Skip to contents

R-CMD-check pkgdown Codecov test coverage License: MIT Lifecycle: experimental

This vignette demonstrates descriptive saddle-pressure analysis using synthetic measurements. Source-specific observations are not universal diagnostic limits.

Load the saddle layout

layout <- pr_layout_saddle("horse")
names(layout$regions)
#> [1] "cranial_left"  "cranial_right" "middle_left"   "middle_right" 
#> [5] "caudal_left"   "caudal_right"

Load a synthetic trial

trial <- pr_example_trial("saddle_horse")
trial
#> 
#> ── pr_trial ────────────────────────────────────────────────────────────────────
#> • System: "saddle_mat"
#> • Layout: "saddle_horse"
#> • Frames: 500
#> • Duration: 9.98 s
#> • Sampling: 50 Hz
#> • Sensors: 248
#> • Subject: "HORSE01"
#> • Date: "2026-11-22"
#> • Condition: "walk"

Inspect the pressure map

Saddle pressure heatmap with 6-region mask

Regional analysis

masks <- pr_mask_saddle_6(layout)
reg <- pr_calc_regional(trial, masks)
reg
#> # A tibble: 6 × 6
#>   region          mpp   mvp max_force contact_area pti_mean
#>   <chr>         <dbl> <dbl>     <dbl>        <dbl>    <dbl>
#> 1 cranial_left   18.2  2.41     105.           320     17.7
#> 2 cranial_right  16.5  2.13      94.5          320     15.4
#> 3 middle_left    11.9  1.71      84.1          370     12.7
#> 4 middle_right   10.5  1.53      72.6          370     11.2
#> 5 caudal_left    18.3  2.03     113.           430     14.4
#> 6 caudal_right   15.4  1.80      98.5          410     12.6

Regional mean-pressure bar chart

Bridge detection

bridge <- pr_calc_saddle_bridge(trial)
bridge
#> 
#> ── Saddle Bridge Analysis ──
#> 
#> • Bridge ratio: 0.774
#> • Bridging: FALSE
#> Pressure distribution appears continuous across the saddle (ratio = 0.77).

Slip / asymmetric loading

slip <- pr_calc_saddle_slip(trial)
slip$symmetry_index_mean
#> [1] 10.97179
slip$dominant_side
#> [1] "balanced"

Composite report

Saddle-fit composite report

Reference context and reproducibility

pr_ref_saddle() now attributes the six source observations to Nyikos (2005). An overlay requires matching species, gait and mask definitions recorded in trial$metadata$reference_context; synthetic examples do not establish that context. Unsupported legacy numeric entries are withdrawn.

Novel Pliance files are detected by their header, including files with a temporary upload name. The Novel layout retains grid coordinates and unknown sensor area: force, physical contact area and millimetre distances cannot be inferred. pr_summary_long() states units and availability explicitly.

Use pr_temporal_profile(trial, "corrected") for measured sampling and stable plateau handling. legacy_method repeats the historical 50 Hz assumptions but cannot guarantee identical cycles across numerical environments. Exact historical cycles must be bound to the recording with pr_lock_cycles() and replayed with pr_temporal_profile(trial, "reference", locked_cycles). Reference phase maps use the central 80 percent of saved cycles and an untrimmed mean.

Source

Nyikos S, et al. (2005). Pferdeheilkunde 21:187-198, abstract p.187. Original article.