Figures and data

Implanted electrodes in the auditory cortex and the primary (A1) and secondary (A2) auditory cortex.
(A) Blue dots denote the implanted electrode contacts from all epilepsy patients within 15 mm to the centroid of the auditory cortex (blue boundary) at the left and right hemispheres. Orange dots are electrode contacts within 5 mm to the centroid of the auditory cortex. (B) Boundary of A1 and A2 from the atlas. Ant.: anterior. Pos.: posterior. (C) Proportions of six cortical layers identified in Big Brain in 10 equi-distance boundaries between pial and gray/white matter in the auditory cortex, A1, and A2.

Correlations between music acoustical features and fMRI signals.
(A) distributions of the t-statistics of correlations between music acoustic envelope and fMRI signals across cortical depths (FDR corrected p<0.05) around the auditory cortices in both hemispheres. Figure 1 provides zoomed-in views of the relevant areas. (B) Regression coefficients between music acoustic envelope and fMRI signals at the auditory cortex (ctx), A1, and A2 across cortical depths. Comparisons are annotated by horizontal square brackets. All are insignificant. (C) The pair-wise comparison of music acoustic envelope-fMRI signal regression coefficients between cortical depths at the auditory cortex, A1, and A2. Values are t-statistics. Significant regression coefficient differences (FDR corrected p<0.05) at data grids at the upper-left triangular part of the data matrix are color-coded. The value at row r and column c represents the regression coefficient difference of (depth r - depth c).

Correlations between intracranial neural activity and fMRI signals across cortical depths.
(A): The correlation between fMRI signals at three cortical depths (normalized depth, n.d., of 0.1, 0.5, and 0.9) and neural oscillations measured by electrodes within 15 mm to the centroid of the auditory cortex. Values are regression coefficients. Significances of regression coefficients (beta’s; false discovery rate, FDR, corrected p<0.05 across frequencies) at cortical depths are shown in horizontal bars at the bottom of the panel. The light blue and red horizontal bars at the top of the panel indicate significant differences between n.d.=0.5 and n.d.=0.9 and between n.d.=0.5 and n.d.=0.1, respectively (FDR corrected p<0.05 across frequencies). (B): Regression coefficients between fMRI signals across cortical depths and neural oscillations. Significant coefficients (FDR corrected p<0.05 across frequencies and cortical depths) are color-coded. (C): Distributions of the t-statistics of correlations between intracranial neural activity and fMRI signals across cortical depths (FDR corrected p<0.05) in alpha, beta, low gamma, and broadband high frequency activity around the auditory cortices in both hemispheres. See Figure 1 for the enlarged areas. (D): Correlations between neural oscillations and fMRI signals across cortical depths at primary (A1) and secondary (A2) auditory cortices and their differences. Significant coefficients (FDR corrected p<0.05 across frequencies and cortical depths) are color-coded.

Correlations between HFA and fMRI signals.
(A): distributions of the t-statistics of correlations between HFA and fMRI signals across cortical depths (FDR corrected p<0.05) around the auditory cortices in both hemispheres. See Figure 1 for the enlarged areas. (B): Regression coefficients between HFA and fMRI signals at the auditory cortex (ctx), A1, and A2 across cortical depths. Comparisons between A1 and A2 are annotated by horizontal square brackets. Significant differences (t-test, FDR corrected p<0.05) are denoted by asterisks. Solid lines and shaded error bounds represent the cubic spline models of the acoustic envelope-fMRI correlations (derived from Figure 2) for comparison. Colored asterisks (blue, orange, yellow) denote depths where the HFA-fMRI correlation (bars) is significantly different from the acoustic-fMRI correlation (lines) for the respective regions. (C): The pair-wise comparison of HFA-fMRI signal regression coefficients between cortical depths at the auditory cortex, A1, and A2. Values are t-statistics. Significant regression coefficient differences (FDR corrected p<0.05) at data grids at the upper-left triangular part of the data matrix are color-coded. The value at row r and column c represents the regression coefficient difference of (depth r - depth c). (D): The centroid of the area with significant correlation between HFA and depth-specific fMRI signals.

Correlations between intracranial neural activity and fMRI signals across cortical depths.
(A): The correlation between fMRI signals at three cortical depths (normalized depth, n.d., of 0.1, 0.5, and 0.9) and neural oscillations measured by electrodes within 5 mm to the centroid of the auditory cortex. Values are regression coefficients. Significances of regression coefficients (beta’s; false discovery rate, FDR, corrected p<0.05 across frequencies) at cortical depths are shown in horizontal bars at the bottom of the panel. The light blue and red horizontal bars at the top of the panel indicate significant differences between n.d.=0.5 and n.d.=0.9 and between n.d.=0.5 and n.d.=0.1, respectively (FDR corrected p<0.05 across frequencies). (B): Regression coefficients between fMRI signals across cortical depths and neural oscillations. Significant coefficients (FDR corrected p<0.05 across frequencies and cortical depths) are color-coded. (C): Distributions of the significance of correlations between intracranial neural activity and fMRI signals across cortical depths (FDR corrected p<0.05) in alpha, beta, low gamma, and broadband high frequency activity around the auditory cortices in both hemispheres. See Figure 1 for the enlarged areas. (D): Correlations between neural oscillations and fMRI signals across cortical depths at primary (A1) and secondary (A2) auditory cortices and their differences. Significant coefficients (FDR corrected p<0.05 across frequencies and cortical depths) are color-coded.

Relationships between intracranial neuronal activity and fMRI signals across cortical depths at A1 (left column) and A2 (right column) between repeated listening.
Significant coefficients (FDR corrected p<0.05 across frequencies and cortical depths) are color-coded.

Average T statistics for the correlation between oscillatory SEEG signals and fMRI signals in the auditory cortex using data from different patient/control participant groups at 7T (yellow) and the same patients at 3T (purple).
Gray area denotes regions of p>0.05 after controlling the False Discovery Rate in multiple comparisons.


MNI coordinates of the implanted electrodes and the distance to the centroid of the auditory cortex in the left (A) and right (B) hemispheres.
Values in Contact column denote the electrode contact from the tip of the implanted electrode. Contacts on the same electrode were separated by 5 mm. The auditory cortex regions was defined from an atlas [33].