Model parameters

This zip file contains a set of gifti files that include the model parameters reported in the manuscript.
The data in separate directories are in the original HCP 59k surface format and upsampled to the fsaverage 164k surface format. The data were resampled using the connectome workbench wb_command -metric-resample command with the optional -largest flag and BARYCENTRIC resampling.

The gifti data arrays are ordered as follows:
1. Mu
The mean of the gaussian function estimated by the CSS model. Defined over log frequency in Hz.

2. Sigma
The standard deviation of the gaussian function.

3. Exponent
The exponent parameter (n).

4. xval_R2
Cross validated R2 of the CSS model

5. FWHM
The full width half maximum of the exponent-adjusted gaussian functions - calculated as reported in the methods section.

6. Speech_xval_R2
Cross-validated R2 of the speech-selective model.

7. Speech_beta_diff
The difference between the speech and non-speech beta weights of the speech selective model.

8. PCA_1
The first PCA component as reported in the manuscript.

9. PCA_2
The second PCA component as reported in the manuscript.

10. PCA_3
The third PCA component as reported in the manuscript.
Note that, as reported in the paper, items 5 onwards are restricted to the tonotopic region of interest that we define (available as another item in the 'Masks and atlases' item of the repository).