Working Memory and Alzheimer's

The founder of MentalMather watched his grandmother lose her ability to do arithmetic before she lost almost anything else. She had been sharp her entire life — the kind of person who calculated restaurant tips in her head without hesitation, who managed household finances with a pencil and paper and never made an error. Then one day, she couldn't split a bill. Then she couldn't make change. Then the numbers stopped making sense altogether.

It was months before anyone used the word Alzheimer's. But looking back, the math was the first thing to go.

This isn't just one family's story. It's a pattern that neuroscience has documented extensively, and understanding why arithmetic is so vulnerable reveals something important about how cognitive decline actually works.

What Happens to Working Memory in Alzheimer's

Alzheimer's disease is most commonly associated with memory loss — forgetting names, misplacing objects, losing track of recent conversations. But the research is clear that working memory is affected early in the disease, often before the more dramatic symptoms of episodic memory loss become apparent.

A 2015 review published in BioMed Research International tracked working memory decline across three stages — normal aging, mild cognitive impairment (MCI), and Alzheimer's disease — and found that the central executive component of working memory shows impairment during MCI, which is often a precursor to Alzheimer's. This means working memory starts degrading before a formal Alzheimer's diagnosis is even on the table.

A peer-reviewed study in the journal Dialogues in Clinical Neuroscience confirmed this pattern: both working memory and long-term declarative memory are affected early in Alzheimer's progression, with the individual pattern of impairment correlating with structural and functional changes in the brain.

But why does this show up in arithmetic specifically?

Why Math Is So Sensitive to Working Memory Decline

Mental arithmetic is one of the most working-memory-intensive everyday tasks a human being can do. When you calculate 86 - 47 in your head, you need to hold the original numbers, track which column you're working on, carry or borrow across columns, and hold partial results — all simultaneously.

Unlike remembering someone's name (which is a retrieval task) or recognizing a familiar face (which is a pattern-matching task), arithmetic requires you to hold information while actively manipulating other information. That's the textbook definition of working memory load.

As the RAM analogy explains: when your brain's working memory capacity starts to shrink, the first things to fail are tasks that require holding multiple items while processing others. Simple addition of small numbers might survive because the load is low. Multi-step problems, carrying digits, division with remainders — those demand more working memory and fail sooner.

This is why the Alzheimer's Association lists difficulty with numbers and following plans as one of its 10 early warning signs. The inability to track monthly bills, follow a familiar recipe, or work with numbers is among the first cognitive changes that patients and caregivers notice.

Having a bad day with mental math does not mean you have Alzheimer's disease. Normal cognitive aging also affects working memory — just much more slowly and mildly. The difference is pattern and trajectory: a gradual, consistent decline that gets progressively worse is different from normal day-to-day variation. This article is about understanding the science, not self-diagnosis. If you're concerned about cognitive decline, talk to a healthcare professional.

The Clock Drawing Test Parallel

One of the most widely used screening tools for cognitive impairment is the Clock Drawing Test — a doctor asks a patient to draw a clock face with the numbers in the right positions and the hands showing a specific time. It seems almost absurdly simple, but it's remarkably sensitive to early cognitive decline because it requires multiple cognitive systems working together: spatial reasoning, number sequencing, and — crucially — working memory to hold the instructions while executing the task.

Mental arithmetic works on the same principle. It's not that math is uniquely important — it's that math is a uniquely sensitive probe for the cognitive systems that decline first. It's like a canary in the coal mine. The math doesn't matter. What the math reveals about the underlying cognitive machinery matters enormously.

What the Research Shows About Tracking Over Time

A key insight from the research on MCI and Alzheimer's is that tracking cognitive performance over time is more informative than any single measurement. One bad score means nothing — everyone has off days. But a trend line that consistently moves downward over months or years is a meaningful signal.

The 2015 BioMed International review specifically suggests that monitoring performance on working memory and executive function tasks could help signal progression from normal cognition to MCI to Alzheimer's. The challenge has always been making that monitoring practical — most people aren't going to visit a neuropsychologist every month for formal testing.

This is where daily self-measurement becomes potentially valuable. Not as a diagnostic tool — we want to be emphatic about that — but as a source of longitudinal data that you can share with a healthcare provider if you notice concerning trends.

Cognitive Decline vs. Cognitive Laziness

Before anyone panics: most people who feel like their mental math has gotten worse are not experiencing pathological decline. They're experiencing cognitive laziness — a natural consequence of living in a world where your phone has a calculator and you haven't needed to do mental arithmetic in years.

Use it or lose it is a real phenomenon in cognition, but it's different from decline. Cognitive laziness means the skill has atrophied from disuse — but the underlying machinery is still intact. It's like not running for a year: your fitness drops, but your legs still work. Pathological decline means the machinery itself is degrading.

The difference shows up in the data. Someone experiencing disuse-related slowdown will typically see their performance improve with a few days of practice and then stabilize. Someone experiencing genuine cognitive decline will see performance that doesn't recover — or recovers briefly and then continues trending downward.

This is another reason why longitudinal tracking matters. A single snapshot can't distinguish between these two scenarios. A trend line over weeks and months can.

What You Can Do

The research on cognitive decline prevention consistently points to the same factors: regular physical exercise, quality sleep, social engagement, cognitive stimulation, and cardiovascular health. None of these are glamorous or novel. All of them have strong evidence behind them.

What we'd add is: pay attention. The people who catch cognitive changes early are the people who are tracking — whether formally or informally. A daily cognitive benchmark gives you data. Data gives you the ability to notice a change while it's still early, and early detection is consistently associated with better outcomes in cognitive health.

The Alzheimer's Association recommends scheduling an appointment with a doctor if you notice any of their 10 warning signs. Having data to bring to that appointment — "here's my cognitive performance trend over the past six months" — gives your doctor something concrete to work with, rather than the much harder question of "do you feel like your memory has gotten worse?"

A Note From Our Founder

We built MentalMather as a cognitive measurement tool, not a medical device. It doesn't diagnose anything. It doesn't claim to prevent anything. What it does is give you a consistent, daily data point on how your brain is performing relative to your own baseline.

For most users, that data point is interesting and occasionally motivating. For some users — especially those with a family history of cognitive decline — it might be something more. It might be the early-warning system they wish their family had had.

My grandmother didn't have a way to track the change happening inside her brain. By the time anyone noticed, the window for early intervention had already closed. I can't change what happened to her, but I can build a tool that gives other families a chance to notice sooner.

If you're worried, please talk to a doctor. If you want data to bring to that conversation, we're here for that.