Anatomical Boundary of the Tympanic Membrane Pars Flaccida of the Meriones unguiculatus (Mongolian Gerbil)

Authors


Abstract

The pars flaccida of the Meriones unguiculatus (Mongolian gerbil) was in previous studies shown to bulge almost spherically when pressurized, a behavior suggesting that it is suspended by a fixed circular boundary. The question arises whether this “functional” boundary is based on an underlying circular anatomical boundary, an important issue for modeling the middle-ear mechanics. In this article, the boundaries of the Mongolian gerbil pars flaccida were visualized in situ with otomicroscopy and in slides with light microscopy and by micro-CT radiology. For the major part of its circumference, the pars flaccida was found to be suspended by rigid bone, that is, the tympanic legs. The remaining boundary is made up of the terminal portion of the handle of the malleus and the soft tissue of the terminal arches. The attachment to these structures is simple and uncomplicated, and the geometry is regular and symmetric: deviating by only 3.5% from a perfect circular shape. The findings justify the use of a fixed circular boundary as a good approximation for the modeling of pars flaccida deformation in the Mongolian gerbil. Anat Rec, 2011. © 2011 Wiley-Liss, Inc.

Modeling the sound transmission through the middle ear is an important tool in predicting the hearing outcome of various designs of ossicle prosthesis and surgical methods. It also contributes to an increased understanding of the middle-ear acoustics. To make relevant and accurate models, the precise anatomical parameters need to be assessed and included. In the development of such models, various mammal ears such as cat and gerbil are used (Ladak and Funnel, 1996; Elkhouri et al., 2006; Tuck-Lee et al., 2008). The tympanic membrane (TM) is one of the major structures involved in the sound transmission. In mammals, it has two distinct parts, that is, the pars tensa (PT) and the pars flaccida (PF). In most species, the PF is smaller than the PT. The PF versus PT area ratio ranges widely among 21 studied mammalian species from 0 in the Talpa europaea (mole) and Chinchilla laniger to 0.68 in the Erinacaeus europaeus (hedgehog) (Kohllöffel, 1984). In man, the ratio varies between 0.027 and 0.05, and in the Mongolian gerbil it is ∼0.13 (Lay, 1972; Chole and Kodama, 1989; Decraemer et al., 1991; Dirckx et al., 1998; Henson et al., 2005), which makes gerbil a particularly interesting model to study this aspect of TM function. For a number of reasons, gerbil is an interesting animal model to study middle-ear mechanics. It has, therefore, served in many studies on sound transmission in the middle ear (e.g., Woolf and Ryan, 1988; Ravicz et al., 1992, 1996, 2007, 2008; Cohen et al., 1993; Ravicz and Rosowski, 1997; Olson, 1998; Overstreet and Ruggero, 2002; Decraemer et al., 2007; de La Rochefoucauld et al., 2008; de La Rochefoucauld and Olson, 2010). The PT collects sound energy and transmits these sound vibrations to the ossicles. The function of the PF is not completely understood. It has been assigned a role in the immunological defense of the middle ear and in the pressure regulation of the middle ear (Widemar et al., 1984; Dirckx et al., 1997). It has also been suggested to reduce the sensitivity of the middle ear to sounds of frequencies below 500 Hz (Kohllöffel, 1984; Teoh et al., 1997). The PF may also play an important role in the pathogenesis of chronic retraction-type middle-ear disease (Ars, 1991). The PT and PF have different embryological origins, which are reflected in the difference in their structural and physical properties (Shrapnell, 1832; Lim, 1968, 1995; Hellström et al., 1983; von Unge et al., 1991). It is not easy to measure the mechanical stiffness of the TM in vivo. To assess these properties, an in vitro model was developed. It was used to measure the PT displacement in response to applied middle-ear pressure in the Mongolian gerbil (von Unge et al., 1993; Dirckx and Decraemer, 1997). In further studies, the PF stiffness was also measured (Dirckx et al., 1998; Larsson et al., 2001, 2003, 2005). These measurements confirmed the hypothesis originally proposed by Shrapnell in 1832 of the PF having quite different mechanical properties than the PT (hence their names).

Shrapnell made his pioneering TM studies on the goat. The Mongolian gerbil currently attains great interest as an animal model for middle-ear studies and has also been used in finite element modeling of the sound transmission (Chole and Kodama, 1989; von Unge et al., 1991, 1993, 1997; Larsson et al., 1999, 2001, 2003, 2005; Rosowski et al., 1999; Elkhouri et al., 2006). In previous studies, interesting findings were made: when the gerbil middle-ear cavity is inflated the PF takes the form of a spherical cap, and the “functional boundary” of the PF displacement is almost circular (Dirckx et al., 1997, 1998). In this study, we aim to assess the anatomical boundary of the PF using otoscopy, microscopy, high-resolution micro-CT radiography, and light microscopy. The findings will be compared with a perfect planar circular “functional boundary.” The anatomy of the PF is essential to its deformation under pressure and, hence, to its functional role in the mechanics of the middle ear.

MATERIALS AND METHODS

Six isolated temporal bones of healthy, female Mongolian gerbils (Meriones unguiculatus) aged between 3 and 6 months were used. The temporal bone is an otosurgical entity that consists of the petrous, the tympanic, and the squamous bones including the middle- and inner-ear structures. The animals were housed in cages with food and water ad libitum in the animal facility of the University of Antwerp.

Stockholm's North Animal Ethics Board (Dnr. N. 170/94) approved the study.

After decapitation of the animal under carbon dioxide anesthesia, the temporal bones were isolated. Two temporal bones were dissected further for anatomical examination under a dissection microscope, two were examined with high-resolution micro-CT radiography, and two were processed for histological examination of the PF boundary and adjacent structures.

Otoscopy of a Lateral Aspect of the TM

The fresh tympanic bulla was observed and photographed under a dissection microscope (Zeiss Stemi 2000-C). The TM was observed through the external bony meatus with the use of transillumination.

Microscopy of a Medial Aspect of the TM

For more detailed assessment of the anatomy of the PF and its boundary, the temporal bone was dissected so that the TM could be studied with the operating microscope from the middle-ear side. Thus, the tympanic bulla was extensively dissected with removal of all structures situated more medial than the TM annulus. The handle of malleus, which is firmly attached to the PT, was, however, not removed because this is hard to accomplish without damaging the TM. The neck of the malleus was cut and the malleus head removed. The dissection thus leaves only the TM with the surrounding bone frame and the attached handle of malleus, the bony external ear canal, and some remnants of the bullar bone. The two parts of the TM, that is, the PT and PF, became fully accessible for anatomical examination and for photodocumentation.

Alignment.

At a relatively high magnification level (∼50 times), the focal depth in the dissection microscope becomes quite low, that is, less than 0.5 mm. This feature was used to align the plane of the PF so that it was orthogonal to the observation direction: only when perpendicularly aligned is there full focus on all portions of the PF boundaries. This observation also shows that the boundary lies within one plane. This alignment was used for the photographs obtained for boundary analysis.

Boundary assessment.

The photographs of the PF were oriented with the anterior direction to the left and the superior direction up. They were digitized and the boundaries of the PF were traced manually with the aid of National Institute of Health (NIH) Image software (version 1.44). The data were used for further calculations and analyses.

Scaling.

The scaling of the photographs was obtained from a recording where the specimen was replaced by a linear calibration scale.

Reference system.

A polar coordinate reference system was adopted to facilitate the description of the PF boundary in the Results section.

PF surface area and boundary shape.

With the aid of the NIH Image computer software, on a digitized image of the PF the PF boundary and its geometrical center point was defined. The PF area was calculated by counting calibrated pixels, and its mean radius was calculated. Subsequently, a circle was superimposed on the representation of the PF, with its geometrical center coinciding with the point previously determined as the center point of the PF and with a radius RPF determined from RPF = √ (APF/π). Subsequently, the area of the overlapping region of both the PF and the circle was directly determined, as well as the sum of areas included by either the PF or the circle. The proportion of these two measures was used as a measure of how much the PF deviates from a perfect circular shape.

Micro-CT Radiography

Specimens were scanned at the CT scanning facility of UGCT at the Ghent University (www.ugct.ugent.be), using a micro-CT radiograph of medium energy (up to 160 kV) with directional X-ray tube and spot size of 10 μm. Specimens were mounted on a controllable rotating table (MICOS, UPR160F-AIR) and fixed in a plastic tube devoid of fluids to increase the contrasts of the reconstructed sections. For each specimen, a series of 1,000 projections of 1,496 × 1,496 pixels was recorded covering 360 degrees. The voxel size was 13.5 μm. Reconstruction of the tomography projection data was done using the Octopus package (Octopus version 8.5, Center for Micro tomography University of Ghent), resulting in reconstructed 3D volumes of 1,496 × 1,496 × 1,800 voxels. Volume and surface rendering was performed using Amira (Amira 4.1, Mercury Computer Systems).

Histology

The temporal bone, which consists of the petrous, the tympanic, and the squamous bones including the middle- and inner-ear structures, was removed en bloc. The tympanic bulla was opened, and the temporal bone was fixed in 4% paraformaldehyde in cacodylate buffer, decalcified in 0.1 M EDTA for 1–2 weeks until the bone was soft, dehydrated, and embedded in Agar resin 100R (Agar Scientific, England). For further details, please refer to von Unge et al. (1991). Sections of the TM of ∼1-μm thickness were made perpendicularly to the annulus plane, in parallel to the direction of the handle of malleus. The sections were stained with toluidine blue and studied and photographed using a Zeiss Axioplan photomicroscope.

RESULTS

Otoscopy of a Lateral Aspect of the TM

The gerbil temporal bone, with its relatively large bulla, somewhat resembles a small bird's egg: it has a thin, hard, rounded shell of bone and is to a great extent hollow. The PF is the most lateral part of the TM and is almost perpendicular to the axis of the external ear canal. Therefore, it is easily accessible for visual examination through the ear canal. The use of transillumination enhances the visibility of the PF and surrounding structures. Figure 1 shows a view of the PF and the superior part of the PT through the bony external meatus. The PT forms the medial end of the external ear canal. The superior portion of the PT is situated more laterally, and its inferior portion is more medial and hardly seen through the ear canal. Figure 1 shows only the superior portion of the PT. The middle portion fades away in a medial direction out of focus and view. The PT is almost perfectly transparent. The handle of the malleus, which is embedded in the PT, can be clearly observed through the external meatus.

Figure 1.

Otoscopy of the tympanic membrane photographed through the external ear canal of an isolated right side temporal bone in a transillumination setup. The pars flaccida is well exposed, its boundary indicated by a few thin arrows, whereas the pars tensa, obliquely positioned, fades away in a medial direction inferiorly to become hidden by the ear canal wall. Some wrinkles are seen on the pars flaccida surface resembling waves on a sea. The pointed posterior tympanic spine (ps) and the more blunt anterior tympanic spine (as) divide the pars flaccida from the posterior portion of pars tensa and give the direction of their intersection toward the malleus, seen as a rounded fold (hollow arrows). The handle of malleus is labeled (h), and the shadows of the malleus head and the incus body are projected onto the TM.

The PF exhibits small wrinkles, but the overall appearance is that of a flat membrane. As the PF is partly transparent, the bony border that suspends the PF is easily seen. It is almost circular and covers around two-thirds of the circumference. Two bony processes, the anterior tympanic spine, and the posterior tympanic spine extend the bony border of the PF. The posterior spine is longer and more pointed than the anterior one. Between these spines the bony boundary is incomplete. This gap is spanned by two soft tissue structures: the anterior and the posterior terminal arches (arcus terminalis) in the junction zone between the PF and the PT. The arches complete the circumference of the PF. Observed from a lateral view the arches are seen at the fold between the PF and the PT (Fig. 1). The outlines of the neck and the head of the malleus in conjunction with the incus body are vaguely seen through the PF. Although the entire surface of the PF can be observed through the external ear canal, a detailed investigation of its circumference is better performed from the middle-ear side where its boundaries can be observed more clearly.

Dissection Microscopy of a Medial Aspect of the TM

The dissections include removal of the ossicles (leaving only the handle of malleus in place). Thereby, only the TM with its suspending structures, the external ear canal, and a portion of the bullar bone remain. The entire TM, that is, the PT and PF, has become fully exposed from the middle-ear side for anatomical examinations. Figure 2a shows a view of an entire right TM and the remaining portion of the temporal bone. In a close-up view (Fig. 2c), the PF again exhibits a wrinkled surface, similar to what was seen in the lateral view (Fig. 1). The major portion of the PF circumference is built up by bone. Visually the PF appears almost circular. For orientation some structures are indicated in Fig. 2d.

Figure 2.

(a) Dissected right side temporal bone at the end stage of the dissection, observed from the middle-ear side. The anterior is to the left in the figure. All structures that were situated medial to the annulus have been removed. The handle of malleus (H) is still in place, attached to the pars tensa. The pars flaccida (PF) is fully visible. (b) 3D model of right side temporal bone reconstructed from a micro-CT radiography recording with Amira. Bone structures such as handle of malleus, tympanic ring, and bony ear canal are visualized. Large parts of the bulla and bone are virtually removed. The malleus head and neck are removed, as well as the incus and stapes. (c) Dissected temporal bone in close-up view of the same ear as in Fig. 2a. The pars flaccida (PF) has an almost circular shape. In the photograph, the contour of the external meatus is hinted through the tympanic membrane just inferiorly to the terminal arches. (d) Schematic indication of structures of Fig. 2c: the terminal arches are hinted with light gray lines and the bone edges with dark lines. (e) Contours and polar coordinate system. The contours of the pars flaccida, the pars tensa, the terminal arches, and the handle of malleus are traced manually with the aid of NIH Image software. The polar coordinate system is superimposed (see text). Angles are given for different structures engaged in the pars flaccida boundary. (f) Close-up view of 3D model of right side temporal bone reconstructed the same scans as Fig. 2b. Note that a large part of the malleus has been removed in the 3D model to obtain a clear view of the malleus handle being part of the PF boundary. (g) Pars flaccida and a fitted circle. Outlining of structure contours from Fig. 2e and the best-fitted circle superimposed. The surface area of the pars flaccida boundary is shaded in gray. The circle line itself is included in the circle area. (h) Reconstructed pars flaccida boundary and a fitted circle. Illustration of the circle fitted through selected 3D point coordinates of the pars flaccida boundary obtained from a right micro-CT radiography model. The points are projected to the plane of the circle with radius 711 μm and deviate only about 2% from it.

A polar coordinate system superimposed on the image of Fig. 2c is shown in Fig. 2e. The origin of the coordinate system is set at the center point of the PF defined with the aid of the NIH Image computer software. The center point is defined as the crossing point between two different diagonals drawn through the PF in a way so that each diagonal divides the PF area into two halves. The 0-degree “reference angle” is defined as the direction toward the center of the mallear handle, and angles increase in an anticlockwise direction in a right side ear (going from inferior direction through posterior, superior, anterior, and again inferior direction) and clockwise in a left side ear.

Inferiorly, at 0-degree angle the boundary of the PF is defined by the junction with the superior end of the handle of malleus (Fig. 2e). Going anticlockwise, this junction ends at 13-degree angle in this right ear. From here, at 13-degree angle there is a gap in the bony boundary until 55-degree angle (Fig. 2f). In this bone gap, the PF boundary is provided by the posterior portion of the terminal arch, which is a patch of soft tissues. This patch can be identified between the bone frame, the PF, and the PT by its difference in light reflection and transparency when compared with the PF and the PT (Fig. 2c). At 55-degree angle the PF meets the apex of the posterior tympanic spine of the posterior tympanic leg (terminology according to Henson, 1961) (Fig. 2d). The bone structures that surround the PF and constitute the external ear canal are the posterior and anterior tympanic legs of the temporal bone (“Tympanicum Schenkeln” according to the terminology of Bondy, 1908) (Fig. 2d). They meet in a bone suture located at ∼130-degree angle in the coordinate system. From the tympanic legs, the posterior and anterior tympanic spines emerge. As seen on the images of Figs. 1 and 2, the posterior spine is long and sharply pointed, whereas its counterpart, the anterior spine, is broad and not pointed. At the site of the suture, between 126- and 135-degree angle, a minor cleft is present (Fig. 2e,f).

Going further anticlockwise, the PF boundary is defined by the anterior leg until the base of its anterior spine at 271-degree angle. From here, the PF boundary is defined by the triangular, anterior portion of the terminal arch, until it reaches the anterior edge of the superior end of the handle of the malleus at 338-degree angle (Fig. 2d). In the remaining 22-degree angle of the circumference, the handle of the malleus constitutes the PF boundary. There might be an overestimation of the 22-degree angle because of the prominent part of the handle blocking the view (Fig. 2c). Data from the left and right ears are shown in Table 1.

Table 1. Circumference in degrees and PF radius
 MicroscopyMicro-CT scan
Right earLeft earLeft earRight ear
  1. Circumference in degrees of the suspending and boundary structures of the Mongolian gerbil pars flaccida and its radius. The remnant of the malleus neck conceals in at least one ear the sight of the boundary at the posterior arch in dissection microscopy, causing an incorrectly high value for the handle and low value for the arch. These readings are therefore put within brackets.

Post-terminal arch (°)42524041
Post-tympanic leg (°)71717065
Suture (°)9151117
Ant-tympanic leg (°)136129139136
Ant-terminal arch (°)(67)(62)8387
Handle of malleus (°)(35)(31)1514
PF radius (mm)0.710.710.760.71

During dissection, the malleus was cut at the neck leaving a portion of the neck overhanging the PF at the upper end of the handle. This portion hides the real PF-to-malleus handle junction when observed from the medial view (Fig. 2e). In the view from the external side on Fig. 1, one can see that this junction is almost straight. Furthermore, Fig. 2b shows this area without such concealing overhang, as the virtual removal of the neck is made with sufficiently high precision in contrast to physical cutting of the neck which leaves portion of it still in place.

In summary, the PF is in ∼60.0% (210-to 216-degree angle) out of its 360-degree angle of circumference attached to the temporal bone, and in the remaining part it is attached to either soft tissues, that is, the terminal arches or to the mobile handle of malleus.

The PF center point defined with the aid of the NIH Image software is demonstrated for one ear in Fig. 2e, and the PF area was directly determined by counting calibrated pixels: APF = 1.56 mm2. Figure 2 demonstrates the superimposed circle on the representation of the PF, with coinciding geometrical center points. The radius RPF was determined from RPF = √ (APF/π). The area of the overlapping region of both the PF and the circle was then measured: 1.45 mm2. Thus, 0.11 mm2 of the PF was not included within the circle, and 0.11 mm2 of the circle was not included within the PF boundary. The area portion of the PF that is excluded by the circle is a measure of the degree of deviation from a perfect circular shape. By calculation the deviation from a perfect circle is (0.11/1.56) = 7.1%. The radius of the circle, and thereby also the mean radius of the PF, is RPF = 0.71 mm.

Micro-CT Radiography

The micro-CT radiography records only bone but no soft tissue, so only those portions of the PF boundary that are composed of bone can be analyzed. The 3D volume model has been virtually trimmed, and parts are cut away to allow a clear view of bony boundaries of the PT and PF (Fig. 2b). The PF region is magnified in Fig. 2f, where it is verified that the attachment to the handle of malleus is less than what Fig. 2e suggests. Based on the polar coordinates, the boundary proportions for the different suspending structures were calculated (Table 1). The proportions are consistent between the four ears, except for the measurements of the width of the handle of malleus, which were distorted in microscopy recordings due to the remnants of the malleus neck.

The deviation from the shape of a perfect planar circle was also calculated from the micro-CT radiography-based models. As only bone structures are recorded in micro-CT, the calculation is limited to those PF boundary portions that are defined by bone. A set of 3D coordinates was selected along the PF ring (similar to the points indicated in Fig. 2e). A circle is fitted through these data points. Figure 2h displays these points projected in the plane of the circle for the right ear. The mean value of the shortest distances to the perimeter of the circle is 13.6 ± 10 μm for the left ear and 13.2 ± 8.6 μm for the right ear. The radius of the fitted circle is, respectively, RPF = 0.76 mm and RPF = 0.71 mm. The deviation of the chosen 3D points on the micro-CT radiography model to the radius of the circle is thus only 2%.

The 3D models of the middle-ear bone obtained from micro-CT radiography furthermore provide the opportunity to measure the deviation from a perfect plane of the bony rim of the PF, as well as the angle between the planes of the bony rims of the PF and the PT. Thus, it was found that the mean orthogonal deviation of the points to their fitted perfect plane is 15.4 ± 6 μm for the left and 16.3 ± 10 μm for the right ear. This is mostly caused by the posterior leg which tends to form a more medial boundary when compared with the anterior one. The angle between the planes of the PF and PT boundaries is 23 degree (±3 degree), giving a slightly more upright position of the PF compared with the PT.

Histology

Figure 3 shows micrographs of three different sections of the PF boundary. The subepithelial, lamina propria layer is the dominating layer in the PF. It is, unlike the situation in the PT where densely packed collagen fibers are present, composed of loose connective tissue (Fig. 3). It has a low density of collagen fiber structures and lacks the strict fiber arrangement that is present in the PT (von Unge et al., 1991). The similar situation has been reported for the human TM (Lim, 1968). The attachment of the PF to the surrounding structures differs from that of the PT, that is, the PF lacks a fibrous annulus and tympanic sulcus morphology. Toward the periphery, near the boundary regions, the relatively sparse fiber structures form a radiating band that inserts directly at the edge and at the ear canal side of the surrounding bone (Fig. 3c).

Figure 3.

(a) Light microscopy section through anterior portion of pars flaccida (PF), the anterior terminal arch (TA), and anterosuperior portion of pars tensa (PT). The middle-ear side (ME) is indicated. The tympanic membrane becomes gradually more condensed as it converts from pars flaccida to terminal arch. The wave-like pattern in pars flaccida may represent remnants of the wrinkles seen on otoscopy. (b) Light microscopy section through middle section of pars flaccida (PF) and its attachment (arrow) to the handle of malleus (H). Pars flaccida is condensed and thin in the region close to the malleus, similar to the appearance of the terminal arch. (c) Light microscopy section of superior portion of pars flaccida (PF) with its attachment (arrow) to the temporal bone (TB). Pars flaccida is condensed in the region close to the bone. The fiber structures of lamina propria have a simple attachment to the bone.

In frontal view, the anterior terminal arch appears triangular and larger than the posterior one, which is formed as a narrow band (Fig. 2c,d). Under light microscopy, the cross section of the arch is somewhat similar to that of the PF, although it is thinner and appears denser (Fig. 3a). The boundary between the PF and the arch is hard to pinpoint exactly in the slides because the structural differences appear gradually. The histology of the PT differs somewhat from that of the arch; it appears more even and smooth and has fewer cells. This boundary is also somewhat diffuse and hard to determine exactly. Close to the handle of the malleus, the PF is thinner and denser (Fig. 3b) resembling the cross section of the terminal arches. Therefore, this narrow region may be regarded as a connecting part of the two arches. The lamina propria inserts bluntly and directly into the handle.

All boundaries between the different TM substructures are relatively well displayed on a transillumination photograph of the medial view of the TM (Fig. 1). In regard of the gradual transition from the morphology of one of these TM substructures to another, it seems relevant to utilize the boundaries that are traced on the transillumination photographs for further analysis. The light microscopy section in Fig. 3a shows a fold in the terminal arch region between the PT and the PF. This fold corresponds to the fold demonstrated by otoscopy (Fig. 1).

DISCUSSION

The exact definition and description of the anatomical middle-ear structures is important for the creation of a computer model to simulate the middle-ear sound transmission. A few mammalian species have become preferred models for such simulation, like the Mongolian gerbil (Elkhouri et al., 2006). In our previous moiré interferometry studies, it was found that the gerbil PF attains the regular and simple shape of a spherical cap when pressurized in a lateral direction (Dirckx et al., 1997). When pressurized in a medial direction, the proximity of the neck or head of the malleus would soon interfere with its displacement and prevent the shaping of a sphere cap. Our intention was to investigate the anatomy of the gerbil PF to define its precise anatomical boundary utilizing different methods: otomicroscopy, dissection microscopy, micro-CT radiography, and histology.

The gross structure of the gerbil PF is quite flat when not exposed to pressure, whereas the epithelial surfaces are wrinkled as is seen by otoscopy and dissection microscopy. Previous moiré interferometry findings showed that pressure variations of only a few tens of Pascal's across the PF suffice to make it flip from bulging to one side over to the other. It gave the impression of too large a membrane for its frame (Larsson et al., 2001). In conjunction with the presently observed epithelial wrinkling phenomenon, it appears that the lamina propria is an elastic sheet covered with less elastic epithelial “wall paper.” The wrinkled surface of the PF observed by otomicroscopy is also apparent in the photomicrographs, although the wrinkles reduce by the bulging of the PF that tend to occur during the embedding procedure. This extra tissue—the PF being too large to fit its boundary—thus allows very easy deformation of the TM in small overpressure or underpressure in the middle ear, which is most likely involved in its function. One possible function of the PF may be to “absorb” or attenuate the effects of small but rapid pressure changes in the ear (Dirckx et al., 1997, 2006; Didyk et al., 2007). It is the authors' personal experience from several previous examinations on anesthetized gerbils that the PF most often is not seen to be flat but bulging inward.

Several authors have investigated the acoustic function of the PF (e.g., Arimoto et al., 1988 in human; e.g., Teoh et al., 1997; Rosowski et al., 1999; Rosowski and Lee, 2002 in gerbil). Experiments using stiffening and closing of the PF showed that the acoustic behavior of the ME changes. It was discussed that the presence of the PF may on the one hand facilitate the motion of the malleus or on the other hand may reduce low-frequency hearing by shunting part of the acoustic energy. The PF acts as an independent path for volume velocity to enter the middle ear, thereby reducing the sound pressure difference across the PT (Teoh et al., 1997; Rosowski et al., 1999). At low frequencies, the volume flow through its stiffness determined. At about 500 Hz, the impedance through the path has a broad minimum, and above 500 Hz, the mass of the tissue causes a decrease in volume flow through the path. The authors conclude that the effect of PF on middle-ear sound flow in gerbil is primarily restricted to frequencies less than 500 Hz, where it acts like a parallel stiffness into the middle-ear air spaces. This stiffness-controlled behavior in parallel with the stiffness of the tensor produces a flat 5–10 dB step-like reduction in tensa motion at frequencies below 500 Hz. Rosowski and Lee (2002) showed a small but significant effect on umbo vibration velocity when the PF was immobilized, but the small size of the effects as function of middle-ear pressure argued against a pressure regulation function of PF. Therefore, also with respect to the acoustic function many questions still remain. The exact functioning of the system is not clear yet, and realistic finite element modeling can bring better insight. The advantage of computer modeling is that situations can be created that are very difficult or even impossible to realize experimentally (for instance making the PF very flaccid by changing its elastic parameters), and to create such a realistic model, correct anatomical data are needed.

The PF is attached to rigid bone, that is, the anterior and posterior legs, for approximately two-thirds of its entire stretch of boundary. The fibers of the lamina propria insert into the bone without a tympanic annulus that suspends the PT. This part of the boundary has a shape that is very close to a perfect planar circle. The average deviation from a perfect plane is only tens of a micrometer. The deviation of the boundary of the PF to a circle is found to be only 7.1% in surface area or 3.5% in radius. From the radiographic images, the deviation in radius between the circumference and a perfect circle is even smaller (2%). The assessed mean radius of the PF (0.71 mm) is consistent with the measurements obtained by Teoh et al. (1997), who reported a mean PF area of 1.4 ± 0.2 mm2 (N = 7).

Approximately one-third of the PF boundary is made up by structures other than rigid bone. The anterior and posterior terminal arches compose most of it. The terminal arches are seldom discussed in the middle-ear literature. They consist of soft tissue, much like the PT and PF, but can be identified by their different transmission and reflection of light (Figs. 1a, 2c). They form a fold in-between the two major parts of the TM (Figs. 2c, 3a). Histology reveals a gradual structural change in the lamina propria in the transition zone between the PF and the arches (Fig. 3a). The lamina propria gradually becomes more condensed at the arches. The epithelial linings continue without significant changes. The function of the terminal arches is most likely to separate the acoustic function of the PT from that of the PF, whose specific function is unclear. The fold running through the terminal arches may have a reinforcing effect at this section of the PF boundary and further contribute to the formation of a sphere cap of the PF, which was shown in our previous report (Dirckx et al., 1997). The angle of 23 degree between the planes of the PF and the PT may contribute to the creation of this fold.

The TM fiber insertion into the top end of the handle of malleus is simple and similar to what was found in the tympanic legs. This part of the PF boundary is not fixed, which allows the malleus to rotate during static pressure across the TM, although the actual rotation necessary is relatively small because it is around the functional axis of the malleus. It is, therefore, unlikely that the lack of fixation of this part of the boundary has any significant function for the PF.

As mentioned above, the gerbil PF boundary is very close to a planar circle, consistent with the functional behavior at static pressures reported by Dirckx et al. (1997). This almost perfectly circular anatomical boundary of the PF, with quite small interindividual variations, is consistent with displacement measurements where an almost circular shape of the “functional boundary” of the PF was shown. It is unclear why the gerbil PF, in contrast to other studied mammals, has such regular shape. It may offer a functional advantage because a circular boundary would yield the largest volume displacement for a given surface area and membrane stretch. This may increase the effectiveness of the PF to act as a high-pass filter, buffering quasi-static pressure differences across the TM, and as such, the PF may play a role in hearing protection against low-frequency sounds in the gerbil. Addition of a correct model of PF to current finite element models of the middle ear will improve the realism of such models. Highly realistic middle-ear models improve our insight in the fundamental functioning of the middle-ear mechanics and are highly useful for the development and optimization of middle-ear hearing implants and ossicle prostheses. To fully understand the role of the PF, details regarding its maximal distension, its compliance, its dynamic behavior, and especially its elastic properties will be needed. Its acoustic behavior and displacement under static pressure have already drawn the attention of several authors. For the gerbil, one of the important animal models in middle-ear research, an accurate anatomical description is, however, still missing. To obtain correct finite element modeling results, this morphological information is essential, before materials parameters, and so forth can be added.

CONCLUSIONS

To obtain realistic functional computer models for the middle-ear system of the Mongolian gerbil, an exact anatomical description of the mechanical components is necessary. It has been shown that the PF has a simple and relatively homogeneous ultrastructure. In this report, it is shown that the PF also has a simple attachment to the suspending structures and that its geometry is quite simple, regular, and symmetric: almost circular and planar as long as pressures on each side are equal. Furthermore, the PF rim and plane are at an angle of 23 degree to the PT plane. The data are derived from a small number of specimens. The variability amongst ears from one animal to another is, however, small. These results justify simple modeling of its deformation. Our accurate anatomical description will allow incorporating the PF in finite element models of the ME. Combined with correct materials parameters this will form an important addition in the realization of realistic functional ME models. Its anatomical features also explain its functional boundaries shown in previous deformation studies.

Ancillary