PhysioNIRS

PhysioNIRS provides governed SNIRF input/output, optical-density, haemoglobin, motion-correction, and short-separation contracts for near-infrared spectroscopy data represented as PhysioExperiment objects.

library(PhysioNIRS)

# A continuous-wave SNIRF file bundled with the package (no download needed).
snirf <- system.file(
  "extdata", "snirf_official_simple_probe.snirf", package = "PhysioNIRS"
)
x <- readSNIRF(snirf)
measurementList(x)
sourceDetectorDistances(x, unit = "m")

x_od <- intensityToOD(x, assay_name = "raw")   # raw intensity -> optical density
x_hb <- mbll(x_od, assay_name = "OD")          # optical density -> haemoglobin

writeSNIRF(x, file.path(tempdir(), "roundtrip.snirf"))

For a governed optical-density assay, short-separation processing is explicit:

short <- identifyShortChannels(x_od, threshold_m = 0.01)

# When the probe includes short channels, regress them out and build a
# physiology nuisance design. (This symmetric example probe has only
# standard-separation channels, so the step below is skipped here.)
if (any(short$is_short) && !all(short$is_short)) {
  corrected <- shortSeparationRegress(
    x_od, assay_name = "OD", short = short
  )
  nuisance <- shortSeparationDesign(
    x_od,
    assay_name = "OD",
    short = short,
    aggregation = "mean",
    physiology_bands = c("mayer", "respiration")
  )
}

Short-channel regression is a signal-processing choice, not evidence that a channel contains only extracerebral physiology. Named physiology bands are nuisance ranges rather than diagnoses, and population-appropriate explicit ranges should be used when the adult defaults are unsuitable.

SNIRF metadata can contain subject identifiers and acquisition dates. Remove or replace identifying metadata before sharing files.

Installation

The ecosystem builds on Bioconductor, so its repositories have to be on the list as well -- without them the install stops at SummarizedExperiment.

install.packages("BiocManager", repos = "https://cloud.r-project.org")
install.packages(
  "PhysioNIRS",
  repos = c("https://x-biosignal.r-universe.dev", BiocManager::repositories())
)

From GitHub instead:

install.packages("remotes", repos = "https://cloud.r-project.org")
remotes::install_github("x-biosignal/PhysioNIRS")