Attention
Landmark Task
Also called: landmark test, Milner landmark task, pre-bisected line task, pre-transected line judgment, tachistoscopic line bisection, perceptual line bisection
Measure spatial bias without a pointing movement: people judge which part of a briefly shown, pre-marked line is shorter or longer.
6-7 minTimed: desktop or laptop with a mouse or trackpad
Free to start: 10 free participants per account, then $0.99 per participant for academic research, or any yearly plan. Preview any test free after you sign up.
- Papers, 2021 to 2025
- 33
- Evidence grade
- C
- Registered clinical studies
- 2
What it measures
A measure of the balance of attention between left and right, with no pointing movement. A long line with a small mark near its middle flashes for 150 ms, and participants press the arrow key for the side whose part is shorter or, in other blocks, longer. The score is the mark position at which both parts look equal, in percent of the line's length.
Constructs
- Spatial attention
- Visuospatial judgment of length
- Attentional bias between left and right
Research uses
- Spatial attention and neglect research
- Pseudoneglect and aging research
- Within-study group comparison
How the paradigm works
The Landmark Task measures the balance of attention between the left and right of space through a judgment, with no pointing movement (Milner, Brechmann and Pagliarini, 1992; Harvey, Milner and Roberts, 1995). Each line is already marked near its middle, and participants say which part is shorter or which is longer. Healthy young adults judge the left part a little longer than it is, so the point that looks central lies left of the true middle: pseudoneglect (Learmonth and colleagues, 2015). The bias moves rightward with age and with short lines (Benwell and colleagues, 2014), and patients with left neglect judge the left part shorter (Harvey, Milner and Roberts, 1995).
More about the paradigm
On each trial a cross shows for 1,000 ms, then for 150 ms a 600 px line with a small mark across it where the cross was, then the cross again until the participant presses the left or right arrow key, with no time limit. The mark lies on the line's true middle or 1, 2, 3 or 4 percent of its length to the left or right; the line moves, so the mark always falls where the cross was (Benwell and colleagues, 2014). In two blocks the question is which part is shorter and in two which part is longer, in the order first, second, second, first, so a preference for one key can be told apart from a bias in seeing the line (Bisiach and colleagues, 1998; Toraldo and colleagues, 2004). Each offset appears 8 times with each question (standard; 4 to 12): 144 trials after 6 practice trials with feedback, about 6 to 7 minutes. Every line is set before the session: no answer changes a later trial, and there is no training mode. Desktop and laptop only.
The main score is the point of subjective equality: the mark's offset at which both parts look equal, in percent of the line's length, from a cumulative Gaussian fitted to the share of "left part longer" judgments at each offset; negative is a leftward bias. The line bisection test measures the same kind of bias by marking the middle with a click or tap, which adds a movement and free viewing; its deviation is in percent of half the line, so double this score to compare. In healthy adults the two tasks correlate only weakly (Learmonth and colleagues, 2015), and the landmark bias was the more stable across sessions (Mitchell and colleagues, 2020). Scores support research and do not diagnose neglect.
Scores
- Point of Subjective Equality%
- The mark's offset from the true middle at which both parts look equal, in percent of the line's length, from a cumulative Gaussian fitted to the judgments; positive to the right.
- Negative means the left part looked longer: a leftward bias, common in healthy young adults
- Just Noticeable Difference%
- Half the distance between the fitted curve's 25% and 75% points, in percent of the line's length.
- Lower means finer judgments
- Left Key Rate%
- Share of answers on the left arrow. Each question asks equally often at every offset, so a bias in seeing the line cancels out.
- 0.5 means no preference for one key
- Left Part Longer by Offset%
- Share of judgments that the left part was longer at each mark offset, keyed by its signed percent (positive: the mark right of the middle).
- Rises from 0 to 1 across the offsets
- Accuracy%
- Share right where the mark was off the true middle.
- Higher means finer judgments; a bias lowers it on one side
- Accuracy at the Largest Offsets%
- Share right with the mark 4 percent from the middle, where one part is plainly longer.
- Under 0.5 is invalid and under 0.8 is flagged for review
- Largest-Offset Trials Scoredtrials
- Answers with the mark 4 percent from the middle that count.
- Informational
- Median Answer Timems
- Median time from the line's end to the answer.
- Informational; the answer has no time limit
- Anticipationscount
- Answers under 100 ms from the line's end, which are not scored.
- Lower is better
- Early Key Pressescount
- Arrow keys pressed while the line showed, which are not taken as answers.
- Informational
- Frame Intervalms
- Median time between screen frames during the trials.
- About 16.7 on a 60 Hz screen
- Completed Trialstrials
- Trials answered in the scored run.
- Fewer than planned is flagged for review
- Planned Trialstrials
- Trials the run set out to show.
- Informational
Use and evidence
33 papers from 2021 to 2025 name this task in their title or abstract, down 41% compared with 2016 to 2020. That ranks it 290 of 395 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 12 |
| 2024 | 5 |
| 2023 | 4 |
| 2022 | 6 |
| 2021 | 6 |
| 2020 | 7 |
| 2019 | 12 |
| 2018 | 11 |
| 2017 | 14 |
| 2016 | 12 |
| 2015 | 6 |
| 2014 | 5 |
| 2013 | 8 |
| 2012 | 12 |
| 2011 | 5 |
| 2010 | 8 |
| 2009 | 3 |
| 2008 | 1 |
| 2007 | 5 |
| 2006 | 3 |
| 2005 | 3 |
| 2004 | 10 |
| 2003 | 2 |
| 2002 | 6 |
| 2001 | 4 |
| 2000 | 3 |
Use grade D
30 to 99 papers from 2021 to 2025.
Evidence grade C
At least 2 test-retest papers, 30 reliability or validity papers or 3 registered clinical studies.
- Papers on test-retest reliability
- 4
- Papers on reliability or validity
- 15
- Registered clinical studies using it as an outcome
- 2
Note on the counts: a few counted papers use the words for navigation or face landmarks.
How the grades work
Use: A is 1,000 or more papers from 2021 to 2025, B 300 to 999, C 100 to 299, D 30 to 99 and E fewer than 30. Evidence: A means at least 25 papers report its test-retest reliability and at least 25 registered clinical studies use it as an outcome; B, at least 8 test-retest papers, and either 5 registered clinical studies or 150 reliability or validity papers; C, at least 2 test-retest papers, 30 reliability or validity papers or 3 registered clinical studies; D, little published measurement evidence found. The counts measure use. They include every paper that names the task in its title or abstract. Grades describe the published research on this task, not ConductCognition's version of it.
Source: ConductCognition test census 0.14.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Display
- On a dark panel: a cross for 1,000 ms, then for 150 ms a line of 600 by 6 px with a mark of 4 by 40 px across it where the cross was, then the cross again with the question.
- Mark positions
- On the line's true middle or 1, 2, 3 or 4 percent of its length to the left or right. The line moves and the mark stays where the cross was, so the cross gives no reference to the middle.
- Questions
- Which part is shorter in two blocks and which part is longer in two, in the order first, second, second, first; the first question comes from the session's seed.
- Answer
- Press the left or right arrow key for the side of the part asked about, with no time limit. Keys pressed while the line shows are counted and not taken as answers.
- Trials
- 8 at each of the 9 mark positions with each question (standard; 4 to 12): 144 trials, each block in a seeded order.
- Practice
- 3 lines with each question, the mark 3 or 4 percent from the middle, with feedback for 1,000 ms.
- Fixed displays
- Every line is set before the session, and no answer changes a later trial: there is no staircase and no training mode.
- Score
- The point of subjective equality: the mark's offset at which both parts look equal, in percent of the line's length, from a cumulative Gaussian fitted to the share of "left part longer" judgments at each offset; negative is a leftward bias. Answers under 100 ms leave the score.
The line is drawn in screen pixels with no calibration. The line bisection test asks for the middle by hand and scores in percent of half the line; double this score to compare.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: landmark_task_test from the scored run, the left or right arrow for the side whose part is shorter or, in other blocks, longer, with no time limit. Lines marked off the middle have a right answer (chance 50%; each offset on each side and each question equally often); lines marked on the middle have none and are left out of accuracy. Below chance invalidates and not above chance is flagged; one arrow on 90% of answers invalidates when accuracy is not above chance. Answer times run from the line's end; responses under 100 ms follow the timed-test rule. Under 50% right with the mark 4 percent from the middle, on at least 8 such trials, invalidates and under 80% is flagged. No point of subjective equality within the offsets shown, or a run short of its planned trials, is flagged for review.
Why these rules
Accuracy significantly below chance (one-sided exact test, p < .05) cannot come from guessing; it means the response mapping was reversed or the task was not followed (Slick, Sherman and Iverson 1999 treat below-chance forced-choice performance as definite response bias). Chance-level accuracy can be a real result in impaired participants, so it is only flagged. One option on at least 90% of answered trials, when the correct options are balanced and accuracy is not above chance, means the answers did not depend on the items (the task counterpart of long-string responding in surveys; Curran 2016). An answer under 100 ms from the line's end was pressed before the line could be judged. With the mark 4 percent of the line from its middle, one part is plainly longer, so errors there mean the question was mixed up or the line was not seen. A shift of the judgments to one side never invalidates, because that shift is what the test measures. The line shows for 150 ms, so a hidden tab withholds the percentile.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Trials at each mark position | 8 | 4 to 12 |
Yearly plans can change settings within these limits. Free and pay-as-you-go studies use the standard settings.
Need a setting this test does not have?
Devices and languages
- Timed test: participants need a desktop or laptop with a mouse or trackpad. Participant devices
- Languages: English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese.
- Available on: Every plan, including the free plan.
- No percentile. Scores and quality flags only. How norms work
- Recruiting online? See the setup guides for Prolific, Sona Systems, CloudResearch Connect and Qualtrics.
Cite and describe
Participants completed the Landmark Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Milner, Brechmann and Pagliarini, 1992; Harvey, Milner and Roberts, 1995), with the standard settings (trials at each mark position: 8). The main score was point of subjective equality (the mark's offset from the true middle at which both parts look equal, in percent of the line's length, from a cumulative Gaussian fitted to the judgments; positive to the right).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Milner, A. D., Brechmann, M., & Pagliarini, L. (1992). To halve and to halve not: An analysis of line bisection judgements in normal subjects. Neuropsychologia, 30(6), 515-526. https://doi.org/10.1016/0028-3932(92)90055-QJudgments of pre-bisected lines in healthy adults, with no bisecting movement.
- Harvey, M., Milner, A. D., & Roberts, R. C. (1995). An investigation of hemispatial neglect using the Landmark Task. Brain and Cognition, 27(1), 59-78. https://doi.org/10.1006/brcg.1995.1004The landmark task: which end of a marked line is nearer the mark, to separate a bias in seeing from a bias in moving.
- Bisiach, E., Ricci, R., Lualdi, M., & Colombo, M. R. (1998). Perceptual and response bias in unilateral neglect: Two modified versions of the Milner landmark task. Brain and Cognition, 37(3), 369-386. https://doi.org/10.1006/brcg.1998.1003Asking for the longer part and for the shorter part, to separate a bias in seeing from a preference for one response.
- Toraldo, A., McIntosh, R. D., Dijkerman, H. C., & Milner, A. D. (2004). A revised method for analysing neglect using the landmark task. Cortex, 40(3), 415-431. https://doi.org/10.1016/S0010-9452(08)70136-9The point of subjective equality as the bias in seeing, apart from the response bias.
- Benwell, C. S. Y., Thut, G., Grant, A., & Harvey, M. (2014). A rightward shift in the visuospatial attention vector with healthy aging. Frontiers in Aging Neuroscience, 6, 113. https://doi.org/10.3389/fnagi.2014.00113The computer version: a cross for 1 s, the line for 150 ms with its mark where the cross was, the cross until the answer, and a cumulative logistic fit; long lines bring out the leftward bias.
- Learmonth, G., Gallagher, A., Gibson, J., Thut, G., & Harvey, M. (2015). Intra- and inter-task reliability of spatial attention measures in pseudoneglect. PLoS ONE, 10(9), e0138379. https://doi.org/10.1371/journal.pone.0138379Test-retest r = 0.595 in 50 young adults, and a weak correlation with manual line bisection.
- Mitchell, A. G., Harris, J. M., Benstock, S. E., & Ales, J. M. (2020). The reliability of pseudoneglect is task dependent. Neuropsychologia, 148, 107618. https://doi.org/10.1016/j.neuropsychologia.2020.107618Across four sessions the landmark bias was reliable (Cronbach's alpha 0.80) and the line bisection bias was not.
Source papers
- Milner AD, Brechmann M, Pagliarini L (1992). To halve and to halve not: an analysis of line bisection judgements in normal subjects. Neuropsychologia, 30(6):515-526.Paradigm source, Adults
- Learmonth G, Gallagher A, Gibson J, Thut G, Harvey M (2015). Intra- and inter-task reliability of spatial attention measures in pseudoneglect. PLoS ONE, 10(9):e0138379.Paradigm source, Young adults
Questions
How long does the Landmark Task take?
About 6 to 7 minutes, as listed in the test library.
What device does a participant need?
Timed test: participants need a desktop or laptop with a mouse or trackpad. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Point of Subjective Equality, Just Noticeable Difference, Left Key Rate, Left Part Longer by Offset, Accuracy, Accuracy at the Largest Offsets, Largest-Offset Trials Scored, Median Answer Time, Anticipations, Early Key Presses, Frame Interval, Completed Trials and Planned Trials. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Landmark Task used in research?
33 papers from 2021 to 2025 name it in their title or abstract, down 41% compared with 2016 to 2020. That ranks it 290 of 395 test families in the ConductCognition test census (use grade D). Source: ConductCognition test census 0.14.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Landmark Task have?
In the published research, 4 papers report its test-retest reliability and 2 registered clinical studies use it as an outcome (evidence grade C in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
Which settings can a researcher change?
Trials at each mark position (4 to 12). Settings can be changed on yearly plans; free and pay-as-you-go studies use the standard settings.
Which languages can participants use?
Participants can take this test in English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese. English is on every plan, and the other languages come with yearly plans. Have a native speaker read the machine-translated text before launch.
Is this a clinical test?
No. ConductCognition is for research use only. It runs, scores and exports the test; clinical interpretation is outside the platform.
Every test: the full test list.
Run this test in your next study
Start free with 10 free participants on any test, then use participant credits or a yearly plan.