The Clock-Drawing Test: Patterns, Scoring, and What They Suggest
Instructions, the executive demand of 10 past 11, command versus copy, and schematic patterns across common clinical syndromes.
Clinical Summary
The clock-drawing test is a brief screen of semantic knowledge, visuospatial layout, and executive recoding. Ten past eleven is used because it pulls the hands toward the numbers 10 and 11. Command and copy dissociate planning from construction. Patterns are typical, not diagnostic, and education, vision, and motor disease change the drawing.
The clock-drawing test is a two-minute screen of whether someone can hold the idea of a clock, lay the numerals out in space, and set the hands to a time that has to be recoded. Medical students meet it inside the MoCA and the SLUMS. Clinicians use it at the bedside because the errors are visible. This chapter is a way to look at those errors: what the instruction is asking, why the usual time is 10 past 11, how command and copy pull on different systems, and what a few typical patterns suggest about the networks involved. The patterns are typical. They are not diagnoses.
Screening, then the rest of the examination. A clock can support a concern for cognitive impairment and can suggest where to look next. It does not establish Alzheimer disease, a stroke, or any other diagnosis. Education, literacy, vision, and motor control all change the drawing. Shulman’s review of studies published through 1998 found a mean sensitivity of 85% and a mean specificity of 85% across scoring systems — an average over heterogeneous methods, not the performance of one cutoff in your clinic.
How to give it
Give the person a blank sheet and a pen. Say the three parts separately, and wait for each one before you add the next. A common bedside script, the one used with Shulman’s scale and with the Mini-Cog, is:
1. “I’d like you to draw a clock.”
2. “Put in all the numbers.”
3. “Set the hands to 10 past 11.”
Some protocols name the time differently. Sunderland’s 10-point scale asked for 2:45. The time is part of the method. If you change it, you change which errors the task can provoke, and you should score with the system that matches the instruction you used.
Why 10 past 11
“Ten past eleven” names two numerals, 10 and 11, and then asks the person not to point at them. A correct clock puts the minute hand on the 2, because 10 minutes is 10/60 of the hour-circle, and it puts the shorter hour hand a short way past the 11, toward the 12 — about one-sixth of the distance from 11 to 12. That recoding is an executive step. The frontal pull is toward the numbers just spoken. When inhibition fails, one or both hands land on 10 and 11. Rouleau and colleagues called this a stimulus-bound response and found it more often in Alzheimer disease than in Huntington disease.
The same time also places both hands in the upper half of the face, so a left-sided spatial omission is easy to see, and it requires two hands of different lengths. Equal-length hands are a separate error from a wrong time.
Command, then copy
The command condition is the unprompted drawing above. The copy condition is a second drawing made while a correct clock stays in view. Royall’s CLOX was built around that split: CLOX1, the unprompted clock, tracked an executive measure (the EXIT25), and CLOX2, the copy, tracked the Mini-Mental State Examination. In their sample the EXIT25 accounted for most of the variance in CLOX1.
The gap between the two drawings is the teaching point.
- In Alzheimer disease, Rouleau’s group found a marked improvement from command to copy. The model supplies the concept the person could not generate. A later two-year follow-up of the same paradigm found a steeper decline on command than on copy, with conceptual errors present early.
- In Huntington disease, the same comparison did not show that marked copy improvement. Graphic difficulty stayed on the page.
- In Lewy body dementia, Gnanalingham and colleagues found the opposite of the Alzheimer pattern: controls, people with Alzheimer disease, and people with Parkinson disease all improved from the draw condition to the copy, and people with Lewy body dementia scored equally poorly on both.
Ask for the copy whenever the command clock is abnormal. A copy that cleans up points toward planning or conceptual failure. A copy that stays broken points toward visuoconstruction or motor execution.
What the drawing is asking the brain to do
Four jobs are packed into one figure. Knowing what a clock is, including that it has twelve hours and two hands, draws on semantic memory, with temporal and temporoparietal cortex doing much of that work. Placing marks around a circle draws on parietal visuospatial and constructional cortex. Choosing the 2 for “ten minutes,” stopping at 12, and resisting the pull toward the spoken numbers draws on frontal and frontostriatal executive control. Making the line at all draws on motor cortex, basal ganglia, and cerebellum. A tremor, chorea, paresis, or severe arthritis can spoil a clock whose idea was intact.
The Cognitive Domains section of this site places the clock under both executive functioning and motor construction, which is why the MoCA gives it 3 points and the SLUMS gives it 4. Those points are a screen inside a screen. The figures below are for pattern recognition after the score, when you want to know what kind of error you are looking at.
A gallery of schematic clocks
Illustrative schematic, not a patient drawing
1. Normal clock
Closed contour, numerals 1–12 evenly spaced, two hands of different length. The gold minute hand points at the 2. The green hour hand sits just past 11. Those colors are a legend for this one figure; a patient’s pen is one color.
Illustrative schematic, not a patient drawing
2. Mild Alzheimer disease
The numerals are all there, with uneven spacing. The hands are about the same length and point at 10 and 11, the numbers named in the instruction, instead of recoding “10 minutes” onto the 2. Stimulus-bound responses of this kind were more frequent in Alzheimer disease than in Huntington disease in Rouleau’s comparison. An early clock often still looks like a clock. See the Alzheimer disease chapter for the rest of the syndrome.
Illustrative schematic, not a patient drawing
3. Moderate to severe Alzheimer disease
The concept loosens. Numerals drift outside the circle, 12 is repeated, several numerals are missing, and the time is written in words with no hands. Conceptual errors in Rouleau’s series were almost exclusive to Alzheimer disease and tracked dementia severity. Sunderland’s group, using a 1-to-10 scale and the time 2:45, found mean scores of 4.9 in Alzheimer disease and 8.7 in controls, with inter-rater reliability of r = 0.86 on the patient clocks. They were explicit that the test is not a definitive indicator of Alzheimer disease.
Illustrative schematic, not a patient drawing
4. Right parietal injury, left hemispatial neglect
Every numeral is crowded onto the right half. The left side is empty. That is the bedside picture of left hemispatial neglect, most often after injury to right parietal attention networks. A clock does not localize a stroke by itself. Field cut, an unfinished drawing, and a page placed too far to the person’s left can each empty one side. Pair the figure with line bisection, a cancellation task, and imaging.
Illustrative schematic, not a patient drawing
5. Vascular disease and subcortical executive failure
Numerals arrive in clusters, with gaps left where the spacing was never evened out. An extra 4 sits inside the face, and one hand is pulled toward 10. This is a planning drawing. On Royall’s CLOX, that planning load sits in the command condition: a command clock that is clearly worse than the copy fits an executive account better than a pure constructional one. Frontal–subcortical circuits and small-vessel white matter are the regions clinicians associate with this pattern. The same look occurs in other dysexecutive states.
Illustrative schematic, not a patient drawing
6. Behavioral variant frontotemporal dementia
The circle is round and 1–12 are in place, which fits the relative sparing of posterior visuospatial cortex early in this illness. Numbering then continues to 13 and 14, a failure to stop. Blair and colleagues found that a frontotemporal group scored higher, meaning better, than an Alzheimer group, and made fewer stimulus-bound, conceptual, and spatial or planning errors. An early drawing is often relatively preserved. Perseveration like this figure is worth recognizing when monitoring fails. It is not the expected clock in every patient, and it is not specific to frontotemporal dementia. The frontotemporal dementia chapter covers the clinical syndrome.
Illustrative schematic, not a patient drawing
7. Lewy body dementia and Parkinson disease dementia
The contour is broken into arcs, and the numerals are very small and scattered off the rim. Gnanalingham and colleagues found copy scores improved over the draw condition in controls, Alzheimer disease, and Parkinson disease, and did not improve in Lewy body dementia, where both were poor. Cahn-Weiner and colleagues, matching groups for age, education, and dementia severity, found no significant difference on a global clock score. Lewy body dementia produced more conceptual errors, and both Lewy body dementia and cognitively impaired Parkinson disease produced more planning errors than Alzheimer disease. Overall classification was 69% for Alzheimer disease versus Lewy body dementia and 68% for Parkinson disease versus Lewy body dementia. Tiny numerals can be the micrographia of parkinsonism showing up in the drawing — a motor contribution, not a cognitive signature. Use the pattern as a reason to look harder, including at fluctuation, REM-sleep behavior, and parkinsonism. See Parkinson disease.
Illustrative schematic, not a patient drawing
8. Huntington disease and other subcortical motor disease
The contour and the hands waver, and the 5 has been pulled into the interior. Rouleau’s comparison found graphic difficulties very common in Huntington disease and virtually absent in Alzheimer disease, with conceptual errors running the other way. Copying a model produced a marked improvement in the Alzheimer group and not in the Huntington group. Striatal and frontostriatal disease is the core of Huntington disease. Chorea, tremor, and poor motor control can distort a line even when the person knows what a clock is.
Illustrative schematic, not a patient drawing
9. Left-hemisphere injury and aphasia
The spatial slots are intact and the hands are set to 10 past 11, but 4 is written as a word, 7 is missing, and 8 is a question mark. This figure is a teaching picture of a language or numeral-retrieval problem with the layout left in place. Bonini and Radanovic found that people with post-stroke aphasia scored lower on clock drawing than non-aphasic patients with hemispheric lesions, so the verbal command itself can pull the score down. A missing numeral in aphasia is a different observation from an empty left half of the clock.
Comparison
| Condition | Typical clock pattern | Region the error stresses |
|---|---|---|
| Normal | Closed circle, 1–12 spaced, minute hand at 2, hour hand just past 11 | The task is intact across semantic, spatial, executive, and motor systems |
| Mild Alzheimer disease | Numerals present with spacing errors; hands of equal length or aimed at 10 and 11 | Temporoparietal semantic and spatial cortex, plus emerging frontal control |
| Moderate to severe Alzheimer disease | Numerals outside the circle, repeated or missing; time written in words; no hands | Widespread cortical disease, including semantic knowledge of what a clock is |
| Right parietal injury / left neglect | Numerals crowded on the right; left side empty | Right parietal attention networks |
| Vascular / subcortical executive | Clustered numerals, gaps, perseveration; command often worse than copy | Frontal–subcortical circuits and small-vessel white matter |
| Behavioral variant frontotemporal dementia | Often relatively preserved early; perseveration past 12 when monitoring fails | Frontal and anterior temporal cortex; posterior visuospatial cortex relatively spared early |
| Lewy body dementia / Parkinson disease dementia | Visuoconstruction poor on command and on copy; planning errors; very small numerals when parkinsonism is active | Posterior visuoperceptual cortex; nigrostriatal motor system adds micrographia |
| Huntington disease | Irregular, poorly controlled lines; planning errors; copy often fails to rescue the drawing | Striatum and frontostriatal circuits |
| Left-hemisphere stroke / aphasia | Spatial frame often preserved; numerals substituted, written as words, or omitted | Left perisylvian language cortex; the command itself is a language task |
Scoring systems, briefly
Several published systems exist. They do not use the same time, the same points, or the same cutoff, and Shulman noted that despite that variability the published psychometric properties looked broadly similar. Learn the one your clinic uses, and name it in the note.
- Shulman (1986, reviewed 2000). A short ordinal scale, commonly used as 0–5, that emphasizes the qualitative look of the drawing: a perfect clock at the top, severe visuospatial disorganization or no recognizable clock at the bottom. The 2000 review is the best single orientation to the menu of systems.
- Sunderland (1989). A 10 (best) to 1 (worst) scale. The requested time was 2:45. Anchors run from a correct clock down to drawings that are no longer recognizable as clocks. Useful as a severity correlate in a research or office setting where full neuropsychological testing is not available.
- Mini-Cog (Borson, 2000). Three-item recall plus a clock, designed as a roughly three-minute screen. In the development sample — 129 people with probable dementia and 120 with no history of decline, 124 of them non-English speakers — sensitivity was 99% and 96% of subjects were correctly classified, and education and language did not move the result the way they moved the MMSE. The authors asked for validation in further clinical and population samples before treating those figures as settled. The clock here is one component of the screen, scored normal or abnormal, alongside recall.
- CLOX (Royall, 1998). Two clocks. CLOX1 is the unprompted command and is the executive clock. CLOX2 is the copy. The difference is the point of the test.
What lowers a score for non-cognitive reasons. Ainslie and Murden asked non-demented older adults to draw a clock at 3:00 and scored it three ways. On the Shulman scale, sensitivity and specificity for dementia in the subgroup with 8 or fewer years of school were 90% and 42%. They concluded that educational effects make clock drawing a poor single screen in a poorly educated population. Limited experience with analog clocks, low literacy, uncorrected vision, and motor disease (tremor, paresis, arthritis, chorea) belong on that same list. Ask whether the person has used a clock face, can see the page, and can control the pen before you interpret the figure as cortex.
Clinical pearls
- Give the three instructions one at a time. Pre-drawing the circle, or placing the 12 for the person, removes part of what you were trying to test.
- Use 10 past 11 when you want the stimulus-bound error to be available. If your scale specifies another time, use that time.
- When the command clock is abnormal, ask for a copy before you decide the error is spatial.
- Write down the kind of error, not only a number: stimulus-bound hands, conceptual loss, left-sided omission, perseveration past 12, graphic distortion, or a clock that is poor on both command and copy.
- A better-preserved clock in a frontal behavioral syndrome is expected often enough that a “good clock” does not argue against behavioral variant frontotemporal dementia. Blair’s frontotemporal group outperformed the Alzheimer group.
- A clock that is poor on both command and copy, in a person with fluctuation, dream enactment, or parkinsonism, belongs in a Lewy-body workup. Gnanalingham described that copy failure. Cahn-Weiner’s classification rates, 69% and 68%, are why the clock stays a clue.
- Palmqvist and colleagues offered a practical combination in MMSE-matched patients, not a clock-only rule. In the MMSE 21–27 range, an orientation subscore multiplied by 3 that was at least as high as the total MMSE, together with an impaired clock, had sensitivity 93% and specificity 70% for Lewy body dementia versus Alzheimer disease. Add a cube copy that is not three-dimensional and the question is stronger. Then use consensus criteria, not the clock.
- In aphasia, rescore the possibility that the command was not understood. A word written in place of a numeral is language. An empty left half is spatial.
- Record the scale you used. “Clock 3/5, Shulman” is interpretable later. “Clock abnormal” is not.
Recommended Videos
These videos are third-party educational content hosted on YouTube.
- Shulman KI, Shedletsky R, Silver IL. The challenge of time: clock-drawing and cognitive function in the elderly. Int J Geriatr Psychiatry. 1986;1(2):135-140. doi:10.1002/gps.930010209
- Shulman KI. Clock-drawing: is it the ideal cognitive screening test? Int J Geriatr Psychiatry. 2000;15(6):548-561. doi:10.1002/1099-1166(200006)15:6<548::aid-gps242>3.0.co;2-u. PMID 10861923
- Sunderland T, Hill JL, Mellow AM, et al. Clock drawing in Alzheimer’s disease: a novel measure of dementia severity. J Am Geriatr Soc. 1989;37(8):725-729. doi:10.1111/j.1532-5415.1989.tb02233.x. PMID 2754157
- Borson S, Scanlan J, Brush M, Vitaliano P, Dokmak A. The Mini-Cog: a cognitive ‘vital signs’ measure for dementia screening in multi-lingual elderly. Int J Geriatr Psychiatry. 2000;15(11):1021-1027. doi:10.1002/1099-1166(200011)15:11<1021::aid-gps234>3.0.co;2-6. PMID 11113982
- Royall DR, Cordes JA, Polk M. CLOX: an executive clock drawing task. J Neurol Neurosurg Psychiatry. 1998;64(5):588-594. doi:10.1136/jnnp.64.5.588. PMID 9598672
- Rouleau I, Salmon DP, Butters N, Kennedy C, McGuire K. Quantitative and qualitative analyses of clock drawings in Alzheimer’s and Huntington’s disease. Brain Cogn. 1992;18(1):70-87. doi:10.1016/0278-2626(92)90112-y. PMID 1543577
- Rouleau I, Salmon DP, Butters N. Longitudinal analysis of clock drawing in Alzheimer’s disease patients. Brain Cogn. 1996;31(1):17-34. doi:10.1006/brcg.1996.0022. PMID 8790932
- Gnanalingham KK, Byrne EJ, Thornton A, Sambrook MA, Bannister P. Motor and cognitive function in Lewy body dementia: comparison with Alzheimer’s and Parkinson’s diseases. J Neurol Neurosurg Psychiatry. 1997;62(3):243-252. doi:10.1136/jnnp.62.3.243. PMID 9069479
- Cahn-Weiner DA, Williams K, Grace J, Tremont G, Westervelt H, Stern RA. Discrimination of dementia with Lewy bodies from Alzheimer disease and Parkinson disease using the clock drawing test. Cogn Behav Neurol. 2003;16(2):85-92. doi:10.1097/00146965-200306000-00001. PMID 12799594
- Blair M, Kertesz A, McMonagle P, Davidson W, Bodi N. Quantitative and qualitative analyses of clock drawing in frontotemporal dementia and Alzheimer’s disease. J Int Neuropsychol Soc. 2006;12(2):159-165. doi:10.1017/S1355617706060255. PMID 16573849
- Palmqvist S, Hansson O, Minthon L, Londos E. Practical suggestions on how to differentiate dementia with Lewy bodies from Alzheimer’s disease with common cognitive tests. Int J Geriatr Psychiatry. 2009;24(12):1405-1412. doi:10.1002/gps.2277. PMID 19347836
- Ainslie NK, Murden RA. Effect of education on the clock-drawing dementia screen in non-demented elderly persons. J Am Geriatr Soc. 1993;41(3):249-252. doi:10.1111/j.1532-5415.1993.tb06701.x. PMID 8440847
- Bonini MV, Radanovic M. Cognitive deficits in post-stroke aphasia. Arq Neuropsiquiatr. 2015;73(10):840-847. doi:10.1590/0004-282X20150133. PMID 26465401