In mne-tools/mne-r: Fast Access to MNE-Python From Within R

This is a basic demo showcasing the reticulate package through which we obtain access to MNE-Python from inside R. We will read in the raw data, epoch it and then compute the evoked response. Subsequently we will use ggplot to visualize the results.

library(tidyverse)
library(mne)

Let's read in the raw data.

data_path <- mne$datasets$sample$data_path()
subject <- "sample"

raw_fname <- paste(data_path,
                   'MEG', 
                   subject,
                   'sample_audvis_filt-0-40_raw.fif',
                   sep = '/')

raw <- mne$io$read_raw_fif(raw_fname, preload = T)

Let's now quickly preprocess the data.

raw$filter(1, 40)

We can now go ahead and compute evokeds.

events <- mne$find_events(raw)
storage.mode(events) <- "integer"  # R gets the events as floats.

tmin <- -0.2
tmax <- 0.5
baseline <- reticulate::tuple(NULL, 0)
event_id <- list("aud/l" = 1L, "aud/r" = 2L,
                 "vis/l" = 3L, "vis/r" = 4L)
picks <- mne$pick_types(raw$info, meg = T, eeg = T)
epochs <- mne$Epochs(raw = raw, events = events, event_id =event_id,
                     tmin = tmin, tmax = tmax,
                     picks = picks %>% as.integer(),
                     baseline = baseline, reject = NULL, preload = T) 

evoked <- epochs$average()

Now it's time to do a bit of R processing for getting doing custom evoked plots.

# use MNE-R function
evoked_df <- mne::get_data_frame(evoked)

Let's plot it using ggplot2.

ggplot(
  data = evoked_df,
  mapping = aes(x = time, color = channel, y = observation)) +
  geom_line(mapping = aes(group = channel)) +
  facet_wrap(
    ~ch_type,
    nrow = 3,
    scales = "free",
    strip.position = "left",
    labeller = as_labeller(c(eeg = "EEG [mV]", grad = "GRAD [fT/cm]",
                             mag = "MAG [fT]"))) +
  theme_minimal() +
  guides(color = F) +
  labs(x = 'time [ms]', y = NULL) +
  theme(text = element_text(family = "Helvetica", size = 24))

Let's see if we can add spatial colors as in the MNE plot.

# we look up a private function from the epochs object
pos <- epochs$`_get_channel_positions`()
# we do the same normalization as in MNE
pos <- pos - apply(pos, 1, min)
pos <- pos / apply(pos, 1, function(x) {max(x, 1e-16)})

# we set rgb values
evoked_df$r <- pos[, 1]
evoked_df$g <- pos[, 2]
evoked_df$b <- pos[, 3]

ggplot(
  data = evoked_df,
  mapping = aes(x = time, color = channel, y = observation)) +
  geom_line(mapping = aes(group = channel,
                          color = rgb(r, g, b)),
            size = 0.8, alpha=0.7) +
  facet_wrap(
    ~ch_type,
    nrow = 3,
    scales = "free",
    strip.position = "left",
    labeller = as_labeller(c(eeg = "EEG [mV]",
                             grad = "GRAD [fT/cm]",
                             mag = "MAG [fT]"))) +
  theme_minimal() +
  guides(color = F) +
  labs(x = 'time [ms]', y = NULL) +
  theme(text = element_text(family = "Helvetica", size = 24))