Temporal bone characterization and cochlear implant feasibility in the common marmoset (Callithrix jacchus)

Abstract

The marmoset (Callithrix jacchus) is a valuable non-human primate model for studying behavioral and neural mechanisms related to vocal communication. It is also well suited for investigating neural mechanisms related to cochlear implants. The purpose of this study was to characterize marmoset temporal bone anatomy and investigate the feasibility of implanting a multi-channel intracochlear electrode into the marmoset scala tympani. Micro computed tomography (microCT) was used to create high-resolution images of marmoset temporal bones. Cochlear fluid spaces, middle ear ossicles, semicircular canals and the surrounding temporal bone were reconstructed in three-dimensional space. Our results show that the marmoset cochlea is ∼16.5 mm in length and has ∼2.8 turns. The cross-sectional area of the scala tympani is greatest (∼0.8 mm²) at ∼1.75 mm from the base of the scala, reduces to ∼0.4 mm² at 5 mm from the base, and decreases at a constant rate for the remaining length. Interestingly, this length–area profile, when scaled 2.5 times, is similar to the scala tympani of the human cochlea. Given these dimensions, a compatible multi-channel implant electrode was identified. In a cadaveric specimen, this electrode was inserted ¾ turn into the scala tympani through a cochleostomy at ∼1 mm apical to the round window. The depth of the most apical electrode band was ∼8 mm. Our study provides detailed structural anatomy data for the middle and inner ear of the marmoset, and suggests the potential of the marmoset as a new non-human primate model for cochlear implant research.

Introduction

The marmoset monkey (Callithrix jacchus) is a small New World primate (300–500 g). Its hearing range and the structure of its auditory cortex are similar to other primates (Fay, 1988; de la Mothe et al., 2006; Osmanski and Wang, 2011), and it has been increasingly used as a model for investigations of central and peripheral auditory functions (Wang, 2000; Wang, 2007; Wang et al., 2008; Valero et al., 2008; Nelson and Young, 2010; Slee and Young, 2010; Bartlett et al., 2011; Watkins and Barbour, 2011). It has emerged as an important model for studying auditory processing of species-specific vocalizations (Brumm et al., 2004, Miller et al., 2009), pitch processing (Bendor and Wang, 2005, Bendor and Wang, 2010) and auditory–vocal interactions (Eliades and Wang, 2003, Eliades and Wang, 2005, Eliades and Wang, 2008). As a non-human primate species, the marmoset shares greater similarity with humans in structural anatomy and brain organization than do feline and rodent species commonly used in cochlear implant (CI) research. Marmosets possess a rich vocal repertoire and highly communicative nature that make them well suited to studying vocal production and mechanisms related to speech processing. Moreover, there has been an increasing body of data on marmoset auditory cortical physiology (Wang et al., 1995; Liang et al., 2002; Sadagopan and Wang, 2008; Bendor and Wang, 2008), making marmoset a valuable non-human primate model to study the cortical mechanisms involved in processing CI stimulation.

The first step in developing the marmoset as a viable animal model for CI research was to study the anatomy of the temporal bone and determine the feasibility of implanting a multi-channel intracochlear electrode into its scala tympani. A detailed three-dimensional anatomical study of the marmoset temporal bone was essential for several reasons. First, there are few published data on the dimensions of the marmoset middle and inner ear structures. The number of turns of the cochlea is reported to be 2.75 (Gray, 1907; West, 1985; Borin et al., 2008). Gray (1907) and Spoor and Zonneveld (1998) reported several semicircular canal measures, and Borin et al. (2008) offered photographic documentation and qualitative description of the micro-dissection of a marmoset temporal bone, but more specific anatomical details have not been quantitatively described. As the marmoset continues to be used as a model for auditory research, dimensions of its ear canal, middle ear ossicles, and cochlear fluid spaces will be valuable. Second, the length, area and volume measurements of the cochlear fluid spaces described here will be useful in future modeling studies related to cochlear implants in marmosets. For example, three-dimensional reconstructions can be used in finite-element models of current flow resulting from intracochlear electric stimulation (Finley et al., 1990; Frijns et al., 1995; Hanekom, 2001; Whiten, 2007). Therefore the results from the present study serve to better establish the marmoset as a model for cochlear implant research.

Several techniques can be employed to determine cochlear dimensions. Serial histological sections can be photographed and images stacked to create three-dimensional reconstructions, but cutting large numbers of serial sections and accurately aligning images is time consuming and labor intensive. Wysocki has conducted several detailed studies of the cochlea of human, cat, dog, cattle, and macaque using latex molds (Wysocki, 1999, Wysocki, 2001). In his technique, liquid latex is injected into preserved temporal bones, and casts are finely sliced and direct measurements are made. Another approach is to perform high-resolution magnetic resonance imaging scans. Thorne et al. (1999) measured cochlear fluid space dimensions for human, guinea pig, bat, rat, mouse, and gerbil using three-dimensional reconstruction of MRI data. An alternative imaging approach that is less costly is X-ray micro computed tomography (microCT), which can be done at high resolution using a temporal bone specimen for characterization of volume spaces. One particular advantage is that it can be used as well for visualization of implanted electrode location (Ketten et al., 1998). We therefore chose to use high-resolution microCT scans to reconstruct the marmoset temporal bone in three-dimensional space and characterize cochlear dimensions.