Skip to the question
AI Certainty / MentalKG
中文

MentalKGResearch uses preview

Mental rotation · question → evidence → use

Does a larger pupil change always mean better cognitive performance?

Read the finding first. Then check what it can be used for.

01

Current answer

Original article ↗

Not necessarily. A larger change can accompany slower, less accurate responses.

This example concerns a mental-rotation task. Its condition contrasts do not turn pupil change into a personal ability score or directly test working-memory capacity.

1 study2 linked cardsSame paper and overall sample
02

What can these cards support?

Each label describes completeness for that use. It is not scientific validation or a quality certificate.

Author report · Limited display

These attributed summaries preserve the checked results and their limits. Missing model N remains visible; it does not by itself hide a faithful summary.

Quantitative pooling · Not ready for pooling

No pooling target, scale, variance/correlation, or cohort-deduplication plan has been assessed for these cards. Display permission does not supply those inputs.

Independent reproduction · Not performed here

Checking the article and table cells is not rerunning original data and analysis code. No such run was performed here; this says nothing about other researchers’ work.

Causal support · Not established

The study compares task conditions; it does not intervene on pupils. It does not establish that changing a pupil changes cognitive performance, even if an analysis were reproduced.

03

Two views of one study

Attributed author report

01 / 02

Stimulus-condition contrast

Bauer and colleagues report that, in the model’s main contrast, cube figures had larger pupil changes, slower responses and lower accuracy than body postures.

Model main contrast · cube figures minus body postures

Model N: unknownModel trial count: unknown

Read the original study ↗
Author values, units & intervals
Selected author-reported coefficients; not re-estimated.
Outcome / unitEstimate / SE95% CI
+1.03SE 0.07
[0.90, 1.17]
Accuracy ↗log odds
− 2.58SE 0.39
[− 3.72, − 1.96]
+0.22SE 0.04
[0.14, 0.30]

95% CIs use 1,000 parametric bootstrap simulations. The finer interval algorithm is not reported in the checked material. These are not relabeled as Wald intervals. SEs are retained from the author’s tables, not inferred from CI widths.

Accuracy coefficients are log odds, not percentage-point changes. Pupil coefficients retain the printed 10⁻¹ mm unit; no scale conversion or standardized effect was computed.

Exact table-row locations

Tab1 · STI(CF-BP)/article/body/sec[7]/sec[1]/table-wrap[1]/table[1]/tbody[1]/tr[13]

Tab2 · STI(CF-BP)/article/body/sec[7]/sec[2]/table-wrap[1]/table[1]/tbody[1]/tr[13]

Tab3 · STI(CF-BP)/article/body/sec[7]/sec[3]/table-wrap[1]/table[1]/tbody[1]/tr[12]

Sample & measurement limits

109 students participated, 29 were excluded for software failure and 80 were retained overall. Final participant and trial counts after filtering are not enumerated for each model in the checked text and tables; 80 is not substituted as model N.

Pupil change is trial maximum minus the baseline from about the first 40 ms after stimulus onset. Carry-over from preceding trials limits interpretation.

  • Only non-mirrored pairs enter these analyses; angle-specific reaction-time outliers are excluded.
  • Response-time and pupil models use correct responses. Accuracy retains the correct/incorrect outcome.
  • Pupil analyses apply additional eye-data and baseline/maximum filters. Whether missing eye data also excludes behavioral observations is unclear.
  • Task-condition effects are not between-person pupil–performance correlations, working-memory-capacity measurements, or pupil interventions.

D405 · Original frozen review status retained; this preview changes no production admission.

Attributed author report

02 / 02

Angular-disparity main term

Bauer and colleagues report an angle main term with larger pupil changes, slower responses and lower accuracy. The stimulus × angle interaction means this is not a common slope for every stimulus.

Model main term · per 100° · stimulus × angle interaction retained

Model N: unknownModel trial count: unknown

Read the original study ↗
Author values, units & intervals
Selected author-reported coefficients; not re-estimated.
Outcome / unitEstimate / SE95% CI
Response time ↗seconds / 100°
+0.91SE 0.04
[0.83, 1.00]
Accuracy ↗log odds / 100°
− 1.95SE 0.18
[− 2.33, − 1.61]
Baseline-corrected pupil change ↗10⁻¹ mm / 100° (as printed)
+0.28SE 0.04
[0.21, 0.35]

95% CIs use 1,000 parametric bootstrap simulations. The finer interval algorithm is not reported in the checked material. These are not relabeled as Wald intervals. SEs are retained from the author’s tables, not inferred from CI widths.

Accuracy coefficients are log odds, not percentage-point changes. Pupil coefficients retain the printed 10⁻¹ mm unit; no scale conversion or standardized effect was computed.

Exact table-row locations

Tab1 · DEG (100°)/article/body/sec[7]/sec[1]/table-wrap[1]/table[1]/tbody[1]/tr[10]

Tab2 · DEG (100°)/article/body/sec[7]/sec[2]/table-wrap[1]/table[1]/tbody[1]/tr[10]

Tab3 · DEG (100°)/article/body/sec[7]/sec[3]/table-wrap[1]/table[1]/tbody[1]/tr[9]

Sample & measurement limits

109 students participated, 29 were excluded for software failure and 80 were retained overall. Final participant and trial counts after filtering are not enumerated for each model in the checked text and tables; 80 is not substituted as model N.

Pupil change is trial maximum minus the baseline from about the first 40 ms after stimulus onset. Carry-over from preceding trials limits interpretation.

  • Only non-mirrored pairs enter these analyses; angle-specific reaction-time outliers are excluded.
  • Response-time and pupil models use correct responses. Accuracy retains the correct/incorrect outcome.
  • Pupil analyses apply additional eye-data and baseline/maximum filters. Whether missing eye data also excludes behavioral observations is unclear.
  • Task-condition effects are not between-person pupil–performance correlations, working-memory-capacity measurements, or pupil interventions.

D406 · Original frozen review status retained; this preview changes no production admission.

04

One original publication

Robert Bauer, Leonardo Jost, Bianca Günther, Petra Jansen

Pupillometry as a measure of cognitive load in mental rotation tasks with abstract and embodied figures

Psychological Research · Online 2021 · volume 86 (2022)

Citation & source-checking scope

Bauer, R.; Jost, L.; Günther, B.; Jansen, P. (2022). Pupillometry as a measure of cognitive load in mental rotation tasks with abstract and embodied figures. Psychological Research 86, 1382–1396. https://doi.org/10.1007/s00426-021-01568-5

Summaries and translations are assistant-written. Six selected rows retain author-reported values; no images, full tables, participant data or article text are reproduced.

The supplied isolated proposal and licensed XML bind the six rows. Article-level CC BY 4.0 was checked in the original permissions. This is text/table checking, not raw-data reproduction or scientific certification.

Frozen source fingerprintSHA-256 1ca5c28a145ef6f7a5a2da96c522077b009ce85218c522dd893a0d7ab30729bd
Candidate version & use boundaries

mentalkg-use-candidate-20261008.1 · mentalkg-isolated-eligibility-proposal-20261008.2

Local candidate only. No production admission, quantitative pooling, original-data run, clinical scoring, or causal upgrade has been performed. No model is called and no new answer is generated.

Whether a pooling method needs model N requires a specified method and scientific review. Missing N is not a permanent ban on every conceivable method; no actual pooling plan has been assessed here.

Checking material structure and locations does not authenticate every scientific claim.