EEGManySteps-Data-Curation

Data curation repository for EEGManySteps

This repository contains helpful resources for curating data sets collected for EEGManySteps.

1. Metadata from individual contributions, including link to the data repository if available.
2. Guideline to curating your data for submitting to EEGManySteps
3. Scripts for converting individual submissions to a standardized format (BIDS + custom metadata).
4. Template and utility scripts for checking metadata integrity and conversion (WIP)

Overview of submitted data sets

Access requests for "available on request" or "controlled access" data sets are to be sent to eegmanysteps@gmail.com

• Submission track A (ds_A...) : auditory oddball with button press
• Submission track B (ds_B...) : auditory oddball counting task
• Submission track C (ds_C...) : walking task

Dataset ID	Contributor	Source EEG Format	Source Motion Format	Link to data set	Status	Notes
ds_A1_TB	Tjeerd Boonstra	.	.	available upon request	curation in progress	treadmill
ds_A2_UK	Daniel Büchel	.	.	n/a	data collection in progress	.
ds_A3_SJBP	Sein Jeung	.xdf	.xdf	n/a	data collection in progress	overground
ds_B1_MK	Melanie Klapprott	.	.	in preparation	curation in progress	.
ds_C1_LSIE	Grant Hanada/Daniel Ferris	.set	.mat	figshare	curation complete	treadmill + overground, curated set available upon request
ds_C2_KUMC	Sodiq Fakorede/Hannes Devos	.xdf	.xdf	in preparation	curation in progress	clinical
ds_C3_HIP	Anna Wunderlich/Klaus Gramann	xdf.	xdf.	controlled access	curation complete	aging/hearing impaired
ds_C4_TW	Nadine Jacobsen/Julius Welzel	.	.	in preparation	curation in progress	'.'

Metadata guideline for contributors

Contributors are asked to convert the data set to a standardized format, following the terminology prescribed by the Brain Imaging Data Structure (BIDS), especially EEG-BIDS and Motion-BIDS.

Additionally, we EEGManySteps define custom metadata fields that are required ("MUST") or recommended for interpretation of gait-EEG data.

Gait metadata

In addition to BIDS motion fields, these custom ManySteps fields MUST be present in the *_motion.json and *_eeg.json file.

Field	Allowed Values	Description
ManySteps_SpaceType	"Indoor", "Outdoor"	Specifies whether the walking environment was indoors or outdoors.
ManySteps_SurfaceType	"Treadmill", "Overground" or user-defined	Type of walking surface. Custom values may be used if necessary.
ManySteps_Footwear	"On", "Off", "n/a"	Indicates whether participants wore footwear during the task.
ManySteps_Gait_Description	Free text	Additional comments about the gait protocol (e.g., instructions for turning, walking speed, experimental constraints).
ManySteps_Secondary_Task	"Yes", "No"	Indicates whether participants performed an additional task while walking.

If "ManySteps_SurfaceType" is set to "Treadmill", the following fields MUST be specified as well.

Field	Allowed Values	Description
ManySteps_Treadmill_Speed_Scheme	"Fixed", "Individualized"	Specifies whether the treadmill speed was fixed for all participants.
ManySteps_Treadmill_Speed	Numerical value	Treadmill speed in m/s

If "ManySteps_SurfaceType" is set to "Overground", the following field MUST be specified as well.

Field	Allowed Values	Description
ManySteps_Overground_Distance	Numerical value or "n/a"	in meters in case the participant was walking back and forth on a straight line

Sensor placement

BIDS-Motion currently does not restrict keywords for body parts for sensor placement. These keywords are entered into column ‘placement’ of ‘*_channels.tsv’ file. The body parts MUST use the vocabulary as defined in this document. The coordinates are recommended to be provided.

placement	placement_coords
Head	50,50,100
LowerBack	50,50,100
LeftFoot	50,70,30
RightFoot	50,70,30
...

Pre-extracted gait events

If the system you recorded data with do not provide raw time series data, gait events MUST be shared in ‘*_events.tsv’ file accompanying EEG or motion data. You may add a custom-event for describing events that are not defined in the table below and provided description for it.

Keyword	Description
RIC	right foot intial contact
RFC	right foot final contact
LIC	left foot intial contact
LFC	left foot final contact

Time synchronisation between motion.tsv and EEG

The type of time synchronisation method should be indicated in the dataset_description.json file, added as a cutom field as follows :

"ManySteps_Timesynch" : "regular_sampling", "latency_channel", "hardware-trigger", "software-trigger"

• regular_sampling : fixed, reliable smapling rate. Latency of each sample can be derived from sample index and SamplingFrequency
• latency_channel : per-sample lateancy provided as an additional channle in motion data (type = "latency")
• hardware-trigger : Hardware-based TTL trigger synchronization between systems
• software-trigger : Software-based trigger synchronization between systems

Please use ‘acq’ column in the ‘*_scans.tsv’ file for aligning the onsets of different data streams if the recording starts at different times.
For instance, if the EEG recording started 1.2 sec earlier than the motion recording, this difference is expressed as the difference in the datetime value in "acq" column.
In case the sampling rate is irregular, please also provide information about sample-by-sample latency as data channel concatenated with the motion data. This channel should have type ‘latency’, expressed as seconds from the onset of the corresponding motion.tsv file.