The Symbol-Digit Test: Processing Speed, Explained
The symbol-digit coding task is one of the most replicated measures in cognitive neuroscience. Here is what it measures, why, and what your score actually tells you.
What you are actually doing
In Keel's symbol-digit test, you are shown a key pairing symbols to digits (1-9) and then presented with a stream of symbols. Your task: enter the correct digit for each symbol as quickly and accurately as possible within a fixed time window.
This is a brief adaptation of the Symbol Digit Modalities Test (SDMT), a neuropsychological instrument developed by Aaron Smith in 1973 that remains one of the most widely used cognitive assessments in clinical and research settings. The SDMT is part of the standard battery used in multiple sclerosis research, traumatic brain injury evaluation, and Alzheimer's disease clinical trials.
What it measures
Despite looking like a simple number-substitution game, the SDMT engages a surprisingly broad network: visual processing (identifying the symbol), working memory (holding the key mapping while scanning), attention (maintaining focus across many items), and executive function (managing the substitution rule). The speed at which you complete accurate substitutions reflects the efficiency of white matter pathways connecting these regions.
Critically, the task is primarily sensitive to processing speed — how fast your brain can perform the core mental operation — rather than whether you know the answer. Unlike a vocabulary test, there is no prior knowledge to draw on. The only variable is neural efficiency.
This property makes symbol-digit tasks excellent for detecting change over time. They are relatively insensitive to education level and are equally fair across different backgrounds. A neuroscientist and a postal worker with equivalent cognitive health will perform comparably.
Why processing speed makes such a good daily signal
Processing speed is one of the first cognitive abilities to show transient effects from common daily factors: sleep deprivation, acute illness, alcohol, certain medications, stress, and blood sugar fluctuation all measurably suppress it within hours. This makes it an ideal daily probe — if something is affecting your cognitive function today, processing speed will likely catch it.
It is also, over a longer time scale, one of the earliest domains to show age-related and disease-related change. Vascular cognitive impairment — caused by small vessel disease in white matter — often presents first as processing speed slowing, before memory changes become apparent. Keel's daily data can reveal a rate of change that a once-yearly clinical screen simply cannot detect.
What affects your score on any given day
Expect normal session-to-session variation of roughly 5-15% in processing speed tasks. A single below-average score is almost never meaningful. The sources of legitimate daily variation include: sleep quality (a single poor night can reduce processing speed by 10-20%), time of day (most people are fastest in the late morning), caffeine (moderate caffeine improves processing speed for regular users; withdrawal suppresses it), acute illness (even a mild cold suppresses processing speed measurably), and practice effects (scores typically improve over the first 7-10 sessions as you learn the format).
Your trend line across weeks accounts for this noise. A consistent downward drift — not a single bad day — is what warrants attention.
Frequently asked questions
Is the symbol-digit test the same as the SDMT used in clinical settings?
Keel's test is adapted from the SDMT format but is not a direct clinical administration. The validated SDMT is administered by a trained clinician under standardized conditions. Keel's version uses the same core mechanism (symbol-to-digit substitution under time pressure) and measures the same underlying construct, but should be understood as a personal tracking tool, not a clinical assessment.
Why do I sometimes score very differently day to day?
Processing speed is one of the most state-sensitive cognitive measures — it responds quickly to sleep, stress, illness, and other daily factors. A 10-15% variation between sessions is normal. The signal you should pay attention to is your rolling trend over weeks, not individual session scores.
My score has been declining for a few weeks. Should I be worried?
A persistent decline over weeks worth discussing with a doctor, especially if it is not explained by a clear cause like illness recovery, a medication change, or sustained poor sleep. Keel is designed to surface exactly this kind of trend — but it cannot tell you why it is happening. That conversation belongs with a healthcare provider.
Related resources
Start tracking your cognitive baseline
Four minutes a day. Five short tests. One trend line that builds over weeks and months so you can see where you stand — and separate a bad day from a real change.
Free to start. No account required. Not a diagnostic tool.