CoChat
Sign Up
Cited by 235Gaze Tracking and Assistive TechnologyNeural and Behavioral Psychology StudiesHuman-Automation Interaction and Safety

The DOI matches its Crossref record; 106 of the 107 references it cites were found in a registry or an index.

Methods in cognitive pupillometry: Design, preprocessing, and statistical analysis

Open access
Mathôt, Sebastiaan, Vilotijević, Ana·2022·Behavior Research Methods
DOI 10.3758/s13428-022-01957-7PMC PMC10556184

Published version of a preprint

Open this paper in CoChat — free accountOpen PDF

Abstract

Cognitive pupillometry is the measurement of pupil size to investigate cognitive processes such as attention, mental effort, working memory, and many others. Currently, there is no commonly agreed-upon methodology for conducting cognitive-pupillometry experiments, and approaches vary widely between research groups and even between different experiments from the same group. This lack of consensus makes it difficult to know which factors to consider when conducting a cognitive-pupillometry experiment. Here we provide a comprehensive, hands-on guide to methods in cognitive pupillometry, with a focus on trial-based experiments in which the measure of interest is the task-evoked pupil response to a stimulus. We cover all methodological aspects of cognitive pupillometry: experimental design, preprocessing of pupil-size data, and statistical techniques to deal with multiple comparisons when testing pupil-size data. In addition, we provide code and toolboxes (in Python) for preprocessing and statistical analysis, and we illustrate all aspects of the proposed workflow through an example experiment and example scripts.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.3758/s13428-022-01957-7

Cites

106 of 106 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 A schematic representation of the proposed workflow for conducting cognitive-pupillometry experiments
Fig. 1 A schematic representation of the proposed workflow for conducting cognitive-pupillometry experimentsCC BY · PubMed Central
Fig. 2 Schematic trial sequence for the example experiment. Participants reported whether a target was a luminance increment or, as in this example, a luminance decrement embedded in a dynamic stream of noise. The probable eccentricity (near, medium, far) of the target was indicated by a symbolic cue (a square in this example)
Fig. 2 Schematic trial sequence for the example experiment. Participants reported whether a target was a luminance increment or, as in this example, a luminance decrement embedded in a dynamic stream of noise. The probable eccentricity (near, medium, far) of the target was indicated by a symbolic cue (a square in this example)CC BY · PubMed Central
Fig. 3 A schematic representation of a DataMatrix object. Each row represents a trial. Each column represents a variable. Pupil-size data (here: pupil_stream) is stored in a special column type (SeriesColumn) that has an additional dimension to store how pupil size changes as a function of time (sample number)
Fig. 3 A schematic representation of a DataMatrix object. Each row represents a trial. Each column represents a variable. Pupil-size data (here: pupil_stream) is stored in a special column type (SeriesColumn) that has an additional dimension to store how pupil size changes as a function of time (sample number)CC BY · PubMed Central
Fig. 4 Two individual trials during which a blink occurred. Blinks are often characterized by a sharp drop in pupil size (due to closing of the eyelid) followed by a period of missing data (due to the eyelid being fully closed), followed in turn by sharp increase in pupil size (due to the eyelid reopening). a An example of cubic-spline interpolation, based on four points around the blink. This results in a smooth curve. b An example of linear interpolation, based on two points around the blink. This results in a straight line
Fig. 4 Two individual trials during which a blink occurred. Blinks are often characterized by a sharp drop in pupil size (due to closing of the eyelid) followed by a period of missing data (due to the eyelid being fully closed), followed in turn by sharp increase in pupil size (due to the eyelid reopening). a An example of cubic-spline interpolation, based on four points around the blink. This results in a smooth curve. b An example of linear interpolation, based on two points around the blink. This results in a straight lineCC BY · PubMed Central
Fig. 5 Pupil size during a single trial from our example experiment. The left y-axis shows pupil size in arbitrary units as recorded by the EyeLink eye tracker. The right y-axis shows pupil size in millimeters of diameter based on a conversion formula. The lines do not overlap perfectly because the conversion is not linear
Fig. 5 Pupil size during a single trial from our example experiment. The left y-axis shows pupil size in arbitrary units as recorded by the EyeLink eye tracker. The right y-axis shows pupil size in millimeters of diameter based on a conversion formula. The lines do not overlap perfectly because the conversion is not linearCC BY · PubMed Central
Fig. 6 Pupil size during a selection of individual trials from our example experiment. a Before baseline correction. b After baseline correction
Fig. 6 Pupil size during a selection of individual trials from our example experiment. a Before baseline correction. b After baseline correctionCC BY · PubMed Central
Fig. 7 Visualization of data quality for a single participant from our example experiment. a A plot with pupil size during all trials plotted as semitransparent lines allows you to check visually for distortions, such as downward spikes. b A histogram of baseline pupil sizes allows you to check visually for clusters of outliers
Fig. 7 Visualization of data quality for a single participant from our example experiment. a A plot with pupil size during all trials plotted as semitransparent lines allows you to check visually for distortions, such as downward spikes. b A histogram of baseline pupil sizes allows you to check visually for clusters of outliersCC BY · PubMed Central
Fig. 8 A plot with the mean number of blinks per trial as a function of experimental condition and participant allows you to check visually for systematic differences in blink rate between conditions. Error bars indicate 95% between-subject bootstrapped confidence intervals. Based on complete data from our example experiment
Fig. 8 A plot with the mean number of blinks per trial as a function of experimental condition and participant allows you to check visually for systematic differences in blink rate between conditions. Error bars indicate 95% between-subject bootstrapped confidence intervals. Based on complete data from our example experimentCC BY · PubMed Central
Fig. 9 A plot with the deviation of horizontal (a) and vertical (b) eye position from the display center as a function of experimental condition allows you to visually check whether there are systematic differences in gaze deviation between conditions. Error shadings indicate grand standard error. Based on complete data from our example experiment
Fig. 9 A plot with the deviation of horizontal (a) and vertical (b) eye position from the display center as a function of experimental condition allows you to visually check whether there are systematic differences in gaze deviation between conditions. Error shadings indicate grand standard error. Based on complete data from our example experimentCC BY · PubMed Central
Fig. 10 Results from our example experiment. Mean baseline-corrected pupil size during the 750–3000 ms window after cue onset as a function of cue eccentricity. Values are negative due to an overall pupil constriction relative to the baseline period (see also Fig. 12). Error bars indicate grand standard error
Fig. 10 Results from our example experiment. Mean baseline-corrected pupil size during the 750–3000 ms window after cue onset as a function of cue eccentricity. Values are negative due to an overall pupil constriction relative to the baseline period (see also Fig. 12). Error bars indicate grand standard errorCC BY · PubMed Central
Fig. 11 A schematic explanation of four-fold cross-validation. See main text for details
Fig. 11 A schematic explanation of four-fold cross-validation. See main text for detailsCC BY · PubMed Central
Fig. 12 Results from our example experiment. The y-axis represents baseline-corrected pupil size. The lower x-axis represents time in milliseconds since cue onset. The upper x-axis represents the order of the events in the experiment. Differently colored lines represent the three cue eccentricities (far, medium, near). Error bands indicate grand standard error. The vertical line indicates the mean of the samples at which the effect of cue eccentricity was tested
Fig. 12 Results from our example experiment. The y-axis represents baseline-corrected pupil size. The lower x-axis represents time in milliseconds since cue onset. The upper x-axis represents the order of the events in the experiment. Differently colored lines represent the three cue eccentricities (far, medium, near). Error bands indicate grand standard error. The vertical line indicates the mean of the samples at which the effect of cue eccentricity was testedCC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 107 references

Research works printed but not linked: 2

  1. Van de Kraats, J., Smit, E. P., & Slooter, J. H. (1977). Objective perimetric measurements by the pupil balance method. Documenta Ophthalmologica Proceedings Series, 14, 213–219.
  2. Vilotijević, A., & Mathôt, S. (2022). Covert shifts of attention towards the visual periphery induce pupil dilation (p. 2022.07.14.500035). bioRxiv. https://doi.org/10.1101/2022.07.14.500035 · DOI

Cited by

Sorted by citations

Related

See all

Metadata

Authors
Mathôt, S., & Vilotijević, A.
Affiliations
University of Groningen (NL)
Corresponding author
Mathôt, Sebastiaan
ORCID
Venue
Behavior Research Methods
Publisher
Springer Science and Business Media LLC
Published
2022-08-26
Volume
55
Issue
6
Pages
3055-3077
ISSN
  • 1554-351X (ISSN-L)
  • 1554-3528
DOI
10.3758/s13428-022-01957-7
PMC
PMC10556184
OpenAlex
W4293794846
Cited by
235
References
107
Citations by year
2026: 65 · 2025: 95 · 2024: 46 · 2023: 22 · 2022: 5
FWCI
20.78
Type
journal article
Language
en
Open access
Licence
Locations
  • Behavior Research Methods (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • University of Groningen research database (University of Groningen / Centre for Information Technology) (submittedVersion, cc-by)
  • OSF Preprints (OSF Preprints) (submittedVersion)
Funders
Nederlandse Organisatie voor Wetenschappelijk Onderzoek (VI.Vidi.191.045)
Topics
  • Physical Sciences › Computer Science › Human-Computer Interaction › Gaze Tracking and Assistive Technology
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Neural and Behavioral Psychology Studies
  • Social Sciences › Human Factors and Ergonomics › Human-Automation Interaction and Safety
Keywords
pupil size, experimental design, multiple comparisons, statistical analysis, attention, cognitive effort, working memory, Python library
MeSH terms
Attention, Cognition, Humans, Memory, Short-Term, Pupil, Research Design

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review