The Sequence Memory Test: Working Memory, Explained

In Keel's sequence memory test, you observe a series of items presented in order — numbers, letters, or tones — and then reproduce the sequence from memory, either in forward order (straightforward recall) or backward order (active manipulation). The sequence length adapts to your performance, increasing when you succeed and decreasing when you fail, converging on your working memory span.

This is the digit span paradigm, codified by psychologist George Miller in his 1956 paper 'The Magical Number Seven, Plus or Minus Two' — one of the most cited papers in cognitive psychology. Digit span (forward and backward) is included in nearly every major neuropsychological battery, including the Wechsler Adult Intelligence Scale (WAIS), the Wechsler Memory Scale (WMS), and the CANTAB battery used in clinical trials.

Forward digit span measures the phonological loop — Alan Baddeley's term for the component of working memory that holds verbal and sequential information via a short-lived acoustic trace that is refreshed by internal rehearsal. This is the mechanism you use when you hear a phone number and repeat it to yourself while dialing.

Backward span adds a manipulation component: you must hold the sequence in memory while simultaneously reversing it. This engages the central executive — the attentional controller that manages the flow of information in working memory. Backward span is more sensitive to prefrontal function than forward span, and more sensitive to early cognitive decline.

Keel's sequence memory test tracks both span length and accuracy across sessions, building a picture of your working memory capacity over time.

Working memory capacity is tightly linked to general fluid intelligence — the ability to reason, solve novel problems, and learn new things. It draws heavily on the dorsolateral prefrontal cortex and the anterior cingulate, both of which are vulnerable in multiple forms of cognitive decline.

In Alzheimer's disease, working memory deficits often appear in the mild cognitive impairment (MCI) stage, before diagnostic criteria for dementia are met. A 2019 study in Neuropsychology Review found that working memory composite scores were among the most sensitive predictors of conversion from MCI to Alzheimer's dementia. In ADHD — a condition with enormous underdiagnosis in adults — working memory deficit is the central cognitive feature, distinct from the age-related decline Keel is primarily tracking.

Working memory capacity is also modifiable, within limits. Cardiovascular exercise, sleep quality, and cognitive stimulation all have evidence-backed effects on working memory performance. If you are tracking a protocol, working memory is one of the domains where a genuine intervention effect should be detectable.

Working memory is acutely sensitive to sleep deprivation. Slow-wave sleep — the deep non-REM stage that dominates the first half of the night — is when synaptic consolidation occurs and the glymphatic system clears metabolic waste including amyloid-beta. A night without adequate slow-wave sleep measurably reduces working memory performance the following day.

Stress has a paradoxical effect on working memory: mild stress (moderate cortisol elevation) can briefly sharpen working memory for immediately relevant information, while chronic stress or very high acute cortisol reliably impairs it. Multitasking and distraction suppress working memory performance by consuming the central executive's attentional resources.