diff --git a/micromed_io/to_mne.py b/micromed_io/to_mne.py index c661081..e6a3771 100644 --- a/micromed_io/to_mne.py +++ b/micromed_io/to_mne.py @@ -1,57 +1,71 @@ """Micromed module to load and transform data from Micromed recordings to mne format""" -from datetime import timezone +from datetime import timedelta, timezone from pathlib import Path from typing import List, Union import mne import numpy as np + from micromed_io.trc import MicromedTRC def create_mne_from_micromed_recording( recording_file: Union[str, Path], sub_channels: List[str] = None, + start_time: float = 0.0, + stop_time: float = None, ch_types: Union[List[str], str] = "eeg", ) -> mne.io.RawArray: - """Create a mne Raw instance from a Micromed recording + """Create a mne Raw instance from a Micromed recording. Parameters ---------- recording_file : Union[str, Path] - The micromed recording file + The micromed recording file. sub_channels : List[str], optional The channels to pick from the recording. If None, all channels are picked. - ch_types: Union[List[str], str], optional - The list of channel types. Types must be in ``['grad', 'mag', 'ref_meg', 'eeg', + start_time : float, optional + The start time (in seconds) of the recording data to include in the mne Raw instance. + Defaults to 0.0. + stop_time : float, optional + The stop time (in seconds) of the recording data to include in the mne Raw instance. + If None, all data until the end of the recording is included. Defaults to None. + ch_types : Union[List[str], str], optional + The list of channel types. Types must be in ['grad', 'mag', 'ref_meg', 'eeg', 'seeg', 'dbs', 'ecog', 'eog', 'emg', 'ecg', 'resp', 'bio', 'misc', 'stim', 'exci', 'syst', 'ias', 'gof', 'dipole', 'chpi', 'fnirs_cw_amplitude', 'fnirs_fd_ac_amplitude', - 'fnirs_fd_phase', 'fnirs_od', 'hbo', 'hbr', 'csd']`` + 'fnirs_fd_phase', 'fnirs_od', 'hbo', 'hbr', 'csd']. Returns ------- mne.io.RawArray - A mne Raw instance containing the requested channels + A mne Raw instance containing the requested channels. Notes ----- Some info are hardcoded, such as: - - ``device_info`` type: Micromed - - ``device_info`` site: Unknown + - `device_info` type: Micromed + - `device_info` site: Unknown - Update the code if needed + Update the code if needed. Examples -------- >>> from micromed_io.to_mne import create_mne_from_micromed_recording - >>> mne_raw = create_mne_from_micromed_recording("path/to/file.TRC") + >>> mne_raw = create_mne_from_micromed_recording("path/to/file.TRC", start_time=10.0, stop_time=20.0) """ micromed_trc = MicromedTRC(recording_file) if sub_channels is None: sub_channels = micromed_trc.micromed_header.ch_names - sub_eegs = micromed_trc.get_data(picks=sub_channels) + sfreq = micromed_trc.get_sfreq() + start_sample = int(start_time * sfreq) + stop_sample = int(stop_time * sfreq) if stop_time is not None else None + sub_eegs = micromed_trc.get_data( + picks=sub_channels, start=start_sample, stop=stop_sample + ) info = mne.create_info( ch_names=sub_channels, @@ -72,20 +86,32 @@ def create_mne_from_micromed_recording( sub_eegs, info, ) - raw.set_meas_date( - micromed_trc.micromed_header.recording_date.replace(tzinfo=timezone.utc) - ) - + # set measurement date + meas_date = micromed_trc.micromed_header.recording_date.replace( + tzinfo=timezone.utc + ) + timedelta(seconds=start_time) + raw.set_meas_date(meas_date) + + # markers and notes should be added only if not in + # start_time-stop_time interval + if stop_sample is None: + stop_sample = np.inf # Add annotations from notes for note_sample, note_val in micromed_trc.get_notes().items(): - raw.annotations.append( - onset=note_sample / info["sfreq"], duration=0, description=note_val - ) + if stop_sample > note_sample >= start_sample: + raw.annotations.append( + onset=(note_sample - start_sample) / info["sfreq"], + duration=0, + description=note_val, + ) # Add annotations from markers for marker_sample, marker_val in micromed_trc.get_markers().items(): - raw.annotations.append( - onset=marker_sample / info["sfreq"], duration=0, description=marker_val - ) + if stop_sample > marker_sample >= start_sample: + raw.annotations.append( + onset=(marker_sample - start_sample) / info["sfreq"], + duration=0, + description=marker_val, + ) return raw