**3. Results**

## *3.1. Population Statistics and Correlation Analysis*

Figure 2 shows a matrix of correlations and histograms of the cochlear parameters where the histograms can be seen to follow a normal distribution [30]. Strong correlations were found between the cochlear volume and all other parameters (B (ρ = 0.82, *p* < 0.05), height (ρ = 0.58, *p* < 0.05), cochlear duct length (ρ = 0.74, *p* < 0.05) and roller coaster height (ρ = 0.53, *p* < 0.05)), except for A (ρ = 0.41, *p* < 0.05) and the wrapping factor (ρ = −0.45, *p* < 0.05). Cochlear volume was negatively correlated with the wrapping factor. In addition to the strong correlation with the cochlear volume, cochlear B also showed a strong positive correlation with LW length (ρ = 0.74, *p* < 0.05) and strong negative correlation with the wrapping factor (ρ = −0.62, *p* < 0.05). Unsurprisingly, cochlear B was only weakly correlated to cochlear height (ρ = 0.39, *p* < 0.05) and roller coaster parameter (ρ = 0.43, *p* < 0.05), as these parameters are related to a dimension orthogonal to the plane where cochlear B was measured. In addition to the strong correlation between cochlear height and volume, cochlear height was also strongly correlated with the roller coaster parameter (ρ = 0.74, *p* < 0.05), which was measured in the same dimension. Cochlear A did not show any strong correlation with the other parameters. The correlation between cochlear B and A was also weak (ρ = 0.39, *p* < 0.05). Figure 3 shows the correlation plots between parameters A, B and the LW length at different cochlear angles (90◦, 180◦, 270◦, 360◦, 450◦, 540◦). In general, B shows a stronger correlation to the LW length than A at all angular insertion depths.

**Figure 2.** Population statistics and correlation plots between anatomical features of the cochlea. Histograms of the parameters with fitted Gaussian curves are shown in the diagonal of the matrix. Scatter plots show the correlation between the parameter indicated in the column titles and the parameter in the row titles. A strong correlation (ρ > |0.50|) between parameters is represented by scatter plots with filled circles and a weak correlation is shown by empty circles. Solid lines indicate the linear regression curves. Note that the scales of the y-axes do not apply to the histograms. ρ: Pearson correlation coefficient; μ: mean; σ: standard deviation; \* depicts a significant correlation (*p*-value < 0.05).

**Figure 3.** Correlation plots between cochlear duct lengths and cochlear size. A strong correlation (ρ > |0.50|) between parameters is represented by scatter plots with filled circles and a weak correlation is shown by empty circles. Solid lines indicate the linear regression curves. ρ: Pearson correlation coefficient, \* depicts significant correlation (*p*-value < 0.05).

Figure 4 shows the evolution of the ST height, area and radius parameters along the cochlea. All investigated ST parameters show a non-monotonic decrease between 0 and 570◦ followed by an approximately linear decrease up to 690◦. Between 0 and 570◦, all ST parameters display notches around 150◦, 360◦ and 510◦, and local peaks around 270◦ and 420◦. These could be due to the presence of the porous bone surrounding the common cochlear artery [39].

**Figure 4.** Scala tympani maximum vertical height (**A**), area (**B**) and radius of largest fitted circle (**C**) as a function of the angular distance. Dots represent individual measurement points. Error bars represent the mean and ±1 standard deviation; dotted lines show the 10th and 90th percentiles. Vertical dotted grid lines indicate the angular distance bands that were used to select the N measurement points indicated at the top of panel A to calculate the statistics.

#### *3.2. Inter-Sex Analysis*

Figure 5 shows the inter-sex differences between each cochlear parameter. For the male population, the following dimensional characteristics were observed: A (mean: 9.11 ± 0.58 mm, median: 9.13 mm, inter-quartile range (IQR): 0.81 mm), B (6.85 ± 0.25 mm, median: 6.83 mm, IQR: 0.37 mm), height (4.32 ± 0.15 mm, median: 4.31 mm, IQR: 0.22 mm), volume (64.93 ± 4.40 mm3, median: 64.70 mm3, IQR: 5.99 mm3), cochlear duct length (41.48 ± 1.06 mm, median: 41.56 mm, IQR: 1.61 mm) and the wrapping factor (81.20 ± 0.69◦, median: 81.190, IQR: 0.970). By comparison, the following dimensions were observed for the female population: A (8.97 ± 0.52 mm, median: 8.92 mm, IQR: 0.63 mm), B (6.73 ± 0.21 mm, median: 6.71 mm, IQR: 0.28 mm), height (4.25 ± 0.15 mm, median: 4.24 mm, IQR: 0.21 mm), volume (62.04 ± 3.91 mm3, median: 41.90 mm3, IQR: 4.70 mm3), cochlear duct length (41.07 ± 0.91 mm, median: 40.96 mm, IQR: 1.13 mm) and the wrapping factor (81.30 ± 0.71◦, median: 81.360, IQR: 0.850). An independent *<sup>t</sup>*-test revealed statistically significant differences for all parameters except the wrapping factor. Generally, female cochleae seem to be smaller and more tightly wound around the modiolus than male cochleae. However, all parameters showed a significant overlap between the two populations.
