
Compare observed, predicted and residual EEG scalp maps through time
Source:R/plot_spatiotemporal_topography.R
plot_spatiotemporal_topography.RdPlots interpolated maps of observed sensor-level means, population-level posterior median expected responses, and residuals (`observed - predicted`). Inferences and clusters remain at the *original sensor locations*, not at interpolated pixels. Interpolation is for visualisation only and uses inverse-distance weighting (IDW) without additional package dependencies.
Usage
plot_spatiotemporal_topography(
results,
observed_data,
observed_data_b = NULL,
outcome_id = "voltage",
sensor_id = "channel",
time_id = "time",
spatial_id = c("xproj", "yproj"),
participant_id = NULL,
times = NULL,
n_times = 8L,
what = c("observed", "predicted", "residual"),
grid_n = 85L,
idw_power = 2,
limits = NULL,
contour = TRUE,
show_sensors = TRUE,
show_labels = FALSE,
head_outline = TRUE,
cluster_overlay = c("predicted", "all", "none"),
cluster_alpha = 0.9,
cluster_size = 2.8
)Arguments
- results
Object returned by [analyse_bgam_spatiotemporal()].
- observed_data
Original long-form observations, or condition A data.
- observed_data_b
Observed condition B if `results` is an A - B contrast.
- outcome_id
Observed outcome column.
- sensor_id
Sensor identifier shared by observed data and predictions.
- time_id
Numeric time column.
- spatial_id
Two projected spatial coordinate columns.
- participant_id
Optional participant ID for an equal-participant- weighted observed average. With `NULL`, observations are pooled.
- times
Optional requested times; closest existing samples are used.
- n_times
Number of evenly distributed panels if `times = NULL`.
- what
One or more of `"observed"`, `"predicted"`, `"residual"`.
- grid_n
Number of interpolated pixels per grid axis.
- idw_power
Positive IDW inverse-distance exponent.
- limits
Symmetric range half-width of a shared color scale, or `NULL` for automatic scaling across all selected panels.
- contour
Whether to draw interpolated iso-value contours.
- show_sensors
Show original electrode locations.
- show_labels
Show sensor names (can be busy with dense montages).
- head_outline
Add schematic scalp circle and nose.
- cluster_overlay
`"predicted"` highlights inference clusters only on predicted maps; `"all"` on all maps; `"none"` disables highlights.
- cluster_alpha
Transparency of cluster-marking rings.
- cluster_size
Size of cluster-marking rings.
Author
Ladislas Nalborczyk ladislas.nalborczyk@cnrs.fr.
Examples
if (FALSE) { # \dontrun{
# Not run: generating `results_st` requires
# computationally intensive Bayesian model fitting.
# `eeg_data2` contains the corresponding observations.
plot_spatiotemporal_topography(
results_st, eeg_data2, outcome_id = "voltage",
sensor_id = "channel", n_times = 8
)
} # }