Skip to content

Explore methods through figures

A researcher should be able to recognize what a method shows before copying code. This visual explorer collects 12 images already produced by the existing deterministic documentation scripts. These are not new data or independently calibrated claims.

Start with an analysis path Complete generated figure gallery Method selector

Read the figure together with its assumptions. Synthetic teaching images illustrate API behavior. Diagnostic images expose sensitivity or fit characteristics; neither category by itself proves instrument validity, statistical coverage, or missingness identification.

12 of 12 figures shown

Planar gaze trajectories — deterministic synthetic documentation figure

Whole-path spatial structure · Synthetic teaching

Planar gaze trajectories

Displays the recorded 2-D path concept; it cannot diagnose psychological attention or event quality.

Guide · Method or evidence

Observed support versus latent truth — deterministic synthetic documentation figure

Missingness and known simulation truth · Synthetic teaching

Observed support versus latent truth

Illustrates how a simulator retains ground truth; real recordings do not expose their latent truth.

Guide · Method or evidence

Functional principal-component shapes — deterministic synthetic documentation figure

FPCA modes of variation · Synthetic teaching

Functional principal-component shapes

Component perturbations summarize variation, not a latent cognitive mechanism.

Guide · Method or evidence

Cumulative explained functional variance — deterministic synthetic documentation figure

Variance explained across components · Synthetic teaching

Cumulative explained functional variance

Retained variance helps describe a fitted representation; it does not choose a confirmatory model automatically.

Guide · Method or evidence

Sparse latent covariance surface — deterministic synthetic documentation figure

Sparse FPCA covariance · Synthetic teaching

Sparse latent covariance surface

Smoothing and PSD decisions affect the estimated covariance and must be reported.

Guide · Method or evidence

Sparse score conditioning — deterministic synthetic documentation figure

Conditional covariance numerical behavior · Diagnostic

Sparse score conditioning

Condition numbers signal sensitivity, not certified uncertainty coverage.

Guide · Method or evidence

Registration phase displacement — deterministic synthetic documentation figure

Temporal warping / alignment · Synthetic teaching

Registration phase displacement

Phase displacement should be reported rather than hidden in preprocessing.

Guide · Method or evidence

Continuous planar curvature — deterministic synthetic documentation figure

Signed local geometry · Synthetic teaching

Continuous planar curvature

Curvature is derived geometry and is sensitive to acquisition noise and smoothing choices.

Guide · Method or evidence

Time-varying predictor coefficient — deterministic synthetic documentation figure

Functional regression and bands · Synthetic teaching

Time-varying predictor coefficient

A displayed band is method- and design-conditional, not universally calibrated.

Guide · Method or evidence

Repeated-trial residual autocorrelation — deterministic synthetic documentation figure

Within-participant dependence · Diagnostic

Repeated-trial residual autocorrelation

Residual structure is a diagnostic; it does not select a covariance family without a declared rule.

Guide · Method or evidence

Multichannel surrogate diagnostics — deterministic synthetic documentation figure

Cross-channel spectrum preservation · Diagnostic

Multichannel surrogate diagnostics

Surrogate diagnostics describe a declared null and algorithmic mismatch, not all nonlinear structure.

Guide · Method or evidence

Fixed-family wild-bootstrap tests — deterministic synthetic documentation figure

Adjusted p-values for declared target families · Synthetic teaching

Fixed-family wild-bootstrap tests

Studentization and family definition govern interpretation; testing many targets post hoc is different.

Guide · Method or evidence

How to use a figure as research evidence

  1. Start with actual observations. Declare sampling support, units, time origins and participant/trial identifiers. See input schemas.
  2. Read the fitted estimand. A component, predictor coefficient, covariance surface and adjusted p-value answer different questions. See mathematical reference.
  3. Check failure-inclusive evidence. Simulation figures are designed for teaching; independent tests and research exclusions govern stronger claims.
  4. Reproduce the figure. The full generated gallery is rebuilt in documentation CI, rather than being hand-drawn or manually altered for this explorer.

The existing research evidence figure atlas remains separately labelled; its experimental results must not be mixed with these API teaching figures.

Full figure atlas · Reproducible workflows · Research evidence