---
title: "Representative Scanpaths and Cognitive Episodes"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Representative Scanpaths and Cognitive Episodes}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(collapse = TRUE, comment = "#>", eval = FALSE)
```


A representative scanpath is accompanied by a distance matrix and dispersion estimate rather than selected visually.

```{r}
representative <- representative_scanpath(
  scanpath_long,
  method = "medoid",
  distance = "edit",
  id_col = "person_trial_id",
  aoi_col = "aoi"
)
plot_representative_scanpath(representative)
plot_scanpath_dispersion(representative)
plot_scanpath_similarity_matrix(representative)
comparison <- compare_scanpath_distributions(
  scanpath_list,
  group = strategy_group,
  distance = "edit"
)
plot_group_scanpath_transport(comparison)
```

Change-point segmentation provides candidate cognitive episodes. Labels are transparent study-specific rules, not latent psychological facts.

```{r}
changes <- detect_process_changepoints(
  multimodal_stream,
  channels = c("gaze_velocity", "pupil_bc", "eda"),
  time_col = "time"
)
plot_changepoint_ribbons(changes)
episodes <- label_process_episodes(segment_process_episodes(changes))
plot_process_episodes(episodes)
plot_episode_transition_graph(episodes)
plot_episode_duration_distribution(episodes)
```