Overview
--------
This is the "Selective auditory attention in normal-hearing and hearing-impaired listeners" dataset. It was collected as part of a EU H2020-ICT grant (number 644732; COCOHA: Cognitive Control of a Hearing Aid). In brief, it contains EEG data for 44 older adults, of which 22 had a steeply sloping hearing impairment and 22 were normal-hearing, who were listening selectively to audiobooks in single-talker and two-talker conditions. In addition to this, the dataset contains scalp EEG data from subjects listening to tone beeps (ERPs) and to repetitive EFR amplitude modulated tone stimuli. Finally, the dataset contains behavioral and clinical data from an audiological assessment. The dataset contains:

- EEG data recorded from subjects selectively listening to one of two simultaneous speech streams or to a single speech stream in quiet
- Event related potentials (ERPs) during passive listening to 1 kHz pure tones
- Envelope-following responses (EFRs) during passively listening to repetitive tone sequences 
- Eyes-open and eyes-closed resting EEG data
- EarEEG data for 19 of the 44 subjects (EEG recorded inside the ear canals)
- Percentage of correctly answered multiple-choice questions from selective attention EEG experiment 
- Task difficulty ratings during selective attention EEG experiment
- Sentence reception thresholds in a speech-in-noise HINT test (DaHINT test)
- Pure-tone audiograms on both ears
- Scores in a digit span working memory test 
- SSQ questionnaire ratings in three domains (speech, spatial and qualities hearing scale questionnaire)
- Scores from a psychoacoustic tone-in-noise detection task (frequency-temporal test, FT test)

For details, see the 'Details about the experiment' section and [1]


Citing this dataset
--------
Please cite [1]. 



Format
------
The dataset is formatted according to BIDS version 1.3.0 and the BIDS standard extension for EEG (BEP006) that has been merged in the main body of the specification. For more details, see https://bids-specification.readthedocs.io/en/latest/06-extensions.html

Behavioural and clinical data are stored in the 'participants.tsv' file. Task-difficulty ratings and multiple choice questionnaire data from the selective auditory attention experiment are stored in the '_events' files that also describe timing of events recorded in during the selective attention experiment (e.g. 'sub-001_task-selectiveattention_events.tsv'). Columns in the 'selectiveattention_events' files are described in more details in 'task-selectiveattention_events.json'. 



Audio stimuli used in the auditory attention experiment
------
The 'derivatives' folder contains envelopes of the audio speech stimuli used in the selective attention experiment. For example, a value under the 'stim_file'  column in 'selectiveattention_events' with entry 'sub001/target/t001.wav' will be accompanied by an envelope feature stored in '/derivatives/stimuli/sub001/target/t001.mat'. The envelope features are described in more details in the README file in the 'derivatives' folder.

For inquiries regarding the raw audio data, please send an email to jensh@drcmr.dk with the subject line "ds-eeg-snhl audio". Raw audio files should be stored in the '/stimuli' folder under the root folder of the dataset. The values under the 'stim_file' column in ' selectiveattention_events' files that accompany EEG data from the selective attention experiment point to the raw audio stimuli files. For example 'sub001/target/t001.wav' will be translated to '/stimuli/sub001/target/t001.wav'.

The stimuli presented to the hearing-impaired listeners were linearly amplified according to the 'Cambridge formula' (CamEQ) [2] in order to account for loss in audibility. The values under the 'stim_file_without_cambridge_eq' column in the 'selectiveattention_events' files point to the audio files without CamEQ equalization. For example 'sub001/target/t001woa.wav' will be translated to '/stimuli/sub001/target/t001woa.wav'. 


Details about the experiment
--------
For a detailed description of the EEG tasks, see [1]. The supplied files 'task-selectiveattention_events.json', 'task-tonestimuli_events.json' and 'task-rest_events.json' contains more information about the task-related events.  


Selective attention experiment:
--------
The main EEG experiment was designed to measure cortical responses to competing speech streams during a selective attention task. EEG data were recorded from subjects selectively listening to one of two simultaneous speech streams or to a single speech stream in quiet. The speech stimuli consisted of two different audiobooks read by a male and a female speaker. The audio files were split into 50 s long trials and the speech streams were spatially separated at +/- 90 degrees. The experiment consisted of 48 trials. Each subject listened to two blocks of 12 trials with the male speaker as the target, and two blocks of 12 trials with the female speaker as the target. Each block of 12 trials consisted of 4 single-talker trials, and 8 two-talker trials. At the onset of each trial, the subject was instructed to attend to either the male or the female talker. As an additional cue, the target speech stream was switched on ~4 s (jittered between 3 and 5 s) before the interfering speech stream. The number of left versus right target trials was balanced across the experiment. After each trial, the subjects were asked to rate how easy or difficult it was to understand the attended speech in that trial on a continuous rating scale marked 'easy' and 'difficult' at the extremes.  



Important information about EEG data
--------
Biosemi reported CMS/DRL errors during experimental breaks in the selective attention experiment for 5 subjects ('sub015', 'sub033', 'sub037', 'sub038', 'sub039') and in the ERP/EFR data for one subject ('sub003'). Data recorded during these segments is not available. This should be noted e.g. if detrending is performed on the raw EEG data. The timing of the CMS/DRL errors is described in the '_events' files.

Due to technicalities, there were additional triggers in the ERP/EFR data for 4 subjects ('sub005', 'sub008', 'sub009' and 'sub011'). These triggers should not be included in the analysis. The additional triggers have been labeled with 'error_triggers' in the 'trigger_type' column in the 'tonestimuli_events'  files. This is also described in 'task-tonestimuli_events.json'.   


References
--------
[1] Fuglsang, S.A., Märcher-Rørsted, J., Dau, T., Hjortkjær, J. (2020). Effects of sensorineural hearing loss on cortical synchronization to competing speech during selective attention." Journal of Neuroscience 

[2] Moore BCJ, Glasberg BR (1998) Use of a loudness model for hearing-aid fitting. I. Linear hearing aids. Br J Audiol 32:317:335.



Acknowledgments
--------
This work was supported by the EU H2020-ICT grant number 644732 (COCOHA: Cognitive Control of a Hearing Aid) and by the Novo Nordisk Foundation synergy grant NNF17OC0027872 (UHeal). The EarEEG systems were kindly provided by Eriksholm Research Centre.

