Tastes and retronasal odours evoke a shared flavour-specific neural code in the human insula

This dataset constitutes files from an fMRI experiment with 2 sessions per subject (apart from one subject who dropped out after the first session). In this experiment, healthy participants attend a familiarisation session with congruent sweet and savoury flavours followed by two neuroimaging sessions where they separately receive the constituent tastants and odorants (SweT – Sweet Taste; SavT – Savoury Taste; SweO – ‘Sweet’ Odour; SavO – ‘Savoury’ Odour). Each session had up to 6 runs, resulting in a maximum of 12 runs per subject. The dataset consists of a zip file of first-level analyses. Univariate analyses modelled taste blocks, odour blocks, artificial saliva (ArtS) blocks and rinse blocks. Scripts provided use first-level smoothed maps to create contrasts of taste against ArtS and odour against ArtS in directed one-sample t-tests. Multivariate analyses use unsmoothed maps to train a decoder or perform representational similarity analysis. To run the scripts and reproduce our results, download all three zip files and extract them to the same directory. Run each scripts from the appropriate directory following its assigned directory hierarchy. Scripts require MATLAB with the following tollboxes: SPM12 CosMoMVPA Statistics and Machine Learning Toolbox File Description: data.zip: zipped file containing: ratings: collated behavioural ratings prior to each session. NB: sub-17 only attended one session and behavioural data for the first session of sub-26 were lost due to a clerical error. smoothed: first-level analyses generated by SPM for smoothed data (6 mm isometric FWHM). Regressors were: Taste (both SweT and SavT) with a 16 s boxcar function Odour (both SweO and SavO) with a 16 s boxcar function ArtS with a 16 s boxcar function Rinse with a 4 s boxcar function unsmoothed: first-level analyses generated by SPM for unsmoothed data. Regressors were: SweT with a 16 s boxcar function SavT with a 16 s boxcar function SweO with a 16 s boxcar function SavO with a 16 s boxcar function ArtS with a 16 s boxcar function Rinse with a 4 s boxcar function ROIs.zip: zipped file containing the ROIs used for the study: granular_insula.nii: ROI consisting of areas 163/164 and 171/172 of the Brainnetome atlas dysgranular_agranular_insula.nii: ROI consisting of the rest of the insula PiriformMask_2mm.nii: parcellated bilateral piriform cortex MNI_GMmask_2mmisotropic.nii: grey-matter mask derived from the SPM tissue probability map (thresholded at >.2) LOO: a directory of the leave-one-subject-out derived ROIs all containing bilateral insula, derived from univariate taste-ArtS contrast at a threshold of p < .01 (unc.) searchlight: a directory containing peaks from the searchlight crossmodal classifier Brainnetome parcellations: a directory containing each parcellation of the insula in the Brainnetome atlas scripts.zip: zipped file containing: univariate: SPM batch script for univariate analyses mvpa: decoding: decoding of tastants and odourants in the respective ROIs as well as control analyses RSA: new RSA technique to study representational drift of flavours For more information on the experimental setup, scanning parameters and analyses, please refer to the publication. Access to the dataset requires an access request, as per GDPR. Questions should be sent to: [e-mail] or [e-mail]