Skip to main content
Dryad

Cortical encoding of melodic expectations in human temporal cortex

Cite this dataset

Di Liberto, Giovanni et al. (2021). Cortical encoding of melodic expectations in human temporal cortex [Dataset]. Dryad. https://doi.org/10.5061/dryad.g1jwstqmh

Abstract

Humans engagement in music rests on underlying elements such as the listeners’ cultural background and interest in music. These factors modulate how listeners anticipate musical events, a process inducing instantaneous neural responses as the music confronts these expectations. Measuring such neural correlates would represent a direct window into high-level brain processing. Here we recorded cortical signals as participants listened to Bach melodies. We assessed the relative contributions of acoustic versus melodic components of the music to the neural signal. Melodic features included information on pitch progressions and their tempo, which were extracted from a predictive model of musical structure based on Markov chains. We related the music to brain activity with temporal response functions demonstrating, for the first time, distinct cortical encoding of pitch and note-onset expectations during naturalistic music listening. This encoding was most pronounced at response latencies up to 350ms, and in both planum temporale and Heschl’s gyrus.

Methods

----------------------------------------------------------------------------------------------------
--------------------------------------- DiliBach EEG dataset ---------------------------------------

Monophonic Bach music EEG Experiment from:
Di Liberto et al., Cortical encoding of melodic expectations in human temporal cortex, eLife, 2020

The EEG experiment was designed by Giovanni M. Di Liberto and Claire Pelofi
and conducted at the Laboratoire des Systèmes Perceptifs, UMR 8248, CNRS, Ecole Normale Supérieure,
PSL University, France.
Data were collected by Giovanni M. Di Liberto, Claire Pelofi, and Gaelle Rouvier
(See below for the list of authors and affiliations)

Version 1.2: 24 June 2021 - Giovanni M. Di Liberto

--- CND data ---------------------------------------------------------------------------------------

EEG signal was recorded as participants listened to monophonic piano Bach music.
Subjects 1-10 were non-musicians and 11-20 were expert pianists.
One Matlab (.mat) file for each subject. Each data file contains two variables: 'fs' (the sampling
frequency) and the EEG data.

--- Stimuli ----------------------------------------------------------------------------------------
Original stimuli can be found on http://www.jsbach.net
and correspond to Bach violin pieces from sonatas and partitas, and partita for flute in A minor.

Audio files original filenames
audio1.mid - fp-1all.mid
audio2.mid - fp-2cou.mid
audio3.mid - fp-3sar.mid
audio4.mid - fp-4bou.mid
audio5.mid - vp2-1all.mid
audio6.mid - vs1-4prs.mid
audio7.mid - vp1-1al_v2.mid
audio8.mid - vp2-4gig_v2.mid
audio9.mid - vp3-2lou.mid
audio10.mid - vp3-3gav_v2.mid

Usage notes

EEG data were saved according to the CND data format (Continuous-events Neural Data format; www.cnspworkshop.net/cndFormat.html).

A single file is saved containing the stimulus data, which is the same across all subjects. This file contains a variable 'stim' including, for example, stimulus features such as the speech envelope.

A single file per subject contains the subject-specific EEG data. The experiment was grouped into 30 trials (each playing one of the ten available pieces). The 30 trials are collectively stored using a cell array. The trial order is the same as in the 'stim' variable (so the same order across participants), and it includes a variable indicating the original presentation order for each subject.

Key variables and key fields in the CND structures:

*** dataStim.mat: Information and data that is common across subjects (in the CND format)
 stim.data{1,2}: speech envelope vectors for the second piece
 stim.data{2,5}: note onset vector for the fifth piece
 stim.data{3,5}: note onset vectors modulated by the absolute pitch value for the fifth piece
 stim.fs: Envelope signal sampling frequency

*** dataSub1.mat, ..., dataSubN.mat: Subject-specific EEG data (in the CND format)
 eeg.fs: EEG sampling frequency (512 Hz)
 eeg.data: cell array with the EEG signal for each of the 30 experimental trials (corresponding
           to the 30 cells in the dataStim.mat file. Each cell has size: timeSamples x channels
 eeg.chanlocs: channel location information
 eeg.extChan{1}.data: EEG external mastoid electrodes
 eeg.paddingStartSample: EEG data were time-locked to the onset of the speech stimulus for each
                         trial. However, note that a padding was added at the start and end of
                         each trial to allow the users to, for example, perform filtering and
                         then remove side filtering artifacts if needed. Of course, this also
                         means that the first sample of the stimulus corresponds to the EEG
                         sample eeg.paddingStartSample+1 (=513).

Notes:
 - There were 64 scalp channels + 2 mastoid electrodes (electrodes 65 and 66 correspond to left and right mastoids respectively);
 - Stimulus vectors were shared at fs=64Hz, as in Di Liberto et al., eLife, 2020. In addition, we have shared the original MIDI files used for the experiment (see README file).

Funding

European Research Council, Award: Neume, Grant agreement ID: 787836