Comprehensive Measurements and Analyses of Ear Canal Geometry From Late Infancy Through Late Adulthood: Age-Related Variations and Implications for Basic Science and Audiological Measurements

Abstract

This study provides a comprehensive analysis of ear canal geometry from 0.7 to 91 years, based on high-resolution computed tomography scans of 221 ears. Quantified features include cross-sectional areas along the canal’s length, total canal length, curvature, and key anatomical landmarks such as the first and second bends and the cartilage-to-bone transition. Significant developmental changes occur during the first 10 years of life, with adult-like characteristics emerging between ages 10 and 15 years, likely coinciding with puberty. Substantial interindividual variability is observed across all ages, particularly in the canal area. The canal becomes fully cartilaginous at and lateral to the second bend by 0.7 years, with further growth occurring only in the bony segment thereafter. These anatomical findings have important implications for audiologic threshold assessments, wideband acoustic immitance measures, age-appropriate hearing aid fitting schedules, and surgical planning, particularly in pediatric populations where anatomical variation is greatest.

Keywords

ear canal anatomy ear canal development wideband acoustic immitance (WAI)hearing-aid fitting

Introduction

A comprehensive geometrical description of the human ear canal, encompassing its variations across all ages, does not exist. Yet, such knowledge is critical for advancing both basic auditory science and clinical applications. Key anatomical features of the ear canal—its length, cross-sectional area along its length, variations in area (expansions and contractions), and curvature (notably at the first and second bends)—influence the sound-pressure transfer function between the tympanic membrane at the medial end and the sound-pressure input at the lateral end (e.g., Stinson & Lawton, 1989). These anatomical characteristics are depicted in Figure 1. Understanding these features through precise physical measurements can improve descriptions of sound propagation through the ear canal, with significant implications for hearing diagnostics, hearing aid fittings, and surgical interventions.

Figure 1.

Schematic diagram representation of the adult ear canal shape, derived from the airspace extracted from a 3D CT scan (left ear, subject 15,154, female, 43.2 years) using the open-source software platform 3D slicer. The diagram highlights the key anatomical features of the annulus, the second bend, the first bend, and the entrance. White lines with double arrows indicate the distances corresponding to measurements reported in this study. Orthogonal vectors drawn with black lines define the directions of superior, anterior, and right, where right indicates toward the right side of the head and is a result of how the canal is oriented in 3D space. Note that the airspace is visualized as a 2D projection of the 3D canal structure, and curves out of the chosen plane are less or not discernible in this representation. 3D CT = three-dimensional computed tomography; 2D = two dimensional.

Given the lack of measurements of ear canal geometry, we used high-resolution computed tomography (CT) scans to collect and analyze data across a broad age range (0.7–91 years). This study quantifies both age-related changes and individual variability, providing insights into the structural maturation of the canal. Additionally, it examines the locations of the bone-to-cartilage transitions and analyzes their distribution across different age groups. The measurements are reported for predefined age groups and are also available at the individual ear level in the Supplemental Material.

Background

Quantifying ear canal geometry—including both maturational changes and intersubject variability—is important for understanding its influence on sound transmission and for advancing clinical applications such as hearing diagnostics, hearing-aid fittings, and surgical interventions.

Balouch et al. (2023) and Voss et al. (2020) reviewed the limited available data on physically measured ear canal areas in adults, citing studies by Egolf et al. (1993), Johansen (1975), Stinson and Lawton (1989), and Voss et al. (2008), along with estimates from historical textbooks. For many years, the most comprehensive data came from Stinson and Lawton (1989), who measured canal areas and lengths in 14 adult cadaver ears. Consistent with Balouch et al. (2023), they observed considerable intersubject variability beyond 5 mm from the tympanic membrane: “Some canals show a steady increase in area with position, some are approximately constant in cross-sectional area over much of their length, and others show a marked constriction halfway along their length, at the transition between the osseous and cartilaginous portions of the canal. Cross-sectional areas, in the middle portion of the canals, can range between $25 {mm}^{2}$ and $70 {mm}^{2}$ .”

Abdala and Keefe (2012) summarize the current knowledge of ear canal maturation from birth through puberty. Canal-based acoustical measurements suggest that the canal grows in length until at least age 7 (Dempster & Mackenzie, 1990) and that newborns and older infants have shorter and narrower canals than adults (e.g., Keefe & Abdala, 2007; Keefe et al., 1993, 1994). However, direct physical measurements of length, area, and the age at which growth slows are lacking.

Acoustically, sound absorption in an unoccluded canal within a diffuse sound field depends on canal area, with larger areas absorbing more sound (Rosowski et al., 1988). As Abdala and Keefe (2012) note, smaller canal areas result in less efficient sound absorption, which may partly explain higher auditory thresholds in younger ears. Canal geometry may also contribute to the lower otoacoustic emission (OAE) levels observed in infants. Future models and experimental studies examining maturational differences in thresholds and OAEs will benefit from physical measurements of ear canal dimensions, leading to a deeper understanding of how canal geometry influences these two fundamental hearing assessments of thresholds and OAEs.

Another widely used tool in both clinical and research settings is WAI (wideband acoustic immittance). Although WAI has shown potential for identifying middle-ear pathologies (e.g., Merchant et al., 2021), a major challenge is the substantial variability observed in measurements across normal ears (e.g., Voss et al., 2013). A better understanding of how ear canal geometry varies across individuals may help account for some of this variability and improve the specificity of WAI in assessing middle-ear function.

WAI measurements are typically computed using reflectance-based transmission line theory, which depends on the canal area at the measurement probe (e.g., Allen, 1985; Keefe et al., 1993; Voss & Allen, 1994). This reliance on canal area—and the assumption that the canal is cylindrical—is built into all currently FDA-approved WAI devices. Despite substantial anatomical variation both within individual canals and across people, most published WAI data, and all FDA-approved devices, assume a uniform canal area of aproximately $44 {mm}^{2}$ for individuals older than 6 months and $17 {mm}^{2}$ for those younger than 6 months.

Several acoustical methods have been developed to estimate canal area at or medial to the probe location, and these have been applied to WAI measures and other pressure-based canal measurements (e.g., Huang et al., 2000; Keefe et al., 1993, 2024; Lewis & Neely, 2015; Merchant et al., 2015; Rasetshwane & Neely, 2011). While these methods offer a valuable starting point for accounting for probe-tip area and acoustic transformations along the canal, they rely on geometric assumptions (see Keefe et al., 2015) and require a broader frequency range than is available in most clinical and commercial systems. Although acoustical estimates of canal area are promising, they have not yet been directly validated against physical measurements.

Beyond applications in hearing science and audiologic diagnostics, a more comprehensive understanding of ear canal geometry has the potential to advance two additional clinical areas: (a) hearing aid fittings and (b) transcanal endoscopic ear surgery (TEES).

For hearing aid users, particularly young children, ear canal molds must be regularly replaced to accommodate growth. Improved knowledge of canal maturation could inform more precise, age-appropriate replacement schedules, which are currently general and not based on known growth patterns. For example, the most specific guidance comes from the Ontario Protocol for the Provision of Amplification (Ontario Ministry of Children, 2019) which recommends follow-up visits every three months for the first year, every 6 months during the second year, and annually thereafter. However, it notes that this schedule “may vary from infant to infant.” The American Academy of Audiology (2013) offers even less specificity, stating that earmold replacement “may be as frequent as monthly” and advising providers to “be proactive regarding earmold replacement due to the child’s growth.”

Similarly, advances in middle-ear surgical techniques, particularly TEES, are influenced by canal geometry. TEES uses a rigid endoscope and surgical instruments inserted directly through the ear canal, avoiding postauricular incisions or mastoidectomy (Ito et al., 2020). Detailed knowledge of canal dimensions is critical for surgical planning, especially in young children and adolescents, since the smallest portions of the canal may limit instrument access and maneuverability. This is increasingly relevant with earlier detection of middle-ear disease and emerging gene therapies delivered via the transcanal route.

Ultimately, improved characterization of age-related and individual variation in canal geometry will support advancements in hearing science, audiologic diagnostics, hearing aid design and surgical techniques.

Methods

Study Overview: Ears and CT Scans

This study analyzed high-resolution, de-identified CT scans from 221 human ears, spanning an age range from 0.67 to 90.13 years. The scans were utilized to obtain detailed geometric and anatomical measurements along the ear canal. The CT scans had resolutions of 0.4 mm, 0.6 mm, or 0.75 mm, with many originating from cochlear implant candidates. Only scans that displayed normal external and middle-ear anatomy were included in the analysis.

The CT scans were provided by the University of Massachusetts Chan Medical School, while the research was conducted at Smith College. Both institutions granted exemptions from Institutional Review Board (IRB) oversight due to the study’s retrospective design and the use of de-identified data.

Study Design

The primary aim of this research was to investigate how canal geometry changes with age. Age groups (cohorts) were predefined before data collection and analyses, and Table 1 provides an overview of the distribution of ears by age group, sex assigned at birth (male and female), and laterality (left and right). The study aimed to include $\sim$ 20 ears per group, with equal representation of males and females. However, some younger age groups included fewer ears due to limited availability of suitable CT scans. Additionally, the balance between left and right ears within an age group was not always symmetric across males and females. These discrepancies were due to factors such as (a) the order in which CT scans were received and (b) availability of a specific ear in a given scan (e.g., partial cuts, excessive wax, and middle-ear fluid).

Table 1.

Distribution of Ears by Age Group, Sex Assigned at Birth (Male and Female), and Laterality (Left and Right).

Age group (years)	Total number of ears	Number of males (left, right)	Number of females (left, right)
0.7 $-$ 2	9	3, 2	4, 0
2 $-$ 5	13	5, 3	1, 4
5 $-$ 10	20	3, 7	7, 3
10 $-$ 15	20	3, 7	7, 3
15 $-$ 20	19	4, 5	4, 6
20 $-$ 30	21	5, 4	4, 8
30 $-$ 40	19	7, 3	4, 5
40 $-$ 50	20	7, 3	4, 6
50 $-$ 60	20	6, 4	5, 5
60 $-$ 70	20	5, 5	5, 5
70 $-$ 80	20	5, 5	4, 6
80 $-$ 91	20	4, 6	6, 4

Note. Each age group is inclusive at the lower age and cutsoff just prior to the upper age; for example, the age group 2–5 years includes 13 ears (8 male, 5 female) that are ages 2.00 through 4.99 years. The youngest and oldest age groups were initially defined as 0–2 and 80–90 years, respectively, prior to obtaining all CT scans; the final groupings of 0.7–2 and 80–91 years reflect the ages represented in the available scans. When available, race and ethnicity are listed in the data spreadsheet within the Supplemental Material. CT = computed tomography.

Anatomical Definitions and General Measurement Methodology

The majority of the definitions and methods used to characterize canal geometry in this study follow those described by Balouch et al. (2023), with some modifications and additions detailed below. Measurements were performed using OsiriX MD (Pixmeo, Version 13.0.2), a software package equipped with multiplanar reconstruction (MPR) functionality. This enabled systematic measurements along each canal. Key measurements and definitions include:

Canal termination: The plane that intersects the bony tympanic annulus.

Canal central axis: The curved spline in 3D space at the center of the canal from its termination to its entrance.

Canal cross-sectional area: The area within the plane perpendicular to the canal wall at each location along the canal’s central axis.

Canal entrance: The most lateral position where the plane normal to the canal wall is fully enclosed by the canal.

First bend location of the canal: The lateral-most position along the canal’s central axis with the largest canal curvature near the first bend.

Bone-to-cartilage transition: Measured at four canal wall locations (superior, anterior, inferior, and posterior) along the canal’s central axis as the location where the canal wall becomes entirely cartilaginous, based on the MPR sagittal section [not measured by Balouch et al. (2023)].

Second bend location of the canal: Position along the canal’s central axis where the canal becomes fully cartilaginous. This location closely corresponds to the region of the greatest curvature between the first bend and the tympanic annulus, which was not measured by Balouch et al. (2023) and is detailed in Figure 2.

Figure 2.

Scatter plot demonstrating the correlation between the location where the canal becomes fully cartilaginous (x-axis) and the location with the maximum change in angle (e.g., curvature) near the second bend (y-axis). Data points are jittered slightly from their measured values in increments of 2 mm so that overlapping points can be discerned.

Balouch et al. (2023) demonstrated that (a) left and right ears from the same CT scan are significantly more similar than ears from different individuals and (b) area and distance measurements from the same scan showed high correlation across different investigators employing the same methods.

For this study, initial area and distance measurements were performed by one of five investigators. These measurements began with identifying the tympanic annulus and proceeded systematically from the medial to the lateral end of the canal. All measurements were subsequently reviewed by author (SEV), who cross-checked and finalized every measurement for consistency. In general, differences in area measurements across investigators rarely exceeded 5%. Additionally, all canal axis coordinates were finalized by author (SEV) to ensure measurement consistency.

Measurements were taken at 2 mm intervals, starting from the tympanic annulus. At 24 mm from the tympanic annulus or within a few millimeters of the first bend, whichever occurred first, the interval was reduced to 1 mm for greater precision. In 13 of the 221 ears, wax along the canal obstructed measurements at one to four locations. For these cases, the affected areas were recorded as NaN (not a number), and the corresponding points on the canal’s central axis were estimated.

Methodological Improvements and Additions Within OsiriX

This section outlines methodological enhancements and additions to the work of Balouch et al. (2023) using OsiriX, including for the identification of the tympanic annulus, an improvement in the method for finding the central axis, an objective definition for the first bend based on the curvature of the central axis, and a definition for the second bend.

Identification of the Tympanic Annulus

The plane in 3D space that is defined by the tympanic annulus, the bone that marks the boundary between the ear canal and the middle ear, serves as the definition for the medial termination of the ear canal. This definition does not account for the conical shape of the tympanic membrane but enables systematic measurements across all ears. The tympanic annulus was identified systematically using the MPR views and tools in OsiriX. The methodology involved aligning the sagittal plane to intersect the bony tympanic annulus in 3D space through an iterative process. Key steps include as follows:

Initial localization: The investigator navigated to the general area of the CT scan containing the tympanic membrane, which typically marks the boundary between the ear canal and middle-ear air space. While the tympanic membrane was not always visible in the scan, the boundary between the canal and middle-ear air space could be identified in the coronal and transverse sections by locating the bony annulus that separates the canal from the middle-ear air space.

Alignment of the sagittal plane: The sagittal plane was iteratively aligned to intersect the entire circumference of the tympanic annulus in 3D space. First, in the transverse MPR window, the sagittal plane was rotated so that its line intersected the bony annulus at two ends. Next, in the sagittal MPR window, the transverse plane was rotated slightly, and then the process in the transverse MPR window was repeated in that the sagittal plane was readjusted to maintain its intersection with the bony annulus. This process was systematically repeated back and forth between the MPR sagittal and transverse windows until the transverse plane completed a full rotation within the sagittal MPR window, ensuring that the sagittal plane intersected the annulus at all points along its perimeter. We note that the process would remain identical if the coronal MPR window were used in place of the transverse MPR window.

Placement of the first central axis point: Once the sagittal plane fully intersected the bony annulus, the center of the plane was located, as described by Balouch et al. (2023). The first spline point defining the canal’s central axis was then placed at this center.

Improvement in Measuring the Canal’s Central Axis

Balouch et al. (2023) defined the sagittal plane in the MPR view to be normal to the canal walls along the transverse and coronal sections. This study introduced an iterative refinement process to improve alignment within the sagittal plane. Specifically, the transverse and coronal planes were rotated within the sagittal plane to bisect the sagittal section as symmetrically as possible. For perfectly circular sagittal sections, this step is unnecessary, as the transverse and coronal planes would naturally define the diameter. However, when the sagittal section is oval or tear-shaped, aligning the transverse and coronal axes with the major axes of the shape ensures that the center of the sagittal section is correctly identified. Without this refinement, the center point may deviate from the true center, affecting the calculated central axis of the canal. This methodological addition had minimal impact on the measured cross-sectional areas.

Area Measurements

The area within the sagittal section was measured at each location using the region of interest fill tool in OsiriX. To ensure consistency across measurements performed by different investigators, a predefined windowing setting of $-$ 200 to 200 HU (Hounsfield unit) was applied for all area measurements. This setting provided optimal contrast, clearly distinguishing the air within the canal from the canal surface, which could consist of either bone or cartilage depending on the location.

Bone-to-Cartilage Transition Locations

The canal surface transitions from bone near the tympanic annulus to cartilage lateral to the second bend, typically over several millimeters along the canal’s central axis. This transition location was measured, with 2 mm resolution, at the four canal-wall locations of superior, anterior, inferior, and posterior, referenced to the orientation of the head in the scanner. At each location along the canal’s central axis, from medial to lateral positions, the air space border in the MPR sagittal section was examined using the OsiriX “bone” windowing setting ( $-$ 450 to 1,050 HU), which provided clear contrast between bone and cartilage surfaces. The transition to cartilage was defined to occur when the thickness of the bone at a given location was <1 mm and surrounded by cartilage.

Objective Definition for the First-Bend Location

The first bend of the ear canal is defined objectively as the point along the central axis, within the first bend region, that exhibits the largest angular change in the 3D space. In order to calculate the angular change at each point, $ϕ$ , the canal’s central axis was extracted from the “curved path” file saved by OsiriX. To calculate $ϕ$ at each point along the canal’s central axis, consider three consecutive points $A$ , $B$ , and $C$ along the axis in the 3D space of the CT scan. Define two unit vectors: $\vec{A B}$ is the vector from point A to point B normalized by its length, and $\vec{B C}$ is the vector from point B to point C normalized by its length. The angle $ϕ$ at point B is the arc-cosine of the dot product between $\vec{A B}$ and $\vec{B C}$ . The first-bend location is defined as the most lateral point in the first bend region where $ϕ$ deviates maximally from 180 degrees (indicative of a straight line).

Objective Definition for the Second-Bend Location

The measurement protocol was not initially designed to identify the second-bend location. However, after data collection, it became clear that the second bend is a significant anatomical landmark for describing canal development. The curvature at the second bend can span several millimeters, and measurements near this region were typically made at 2 mm resolution. Observations during data collection and analysis indicated that the second bend corresponds to the point where the canal becomes fully cartilaginous. Two methods were considered to define the second-bend location:

Start of cartilaginous-only canal: The location where the canal transitions to being fully cartilaginous, defined as the point with no surrounding bone along the sagittal section’s entire perimeter. This measure can overestimate the location by up to 2 mm due to the resolution of the measurements.

Maximum angle change of canal’s central axis: The location of the greatest angle change near the second bend, calculated using the method described above for the first-bend location. This method has limitations as the second bend spans multiple points, varies in geometry across ears, and depends on angle calculations using three points measured at 2 mm resolution.

Despite the limitations of these two measures, Figure 2 suggests strong agreement between them, with a Spearman correlation coefficient of 0.87 (95% confidence interval ranges from 0.85 to 0.91). A paired t-test of the differences between the measures yields a 95% confidence interval of 0.21–0.57 mm, indicating that the measure of a fully cartilaginous canal results in a slightly larger distance for the second bend, yet smaller than the measurement resolution. Given this high correlation, the location at the start of the cartilaginous-only canal is chosen to define the second-bend location, as it relies solely on 2 mm resolution measurements.

The measurements in Figure 2 show that the second bend marks the location where the canal becomes fully cartilaginous. While this definition for the second-bend location is also a result, it is presented in the Methods section because the result it is used to define the location of the second bend of the canal.

Statistical Methods

Multiple linear regression models including main effects were used to assess the associations of both age group and sex on area and distance measurements along the canal (R Version 4.4.1, 14 June 2024).

For area and distance measures where the model’s variance explained by age group was substantially greater than the variance explained by sex, a one-way ANOVA (analysis of variance) (Matlab Version R2024b, 24.2.0.2712019) was used to make comparisons between all age-group combinations. For each area or distance measure, the Matlab function “anova1” performed a one-way ANOVA to compare the population means for each of the 12 age groups. Next, the Matlab function “multcompare,” which uses the output from the ANOVA analysis, performed pairwise comparisons between pairs of specific age-group categories; we chose the option for the Tukey-Kramer test for honestly significant differences, which accounts for multiple pairwise comparisons with unequal numbers of samples in each group. Beyond the Tukey-Kramer test for pairwise comparisons, no further adjustment for multiple comparisons was undertaken. An alpha level of 0.01 was used to determine statistical significance.

Results

Canal Geometry: Big Picture View

Measurements of area along the canal axis, the canal’s central axis in the 3D space, and key landmarks along the canal are described in this section. Individual measurements of all of these quantities from each of the 221 ears are available in the Supplemental Data (https://osf.io/ecpxf/?view_only=1e975bd3461b40c1bf92672692b2aa03).

Effect of Age on Canal Areas and Lengths

Figure 3 presents the cross-sectional areas measured along the canal length in ears from individuals aged 0.7 to 91 years, with age groups specified in the legend. The left panel shows area versus distance from the tympanic annulus, illustrating age-related changes near the tympanic membrane. Given that canal length increases within the first 10 years of life, comparisons across age groups near the canal entrance are more accurately represented when referenced to the canal entrance rather than its termination, as shown in the right panel. Collectively, these panels reveal a systematic increase in canal area and length with age.

Figure 3.

Cross-sectional ear canal areas from ears ranging in age from 0.7 to 91 years. Thin lines are individual measurements and thick lines are means, which are computed at locations that include four or more ears. The measurement resolution was 2 mm for more medial parts of the canal and 1 mm near and lateral to the first-bend location; in order to reference the area measurements to the entrance, measurements with a 2 mm resolution were interpolated to a 1 mm resolution. Left: areas along the canal’s central axis, referenced to the tympanic annulus which is at 0 mm. Because some canals are longer than others, there are fewer points at the lateral most end of the canal in each age group. Right: areas along the canal’s central axis, referenced to the entrance of the canal, which is at 0 mm. Because some canals are longer than others, there are fewer points at the medial most end of the canal in each age group.

Effect of Sex Assigned at Birth on Canal Areas

Figures 4 and 5 compare the male and female canal areas across age groups. Figure 4 demonstrates that both sexes exhibit similar age-related trends in canal area with comparable variability, although male areas tend to be slightly larger at some locations. High variability, particularly in younger groups with fewer samples, limits definitive conclusions on sex-based differences. Figure 5 provides a direct comparison of canal areas, referenced from the canal entrance, by age group for both sexes. As detailed in the next section, age-based differences in canal areas and distances are generally more pronounced than sex-based differences.

Figure 4.

Mean and standard errors of ear canal areas from ears ranging in age from 0.7 to 91 years, separated by sex (left: female ears and right: male ears). Means are computed at locations that include four or more ears. The measurement resolution was 2 mm for more medial parts of the canal and 1 mm near and lateral to the first-bend location; in order to reference the area measurements to the entrance, measurements with a 2 mm resolution were interpolated to a 1 mm resolution.

Figure 5.

Mean and standard errors of male and female ear canal areas for each age group. Corresponding p-values show results of t-tests between male and female groups at each canal location for each age group. Tests are computed at locations that include four or more ears in both groups (male and female). The measurement resolution was 2 mm for more medial parts of the canal and 1 mm near and lateral to the first-bend location; in order to reference the area measurements to the entrance, measurements with a 2 mm resolution were interpolated to a 1 mm resolution.

Effect of Right Versus Left Ear on Canal Areas

There is no evidence for a difference in the canal areas or distances of left and right ears at any canal location for any age group (results not shown).

Curvature of Canal and Key Landmarks in 3D Space

Figure 6 illustrates the curves representing the canal’s central axis for each ear, defined in the 3D space. For each age group, 2D projections of the 3D central-axis curve in both the sagittal (inferior-to-superior and upper plots) and transverse (anterior-to-posterior and lower plots) planes are displayed. Overall, these curves exhibit consistent features across age groups and ears. In the sagittal plane, as the canal extends from its most medial point (0 mm) toward its lateral end ( $\pm 40 mm$ ), it first moves superiorly, then inferiorly, and reverses back to a superior direction at the first bend. The transverse plane reveals a posterior-to-anterior shift, from the tympanic annulus to the canal entrance. Figure 6 also indicates three key landmarks that systematically progress from the annulus to the entrance: the location where the all-bony canal ends, the location where the canal becomes cartilaginous only, and the first-bend location. The curves in Figure 6 also confirm that the start of the cartilaginous-only canal closely aligns with the second-bend location, as previously shown in Figure 2.

Figure 6.

The canal’s central axis is defined within the 3D space of each CT scan with the center of the tympanic annulus defined as point (0, 0, 0). For each age group, the upper plot shows the curve’s projection on the sagittal (inferior-to-superior) plane, and the lower plot shows its projection on the transverse (anterior-to-posterior) plane. Each curve indicates three key landmarks: the point where the all-bony canal ends as it extends from the tympanic annulus (dark purple, right-pointing triangle), the point where the canal becomes cartilaginous only to the entrance (dark gold left-pointing triangle), and the location of the first bend (lime green asterisk). Note that these planes are defined by the subject’s orientation in the CT scanning equipment; if the head position deviates, as it may especially in younger subjects, the curves will show corresponding rotations. 3D = three-dimensional; CT = computed tomography.

Shape of the Canal’s Cross-Section in the MPR Sagittal Section

Figure 7 presents the MPR sagittal section at the first bend for each of the 221 CT scans, illustrating the cross-sectional area and shape normal to the canal’s central axis at the first bend for each ear. The images are arranged in chronological age order and grouped by the predefined age categories. Visual inspection suggests that the cross-sectional areas are smallest in the youngest age group and increase progressively through adolescence. Across all age groups, there is substantial variability in the cross-sectional area. Furthermore, the majority of canals exhibit an elliptical rather than circular shape, with most, but not all, showing symmetry between the longer and shorter axes. Overall, the data highlight significant diversity in ear canal shapes and sizes.

Figure 7.

The multiplanar reconstruction (MPR) sagittal plane at the first bend for each of the 221 ears, organized by age group in ascending age order. The scale is the same for all images, with a measurement bar of 10 mm shown in the upper left in green.

Area and Distance Measures Along the Canal

This section presents a statistical analysis of 10 canal measures, focusing first on five area measures and then on five distance measures. The area measures include (a) area within the tympanic annulus, (b) area at the second bend, (c) area at 12 mm medial to the canal entrance, (d) area at the first bend, and (e) area at the canal entrance. The length measures include (a) distance from the annulus to the second bend, (b) distance from the annulus to the first bend, (c) distance from the annulus to the entrance (i.e., canal length), (d) distance from the second bend to the first bend, and (e) distance from the first bend to the entrance.

To assess the influences of sex and age group, multiple linear regression models were applied to each of the area and distance measures. These models revealed no significant interaction between sex and age. While some measures exhibited statistically discernible sex differences, these differences were generally minor compared to the effect of age (Table 2). Partial $η^{2}$ (Richardson, 2011) was calculated to describe the variances explained by age group alone and sex alone. With one exception (annulus area), the variance explained by age group was substantially greater than that explained by sex. We also note that the male minus female predicted outcomes (after controlling for age) indicate male ears are generally larger in the tested area and distance measures.

Table 2.

Results From Multiple Linear Regression Models.

Measure	Age group (df = 11)		Sex (df = 1)
	p value	% variance	p value	% variance	Estimated M-F
Area: annulus	$.34$	$6.4 %$	$< .001$	$9.2 %$	$3.8 {mm}^{2}$
Area: second bend	$< .001$	$36.9 %$	$< .001$	$6.4 %$	$6.9 {mm}^{2}$
Area: 12 mm from entrance	$< .001$	$43.6 %$	$.004$	$4.1 %$	$5.0 {mm}^{2}$
Area: first bend	$< .001$	$49.9 %$	$.13$	$1.1 %$	$3.1 {mm}^{2}$
Area: entrance	$< .001$	$39.6 %$	$.38$	$0.4 %$	$2.5 {mm}^{2}$
Distance: annulus to second bend	$< .001$	$54.1 %$	$< .001$	$7.9 %$	$1.1 mm$
Distance: annulus to first bend	$< .001$	$65.2 %$	$< .001$	$15.4 %$	$1.7 mm$
Distance: annulus to entrance	$< .001$	$66.7 %$	$< .001$	$15.8 %$	$1.8 mm$
Distance: second bend to first bend	$.007$	$11.8 %$	$.03$	$2.1 %$	$0.6 mm$
Distance: first bend to entrance	$.008$	$11.6 %$	$.57$	$0.2 %$	$0.1 mm$

Note. No significant interaction between sex and age group existed for any of the models; results here are the output of models with main effects only. The % variance columns are the percent variance explained which is the partial $η^{2}$ statistic. The estimated M-F column indicates the predicted difference in the outcome for males minus females after controlling for age.

To further focus on the larger age-related differences, a one-way ANOVA, as outlined in the methods, was used to evaluate the area and distance measures across age groups. Male and female ears were combined within each age group for these analyses, justified by the balanced representation of males and females with no evidence of interaction between sex and age.

Figure 8 summarizes the canal areas at the five locations using boxplots across 12 age groups (upper row). The locations, from left to right, progress laterally from (a) the tympanic annulus, (b) the second bend, (c) 12 mm medial to the entrance, (d) the first bend, to (e) the entrance. The lower row visually summarizes the one-way ANOVA with multiple comparisons between age group pairs. Blue-shaded boxes represent no statistically discernible difference in mean areas ( $p \geq .01$ ), while red-shaded boxes indicate statistically discernible differences ( $p < .01$ ). Results are symmetric along the diagonal, which is shaded gray.

Figure 8.

Summary of canal areas for each age group at specific locations along the canal (upper plots) and pair-wise comparisons between each age group (lower plots) (left to right columns: area within the tympanic annulus, area at the second bend, area 12 mm medial to the Canal's entrance, area at the first bend, area at the canal entrance). Upper: box plots show median (red line), interquartile ranges (box), whiskers that extend up to 1.5 times the interquartile range, and data points outside the whisker range that are outliers (+). The overall F statistic and corresponding p-value from the one-way analysis of variance (ANOVA) is indicated in the upper left of each plot. Lower: visual representation of the output from the multiple comparison tests across age groups for each area measure. The images are symmetrical around the gray diagonal. Age group pairs filled with red indicate a statistically discernible difference between the corresponding age groups at the corresponding canal location, whereas pairs filled with blue are not discernible.

The area within the tympanic annulus (Figure 8, left column) is not associated with age from 0.7 to 91 years, suggesting that the annulus reaches full development by 0.7 years. Neither ANOVA nor multiple linear regression indicate any discernible age group effects; however, there is a discernible effect of sex. This is the only area measurement that shows no age dependency. Across all 221 ears, the mean area $\pm$ standard error is $63.1 \pm 0.4 {mm}^{2}$ (female $61.2 \pm 0.6 {mm}^{2}$ ; male $64.9 \pm 0.6 {mm}^{2}$ ).

The areas at the other four canal locations show statistically discernible age-group dependence. The lower plots reveal that this age dependence generally follows a stratified pattern: there are minimal discernible differences in area comparisons within age groups spanning 0.7–10 years or 15–91 years. However, most statistically discernible differences occur between ears younger than 10 years and those older than 10–15 years. Additionally, the 80–91 age group exhibits discernible differences in area at the first bend compared to ears up to 50 years old, and at the canal entrance compared to ears up to 20 years old. These findings align with the trends shown in Figure 3, where areas systematically increase with age, yet the large variability within each age group makes these trends not statistically discernible between most adult age groups with 20 ears per group.

Figure 9 presents five measured distances along the canal using boxplots across 12 age groups (upper row), along with pairwise statistical comparisons by age group (lower row). The first three measures—distances from the tympanic annulus to the second bend, from the annulus to the first bend, and from the annulus to the entrance (i.e., canal length)—increase within the age range of 0.7–10 years, reaching full maturity with no age-related differences above 10 years. Notably, there are no discernible differences across any age groups from 0.7 to 91 years for the distance between the second bend and first bend or between the first bend and the canal entrance. This pattern suggests that canal growth primarily occurs along the segment from the annulus to the second bend, which constitutes the portion of the canal that includes some bone around its perimeter (further detailed in the next section). Conversely, the canal length from the second bend to the entrance, composed entirely of cartilage, appears to remain largely unchanged after 0.7 years.

Figure 9.

Summary of canal distances for each age group at specific locations along the canal (upper plots) and pair-wise comparisons between each age group (lower plots) (left to right columns: distance from the annulus to the second bend, distance from the annulus to the first bend, canal length, distance between the second and first bends, and distance between the first bend and canal entrance). Upper: box plots show median (red line), interquartile ranges (box), whiskers that extend up to 1.5 times the interquartile range, and data points outside the whisker range that are outliers (+). The overall F statistic and corresponding p-value from the one-way analysis of variance (ANOVA) is indicated in the upper left of each plot. Lower: visual representation of the output from the multiple comparison tests across age groups for each area measure. The images are symmetrical around the gray diagonal. Age group pairs filled with red indicate a statistically discernible difference between the corresponding age groups at the corresponding canal location, whereas pairs filled with blue are not discernible.

Location of Bone-to-Cartilage Transition

Figure 10 (upper panel) illustrates the variation in distance from the tympanic annulus to the bone-to-cartilage transition along the canal’s central axis across age groups for four canal-wall locations (superior, anterior, inferior, and posterior). For all four locations, the bone-to-cartilage transition is closest to the tympanic annulus in the youngest age group and progressively shifts laterally toward the canal entrance as age increases, up to 10 years. This pattern suggests that the bony portion of the canal extends laterally with age, reaching stability by 10 years.

Figure 10.

Summary of bone-to-cartilage transition for each age group at a given location on the canal wall (left: superior; left-center: anterior; right-center: inferior; and right: posterior). Upper: box plots show median (red line), interquartile ranges (box), whiskers that extend up to 1.5 times the interquartile range, and data points outside the whisker range that are outliers (+). The overall F statistic and corresponding p-value from the one-way analysis of variance (ANOVA) is indicated in the upper left of each plot. Lower: visual representation of the output from the multiple comparison tests across age groups for each wall location. The images are symmetrical around the gray diagonal. Age group pairs filled with red indicate a discernible difference between the corresponding age groups at the corresponding canal location, whereas pairs filled with blue are not discernible.

The lower panel of Figure 10 tests for statistically discernible differences in bone-to-cartilage transition distances between age groups. Discernible differences are observed in the anterior, inferior, and posterior canal walls for ages 0.7–10 years, with fewer discernible age-based differences at the superior position. The bone-to-cartilage transition generally begins along the superior canal wall, with the superior-wall distance from the annulus reaching stability by age 2 years. In contrast, the bone-to-cartilage transition distances stabilize for the anterior, inferior, and posterior canal walls by age 10 years.

Figure 11 (upper) summarizes the distances from the tympanic annulus to the bone-to-cartilage transition at four canal wall locations (superior, anterior, inferior, and posterior) across four age groups: 0.7–2 years (n = 9), 2–5 years (n = 13), 5–10 years (n = 20), and 10–91 years (n = 179). These groups were chosen based on Figure 10, which showed variations below 10 years but no significant differences above this age. For all four age groups, there was a statistically discernible difference ( $p < .0001$ ) between locations; models which accounted for the repeated measurements for each ear, not reported here, yielded the same conclusions as the overall tests reported in Figure 11 (upper). The bone-to-cartilage transition spans several millimeters along the canal, with greater variability across canal location in younger groups. In individuals over 10 years, the transition generally moves from medial to lateral canal locations, systematically going from the superior canal wall to the anterior wall followed by the inferior wall and ultimately the transition completes along the posterior canal wall. This sequence varies among individuals, and the transition length ranges from 2 mm to several millimeters. The lower panel of Figure 11 highlights statistically discernible differences in transitions, with greater consistency in the older age group possibly due to its larger sample size. The exact transition locations for each of the 221 ears are detailed in the Supplemental Material.

Figure 11.

Summary of bone-to-cartilage transition for each location along the canal wall within the indicated age group (left: 0.7–2 years; left-center: 2–5 years; right-center: 5–10 years; and right: 10–91 years). Upper: box plots show median (red line), interquartile ranges (box), whiskers that extend up to 1.5 times the interquartile range, and data points outside the whisker range that are outliers (+). The overall F statistic and corresponding p-value from the one-way analysis of variance (ANOVA) is indicated in the upper left of each plot. Lower: visual representation of the output from the multiple comparison tests for each age group. The images are symmetrical around the gray diagonal. Canal location pairs filled with red indicate a statistically discernible difference between the corresponding canal locations for the corresponding age group, whereas pairs filled with blue are not discernible.

Table 3 addresses the male-female differences for the location of the bone-to-cartilage transition for the 179 ears aged 10–91. The male distances are systematically larger than the female distances, which is consistent with the results reported in Table 2 where the distance from the annulus to the second bend is 1.1 mm larger for males than for females ( $p < .001$ ). Within both the female-only and male-only pair-wise age comparisons, there are not discernible differences between the inferior and posterior locations but there are differences between all other canal-wall location comparisons (not shown).

Table 3.

Mean $\pm$ Standard Error of the Distance From the Annulus of the Bone-to-Cartilage Transition for Each Canal-Wall Location for Ears 10 Years and Older for Female and Male Ears.

Canal-wall location	Female (n = 91)	Male (n = 88)	p value	Confidence interval
Superior	$10.7 \pm 0.2$	$11.1 \pm 0.2$	.19	[ $-$ 0.4, 1.3]
Anterior	$11.7 \pm 0.2$	$13.2 \pm 0.2$	$< .001$	[0.7, 2.4]
Inferior	$14.0 \pm 0.2$	$15.3 \pm 0.3$	$< .001$	[0.4, 2.3]
Posterior	$14.9 \pm 0.2$	$16.2 \pm 0.2$	$< .001$	[0.6, 2.1]

Note. A two-sample t-test reports p-values and confidence intervals to summarize differences between these mature male versus female ears. Confidence intervals indicate how much larger distances are for male ears relative to female ears.

Discussion

Comparison to Published Canal Geometry Results

Figure 12 (left) compares the measurements from this study with those from Stinson and Lawton (1989), demonstrating similar ranges of canal areas along the canal’s length. The entire range from Stinson and Lawton’s 14 ears generally falls within the 5%–95% range of the present study. The 25%–75% range of the present study provides a relatively narrow measurement range compared to the entire range of the minimum to the maximum. At the same time, this plot shows that the 10% of ears outside of the 5%–95% range add substantially to the entire range. It is important to note that Stinson and Lawton (1989) lacked an objective measure of canal length, as they estimated its entrance based on the point where the cross-sectional area was judged to increase most rapidly. Additionally, they standardized all canals to a length of 30 mm. In contrast, our work employs a consistent, objective definition of the canal entrance and length; the summary statistics in Figure 12 are truncated somewhat arbitrarily at 32 mm, chosen as the largest canal length that includes more than half the ears in the dataset.

Figure 12.

Left: comparison of the area measurements along the canal made by Stinson and Lawton (1989) (N = 14, ages 41 -87 years) to those made here on adult canals (N = 140, ages 20 -91 years). Stinson and Lawton (1989) report their total range with the length of each canal adjusted to be 30 mm. For the current work, the quantities listed in the legend were calculated and plotted at each canal location up to 32 mm, which is the largest distance that included 50% of all adult ears. Right: comparison of the area measurements made here for the combined adult age group of 20 -91 years old (N = 140) and the combined children age group of ears <10 years old (N = 42). The dashed lines indicate the area assumed by interacoustics titan wideband acoustic immittance (WAI) instrument of 44.2 mm² for all ages above 0.5 years (blue) and 17.3 mm² for ages <0.5 years (purple).

Balouch et al. (2023) developed the methodology used in this comprehensive study of ear canals across most of the human lifespan. Their work demonstrated that measurements from the left and right ears of the same CT scan are generally consistent, allowing this study to focus on one ear per scan. Balouch et al. (2023) presented results from the CT scans of 47 adults, a subset of the scans analyzed in the present study. To improve figure clarity, the results from Balouch et al. (2023) are not included in Figure 12; their upper range aligns closely with the 95th percentile of this study, while their lower range is nearly identical to the minimum values reported here.

Balouch et al. (2023) divided their smaller dataset into three age groups (18–30, 30–60, and 60–91 years). In order to directly compare the results of the present study to their work, we reanalyzed our larger dataset using these same three adult age groups. With the three more expansive age groups, with larger number of ears in each, there remained no significant differences between adult age groups for canal distances (Figure 9), bone-to-cartilage transitions (Figure 10), or areas at the annulus, second bend, or canal entrance (Figure 8). However, the reanalysis revealed that the area at the first bend in the 18–30 and 30–60 year groups was not significantly different, but both were significantly smaller than in the 60–91 year group. This contrasts slightly with Balouch et al. (2023), who reported that the 18–30 year group had significantly smaller areas than both older groups. Collectively, these results suggest there may be an age-related change in the first-bend area across the entire life span. At the same time, this change may not be clinically significant compared to the substantial range in areas across a given age group. It is possible that such an area change could result from changes in the cartilage structure with age (Ito et al., 2001).

Comparing the current study to Voss et al. (2020) presents challenges, as their methodology involved using silicone ear canal molds to measure the area of the most lateral 13.2 mm of the canal. A key difference lies in the definition of the canal entrance because Voss et al. (2020) employed a different entrance reference than the one used in the present CT scan-based study, making direct alignment of measurements impossible. Nevertheless, the findings of Voss et al. (2020) served as motivation for this present study, as they reported significant age-related increases in canal areas at the entrance, the first bend, and at 12 mm insertion locations. The current CT scan measurements corroborate the trend of increased areas at the first bend and 12 mm insertion locations but do not support age-related area changes at the entrance.

It is likely that the entrance defined by Voss et al. (2020), located at the intertragal notch, corresponds to a point lateral to the entrance identified in the CT scans. In the present study, the canal entrance typically appeared in the superior-anterior quadrant of the ear for most CT scans, with some occurring in the inferior-anterior quadrant. The intertragal notch, positioned inferiorly, may align with canals opening in the inferior-anterior quadrant but is inconsistent with the majority of observed canals. This discrepancy suggests that the silicone mold measurements by Voss et al. (2020) may have inadvertently included part of the concha. While conchal growth with age has not been directly quantified, increases in auricular dimensions with age, as reported by Ito et al. (2001), provide indirect support for this possibility. Additionally, previous studies (e.g., Ito et al., 2001) suggest that the cartilage structure of the canal may deteriorate with age, potentially resulting in greater distortion with a silicone mold. In older ears, this cartilage breakdown could lead to increased expansion of the mold if the cartilage is more compressible, producing larger area measurements compared to younger ears. In summary, the discrepancy between the findings of Voss et al. (2020) regarding increased entrance areas in older ears and the current CT scan measurements may stem from methodological differences and anatomical changes occurring lateral to the canal entrance.

Developmental Changes: Canal Areas, Canal Lengths, and the Bone-to-Cartilage Transition

The lack of published physical measurements of ear canal geometry in infants and children was highlighted in the Introduction and Background sections. Figures 8 and 9 show significant age-related changes in the youngest age groups (0.7–2, 2–5, and 5–10 years) compared to those over 10 years for seven of the 10 analyzed areas and lengths. The remaining three measurements—area within the tympanic annulus, and the distances from the entrance to the first and second bends—appear adult-like by 0.7 years. The finding that tympanic-annulus area stabilizes by 0.7 years aligns with previous research suggesting the annulus reaches near-adult size at birth but undergoes postnatal fusion, ossification, and orientation shifts (Donaldson et al., 1992).

Figures 8 and 9 also suggest that canal diameter and length continue to grow until at least age 10, with length primarily increasing between the annulus and the second bend. At birth, the canal is fully cartilaginous, with no bone lateral to the annulus. By 6 months, tympanic bone forms around the canal at and lateral to the annulus (Donaldson et al., 1992). Our results indicate that after 6 months, the cartilaginous portion stabilizes in length, suggesting that all further canal growth occurs in the bony section. We hypothesize that tympanic bone growth pushes the cartilaginous portion laterally, forming the isthmus or second bend. This bony segment continues to grow until the canal reaches its mature length, while the cartilaginous section remains unchanged after 6 months. Canal area also increases along its length during this growth phase.

Figure 12 (right) compares canal areas between adult ears and ears under 10 years, grouped together. The data show clear differences in areas and lengths between the two groups. The shaded regions representing the interquartile range (25%–75%) indicate that younger ears have shorter canals and smaller areas at most locations.

Further studies with finer age resolution are needed to identify when canal dimensions reach maturity. Due to the current study design—with $\sim$ 20 subjects per age group (5–10 and 10–15 years)—the exact age range of maturity cannot be determined. However, the results suggest that canal maturity likely occurs between 10 and 15 years, coinciding with the typical age range for puberty.

Canal Area Measurements and Maturational Effects on Threshold and OAEs Measurements

Future modeling and experimental studies examining maturational differences in hearing thresholds and OAEs will benefit from the canal area and length measurements presented in this study. For hearing sensitivity assessments conducted with earphones (i.e., not in the free field), the volume of the ear canal between the sound probe and the tympanic membrane influences both the sound pressure generated by the earphone and the OAEs that travel in reverse through the middle ear into the canal.

In most audiologic threshold testing, canal pressures are not measured directly. Instead, calibrations are typically performed using 2 cc or 6 cc couplers, which do not account for anatomical and maturational differences throughout childhood. To better understand how these assumptions may affect threshold measurements, canal volumes can be estimated by age group using data from the present study. These volumes can then be used to estimate how variations in canal impedance may influence the calibration of a given instrument when applied across age groups.

These same canal volume estimates may also help explain differences in OAEs between children and adults. While stimulus levels for evoked OAEs are measured in the ear canal by a microphone (unlike threshold measurements), the emissions themselves depend on reverse sound transmission through the middle ear. In children, the middle ear is terminated by a substantially smaller residual canal volume than in adults, resulting in a different terminating load that can affect OAE measurements. This load can be characterized using the sound source and the estimated residual canal volume (e.g., Voss & Shera, 2004).

Although modeling the effects of canal volume on either threshold measurements or the reverse transfer function of the middle ear is beyond the scope of the present study, the geometric measurements provided here will be essential for such future analyses. More broadly, the canal volumes derived from this study have strong potential to inform how canal pressures vary across development, ultimately supporting more accurate and age-appropriate audiological testing in infants and children.

Effects of Ear Canal Area and Shape on WAI Absorbance Calculations

WAI aims to isolate middle-ear function by minimizing the influence of ear canal acoustics. This is typically achieved by modeling the canal as a cylinder with constant cross-sectional area and applying transmission-line theory to calculate absorbance. One motivation for the present study was to determine the range of canal areas across individuals, with the aim of evaluating the current clinical practice—used in FDA-approved systems—of applying a single assumed canal area for all individuals older than 6 months.

Figure 12 (right) illustrates the two fixed canal areas assumed by the Titan system (marked with dashed lines) alongside data from the current study. This comparison highlights that the assumed values do not adequately reflect the range of actual canal areas, particularly across different ages and probe insertion depths. For example, canal areas in ears from individuals aged 6 months to 10 years are typically smaller than the assumed $44.2 {mm}^{2}$ , while canal areas in adults aged 20–91 years tend to be larger.

Figure 13 demonstrates the impact of canal area assumptions on absorbance calculations. In the upper panels, we used published WAI data from three children (Merchant et al., 2021) and recalculated absorbance using the minimum, median, and maximum canal areas, as determined from the present study, for each subject’s age group. When the area was close to the assumed $44.2 {mm}^{2}$ , absorbance closely matched the values reported by the Titan system. However, when smaller areas were used, absorbance increased notably at low frequencies. In contrast, larger canal areas led to reduced absorbance, at frequencies below 2,000 Hz.

Figure 13.

Absorbance from six individual left ears, each recalculated using multiple assumed canal areas at the probe tip. In all panels, the dark blue dashed line shows the titan instrument's absorbance (interacoustics), based on a fixed canal area of 44.2 mm². Upper: three ears from Merchant et al. (2021), obtained from the online wideband acoustic immittance (WAI) database (Voss et al., 2024), as the youngest, median-aged, and oldest left-ear subjects. As probe-tip canal areas were unknown, absorbance was recalculated using the minimum, median, and maximum areas measured at the first bend for each age group in the present study. Lower: three ears from Voss et al. (2017), with canal areas obtained from silicone mold measurements. Recalculated absorbance used areas measured at a 12 mm insertion depth, defined relative to the canal entrance as described by Voss et al. (2020), and estimated to correspond to the WAI probe-tip location.

In the lower panels of Figure 13, we present a similar analysis using WAI data from three adult ears (Voss et al., 2017), for which physical canal areas were estimated from silicone ear molds. The selected examples span canal areas approximately equal to, 1.5 times larger than, and twice as large as the assumed value of $44.2 {mm}^{2}$ . Again, we observe that larger canal areas systematically reduce absorbance at low frequencies, with less consistent effects at higher frequencies.

These findings underscore that absorbance calculations are highly sensitive to assumed canal area, especially at frequencies below 2,000 Hz where the canal impedance is dominated by compliance and spatial pressure variation is minimal. In this range, the assumed area largely determines the resulting absorbance value, which may limit its utility for assessing middle-ear function. At higher frequencies, where spatial pressure variations become more complex, absorbance is less dominated by canal area and more likely to reflect true middle-ear behavior.

These observations also highlight the limitations of using idealized canal models—specifically, the assumption of a uniform cylinder—for absorbance calculations. The present study demonstrates that real ear canals are neither cylindrical nor horn-shaped, but instead show complex variations in area and curvature along their length. This raises questions about how much intersubject variability is introduced by the cylindrical assumption, and whether this variability significantly impacts absorbance calculations.

Several fundamental questions arise from these findings, including: (a) What are the limitations of absorbance calculations in the context of a more complicated canal shape? (b) If absorbance is calculated using transmission line theory, what is the most appropriate canal area to assume—at the probe tip, a medial average, or another reference point? (c) Do canal curvature and variable cross-sectional areas need to be incorporated into WAI absorbance calculations?

These questions also motivate the exploration of alternative measurement approaches that either avoid assumptions about canal area (e.g., Farmer-Fedor & Rabbit, 2002; Nørgaard et al., 2020; Stinson, 1990) or address challenges related to probe positioning, such as when the probe sits at a sharp bend or is inserted obliquely within the canal (e.g., Nørgaard et al., 2019). However, many of these methods still rely on the assumption of gradually varying canal geometry, which may not hold true for all ears.

Conclusions and Future Work

This study provides comprehensive anatomical measurements of ear canal geometry across ages 0.7–91 years. The dataset includes cross-sectional area profiles along the canal, total canal length, center spline curves representing curvature, and measurements of the area and location of the first and second bends. The transition from cartilage to bone occurs at the second bend, with the length of the cartilaginous portion remaining consistent across all ages beyond 0.7 years, while the bony portion continues to grow with age.

Significant developmental changes in canal geometry occur during the first 10 years of life, with adult-like features typically emerging between ages 10 and 15, likely coinciding with the onset of puberty. This ongoing development of the ear canal during childhood anatomical maturation highlights the importance of accounting for age-related differences between children and adults in audiological assessments, hearing aid fittings, surgical interventions, and basic auditory research involving the external and middle ear.

Among both children and adults, there is substantial intersubject variability in canal geometry, including differences in cross-sectional area, length, and curvature. The marked nonuniformity of the canal—particularly in area variations along its length—suggests that the canal likely has a greater effect on the WAI measure of absorbance than previously appreciated. These findings support the need for continued work to isolate the impedance of the canal from that of the middle ear for diagnostic purposes.

Future studies should extend anatomical measurements to include ear canal development from birth, as this study begins measurements at 0.7 years, and capture canal development with finer age resolution throughout childhood. Such findings will be important for informing hearing-aid mold replacement and for understanding the feasibility of gene therapy strategies for sensorineural hearing loss that rely on transcanal (TEES) delivery to the inner ear during early infancy. Additional pediatric data will also enhance our understanding of anatomical variation during development and enable more age-appropriate approaches to auditory care.

Supplemental Material

sj-pdf-1-tia-10.1177_23312165251345572 - Supplemental material for Comprehensive Measurements and Analyses of Ear Canal Geometry From Late Infancy Through Late Adulthood: Age-Related Variations and Implications for Basic Science and Audiological Measurements

Supplemental material, sj-pdf-1-tia-10.1177_23312165251345572 for Comprehensive Measurements and Analyses of Ear Canal Geometry From Late Infancy Through Late Adulthood: Age-Related Variations and Implications for Basic Science and Audiological Measurements by Susan E Voss, Aaron K Remenschneider, Rebecca M Farrar, Soomin Myoung and Nicholas J Horton in Trends in Hearing

Supplemental Material

sj-csv-2-tia-10.1177_23312165251345572 - Supplemental material for Comprehensive Measurements and Analyses of Ear Canal Geometry From Late Infancy Through Late Adulthood: Age-Related Variations and Implications for Basic Science and Audiological Measurements

Supplemental material, sj-csv-2-tia-10.1177_23312165251345572 for Comprehensive Measurements and Analyses of Ear Canal Geometry From Late Infancy Through Late Adulthood: Age-Related Variations and Implications for Basic Science and Audiological Measurements by Susan E Voss, Aaron K Remenschneider, Rebecca M Farrar, Soomin Myoung and Nicholas J Horton in Trends in Hearing

Footnotes

Acknowledgements

We thank Smith College students Mealaktey Sok (N = 13) and Karen Bekhazi (N = 17) for helping make initial measurements on 30 of the 221 ears presented here. We thank Erin Voss for writing a script that reads the .curvedpath output file from OsiriX and extracts the corresponding 3D coordinates in order to plot the spline curves. We thank Jasper Voss for helpful conversations related to the method to calculate the angles in 3D space between adjacent points on the spline curve. We also thank two anonymous reviewers for helpful suggestions.

ORCID iDs

Susan E Voss

Aaron K Remenschneider

Soomin Myoung

Nicholas J Horton

Ethical Considerations

Exemptions from Institutional Review Board (IRB) oversight were obtained from the University of Massachusetts Chan Medical School, where the CT scans originated, and Smith College, where the research was conducted. The exemptions were granted due to the retrospective nature of the study and the use of de-identified data. We do not have permission to share the CT scans themselves.

Author Contributions

Susan E. Voss led the project at all stages, coordinated the work of coauthors, verified measurements for all 221 ears, compiled data and designed the analyses, and drafted the article. Aaron K. Remenschneider helped define the project, facilitated CT scan acquisition, provided radiological expertise, reviewed questionable scans, monitored project progress, and contributed to drafting and editing the article. Rebecca M. Farrar conducted initial measurements on 75 CT scans, tabulated preliminary data on cartilage-to-bone transitions, and refined the measurement protocol. She contributed suggestions, improvements, and article edits. Soomin Myoung managed CT scan procurement and review, performed initial measurements on 165 younger ears that were later redone with an updated measurement protocol, and contributed to refining the measurement protocol. He also conducted initial measurements on 33 of the ears published here and participated in article editing. Nicholas J. Horton designed the data structure and statistical framework, performed multiple regression analyses, and provided substantial editorial input.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work is funded by Grant R15 DC014129 from the National Institutes of Health NIDCD.

Declaration of Conflicting Interest

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

Supplemental data are available at . This link includes a .pdf file with plots of all individual measurements of area along the canal axis, spline curves measured in 3D space, and key landmarks along the canal from each of the 221 ears. Additionally, a spread sheet (.csv) contains all raw data of canal measurements. The original CT scans may not be shared due to limitations within the exemption granted through our IRB.

Supplemental Material

Supplemental materials for this article are available online.

References

Abdala

Keefe

D. H.

(2012). Morphological and functional ear development. In L. Werner, R. R. Fay, & A. N. Popper (Eds.), Human auditory development (pp. 19–59). Springer. https://doi.org/10.1007/978-1-4614-1421-6_2

Allen

(1985). Measurement of eardrum acoustic impedance. In J. B. Allen & J. L. Hall (Eds.), Peripheral auditory mechanisms (pp. 44–51). Springer. https://doi.org/10.1007/978-3-642-50038-1_6

American Academy of Audiology (2013). Clinical practice guidelines: Pediatric amplification. https://audiology-web.s3.amazonaws.com/migrated/PediatricAmplificationGuidelines.pdf_539975b3e7e9f1.74471798.pdf. Accessed 21 March 2025.

Balouch

A. P.

Bekhazi

Durkee

H. E.

Farrar

R. M.

Sok

Keefe

D. H.

Remenschneider

A. K.

Horton

N. J.

Voss

S. E.

(2023). Measurements of ear-canal geometry from high-resolution CT scans of human adult ears. Hearing Research, 434, 108782. https://doi.org/10.1016/j.heares.2023.108782

Dempster

J. H.

Mackenzie

(1990). The resonance frequency of the external auditory canal in children. Ear and Hearing, 11(4), 296–298. https://doi.org/10.1097/00003446-199008000-00007

Donaldson

J. A.

Duckert

L. G.

Lambert

P. M.

Rubel

E. W.

(1992). Anson Donaldson surgical anatomy of the temporal bone. 4 edition Raven Press.

Egolf

D. P.

Nelson

D. K.

Howell

H. C.

Larson

V. D.

(1993). Quantifying ear-canal geometry with multiple computer-assisted tomographic scans. The Journal of the Acoustical Society of America, 93(5), 2809–2819. https://doi.org/10.1121/1.405802

Farmer-Fedor

B. L.

Rabbit

R. D.

(2002). Acoustic intensity, impedance and reflection coefficient in the human ear canal. The Journal of the Acoustical Society of America, 112, 600–620. https://doi.org/10.1121/1.1494445

Huang

G. T.

Rosowski

J. J.

Puria

Peake

W. T.

(2000). A noninvasive method for estimating acoustic admittance at the tympanic membrane. The Journal of the Acoustical Society of America, 108(3), 1128–1146. https://doi.org/10.1121/1.1287024

10.

Ito

Imada

Ikeda

Sueno

Arikuni

Kida

(2001). A morphological study of age changes in adult human auricular cartilage with special emphasis on elastic fibers. The Laryngoscope, 111(5), 881–886. https://doi.org/10.1097/00005537-200105000-00023

11.

Ito

Kubota

Furukawa

Kakehata

(2020). Measurement of the pediatric and adult osseous external auditory canal: Implications for transcanal endoscopic ear surgery. Otology & Neurotology, 41(6), e712–e719. https://doi.org/10.1097/MAO.0000000000002653

12.

Johansen

(1975). Measurement of the human ear canal. Acustica, 33(5), 349–351.

13.

Keefe

D. H.

Abdala

(2007). Theory of forward and reverse middle-ear transmission applied to otoacoustic emissions in infant and adult ears. The Journal of the Acoustical Society of America, 121(2), 978–993. https://doi.org/10.1121/1.2427128

14.

Keefe

D. H.

Bulen

J. C.

Arehart

K. H.

Burns

E. M.

(1993). Ear-canal impedance and reflection coefficient in human infants and adults. The Journal of the Acoustical Society of America, 94(5), 2617–2638. https://doi.org/10.1121/1.407347

15.

Keefe

D. H.

Bulen

J. C.

Campbell

S. L.

Burns

E. M.

(1994). Pressure transfer function and absorption cross section from the diffuse field to the human infant ear canal. The Journal of the Acoustical Society of America, 95(1), 355–371. https://doi.org/10.1121/1.408380

16.

Keefe

D. H.

Fitzpatrick

D. F.

Porter

H. L.

Chen

(2024). Reflection function, reflectance, and area function measurements in ears of children and adults. The Journal of the Acoustical Society of America, 156(4), 2709–2726. https://doi.org/10.1121/10.0032455

17.

Keefe

D. H.

Hunter

L. L.

Feeney

M. P.

Fitzpatrick

D. F.

(2015). Procedures for ambient-pressure and tympanometric tests of aural acoustic reflectance and admittance in human infants and adults. The Journal of the Acoustical Society of America, 138(6), 3625–3653. https://doi.org/10.1121/1.4936946

18.

Lewis

J. D.

Neely

S. T.

(2015). Non-invasive estimation of middle-ear input impedance and efficiency). The Journal of the Acoustical Society of America, 138(2), 977–993. https://doi.org/10.1121/1.4927408

19.

Merchant

G. R.

Al-Salim

Tempero

R. M.

Fitzpatrick

Neely

S. T.

(2021). Improving the differential diagnosis of otitis media with effusion using wideband acoustic immittance. Ear and Hearing, 42, 1183–1194. https://doi.org/10.1097/AUD.0000000000001037

20.

Merchant

G. R.

Siegel

J. H.

Neely

S. T.

Rosowski

J. J.

Nakajima

H. H.

(2015). Effect of middle-ear pathology on high-frequency ear-canal reflectance measurements in the frequency and time domains. In AIP conference proceedings (vol. 060003). https://doi.org/10.1063/1.4939358

21.

Nørgaard

K. R.

Fernandez-Grande

Laugesen

(2019). Compensating for oblique ear-probe insertions in ear-canal reflectance measurements. The Journal of the Acoustical Society of America, 145(6), 3499–3509. https://doi.org/10.1121/1.5109894

22.

Nørgaard

K. M.

Fernandez-Grande

Schmuck

Laugesen

(2020). Reproducing ear-canal reflectance using two measurement techniques in adult ears. The Journal of the Acoustical Society of America, 147(4), 2334–2344. https://doi.org/10.1121/10.0000946

23.

Ontario Ministry of Children (2019). Infant hearing program: Amplification protocol. https://www.uwo.ca/nca/pdfs/clinical_protocols/IHP_Amplification%20Protocol_2019.01.pdf. Accessed 21 March 2025.

24.

Rasetshwane

D. M.

Neely

S. T.

(2011). Inverse solution of ear-canal area function from reflectance. The Journal of the Acoustical Society of America, 130(6), 3873–3881. https://doi.org/10.1121/1.3654019

25.

Richardson

J. T.

(2011). Eta squared and partial eta squared as measures of effect size in educational research. Educational Research Review, 6(2), 135–147. https://doi.org/10.1016/j.edurev.2010.12.001

26.

Stinson

M. R.

(1990). Revision of estimates of acoustic energy reflectance at the human eardrum. The Journal of the Acoustical Society of America, 88(4), 1773–1778. https://doi.org/10.1121/1.400198

27.

Stinson

M. R.

Lawton

B. W.

(1989). Specification of the geometry of the human ear canal for the prediction of sound-pressure level distribution. The Journal of the Acoustical Society of America, 85, 2492–2503. https://doi.org/10.1121/1.397744

28.

Voss

S. E.

Allen

J. B.

(1994). Measurement of acoustic impedance and reflectance in the human ear canal. The Journal of the Acoustical Society of America, 95(1), 372–384. https://doi.org/10.1121/1.408329

29.

Voss

S. E.

Girardin

K. D.

Zhang

Xia

Nei

Horton

N. J.

(2017). Wideband acoustic immittance: Effects of measurement equipment, age, gender, and ear-canal area. In Abstracts of the American Auditory Society Annual Meeting, Scottsdale AZ.

30.

Voss

S. E.

Horton

N. J.

Fairbank

K. E.

Xia

Tinglin

L. R. K.

Girardin

K. D.

(2020). Measurements of ear-canal cross-sectional areas from live human ears with implications for wideband acoustic immittance measurements. The Journal of the Acoustical Society of America, 148, 3042–3051. https://doi.org/10.1121/10.0002358

31.

Voss

S. E.

Horton

N. J.

Woodbury

R. R.

Sheffield

K. N.

(2008). Sources of variability in reflectance measurements on normal cadaver ears. Ear and Hearing, 29(4), 651–665. https://doi.org/10.1097/AUD.0b013e318174f07c

32.

Voss

S. E.

Shera

C. A.

(2004). Simultaneous measurement of middle-ear input impedance and forward/reverse transmission in cat. The Journal of the Acoustical Society of America, 116(4 Pt 1), 2187–2198. https://doi.org/10.1121/1.1785832

33.

Voss

S. E.

Stenfelt

Neele

S. T.

Rosowski

J. J.

(2013). Factors that introduce intrasubject variability into ear-canal absorbance measurements. Ear and Hearing, 34, 60s–64s. https://doi.org/10.1097/AUD.0b013e31829cfd64

34.

Voss

S. E.

Sun

Horton

N. J.

(2024). Online database of published wideband acoustic immittance (WAI) measurements. In Mechanics of Hearing Workshop, University of Michigan. https://doi.org/10.5281/zenodo.13334749

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.06 MB

172.25 MB