Data structure The data was recorded in the programme 'Signal' version 4 made by Cambridge Electronic Design: ced.co.uk. There are 16 subjects. For subjects 1-9 and 15, the files are in CED format (.cfs), so these would need to be opened in 'Signal'. For subjects 10-14 and 16 the data have been exported as .txt files so that they can be loaded into Matlab directly. (sorry - I didn't have time to export all the files as text data, but I might be able to do this if it is requested). The first row is time and the second row is voltage. Note that an amplified with gain 10,000 was used, so to convert to voltage at the scalp, divide all values by 10,000. See below for more information on the data. P is for participant, ck indicates click stimulation, Da indicated Da stimulation, 1 and 2 indicate repeats, N65 indicates a recording in background noise and NN indicates no stimulation level. e.g.for Participant 12 there are 7 files... P12ck1.txt - 1st click recording in quiet P12ck2.txt - repeat click in quiet P13ckN65.txt - click in 65 dB noise P12Da1.txt - 1st Da in quiet P12Da2.txt - 2nd Da in quiet P12CkN65.txt - Da in 65 dB noise P12NN.txt - no stimulation data An example matlab analysis file 'Time_and_Freq_analysis_ABR_speech.m' is included. This is not the best written or commented Matlab (apologies), but it shows how to load in data, how it was filtered, averaged, artefact rejected and then time and spectral bootstrapping is carried out. Materials and Methods Participants Normally hearing subjects, displaying pure tone hearing thresholds of less than 20 dB HL at 250, 500, 1000, 2000, 4000, and 8000 Hz, were recruited for the study. Participants? eligibility for inclusion was assessed via a series of audiometric screening procedures; Otoscopy, Pure Tone Audiometry and Tympanometry, each carried out according to the relevant British Society of Audiology (BSA) recommended procedures. Participants also completed a brief screening questionnaire designed to assess various aspects of their otological history including previous noise exposure (particularly within the last 48 hours), ear surgery, family history of hearing loss, etc. Based on the results of the screening procedure, a total of 16 normally-hearing volunteers (3 male and 13 female), ranging in age from 22 to 30 years, were considered appropriate for inclusion. As musical training has been suggested to enhance temporal and frequency encoding in the auditory brainstem (Musacchia et al, 2007), all included participants were also asked to provide details of any previous musical experience. 7/16 participants (44%) had received previous, formal musical training (ranging from UK Grade 2-8), while the remaining 56% disclosed no particular previous experience, giving a reasonably balanced representation of musical versus non-musical training amongst participants. Ethical approval was sought prior to initiation of the study via the University of Southampton?s Ethics and Research Governance Online (ERGO) system. Stimulus and Recording Parameters Stimulus/recording parameters aimed to replicate those applied by previous related research (Skoe & Kraus 2010). ABRs were elicited by two different acoustic stimuli presented at 70 dB nHL; a 40ms-duration synthesised /da/ syllable, and a 100?s-duration broadband click. Peak equivalent sound pressure levels (p.e.SPL), established using a Wavetek oscilloscope, were 105.6 dB p.e.SPL (for the voiced region of the /da/ stimulus), 81.9 dB p.e.SPL (for the /da/ onset burst), and 113.5 dB p.e.SPL (click). Stimuli were generated by a Cambridge Electronic Design (CED) ?1401 laboratory interface system, connected to a laptop via USB. The /da/ stimulus consisted of the first 40 ms of a five-formant synthetic speech syllable digitally synthesised using a Klatt cascade parallel formant synthesizer (Klatt, 1980) (at a sampling rate of 10 kHz). All stimuli were delivered monaurally (into the participant?s right ear) in alternating polarities, via 3M EAR-tone 5a insert phones at a rate of 11.1/second. A constant rate, rather than a constant off-time, was chosen to ensure the same rate-adaptation effects for the two stimuli (with an assumption that adaptation is due to the rate, although it could be debated that off-time is more important). A 90 ms analysis window was utilised, with stimulation starting at 10 ms to allow a baseline measurement. The analysis window was intended to be long enough to allow the response to the 40 ms /da/ stimulus to conclude/return to baseline before initiation of a subsequent stimulus presentation. Stimuli were presented in quiet, with a repetition carried out in order to examine the test-retest variability of specific response characteristics. A ?no-stimulus? condition was additionally included (involving electrophysiological recording in the absence of acoustic stimulus presentation) as a means of quantifying levels of residual, background activity and to check the false positive rates of the statistical methods used. The sequence of stimulus presentation was randomised for each individual using a Latin Square technique, aiming to reduce the influence of varying participant state. Responses were recorded differentially between scalp electrodes placed on the midline of the forehead at the hairline (active), and nape of the neck (reference), with a further electrode, placed on the midline of the lower forehead, serving as the common ground. All impedances, as measured between each electrode pair, were maintained below 5k? for all participants throughout recording. Data was sampled at 10 kHz, and filters were 100-3000 Hz. A total of 6000 sweeps (3000 of each polarity) were collected, as recommended by Skoe & Kraus (2010).