Skip to main content

MNERawData

Namespace: MNELIB  ·  Library: MNE Library

#include <mne/mne_raw_data.h>

class MNELIB::MNERawData

Implements the MNE Raw Data (Replaces *mneRawData,mneRawDataRec; struct of MNE-C mne_types.h).

A comprehensive raw data structure

src/examples/ex_mne_api/main.cpp
// MNE-C style reader: buffer directory, channel selection, projection and overlap-add filter
MNEFilterDef filter; // 40 Hz lowpass, 5 Hz transition, 4096-sample blocks
filter.filter_on = true;
filter.size = 4096;
filter.taper_size = 2048;
filter.lowpass = filter.eog_lowpass = 40.0f;
filter.lowpass_width = filter.eog_lowpass_width = 5.0f;
std::unique_ptr<MNERawData> mneRaw(MNERawData::open_file(rawFile.fileName(), false, false, filter));
const MNERawInfo& rawInfo = *mneRaw->info;
const MNERawBufDef& firstBuffer = mneRaw->bufs.front();

MNEChSelection sel; // channel names are stored without spaces
sel.chspick = sel.chspick_nospace = {"MEG0113", "EEG001", "EOG061"};
sel.nchan = sel.ndef = static_cast<int>(sel.chspick.size());
sel.pick_deriv = VectorXi::Constant(sel.nchan, -1);
sel.pick.resize(sel.nchan);
for (int c = 0; c < sel.nchan; ++c) {
    sel.pick[c] = static_cast<int>(mneRaw->ch_names.indexOf(sel.chspick[c]));
}

const int ns = 25;
Matrix<float, Dynamic, Dynamic, RowMajor> segment(sel.nchan, ns);
std::vector<float*> rows{segment.row(0).data(), segment.row(1).data(), segment.row(2).data()};
mneRaw->pick_data(&sel, mneRaw->first_samp + 2990, ns, rows.data()); // calibrated, across a buffer boundary
const double eegRaw = segment(1, 0);
mneRaw->pick_data_filt(&sel, mneRaw->first_samp + 2990, ns, rows.data());
const double eegFilt = segment(1, 0);

// The file stores its 4 SSP projectors as inactive; makeProjection switches them on
std::unique_ptr<MNEProjOp> proj;
MNEProjOp::makeProjection({rawFile.fileName()}, rawInfo.chInfo, rawInfo.nchan, proj);
proj->assign_channels(mneRaw->ch_names, rawInfo.nchan);
proj->make_proj();
const MNEProjItem& firstItem = proj->items[0];
mneRaw->proj = std::move(proj);
mneRaw->pick_data_proj(&sel, mneRaw->first_samp + 2990, ns, rows.data());
const double eegProj = segment(1, 0);

Public Methods​

MNERawData()​

Constructs the MNE Raw Data Refactored: new_raw_data (mne_raw_data.c).

~MNERawData()​

Destroys the MNE Raw Data Refactored: mne_raw_free_data (mne_raw_data.c).

add_filter_response(highpass_effective)​

Create a frequency-domain filter response from the current filter definition and sampling frequency, replacing any existing response.

Parameters:

  • highpass_effective : *int ** Set to non-zero if the highpass filter is active.

setup_filter_bufs()​

Allocate filter output buffers spanning the entire raw data range (with taper overlap), initialize a ring buffer for filtered data, and create the filter response.

setFilter(newFilter)​

Switch to another filter.

Already filtered data is discarded, so later pick_data_filt calls return data filtered with the new definition. Refactored: filter change handling of mne_browse_raw (analyze.c, mne_compare_filters, mne_raw_forget_filtered).

Parameters:

  • newFilter : const MNEFilterDef & The new filter definition.

Returns:

  • bool — true if the filter changed, false if it equals the current one.

load_one_buffer(buf)​

Read a single raw data buffer from the FIFF file into memory, allocating from the ring buffer if needed.

Parameters:

Returns:

  • int — OK on success, FAIL on read error.

compensate_buffer(buf)​

Apply CTF compensation to a loaded raw data buffer if its compensation status differs from the desired one.

Parameters:

Returns:

  • int — OK on success, FAIL on error.

pick_data(sel, firsts, ns, picked)​

Extract raw data samples for selected channels (including derived channels) with compensation applied, zero-padding or edge-extending as needed at boundaries.

Parameters:

  • sel : mneChSelection Channel selection descriptor.

  • firsts : int First sample index (absolute).

  • ns : int Number of samples to extract.

  • picked : *float *** Array of per-channel data vectors (pre-allocated).

Returns:

  • int — OK on success, FAIL on error.

pick_data_proj(sel, firsts, ns, picked)​

Extract raw data with SSP projection applied to each time sample before channel selection.

Parameters:

  • sel : mneChSelection Channel selection descriptor.

  • firsts : int First sample index (absolute).

  • ns : int Number of samples to extract.

  • picked : *float *** Array of per-channel data vectors (pre-allocated).

Returns:

  • int — OK on success, FAIL on error.

load_one_filt_buf(buf)​

Load a single filter-sized buffer by picking projected unfiltered data with taper margins for overlap-add filtering.

Parameters:

  • buf : *MNERawBufDef ** The filter buffer definition to load.

Returns:

  • int — OK on success, FAIL on error.

pick_data_filt(sel, firsts, ns, picked)​

Pick filtered data for a channel selection by loading filter buffers, applying the frequency-domain filter per channel (skipping stimulus channels), and accumulating overlapping results.

Parameters:

  • sel : mneChSelection Channel selection descriptor.

  • firsts : int First sample index (absolute).

  • ns : int Number of samples to extract.

  • picked : *float *** Array of per-channel data vectors (pre-allocated).

Returns:

  • int — OK on success, FAIL on error.

attachDerivations(derivations, keepPrevious)​

Attaches channel derivations (MNE-C mne_raw_attach_derivations): merges them into deriv, matches the merged set to the channels of the file into deriv_matched and validates it against the channel info.

Parameters:

  • derivations : const MNEDerivSet & The derivations to add.

  • keepPrevious : bool Keep the derivations attached before instead of replacing them.

Returns:

  • int — The number of derived channels that are valid for these data.

Static Methods​

open_file_comp(name, omit_skip, allow_maxshield, filter, comp_set)​

Open a raw FIFF file with explicit compensation grade.

Reads channel info, compensation data, SSS data, sets up data/filter buffers and bad channels, and returns a fully initialized raw data object.

Parameters:

  • name : const QString & Path to the raw FIFF file.

  • omit_skip : int If non-zero, discard initial skip samples.

  • allow_maxshield : int If non-zero, allow MaxShield data.

  • filter : const MNEFilterDef & Filter definition to apply.

  • comp_set : int Desired compensation grade (-1 for file native).

Returns:

  • *MNERawData ** — A new raw data object, or NULL on failure. Caller takes ownership.

open_file(name, omit_skip, allow_maxshield, filter)​

Open a raw FIFF file using the file's native compensation grade.

Convenience wrapper for open_file_comp() with comp_set = -1.

Parameters:

  • name : const QString & Path to the raw FIFF file.

  • omit_skip : int If non-zero, discard initial skip samples.

  • allow_maxshield : int If non-zero, allow MaxShield data.

  • filter : const MNEFilterDef & Filter definition to apply.

Returns:

  • *MNERawData ** — A new raw data object, or NULL on failure. Caller takes ownership.

Example​

Used in examples​

ex_mne_api​

MNE library: BEM surfaces, surface projection, ICP, source spaces and epochs.

src/examples/ex_mne_api/main.cpp, line 269
filter.taper_size = 2048;
filter.lowpass = filter.eog_lowpass = 40.0f;
filter.lowpass_width = filter.eog_lowpass_width = 5.0f;
std::unique_ptr<MNERawData> mneRaw(MNERawData::open_file(rawFile.fileName(), false, false, filter));
const MNERawInfo& rawInfo = *mneRaw->info;
const MNERawBufDef& firstBuffer = mneRaw->bufs.front();

Authors of this file​