Cross-modal plasticity following peripheral sensory loss enables deprived cortex to provide enhanced abilities in remaining sensory systems. These functional adaptations have been demonstrated in cat auditory cortex following early-onset deafness in electrophysiological and psychophysical studies. However, little information is available concerning any accompanying structural compensations. To examine the influence of sound experience on areal cartography, auditory cytoarchitecture was examined in hearing cats, early-deaf cats, and cats with late-onset deafness. Cats were deafened shortly after hearing onset or in adulthood. Cerebral cytoarchitecture was revealed immunohistochemically using SMI-32, a monoclonal antibody used to distinguish auditory areas in many species. Auditory areas were delineated in coronal sections and their volumes measured. Staining profiles observed in hearing cats were conserved in early- and late-deaf cats. In all deaf cats, dorsal auditory areas were the most mutable. Early-deaf cats showed further modifications, with significant expansions in second auditory cortex and ventral auditory field. Borders between dorsal auditory areas and adjacent visual and somatosensory areas were shifted ventrally, suggesting expanded visual and somatosensory cortical representation. Overall, this study shows the influence of acoustic experience in cortical development, and suggests that the age of auditory deprivation may significantly affect auditory areal cartography.
Cross-modal plasticity allows the developing brain to adapt to the loss of a sensory modality by maximizing the performance of the remaining senses. In deaf individuals, other senses, such as vision, become imperative for daily interaction with the environment. Furthermore, deaf individuals often show greater detection and localization of visual stimuli when compared with hearing subjects (Reynolds 1993; Bottari et al. 2011). These superior visual abilities in deaf individuals have been shown to involve auditory cortex activation (Neville and Lawson 1987; Nishimura et al. 1999; Finney et al. 2001). However, the ability of the brain to rearrange cortical inputs may depend on when sensory deprivation occurred, an observation consistent with classical theories of critical developmental periods (Xerri 2008). When presented with a moving visual stimulus, prelingually deaf subjects exhibit greater cortical activity in visual processing areas than postlingually deaf subjects (Buckley and Tobey 2011). Limited efficacy of cochlear implants in subjects with prelingual deafness has been attributed to cross-modal reorganization of the early deprived auditory cortex. This reorganization is inferred to be absent from the postlingually deaf who lose their hearing outside their critical period of auditory development (Sharma et al. 2009; Teoh et al. 2004; Buckley and Tobey 2011). Various experimental animal models have provided further insight into the limitations to plasticity throughout development. As pioneers in the field of sensory processing, Wiesel and Hubel (1963) first established the basic concept of critical periods limiting cortical plasticity in blind cats. Monocularly deprived cats were able to resume normal development of visual ocular dominance columns if sensory deprivation was induced at birth, but not if induced in adulthood. Similarly, Rebillard et al. (1977) found the primary auditory cortex (A1) to be visually responsive only if deafness occurred during the first postnatal week in cats, and was unresponsive if deafness was induced 2 months after birth.
However, there is increasing functional evidence that plasticity persists into adulthood, permitting cortical reorganization if faced with sensory deprivation at a late age. Both early- and late-blind adults show recruitment of visual areas underlying superior sound detection abilities compared with hearing controls (Kujala et al. 1997; Voss et al. 2004; Collignon et al. 2009). In the auditory system, cortical areas of the adult deafened ferret were recently shown to be responsive to somatosensory stimulation (Allman et al. 2009). Park et al. (2010) reported compensatory metabolic hyperactivity of the visual cortex and a resurgence in basal metabolic activity in A1 33 months after deafening in adult cats. However, with most studies focussing on congenitally or early-deaf subjects, cross-modal adaptability of adult auditory cortex remains under debate. Multisensory processing has been shown to exist in the adult cat within the auditory field of the anterior ectosylvian sulcus (fAES; Meredith and Allman 2009) and the rostral suprasylvian cortex (Clemo et al. 2007), and multiple auditory areas in ferrets (Bizley and King 2009). Therefore, it seems possible that a multisensory neuron within an auditory area devoid of auditory input could process information from other modalities more efficiently in an adult animal. A direct comparison of cartographical changes following early- versus late-onset deafness may provide insight into the functional reorganization that occurs, but little anatomical information about such compensation is available as yet.
Similarities in cerebral organization substantiate the cat cerebral cortex as an excellent experimental model of the human cerebrum (Payne and Lomber 2002). The cartography of the cat auditory cortex has been extensively examined electrophysiologically, anatomically, and behaviorally. At least 10 discrete areas of auditory cortex have been defined by previous assessments of tonotopy (Knight 1977; Imig and Reale 1980; Reale and Imig 1980; Merzenich et al. 1975; Tian and Rauschecker 1998), anatomical connectivity (Clasca et al. 1997; Lee and Winer 2008a, 2008b), and deactivation of specific loci (Read et al. 2002; Malhotra and Lomber 2007) in hearing cats (Fig. 1). Four tonotopic areas with distinct thalamocortical afferents (Morel and Imig 1987; Lee and Winer 2008a) span across cat auditory cortex, with the anterior auditory field (AAF) at the most dorso-anterior limit. A1 lies on the middle ectosylvian gyrus, dorsal to the posterior auditory field (PAF). On the posterior ectosylvian gyrus, the ventral PAF (VPAF) lies ventral to PAF. fAES, the second auditory cortex (A2), and ventral auditory field (VAF) are 3 nontonotopic areas ventral to AAF and A1. A fourth nontonotopic area, the dorsal zone (DZ), lies at the dorsal limit of the auditory cortex. With contributions to auditory localization (Strominger 1969), the limbic temporal zone (T) and associative insular cortex (IN) comprise the ventral limits of auditory cortex.
Electrophysiological recordings and psychophysical studies in cats with early-onset or congenital deafness have localized visual improvements to discrete auditory areas. In cats with early-onset or congenital deafness, PAF has been shown to play a role in visual localization, while DZ contributes to enhanced visual motion detection (Lomber et al. 2010). Neurons in fAES can be activated by visual stimuli (Meredith et al. 2011), and AAF has also been shown to respond to both visual and somatosensory stimuli (Meredith and Lomber 2011). However, cross-modal plasticity does not seem to be a generalized feature of the entire auditory cortex. Primary auditory cortex (A1) shows little (Rebillard et al. 1977) or no response (Kral et al. 2003) to nonauditory stimuli and has been described as lacking cross-modal plasticity (Stewart and Starr 1970; Kral et al. 2003; Lomber et al. 2010).
|A1—primary auditory cortex|
|A2—second auditory cortex|
|AAF—anterior auditory field|
|DZ—dorsal zone of auditory cortex|
|fAES—auditory field of anterior ectosylvian sulcus|
|PAF—posterior auditory field|
|VAF—ventral auditory field|
|VPAF—ventral posterior auditory field|
|A1—primary auditory cortex|
|A2—second auditory cortex|
|AAF—anterior auditory field|
|DZ—dorsal zone of auditory cortex|
|fAES—auditory field of anterior ectosylvian sulcus|
|PAF—posterior auditory field|
|VAF—ventral auditory field|
|VPAF—ventral posterior auditory field|
As anatomical organization has direct implications on cortical functionality (Schreiner and Winer 2007), an examination of auditory areal cartography provides insight into the limitations of cross-modal reorganization. Histochemical staining can be used to reveal cytoarchitectonic differences, which can differentiate discrete anatomical areas of the cerebral cortex (Rosenquist 1985). The monoclonal antibody SMI-32 (Sternberger Monoclonal, Inc.) is one tool that can be used to describe the cytoarchitecture of the auditory cortex. As a selective marker of nonphosphorylated epitopes on high- and medium-molecular-weight subunits of neurofilament proteins, SMI-32 primarily labels type I neocortical pyramidal neurons with subcortical axonal projections (Sternberger and Sternberger 1983; Campbell and Morrison 1989; Voelker et al. 2004). SMI-32 has been used to demarcate cortical areas in cats (Van der Gucht et al. 2001; Lee and Winer 2008a, 2008b; Mellott et al. 2010), rats (Ouda et al. 2012), monkeys (Chaudhuri et al. 1996; Luppino et al. 2005; Soares et al. 2008; Katsuyama et al. 2010), and humans (Hof et al. 1992; Ding and Van Hoesen 2010). Mellott et al. (2010) recently used SMI-32 to delineate 10 areas of auditory cortex in hearing cats. Each area was characterized by its unique distribution of SMI-32-labeled neurons across the 6 cortical layers. These anatomical boundaries were shown to coincide with functional divisions revealed by electrophysiological studies, demonstrating the validity of SMI-32 immunoreactivity for examining auditory areal cartography in hearing cats (Mellott et al. 2010).
Using the monoclonal antibody SMI-32, the aim of this study was to examine cartographical changes in the auditory cortex of cats deafened early in neonatal development (designated as early-deaf) and cats deafened in adulthood (designated as late-deaf). All 10 previously described regions within the auditory cortex were examined after immunohistochemical processing. With documented reallocation to visual processing following early deafness, PAF, DZ, fAES, and AAF were hypothesized to show expansion in early-deaf cats. With its lack of cross-modality, A1 was hypothesized to have reduced volume in early-deaf cats. All areas of auditory cortex in late-deaf cats were hypothesized to show limited changes in volume.
All 10 auditory areas were successfully delineated based on previously established SMI-32-staining profiles in all cats. This conservation of SMI-32 immunoreactivity demonstrates the utility of SMI-32 staining for determining auditory cortex cartography in hearing cats, and cats with early- or late-onset deafness. Overall, the volume of A1 was significantly reduced in both early- and late-deaf cats. Interestingly, while PAF, DZ, and AAF did not show any significant alterations in volume, early-deaf cats showed further reductions in fAES, and expansions in A2 and VAF. Dorsal auditory areas appeared to be the most mutable, with borders between auditory areas and adjacent visual and somatosensory areas shifted ventrally in early-deaf cats. These results demonstrate the influence of acoustic experience on shaping areal cartography.
Materials and Methods
Auditory areal cartography was examined in the left auditory cortex of 15 adult domestic cats (>5 months; Fig. 2). Cats were housed in colonies to promote social interaction and had free access to water, food, and various enrichment devices, including toys, perches, and scratching posts. The hearing control group consisted of 5 mature female cats (>5 months) obtained from a commercial laboratory animal breeding facility (Liberty Labs, Waverly, NY, USA). Ten deaf cats originating from 6 different litters were born in-house from pregnant dams acquired from Liberty Labs (Waverly, NY, USA). Five cats were ototoxically deafened around the time of hearing onset to constitute the early-deaf group (Fig. 2). Five cats were similarly deafened in adulthood (>5 months) to constitute the late-deaf group (Fig. 2). The early-deaf and late-deaf groups were each comprised of 3 males and 2 females. At least 6 months following deafening, cats were perfused and auditory cortex cytoarchitecture was revealed immunohistochemically with SMI-32. Mean body weights at the time of perfusion did not differ among groups, with average weights of 3.3 ± 0.1, 3.8 ± 0.3, and 4.4 ± 0.5 kg in hearing, early-deaf, and late-deaf cats, respectively. Ten auditory areas were delineated in coronal sections under light microscopy. Layer-specific-labeling densities, the volume of total auditory cortex examined, fractional areal volumes, and areal border positions were compared among the 3 experimental groups. All animals were concurrently used for cortical connectivity studies using biotinylated dextran amine (BDA) tracer methods (Chabot et al. 2012; Kok et al. 2012). All procedures were approved by the University of Western Ontario Animal Use Subcommittee of the University Council on Animal Care and were conducted in accordance with the Canadian Council on Animal Care's Guide to the Care and Use of Experimental Animals (Olfert et al. 1993).
Five newborn cats were given bilateral ototoxic cochlear lesions around the time of hearing onset (14 days postnatal; Shipley et al. 1980) or when the cat was at least 300 g (to support anesthesia and apparatus). For each cat, general anesthesia was induced by inhalation of 4% isofluorane, administered through a mask. An intravenous catheter was inserted in the jugular vein at a shaved and betadine/alcohol scrubbed section of the neck. Auditory brainstem responses (ABRs) were recorded in each animal in the following manner. Ear inserts were placed deep into each ear canal. Sub-dermal electroencephalography (EEG) recording leads were placed in a standard vertex-to-mastoid configuration. Speakers were positioned in front of each ear to deliver auditory stimuli (1–80 dB, 0.1 ms squarewave clicks). Auditory-evoked electrical activity of the brain was detected by the EEG leads and routed to a computer for amplification, signal averaging, and data storage (Nicolet Spirit, 2000 Evoked Potential System, Nicolet Biomedical, Inc., Madison, WI, USA). A baseline ABR was performed before drug administration. To deafen each cat, a loop diuretic, sodium edecrine (40 mg/kg), was infused intravenously at a rate of 20 mg/min. At the onset of diuretic infusion, kanamycin was injected (300 mg/kg, s.c.) to selectively destroy hair cells of the cochlea (Xu et al. 1993). As the drugs were administered, an ABR was recorded to assess hearing function. According to hearing thresholds defined by the World Health Organization (1998), a hearing deficit >80 dB is classified as profound impairment including deafness. Thus when a hearing deficit (flat ABR) at 80 dB was obtained (Fig. 3), the animal was designated as deaf. Diuretic infusion was terminated and lactated ringers were administered intravenously. An additional ABR was performed immediately after drug administration was completed to confirm deafness. The catheter was then removed and the cat was allowed to recover from anesthesia before returning to its mother. A follow-up ABR was performed 3 months later to confirm deafness. Early-deaf cats were allowed to mature and reach adulthood before subsequent tissue analysis (Fig. 2). Five adult cats (>5 months) were similarly deafened to constitute the late-deaf group (Fig. 2). All 5 late-deaf cats demonstrated hearing deficits of >80 dB with ABR waveforms similar to all early-deaf cats (Fig. 3B). ABRs were also performed on 5 adult intact cats to ensure hearing in the control population. All 5 hearing cats demonstrated normal ABR waveforms, with typical 5 vertex morphology, and decreasing peak magnitude and increasing interwave latencies with decreasing sound intensity (Kraus and McGee 1992; Figure 3A).
At least 6 months following deafening, cats were perfused for subsequent SMI-32 processing of auditory cortex. Each cat was anesthetized with ketamine (4 mg/kg, i.m.) and dexdomitor (0.03 mg/kg, i.m.) the afternoon before perfusion to insert a catheter into the cephalic vein for drug administration. On the following day, sodium pentobarbital (40 mg/kg, i.v.) was administered to induce general anesthesia. Heparin (an anti-coagulant) and 1% sodium nitrite (a vasodilator) were co-administered intravenously. For exsanguination, each animal was perfused transcardially with physiological saline (1 L), followed by 4% paraformaldehyde (2 L), and finally 10% sucrose (2 L), all at a rate of 100 mL/min. The brain was exposed and the head was placed in a stereotaxic apparatus. To ensure correct orientation for sectioning and inclusion of the entire auditory cortex, the brain was blocked in the coronal plane at Horsley and Clark (1908) level A23, before removal from the cranium. To cryoprotect the tissue, each brain was placed in 30% sucrose until it sank. The brain was then cut in 60 µm coronal sections and collected serially. Sections from the first of 6 series, at 360 µm intervals, were processed for SMI-32 immunoreactivity.
Sections were rinsed with 0.1 M phosphate buffer (PB). Endogenous peroxidase was blocked with 0.5% hydrogen peroxide treatment for 30 min, after which sections were rinsed with 0.1 M PB. Free-floating sections were incubated in 1 mL 5% normal goat serum (NGS) for 45 min. Sections were then incubated overnight in primary antibody (mouse-anti-SMI-32; 1/2000; Vector Laboratories, Burlingame, CA, USA) at 4°C with 2% NGS. The following morning, tissue was rinsed with 0.1 M PB before incubation in biotinylated secondary antibody (goat anti-mouse IgG; 1/200; Vector Laboratories) for 30 min with 2% NGS. After rinsing with 0.1 M PB, sections were incubated with an avidin–biotin–horseradish peroxidase solution (Vectastain Elite ABC, Vector Laboratories) for 1.5 h. Sections were rinsed with 0.1 M PB, before incubation with a diaminobenzene–nickel chromogen solution for 15 min. After a final rinse with 0.1 M PB, sections were mounted on gelatinized slides, cleared, and coverslipped.
Of the remaining 5 tissue series, series 2 and 3 were stained with Cresyl violet and cytochrome oxidase, respectively, to assist in visualization of cytoarchitecture. Series 4 was processed for visualization of the anterograde and retrograde tracer BDA. Series 5 and 6 were retained as spare series and processed with any of the above methods as needed.
The left auditory cortex was identified by superficial anatomical landmarks in coronal sections. The anterior and dorsal limits of auditory cortex were bound by the rostral and middle suprasylvian sulcus (SSS), respectively. The posterior ectosylvian sulcus was used to localize the posterior limits of auditory cortex, located on the posterior ectosylvian gyrus. Sections containing left auditory cortex were analyzed under a Nikon E600 microscope (Nikon, Melville, New York, USA) with a ×10 objective and a ×10 magnifier. Investigators were blind with respect to which experimental group a subject belonged to during tissue examination. To examine the entire depth of a section, numerous focal levels were taken throughout the z-plane. Neurolucida software (MBF Bioscience, Inc., Williston, VT) was used to identify and quantify SMI-32-labeled neurons. To prevent misidentification of reaction artefact as a neuron, the presence of the nucleus and entire soma membrane was required to count a pyramidal neuron. Ten auditory areas were delineated according to areal profiles and boundaries defined by Mellott et al. (2010). Within the anterior ectosylvian sulcus (AES), auditory areas were distinguished from bordering visual areas by previously described staining profiles (Van der Gucht et al. 2001). Adjacent somatosensory areas were discernible by the greater intensity of SMI-32 immunostaining and higher density of labeled somata, common to all cases (Van der Gucht et al. 2001).
Areal volumes were determined using NeuroExplorer software (MBF Bioscience, Inc.). To account for tissue shrinkage due to immunohistochemical processing, cortical thickness was measured by uniform random sampling with a constant sampling fraction in all animals under ×1000 magnification using an oil immersion objective (Dorph-Petersen et al. 2001). For an auditory area within an individual section, cortical thickness was determined from the difference between the Z focus levels when focussed at the top and bottom surfaces of the tissue. For each case, the mean cortical thickness was determined from all measured thicknesses across examined sections. The shrinkage correction factor was calculated as cut thickness of 60 µm over the average measured thickness for each case. The correction factor was then multiplied by the measured areal volumes to obtain corrected volume measures.
To ensure delineation criteria for each auditory area was consistent across all 3 experimental groups and to assess the possibility of deafness-induced modifications to SMI-32 immunoreactivity, cortical-layer-specific somata counts were determined per unit corrected volume. Layer-specific labeling density was calculated using the following formula:
Borders between the 10 auditory areas, and between auditory areas and adjacent visual or somatosensory areas appeared to have shifted substantially in the deaf cats. Borders were mapped relative to the fundus of the SSS (Fig. 11B). To quantify and compare border position, SSS depth was measured from the fundus to a tangent drawn at the crest of the suprasylvian gyrus. A line parallel to this tangent was drawn perpendicular to the sulcus, through the 6 cortical layers. Measurements to the borders of interest were made from this starting line along cortical layer IV. Absolute distances were normalized to the SSS depth measured for the given section to account for inter-subject variations in sulcus depth and tissue shrinkage due to processing. Borders between auditory areas, and between auditory areas and adjacent visual or somatosensory areas were measured repeatedly at sections throughout the anterior–posterior plane, corresponding to positions A3 though A10 (Snider and Niemer 1961).
Due to the repeated sampling of border positions, neuronal labeling and volumes, one-way ANOVAs with repeated-measures were used to compare normalized border measurements, layer-specific-labeling densities, fractional areal volumes, and volumes of total auditory cortex examined among the 3 groups. Mauchly's test of sphericity confirmed that the variances between groups were equal, validating the repeated-measures ANOVA, and permitting post hoc Bonferroni tests.
SMI-32 staining analysis was used to examine cartographical changes in cat auditory cortex as a result of induced deafness. Here, we report on the alterations in areal cartography revealed immunohistochemically following early- or late-onset deafness.
General Features of SMI-32 Immunohistochemistry
Previously described SMI-32 staining profiles for the 10 auditory areas were present in all 5 hearing animals and conserved in both early- and late-deaf cats. Upon qualitative assessment, relative stain intensity and cortical layer immunoreactivity were consistent for each auditory area across the 3 experimental groups (Fig. 4). Layer-specific-labeling densities did not differ among groups for any of the 10 auditory areas, demonstrating consistent delineation criteria used in all cases (Figs 5 and 6).
In all hearing and deaf animals, there was a consistent lack of labeling in layers I and IV across all areas (Fig. 4). As a result, these layers were omitted from subsequent labeling-density analyses. Overall, there was a dorso-anterior to ventro-posterior staining gradient, such that areas more dorsal and/or anterior tended to be more darkly stained, while areas more ventral and/or posterior were lightly stained. As the most dorsal and anterior area, AAF was the most immunoreactive region in hearing cats, with layers III, V and VI darkly stained (Fig. 4B), and high labeling densities within the same layers (Fig. 5A). Similar labeling patterns were identified in early- and late-deaf cats. Within the anterior ectoslvian gyrus, AAF was discernible from fAES with a wide band of immunoreactivity evident even under weak magnification of the latter. This was due to the widening of supragranular layers (Fig. 4B) and high labeling density of fAES layer III (Fig. 5C) evident across all experimental groups. Ventro-posteriorly to AAF and fAES, over areas spanning most of the auditory cortex, the staining gradient was especially evident in DZ, A1, and A2. Staining intensity within these 3 areas gradually lightened in increasingly posterior sections. In all cats, layer III was strongly immunoreactive in DZ, fAES, and A2 (Fig. 4B, C, E) and was used to delineate the boundary of each region with AI, given the lower layer III-labeling density of AI (Fig. 5D). SMI-32 immunoreactivity continued to decline in more posterior sections, with only moderate labeling in the VAF and PAF (Fig. 4D). The posterior ectosylvian sulcus was used to help localize the VAF/PAF border. VAF was ultimately defined by lower layer VI labeling density (Fig. 5F), consistent in all cats. VPAF was weakly stained (Fig. 4D), and was distinguishable from PAF in hearing cats due to somata labeling restricted to layer V of the former (Fig. 6B). The same criteria were used to delineate PAF and VPAF in early- and late-deaf cats. Ventrally, the limbic and association areas were weakly stained and easily discernible from each other due to the extremely light immunoreactivity of the temporal cortex (T) in all cats (Fig. 4E). Furthermore, the moderate to low labeling density of layer III in the IN (Fig. 6C), compared with the complete lack of layer III labeling in T (Fig. 6D), was used to distinguish between these 2 distinct areas in the ventral regions of the auditory cortex.
Comparisons of Total Auditory Cortex Volumes
The mean total volume of examined auditory cortex did not differ significantly among hearing, early-deaf, and late-deaf cats (Fig. 7). Because only one of the 6 series was assessed for SMI-32 immunoreactivity, areas delineated in consecutive coronal sections within the series were used to estimate the volume of total auditory cortex examined. In early-deaf cats, the mean total auditory cortex volume was reduced to 71% of control levels (684.4 ± 99.3 mm3), but did not differ significantly from hearing or late-deaf cats with mean volumes of 962.5 ± 79.2 and 868.1 ± 72.9 mm3, respectively (Fig. 7). Total volumes varied considerably within all groups ranging from 755.7 to 1182 mm3 in early-deaf cats, and from 326 to 904 mm3 in hearing cats. Late-deaf volumes were more similar to hearing values, ranging from 663 to 1088 mm3. Overall, deafness did not appear to induce any significant changes in the total volume of examined auditory cortex. However, there was a positive correlation between the age at deafening and total auditory cortex volume in early-deaf cats, such that earlier acoustic deprivation resulted in greater reductions in total volume (Fig. 8A). There was no correlation between the total auditory cortex volume and age at deafening in late-deaf cats (Fig. 8B).
Early- and Late-Deaf Cats Showed Reduced A1
The mean fractional areal volumes of total auditory cortex examined were subsequently compared among the 3 experimental groups (Fig. 9). In hearing cats, A1 was one of the largest areas of the auditory cortex, with a mean fractional volume of 14.8 ± 1.0%. However, following early-onset deafness, the mean fractional volume of A1 was significantly reduced to 11.8 ± 0.3% (P < 0.05). Even greater reductions were observed in late-deaf cats, with a mean fractional volume of 10.3 ± 0.4% (P < 0.001).
Expansion of A2 and Reduction of fAES in Early-Deaf Cats
In early-deaf cats, the mean fractional A2 volume of 26.7 ± 1.5% was significantly higher than the fractional volume of 21.0 ± 1.3% in hearing cats (P < 0.05; Fig. 9). Late-deaf cats did not differ significantly from controls, with a mean fractional A2 volume of 21.8 ± 0.9%. Early-deaf cats also exclusively showed significant reductions in fAES volumes, with a mean value of 1.4 ± 0.4% being significantly lower (P < 0.05) than mean volumes of 6.3 ± 0.7 and 3.9 ± 1.1% in hearing and late-deaf cats, respectively. In early-deaf cats, borders between fAES and the adjacent anterior ectosylvian visual area (AEV) were shifted dorsally within the AES (Fig. 10B), though did not differ significantly from hearing controls (Fig. 11I).
The mean fractional volume of AAF did not differ significantly among groups, with mean volumes of 13.0 ± 1.0, 12.1 ± 1.1, and 15.8 ± 1.7% in hearing, early-deaf, and late-deaf cats, respectively (Fig. 9). Despite the maintenance of control volumes, borders between dorso-anterior somatosensory areas and AAF were shifted ventrally (Fig. 10B), with a significantly greater mean normalized border position in early-deaf cats (Fig. 11H). Similarly, while fractional DZ volumes did not differ among groups, borders between DZ and adjacent dorsal visual areas were likewise shifted ventrally in early-deaf cats (Fig. 10B), and had significantly greater mean normalized border positions relative to the SSS (Fig. 11C). In total, dorso-anterior regions demonstrated extensive cartographical changes following early- or late-onset deafness.
Ventro-Posterior Areas Were Minimally Affected in Early- and Late-deaf Cats
PAF was significantly expanded in late-deaf cats, with a mean fractional volume of 11.3 ± 1.2% (Fig. 9; P < 0.05). In contrast, PAF had mean values of 5.6 ± 1.1 and 8.6 ± 1.3% in hearing and early-deaf cats. Accordingly, the A1/PAF border was significantly shifted dorsally in late-deaf cats only (Fig. 11E; P < 0.05). While no difference was observed in late-deaf cats, the mean fractional volume of VAF was significantly higher in early-deaf cats, rising to 11.2 ± 1.2% over the mean value of 6.1 ± 1.1% in hearing animals (Fig. 9; P < 0.05). Similarly, the A1/VAF border was significantly shifted dorsally in early-deaf cats only (Fig. 11F; P < 0.05). The mean fractional volumes of VPAF and limbic and association areas T and IN were unchanged following either early- or late-onset deafness. Therefore, ventral regions were generally less affected than dorso-anterior regions following deafness.
These results demonstrate the distinct effects of deafness on shaping auditory cortex cartography, revealed with the SMI-32 antibody. Hearing cats showed typical SMI-32 immunoreactivity, with areal profiles corresponding to previously published descriptions (Mellott et al. 2010). Furthermore, SMI-32 staining profiles of the 10 auditory areas were evident in both early- and late-deaf cats. Although deafness did not produce significant changes in the volume of total auditory cortex examined, A1 was significantly reduced in all deaf animals. Late-deaf cats showed an additional expansion in the volume of PAF, which corresponded with a significant dorsal shift in the A1/PAF border. In early-deaf cats, dorso-anterior auditory areas appear to be most mutable, given the dorsally shifted borders between A1 and VAF, and ventrally shifted medial borders of AAF and DZ. Early-deaf cats showed further reductions in fAES volumes, and expansions in A2 and VAF. Overall, this study suggests that acoustic experience plays a critical role in auditory cortex development, and that the age at which auditory deprivation occurs appears to significantly affect areal cartography.
Staining Profiles Observed in Hearing Cats Were Conserved in All Deaf Cats
Qualitative and quantitative analyses found that staining profiles of the 10 auditory areas described previously in hearing cats (Lee and Winer 2008b; Mellott et al. 2010) were conserved in cats with early- or late-onset deafness. Similar to previous studies employing SMI-32 for cortical delineation in cats (Mellott et al. 2010) and monkeys (Geyer et al. 2000; Katsuyama et al. 2010), interindividual variation was evident, with density values covering a considerable range within and across all groups. However, within each case, staining profiles were distinct between areas, demonstrating that cortical SMI-32 reactivity can provide consistent delineation criteria for auditory areas in hearing and deaf cats. Interestingly, SMI-32 has been suggested as a marker of various pathological processes (Ouda et al. 2012), with reduced SMI-32 positivity in the lateral geniculate nucleus in cats with early visual deprivation, indicative of axonal disintegration (Bickford et al. 1998; Duffy and Slusar 2009). However, we did not observe any change in immunoreactivity in the auditory cortex following deafness, as labeling densities within a specific layer of a given area did not differ significantly among groups. The normal cytoarchitecture observed in the present study suggests that cats with early- or late-onset deafness adapted to auditory deprivation, rather than having succumbed to pathological damage.
Alterations in Areal Volumes and Borders in Early- and Late-Deaf Cats
Because the volumes of total auditory cortex examined were not significantly reduced in early- or late-deaf cases, deafness did not seem to induce a singular, global response on auditory cortical representation. However, significant areal modifications in terms of size and position in more dorso-anterior regions suggest that certain auditory areas may be more labile following early or late sensory deprivation. In early-deaf cats, borders between dorso-anterior auditory areas and adjacent visual or somatosensory regions were significantly shifted ventrally, suggesting expanded visual and somatosensory systems. The reallocation of an auditory area devoid of acoustic stimulation to process other sensory information would be a sensible and efficient use of cortical space, and would complement functional reorganization of auditory areas observed in previous studies of cross-modal plasticity. Thus cross-modal reorganization may allow for a sensory area to be anatomically maintained in the deprived cortex. Differentially affected fractional volumes may suggest that some areas are more cross-modally reorganized than others. This can be shown with the sustained AAF and DZ volumes, which complement previous findings of visual and somatosensory responsitivity (Meredith and Lomber 2011) and contributions to visual motion detection (Lomber et al. 2010), respectively, in early-deaf cats. In contrast, feline A1 has been found to have limited or no cross-modal plasticity from electrophysiological recordings (Stewart and Starr 1970; Kral et al. 2003) and psychophysical studies (Lomber et al. 2010). In this present study, reduced fractional A1 volume and shifted borders between A1 and adjacent auditory areas demonstrate anatomical consequences of deafening that may complement the previously reported lack of functional cross-modal reorganization.
Reduction in A1 volume appeared complementary to significant expansions in A2 and VAF, and dorsally shifted A1/VAF border in early-deaf cats. This apparent adaptability seems to be characteristic of A2, which has been suggested to possess greater adaptability to environmental influences based on variability in multiple functional parameters (Schreiner and Cynader 1984). However, the exact role A2 plays in acoustical processing is unknown in hearing cats. Simple stimuli used to functionally map and study responsitivity in primary areas have been ineffective for evaluating A2 (Schreiner and Winer 2007). Similarly, there have been limited functional assessments of VAF due to its ventral location that can be problematic for electrode placement. Thus while we report anatomical changes of A2 and VAF following early-deafness, further studies in both hearing and deaf cats will be required to elucidate the functional contributions of these cortical expansions.
In contrast, “deaf” fAES has been shown to play a role in contralateral visual orientation (Meredith et al. 2011). Despite its known cross-modal capacity, the reduced volume observed in early-deaf cats does not support the previous idea that cross-modal reorganization preserves the cortical space of a deprived area. Additionally, the dorsally shifted border between fAES and AEV observed in early-deaf cats suggests an anatomical expansion of AEV into fAES. In their evaluation of the cross-modal properties of fAES, Meredith et al. (2011) were careful to distinguish between 2 distinct results of plasticity. Compensatory hypertrophy describes the expansion or increased activity of visual and/or somatosensory cortices, and has been documented in the visual cortex in deaf humans (Neville and Lawson 1987) and cats (Park et al. 2010). Furthermore, Rauschecker and Korte (1993) suggest that this expansion of the nondeprived sensory modality may occur at the expense of the deprived modality, particularly in regions where 2 separate modalities directly border one another. In contrast, cross-modal plasticity refers to the processing of nonauditory information by the auditory cortex. While auditory inputs into fAES control auditory-orienting behavior in hearing cats, visual inputs into fAES maintain similar orienting behavior to visual stimuli in early-deaf cats (Meredith et al. 2011). Importantly, cooling AEV, the visual cortical area adjacent to fAES, does not produce visual-orienting deficits, whereas cooling cross-modally reorganized fAES in early-deaf cats does. Therefore, the authors argue that this unaltered output represents cross-modal reorganization and not expansion of AEV into fAES. Similarly, Fine et al. (2005) found clear evidence of cross-modal reorganization of the auditory cortex, and failed to observe volumetric changes in visual areas through functional magnetic resonance imaging (MRI) of deaf human subjects.
The exact circumstances or properties that would lead a deprived sensory area to reorganize to receive input from another modality, or atrophy in response to the expansion of another sensory area remains to be elucidated. While the inherent properties of a given area may govern its adaptability, interactions with surrounding areas likely have direct implications for cortical reorganization. In early-deaf cats, the dorsally shifted fAES/AEV border and ventrally shifted AAF and DZ appear to permit visual and somatosensory expansion. AAF and DZ did not demonstrate any volumetric reduction; instead, they appeared to have shifted ventrally into A1, a region that has shown to have less plasticity. In contrast, fAES borders AAF and A2, which were maintained or expanded, respectively, in early-deaf auditory cortex. Thus compensatory hypertrophy of adjacent visual area AEV may have encroached into fAES, which could not move posteriorly into A2 or dorsally into AAF and was therefore diminished in volume. Thus expansion of preserved sensory cortices may be the predominant response, while cross-modal plasticity may work to preserve volumetric representation. The combination of these responses may be most advantageous following sensory loss. Therefore, from our data, we suggest that it is plausible that the present anatomical findings and previous functional observations are not mutually exclusive, as “deaf” fAES may be cross-modally reorganized and have a lower volume than that of a “hearing” fAES. Park et al. (2010) also reported the coexistence of compensatory hypertrophy in the visual cortex and cross-modally reorganized auditory cortex through micro-positron emission tomography techniques. Additional examination of visual and somatosensory cortices in our hearing and deaf animals may provide insight into the compensatory reorganization of sensory cortices following deafness. Future studies combining electrophysiological recording techniques and cartographical assessments are required to thoroughly understand the cortical effects of auditory deprivation.
As expected, auditory cortex cartography was significantly modified in early-deaf cats. However, late-deaf cats also showed pronounced alterations to overall auditory cortex organization. PAF was significantly expanded and shifted dorsally into A1, contributing to the reduction in A1 volume. Similar findings were observed in human patients with adult hearing loss who showed significant reductions in gray matter volume in primary auditory cortex (Peelle et al. 2011). Thus the auditory cortex of late-deaf cats did not appear to have restricted adaptability due to the existence of a critical period limiting plasticity, as found in prior experiments (Kral et al. 2001). Our observations of border movement and volumetric changes in A1 and PAF following late-onset deafness contribute to mounting evidence for persisting cross-modal plasticity in auditory cortices of adult cats (Rebillard et al. 1977; Park et al. 2010) and ferrets (Allman et al. 2009), and in visual cortices of adult humans (Kujala et al. 1997) and mice (Van Brussel et al. 2011). Similarly, Leporé et al. (2010) saw significant changes in cortical structures and white matter connectivity in early blind human subjects, but found less widespread changes in those with late-onset blindness. Here we provide further structural evidence of reduced, yet persisting, plasticity in adulthood, as late-deaf cats show modified auditory cortical representation, albeit fewer alterations than in early-deaf cats.
Areal Modifications Following Sensory Loss in Other Modalities
Similarities in sensory system development allow the findings from various unimodal deprivation models to be pooled to generate a more thorough understanding of the adaptability of the mammalian brain. The blind visual system has extensively been studied, with similar anatomical approaches to examine the organization of visual areas. Myelin-stained sections from neonatally enucleated opossums revealed reduced primary visual cortex (V1), and enhanced primary somatosensory representation in the neocortex (Karlen and Krubitzer 2009). Blind monkeys subjected to in utero bilateral enucleation also showed reduced V1 in Nissl-stained sections (Dehay et al. 1991). Unlike our observed reductions in A1 of early- and late-deaf cats, Dehay et al. (1991) observed more severe areal reductions following enucleation performed earlier rather than later in embryonic development. In the current study, late-deaf cats showed greater reductions in A1 volume, suggesting a possible difference in cortical adaptability following auditory or visual deprivation. Furthermore, Rakic et al. (1991) identified a novel cytoarchitechtonic area within the visual cortex in monkeys subjected to in utero bilateral enucleation. Similarly, Nissl and myelin staining revealed the presence of a new visual cortical area in bilaterally enucleated opossums (Kahn and Krubitzer 2002). However, the acoustically deprived feline auditory cortex does not appear to adapt with a similar response, as every area encountered matched one of the ten established SMI-32-staining profiles.
SMI-32 as Tool for Mapping Auditory Cortex
SMI-32-staining profiles were used to successfully identify 10 auditory areas in all 15 animals, demonstrating that neither early- nor late-onset deafness resulted in the loss or formation of novel SMI-32 immunoreactive auditory areas. This conservation of immunoreactivity ultimately shows the utility of SMI-32 in the objective confirmation of areal boundaries in hearing, early- and late-deaf cats for future electrophysiological, behavioral, and imaging studies. SMI-32 processing can be conducted to verify an accurate placement of recording electrodes or other experimental implements over a specific auditory area of interest. Imaging techniques such as positron emission tomography and functional MRI techniques can be limited in anatomical precision (Bavelier and Neville 2002), with activation in primary cortices indistinct from activation in higher cortices (Rademacher et al. 1993). Moreover, the use of reference atlases based on healthy, hearing individuals to locate auditory areas in deaf human subjects (Lambertz et al. 2005; Auer et al. 2007; Karns et al. 2012) may be inadequate given the changes in areal organization observed in deaf animals. Therefore, deafness-induced modifications to auditory cortex cartography must be considered in future imaging studies on human subjects.
Future Studies to Evaluate Mechanisms of Cross-Modal Plasticity
Conservation of staining profiles in deaf animals suggests that areal neuronal circuitry was unaffected following deafness, maintaining the output function of a given area. Furthermore, tissue series from early- and late-deaf cats stained with Cresyl violet or cytochrome oxidase also appeared normal and similar to sections from hearing cats. Consistent labeling densities in all areas among all 3 groups indicated that volumetric changes could be attributed to changes in the number of columns within an area, rather than changes in neuron number, size, or cortical thickness. This supports previous work attributing enhanced visual and/or somatosensory abilities to the switch in input modality while maintaining the established behavioral program for a deaf auditory area (Lomber et al. 2010). Such changes in anatomical connectivity are known to have direct implications for governing cortex cartography (Schreiner and Winer 2007). However, the exact mechanisms responsible for mediating cross-modal connectivity remain elusive (Bavelier and Neville 2002), with proposed structural and physiological theories. Structural mechanisms attribute cross-modal plasticity to the formation of new synapses, garnering support from reports of increased white matter connectivity between visual cortices (Kim et al. 2009) and in the multimodal processing insula of deaf subjects (Allen et al. 2008), and reduced myelination of axons projecting to and from auditory cortices (Emmorey et al. 2003; Shibata 2007; Li et al. 2012). Numerous existing connections between visual and auditory cortical areas (Hall and Lomber 2008) facilitate multimodal convergence onto individual neurons (Keniston et al. 2010) and may be physiologically enhanced or “unmasked” for 1 sensory modality given deprivation of the other (Feldman 2009). Yu et al. (2012) found that cross-modal stimuli exposure revealed previously ineffective or “silent” inputs into multisensory neurons within the superior colliculus of the cat. Furthermore, the proportion of multisensory cortical neurons was doubled in partially deafened adult ferrets (Meredith et al. 2012). With multimodal processing observed in many auditory areas (Clemo et al. 2007; Bizley and King 2009; Meredith and Allman 2009), multisensory neurons may be involved in cross-modal reorganization in deprived auditory cortex.
Evidently, whether these changes in anatomical connectivity occur at the corticocortical, corticothalamic, or subthalamic level also remains to be elucidated. Allman et al. (2009) proposed that subcortical changes in the brainstem could introduce a different input modality into the ascending auditory pathway, without the need to rewire or form new corticocortical or thalamocortical circuitry. Since novel projections between the inferior colliculus and the dorsal lateral geniculate nucleus have been identified in the blind mole rat (Doron and Wollberg 1994) and mice (Chabot et al. 2007), it is possible that the deaf auditory system can also develop new thalamocortical projections. Partial cochlear lesions resulting in expanded representation of intact regions of the damaged cochlea in cats could not be attributed to the relatively normal projections from the medial geniculate body to A1 (Stanton and Harrison 2000). Instead, Stanton and Harrison (2000) proposed that cross-modal reorganization must occur at subthalamic levels. In support of this, Shore et al. (2008) reported trigeminal input into the dorsal cochlear nucleus of deafened guinea pigs, which may provide somatosensory information to the auditory cortex. Our ongoing study of corticocortical connectivity in early- and late-deaf cats may provide insight into the role of anatomical connectivity in shaping auditory cortex cartography. Further anatomical studies comparing thalamocortical and subthalamic connectivity following early- or late-onset deafness will be useful in elucidating the mechanisms for cross-modal reorganization.
This study demonstrates the use of SMI-32 as a robust marker for distinguishing auditory areas in hearing, early-deaf, and late-deaf cats. It is a reliable tool for revealing significant changes in auditory areal cartography underlying cross-modal plasticity following deafness. For future electrophysiological and behavioral studies, SMI-32 processing is recommended to ensure accurate recording and/or isolation of intended areas. Cortical reorganization following either early- or late-auditory deprivation must be considered, as the present study has demonstrated that the cat auditory cortex remains quite plastic even into adulthood, permitting substantial structural changes following late-onset deafness. However, the young brain ultimately demonstrates greater adaptability to auditory deprivation, with further modifications to areal cartography. This study reveals cartographical consequences of cortical plasticity that may enable or limit functional reorganization in auditory areas observed in electrophysiological and psychophysical studies of the deaf.
This work was supported by grants from the Canadian Institutes of Health Research, the Natural Sciences and Engineering Research Council of Canada, and the Canada Foundation for Innovation.
We thank Pam Nixon, Amee Hall, and Dr. Trecia Brown for technical and surgical assistance during this study. Conflict of Interest: None declared.