Authors: Stephen G. Gordon, Alessandra Sacco, Stephen G. Lomber
Categories: Research Article, Structural MRI, Compensatory crossmodal plasticity, Gray matter, Feline, Sensory cortex
Source: Neuroimage: Reports
Authors: Stephen G. Gordon, Alessandra Sacco, Stephen G. Lomber
In the absence of hearing, the plastic nature of the cerebral cortex allows select regions to be repurposed to serve the processing of remaining sensory modalities. This plasticity can be observed in many ways, including measuring the thickness differences of cortical gray matter between hearing and deaf populations to detect regional adaptations. In this study, T1-weighted images were acquired for hearing (n = 38) and perinatally-deafened (n = 31) cats using an ultra-high field 7T MRI scanner to identify normative feline cortical thickness, as well as areas of differing thickness between groups. Most significant changes to sensory-related regions demonstrated thicker cortices in the deaf compared to the hearing group, while specific non-sensory regions were found to be thinner. Furthermore, there was a modest lateralized component, finding that the gray matter of the left hemisphere was more susceptible to thickness changes following auditory deprivation. These results suggest distinct factors driving the adaptations in sensory versus non-sensory cortices in the brain following deafness, and reinforces the task-retainment model of crossmodal plasticity.
The brain is constantly adapting to its environment, with every experience, or lack thereof, modifying its overall organization in some small way. Dramatic changes in sensory input can cause large, measurable alterations within the brain with diverse implications. In the case of early-onset profound deafness, the auditory cortex is deprived of acoustic stimulation, allowing the remaining senses to employ this region in a process known as compensatory crossmodal plasticity (Bavelier and Neville, 2002). This effect is not limited to auditory deprivation, as blindness also leads to parallel optimizations in cerebral structure and function, with both conditions occasionally enhancing sensory processing which eases the navigation of a modified perceptual environment (reviewed Bavelier et al., 2006; Kupers and Ptito, 2014; Sabourin et al., 2022).
Although crossmodal plasticity has been studied using many methods in the past (Butler and Lomber, 2013), there is still no consensus on the driving forces underlying these plastic changes. One idea suggests the possibility of novel connections forming and infiltrating unused sensory regions (Barone et al., 2013), however this theory is contentious and may be limited specifically to developing brains (Angelucci et al., 1998; Morrone, 2010). Contrastingly, another theory is that there are no new pathways being formed in the brain, rather this plasticity is the result of modifications in the effectiveness of preexisting connections (Kok et al., 2014; Wong et al., 2015; Meredith et al., 2016; Butler et al., 2018). There has been growing evidence of communicative adjustments involving auditory cortical regions following deafness via tissue-based analyses, finding a largely conserved afferent framework (Butler et al., 2017; Butler et al., 2018) with limited novel inputs (Barone et al., 2013; Chabot et al., 2015) and minor reweighting of specific connections (Kok et al., 2014; Chabot et al., 2015; Wong et al., 2015; Butler et al., 2017; Clemo et al., 2016; Meredith et al., 2016; Clemo et al., 2017). For instance, the auditory field of the anterior ectosylvian sulcus (FAES), a plastic area previously found to have comparable tracer projections between early deaf and control felines (Meredith et al., 2016), demonstrated increased dendritic spine diameter and density for multisensory afferents (Clemo et al., 2016). FAES is typically involved in auditory localization in hearing cats (Malhotra and Lomber, 2007) but becomes necessary for the task of visual localization following deafness – shifting senses but retaining its functional role (Meredith et al., 2011).
Additionally, a cortical deactivation study using cooling loops (Lomber et al., 1999) was performed to investigate the causal link between select auditory regions and improved visual processing in congenitally deaf cats. The dorsal zone (DZ) was found to be responsible for enhanced visual movement detection, and the posterior auditory field (PAF) was implicated in improved peripheral visual localization in the deaf (Lomber et al., 2010). Furthermore, multiple studies have shown enhanced visual performance (Bavelier et al., 2006; Lomber et al., 2010; Megreya and Bindemann, 2017) or somatosensory vibrotactile discrimination (Levänen and Hamdorf, 2001) following deafness, alongside altered crossmodal activation in the auditory cortex (Levänen et al., 1998; Finney et al., 2001; Auer et al., 2007; Almeida et al., 2015) and modified resting cortical metabolism (Park et al., 2010, 2023) of deprived individuals. Overall, these findings suggest that following auditory deprivation, sensory brain regions experience altered communication and organization resulting in compensatory functional adaptations.
In addition to behavior, histology, and functional imaging, another method to probe crossmodal plasticity following sensory loss is to measure cortical thickness via magnetic resonance imaging (MRI). As opposed to histological work – the gold standard for tissue measurements – MRI-derived cortical thickness analysis is a non-invasive technique to investigate differences in brain structure across populations. In this method, tissue contrasts within structural MR images are used to define the gray matter/white matter and the gray matter/cerebrospinal fluid borders in order to estimate the thickness of the cortical sheet. In a T1-weighted image, each voxel's intensity results from a plethora of local factors, including synaptic density (Kassem et al., 2013), myelo- and cytoarchitecture (Eickhoff et al., 2005), and the remodeling of neuronal processes (Lerch et al., 2011). Along with their respective unique tissue microstructures, the unequal distributions of cell bodies and myelinated axons in these two tissues allow for reproducible delineations of the gray matter/white matter border throughout the cortex. Therefore, a region that is measured to be thicker when using MRI could simply have reduced myelin infiltration in its lower cortical layers and/or increased synaptic density, with no genuine change in gray matter thickness when evaluated histologically. Although it is assumptive to posit the true cause of any differences in MRI-derived thicknesses found between groups, this method conveniently indicates regions of altered micro- and/or macrostructure for subsequent targeted investigations.
MRI-derived gray matter thickness changes have been documented in a variety of experience-based situations and conditions in humans. Some examples include prelingual deafness causing an overall cortical thinning in adolescents (Li et al., 2012), long-term use of spatial navigation resulting in larger hippocampal volumes (Maguire et al., 2000), and meditation experience increasing the thicknesses of regions associated with attention, interoception and sensory processing (Lazar et al., 2005). While these studies are very useful in terms of finding correlational links between brain regions and possible functions, they are limited by the confounding variables within the populations being studied. In contrast, animals may be used to conduct more controlled experiments due to the ability to regulate environmental variables and testing paradigms more extensively. In addition to human studies, gray matter thickness has also been explored in multiple animal models, including non-human primates (Lecoeur et al., 2011; Rohlfing et al., 2012; Wagstyl et al., 2015), birds (Van Der Linden et al., 1998), and rodents (Markham et al., 2003; Dorr et al., 2008; Lerch et al., 2008).
One animal model which is yet to be thoroughly investigated in MRI-derived cortical thickness studies is the domestic cat. Recently, a study examined MRI-derived thickness measurements in the auditory cortex of the perinatally-deafened cat, finding increases in cortical thickness of some of the more dorsal regions (including DZ as well as primary (A1) and secondary (A2) auditory cortices), and decreases in select ventral areas (Gordon et al., 2023). These results are in contrast with histological work concerning congenitally deaf cats, in which DZ, A1 and A2 were all found to be thinner in the deaf when measured with light microscopy (Berger et al., 2017). Although this opposing finding could be attributed to different feline models of deafness, it is indicative of the relative incompatibility between MRI- and histology-based cortical thickness measurements. Regardless, both of these methodologies are valid for discovering altered brain structure, but the directed nature of histological work, along with additional constraints, precludes it from being realistically applied to the entirety of the cat brain. Expanding upon the investigations within auditory cortex (Gordon et al., 2023), a whole brain MRI gray matter cortical thickness analysis of the deaf cat could highlight plastic regions and direct future hypotheses.
The purpose of the present work is to investigate cortical thickness in control and perinatally-deafened cats throughout the entirety of the brain. Current feline MRI literature demonstrates regional thickness changes as a result of deafness, but results are limited solely to the auditory cortex (Gordon et al., 2023). Using structural T1-weighted images, median thicknesses of 146 regions were extracted using a feline gray matter neuroimaging atlas (Stolzberg et al., 2017) and compared between groups. This exploratory study found that regional cortical thickness differences following perinatal deafness were differentially driven by sensory relevance. The results of this paper could lay the groundwork for future studies to look more deeply at these affected areas, elucidating the microstructural alterations motivating these observed measurements.
All procedures were conducted in compliance with the National Research Council's Guide for the Care and Use of Laboratory Animals (8th Edition, 2011) and the Canadian Council on Animal Care's Guide to the Care and Use of Experimental Animals (1993). Moreover, the procedures were approved by the University of Western Ontario's Animal Use Subcommittee of the University Council on Animal Care.
This study's population consisted of 69 cats, wherein 38 were hearing controls (30 female, 6.4–68.1 months of age) and 31 were perinatally-deafened (14 female, 6.7–75.9 months of age). All animals were obtained from a commercial breeding facility (Marshall BioResources, North Rose, NY). Each animal was group-housed with 6–10 other cats in the Psychology Animal Facility at the University of Western Ontario. The animals lived in an enriched environment (12-h light cycle) with water provided ad libitum and free access to dry food for at least 1 h per day. Animal health was monitored daily by a veterinary technician and weekly by a veterinarian.
Prior to deafening or other subsequent experiments, the baseline hearing threshold of each cat was tested by the recording of an auditory brainstem response (ABR). During this procedure, the animal was sedated with the administration of 0.04 mg/kg dexmedetomidine, commonly known as dexdomitor, intramuscularly (i.m.), and then induced and maintained under gaseous general anesthesia of 1–4% isoflurane and medical oxygen (∼1.0 L/min) delivered through a mask. This non-invasive procedure recorded auditory-evoked electrical activity indirectly in the auditory nerve and brainstem pathways through the placement of four subdermal electroencephalography (EEG) recording leads on the one above each ear, at the mid-cranial scalp, and on the lower back. Auditory stimuli were presented to each ear simultaneously via earphones placed in the ear canals, delivering 0.1ms squarewave clicks ranging between 1 and 80 dB. All animals were considered hearing, without impairment, when a characteristic ABR was present at or below 20 dB (World Health Organization, 2019; Institute for Quality and Efficiency in Health Care, 2006). Once the ABR was recorded, the effects of dexdomitor were reversed by administration of 0.04 mg/kg atipamezole (i.m.). For all animals, recovery was uneventful.
The population of deaf animals (n = 31) had a profound hearing loss induced early in development using one of two methods. In both cases, the mechanism of deafening is the chemical destruction of cochlear hair cells using the ototoxic antibiotics neomycin (Leake et al., 1991) or kanamycin (Xu et al., 1993) prior to cochlear maturation (Brugge et al., 1978). In the first method, 23 kittens (median ages at time of 215 ± 20 days) were subcutaneously (s.c.) administered neomycin daily from postnatal day 1 (P1) until hearing thresholds were above 80 dB as determined by an ABR. This was typically observed by P25–P30. In the second method, kanamycin was injected simultaneously with the infusion of a loop diuretic to 8 kittens (median ages at time of 226 ± 1 days) after P14 and once they had reached 350g body mass to cause deafness in a single procedure. These deafening methods induce extensive loss of cochlear hair cells in cats without damaging spiral ganglion cells, and result in animals that are deaf but otherwise neurotypical. To confirm a permanent threshold shift, all deafened animals received a follow-up ABR at least 3 months after the deafening procedure.
In order to minimize any possible movement in the scanner, cats were anesthetized according to one of two protocols (Sacco et al., 2024): ketamine/isoflurane (Brown et al., 2013; Hall et al., 2014; Gordon et al., 2023) or alfaxalone (Levine et al., 2020). In the ketamine/isoflurane protocol, prior to each imaging session, cats were pre-medicated with a mixture of atropine (0.02 mg/kg s.c.) and acepromazine (0.02 mg/kg s.c.), then anesthetized ∼30 min later with a solution of ketamine (4 mg/kg i.m.) and dexdomitor (0.022 mg/kg i.m.). Upon confirmation of an absent gag reflex, each animal was intubated, and an indwelling catheter was placed in the saphenous vein to facilitate intravenous (i.v.) delivery of fluids and anesthesia. In the alternative protocol, animals were premedicated with dexdomitor (0.04 mg/kg i.m.) prior to intubation and catheter insertion, then anesthetized with alfaxalone (2–4 mg/kg i.v.). Following the induction of anesthesia in either protocol, the animal was placed in a sphinx position within a custom-built coil and sled apparatus (Gilbert et al., 2016). Anesthesia was maintained during each session with a ketamine infusion (1.2–1.8 mg/kg/h i.v.) and spontaneously inhaled isoflurane (∼0.5%) with medical oxygen (∼1.5 L/m) or an alfaxalone infusion (5–10 mg/kg/h i.v.). Throughout the scanning session, heart rate, SpO2, respiration rate, expired CO2, blood pressure, and rectal body temperature were monitored. After scanning, buprenorphine (0.01 mg/kg i.v.) was administered to aid in recovery for the alfaxalone sessions. Animals were monitored until sternal recumbency was achieved, at which point they were placed into individual housing until the next morning, when they were returned to the colony. In all cases, recovery was uneventful.
Structural MRIs were acquired using a custom-manufactured 8/24 channel transmit/receive radio-frequency coil (Gilbert et al., 2016) and a 7 T Siemens Magnetom MRI scanner (68 cm bore diameter, 350 mT/m/s slew rate) located at the Centre for Functional and Metabolic Mapping at the Robarts Research Institute. A high-resolution T1-weighted MP2RAGE image stack was acquired for each subject (TR = 6500ms, TE = 3.93ms, flip angle 1 = 4°, flip angle 2 = 5°, 96 slices, 0.5 mm isotropic voxel size, ∼10.5 min scan time).
Subject images were all skull-stripped using an automated in-house tool for feline images and individually verified by two neuroimaging researchers prior to subsequent processing steps. These brain-only images were analyzed using the Diffeomorphic Registration-based Cortical Thickness (DiReCT; Das et al., 2009) algorithm within the Advanced Normalization Tools software (ANTs; Avants et al., 2014) to generate subject thickness maps. DiReCT first determines the gray matter/white matter and gray matter/cerebrospinal fluid borders, then three-dimensionally draws the gray matter/white matter surface and grows it in the direction of the gray matter/cerebrospinal fluid border. Once the gray matter/cerebrospinal fluid border is reached, all voxels along the growth line are filled with an intensity equal to the measured thickness of the tissue in millimeters. In this way, a gray matter thickness map is generated in which voxel intensity is related to local thickness of the cortical sheet.
Subjects' thickness maps were then imported into MATLAB R2021a (MATLAB, 2021) and the effects of age and sex were linearly regressed out of the data, followed by a normalization of median thickness across subjects. The thickness differences of 73 bilateral (146 total) regions of interest (ROIs; Stolzberg et al., 2017) were analyzed across groups by comparing the true median differences to those of ten million random permutations of the two groups. The p-value resulted from the ratio of comparisons in which the absolute value of the permuted difference was greater than the true difference. To investigate how regional thickness in the control group correlated with that of the deaf population, Pearson's R was calculated. This was also measured for the relationship between control cortical thickness values and the change in thickness following deafness across all ROIs. Additionally, ROIs were grouped by primary function (Stolzberg et al., 2018) and ten million random permutations of the groups were used to investigate the significance of modal-level differences within and between the hearing and deaf populations.
Cortical thickness measurements were compared between hearing and perinatally-deafened populations, exploring group differences regionally across 146 ROIs, as well as modally across 14 functional subdivisions of the cortex.
The median thickness measurements for all 146 ROIs in the hearing and deaf groups are presented in Table 1 and Table 2, respectively, with lateral visualizations in Fig. 1 (hearing) and Fig. 2 (deaf). The median thickness of ROIs in the hearing group was 3.81 ± 0.57 mm, with the thinnest region (left retrosplenial area (RSL)) measuring in at 2.13 mm and the thickest (the ventral division of the left prefrontal cortex (PFvL)) being 7.91 mm. The thicknesses of ROIs appeared to be influenced by relative position, with the thinnest ROIs existing dorsoposteriorly in the brain, and the thickest being at the anteroventral extremes (Fig. 1). Additionally, the thicknesses of bilateral homologs (for example, A1 in the left hemisphere versus A1 in the right hemisphere) differ by only 2% on average, a relationship that can be appreciated by comparing the left and right hemispheres shown in Fig. 1. These trends and relationships are shared by the deaf group's thickness values, which are nearly identical to the hearing when visually comparing Fig. 1, Fig. 2. In fact, the correlation between hearing and deaf regional thickness measurements was highly significant (p = 6.85 × 10^−132^), demonstrating an R^2^ value of 0.98.Table 1Thickness values of all ROIs analyzed in the hearing group. Data is presented as group median thickness ± median absolute deviation (MAD) in mm. ROIs are sorted into functional groups (Stolzberg et al., 2018): visual (yellow), auditory (blue), prefrontal (pink), frontal (gray), somatosensory (brown), motor (green), and other (purple). A visual representation of the locations of the modal groupings is displayed in the bottom right.Table 1Table 2Thickness values of all ROIs analyzed in the deaf group. Data is presented as group median thickness ± median absolute deviation (MAD) in mm. ROIs are sorted into functional groups (Stolzberg et al., 2018): visual (yellow), auditory (blue), prefrontal (pink), frontal (gray), somatosensory (brown), motor (green), and other (purple). A visual representation of the locations of the modal groupings is displayed in the bottom right.Table 2Fig. 1Heatmap of regional thickness values in the hearing group. Median thicknesses (mm) of ROIs are visualized on the cat brain, shown from inverted midsagittal (top) and lateral (bottom) perspectives of each hemisphere.Fig. 1Fig. 2Heatmap of regional thickness values in the deaf group. Median thicknesses (mm) of ROIs are visualized on the cat brain, shown from inverted midsagittal (top) and lateral (bottom) perspectives of each hemisphere.Fig. 2
There were 42 ROIs that demonstrated significantly altered thicknesses between groups (Table 3; Fig. 3; Fig. 4; Supplementary Table 1). Of these, the 6 visual, 13 auditory, and 5 of the 6 somatosensory (with the exception of left are 3a (3aL)) ROIs were thicker in the deaf group. Contrastingly, the remaining 5 prefrontal, 2 frontal, and 9 of 10 other (with the exception of the left perirhinal cortex (36L)) ROIs were all significantly thinner. Lastly, there was a significant anticorrelation between the thicknesses of ROIs in the hearing group and the corresponding thickness differences between the groups, with the inflection point of ∼4 mm in the hearing group (R^2^ = 0.33, p = 3.52 × 10^−14^; Fig. 5).Table 3Table of significant regional thickness changes following deafness. Data is presented as group median thickness ± median absolute deviation (MAD) in mm, with differences calculated as deaf minus hearing. Rows are highlighted according to modal groupings (Stolzberg et al., 2018): visual (yellow), auditory (blue), somatosensory (brown), prefrontal (pink), frontal (gray), and other (purple). For all ROI thicknesses and differences, see Supplementary Table 1.Table 3Fig. 3Heatmap of regional thickness differences between groups. Median thickness differences (mm) of ROIs are visualized on the cat brain, shown from inverted midsagittal (top) and lateral (bottom) perspectives of each hemisphere. Differences were calculated as deaf minus hearing.Fig. 3Fig. 4Heatmap illustrating only the significant regional thickness differences between groups. Significant (p < 0.05) median thickness differences (mm) of ROIs are visualized on the cat brain, shown from inverted midsagittal (top) and lateral (bottom) perspectives of each hemisphere. Differences were calculated as deaf minus hearing.Fig. 4Fig. 5Scatterplot demonstrating the anticorrelation between hearing group thickness and thickness differences. Individual ROIs are represented by black points, while the line of best fit is drawn in red. Thickness differences were calculated as deaf minus hearing, and the anticorrelation has a Pearson's R of −0.5742 (p = 3.52 × 10^−14^). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)Fig. 5
When exploring lateralization within regional results, 24 of 42 differences were found on the left and 11 of these differences were unilateral, as opposed to only 5 which were localized solely in the right hemisphere. Additionally, of the 13 auditory ROIs that were significantly thicker, 10 were left/right paired regions (bilateral A1, A2, DZ, as well as the dorsal (dPE) and posterior (pPE) divisions of the posterior ectosylvian gyrus), with the remaining 3 unilateral changes (the anterior auditory field (AAF) and PAF, as well as the intermediate division of the posterior ectosylvian gyrus (iPE)) all in the left hemisphere (Table 3; Fig. 4).
The thinnest functional group in the hearing brain was visual cortex, with a left/right combined median thickness of 3.09 mm. This is followed, in order, by somatosensory cortex (3.65 mm), auditory cortex (4.05 mm), motor cortex (4.56 mm), frontal cortex (4.66 mm), other regions (insular, limbic and miscellaneous; 4.87 mm), and prefrontal cortex (7.08 mm; Fig. 6A). The corresponding thicknesses for the deaf group were 3.19 mm, 3.76 mm, 4.17 mm, 4.58 mm, 4.59 mm, 4.77 mm, and 6.86 mm, respectively. When analyzed using permutation testing, the bilateral visual (right p = 0.007, left p = 0.005) and somatosensory (right p = 0.015, left p = 0.003), along with the left hemisphere's auditory (p = 0.008) groupings were significantly thicker in the deaf relative to controls, while the ‘other’ grouping was significantly thinner bilaterally (right p = 0.012, left p = 0.036; Fig. 6B along the diagonal).Fig. 6Median group thickness per modality, along with inter- and intramodal comparisons within and across groups. (A) Bar graph showing group median thickness (mm) ± median absolute deviation (MAD) per modality in each hemisphere, with only the significant differences for within-group intramodal and between-group intramodal-intrahemispheric comparisons indicated. (B) Asymmetric matrix of p-values with the top half corresponding to inter- and intramodal comparisons within the hearing group (green) and the bottom half to the deaf group (red). Group comparisons within modalities are shown on the diagonal (purple). Non-significant values are highlighted in a lighter shade of red, green or purple. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)Fig. 6
Following between-group comparisons, the left and right functional group thicknesses were compared within populations. For both hearing and deaf groups, the left and the right subsets of the predominantly sensory (visual, auditory, and somatosensory) groupings, as well as the motor cortices, were not significantly different from each other. Additionally, 3 other comparisons in the hearing group (right Frontal – right Motor, right Frontal – left Motor, right Other – left Frontal), and 3 in the deaf group (right Frontal – right Motor, right Other – left Motor, right Other – left Frontal; Fig. 6B) failed to reach significance. The remaining 84 comparisons per group were significantly different.
The purpose of this study was to measure the baseline cortical thickness of the cat brain, as well as to quantify the differences in thickness following perinatal deafness. Using MRI-derived thickness measurements, analyses demonstrated various changes within and beyond sensory cortices. The direction of these differences appears to be motivated by sensory processing (Table 3; Fig. 6), as well as the normative thickness defined by the hearing group (Fig. 5).
In the present study, MRI-derived cortical thickness values for the control feline brain measured 2–8 mm (Table 1). When looking at modal thicknesses, most values lie within the 2–6 mm range, with the notable exception of prefrontal ROIs (Fig. 6). While it may be the case that prefrontal ROIs in the cat cortex are indeed thicker, it is more likely that these results are influenced by cranial sinuses and coil drop-off decreasing the signal-to-noise ratio (SNR) anteriorly and ventrally in the acquired images. This idea is supported by Fig. 1, Fig. 2, where the thickest measurements are found mainly at the anteroventral extremes of the brain. While this may be considered an issue for the use of MRI when measuring gray matter cortical thickness, the SNR discrepancies near sinuses are consistent across subjects and groups, and thus a non-invasive investigation of the effects of early-onset deafness on feline gray matter is still possible.
There appears to be a distinction between sensory and non-sensory ROIs when considering cortical thickness in the feline brain. The visual, auditory, and somatosensory cortices each exhibit lower overall thicknesses than the non-sensory regional groupings (Fig. 6A). As they are the most dorsal and posterior of all regions and are furthest from cranial sinuses, feline sensory cortices have the highest relative SNRs, therefore this apparent thickness trend for sensory cortices may have something to do with their general locations. There is also the consideration of a possible difference in cytoarchitecture and connectional organization of sensory versus non-sensory regions. Although the current results establish cortical thickness baselines to guide future studies examining the cat cortex, more directed work examining the underlying cytoarchitecture could uncover the macrostructural contributions driving MRI-based cortical thickness in the control cat.
The thicknesses of the ROIs in the deaf cat brain do not differ greatly from those of the hearing, with all regional differences being less than 7% of the corresponding control thicknesses (Supplementary Table 1). While the overall patterns of thicknesses between the two groups are highly similar (R^2^ = 0.98; Fig. 1; Fig. 2), an additional analysis revealed a hidden correlation. Interestingly, the thinner a region was measured to be in the hearing group, the larger the increase in thickness was in the deaf. This is true up to a thickness of ∼4 mm, beyond which the deaf population demonstrates control-thickness-dependent regional thinning (Fig. 5). This was found to be a highly significant anticorrelation with a p-value of less than 10^−13^, yet the implications of this relationship are unknown at this time. A related trend is that for nearly all significant differences in cortical thickness following perinatal deafness, the thickness was increased for sensory and decreased for non-sensory ROIs (Table 3; Fig. 4; Supplementary Table 1). Therefore, when an entire sense is removed or attenuated, it appears that predominantly sensory regions adapt in a distinctive way relative to non-sensory regions due to crossmodal plasticity.
An effect of sensory involvement on cortical thickness alterations following hearing loss agrees with the notion of behavioral role conservation in compensatory crossmodal plasticity (Meredith et al., 2011). Although there has been limited evidence of long-range axonal branching driving crossmodal plasticity following sensory loss (Barone et al., 2013; Chabot et al., 2015), this mechanism appears to be limited to relatively early stages of cortical maturation and spatially coincident axon paths (Angelucci et al., 1998; Sharma et al., 2000). The remaining mechanistic possibilities motivating this plasticity fall largely into two the unmasking and/or reweighting of preexisting synapses (Clemo et al., 2016), as well as local infiltration (Darian-Smith and Gilbert, 1995; Rauschecker, 1995). Unfortunately, none of these possibilities are easily verifiable via current T1-weighted MR sequences, although downstream effects of this plasticity can still be observed when measuring cortical thickness. Regardless of the actual mechanisms underlying compensatory crossmodal plasticity, the current results indicate the possibility of distinct factors driving the adaptations in sensory versus non-sensory cortices, reinforcing conclusions drawn from reviews on deafness (Hribar et al., 2020; Manno et al., 2021) and blindness (Kupers and Ptito, 2014; Sabourin et al., 2022) literature. Presumptively, this could point to increased local branching and synapse density within auditory and proximal extrasensory regions, and a concurrent small loss of previously auditory afferents to the remainder of the brain.
There is a modest yet persistent trend of lateralized differences as a result of deafness seen in this study. The majority (11) of the 16 single-sided regional differences were localized within the left hemisphere, including all 3 significant unilateral auditory differences (Table 3). When grouping the ROIs modally, the left hemisphere's visual, auditory, and somatosensory cortices were all significantly thicker in the deaf compared to controls (Fig. 6). The corresponding parcellations of cortex in the right hemisphere demonstrated a similar increase in thickness, although the right auditory cortex difference failed to reach significance (p = 0.15). Lateralized cortical functioning has been researched and accepted in humans, but less focus has been directed to other species. While paw preference and some laterality of function has been noted in the past for cats (Fabre-Thorpe et al., 1993; Ocklenburg et al., 2019), the modulation of thickness differences seen here point to more complexity than originally acknowledged. Such clear indicators of laterality suggest a large-scale asymmetric organization of feline cortical processing that is discernible through the unbalanced reaction of the overall cortical network to an insult such as hearing loss.
While there is currently only one study that has examined cortical thickness changes as a result of deafness in the cat using MRI (Gordon et al., 2023), there is an absence of analogous investigations using other deaf animal models. Therefore, the current results are only comparable to the small number of deaf human MRI cortical thickness investigations. Many studies have published varying results regarding macrostructural cortical plasticity of the deaf human brain, however most of these studies analyzed volume as opposed to thickness (reviewed Hribar et al., 2020; Ratnanather, 2020; Simon et al., 2020). In terms of cortical thickness, one study found global and regional thinning within the brains of adolescents with prelingual profound deafness (Li et al., 2012), while another study found various loci of thinning as well as thickening throughout the adult prelingually-deaf brain, although none of the results overlapped between these studies (Pereira-Jorge et al., 2018). Unfortunately, all of the regions found to be thinner in the study on deaf adolescents and a subset of those from the adult study were related to the senses, which opposes the present results on deafened cats. The cause for this discrepancy is unknown, however more studies will need to be performed in both species to truly ascertain the underlying mechanisms at play.
Although thickness studies on deaf humans do not seem to generally align with the present work, some of the results from MRI-based investigations of the human brain following deafness point to limited overlap in plastic adaptations across species. Multiple human studies have concluded that the motion-selective area MT/MST has a greater functional connectivity to the visual cortex (Bavelier et al., 2000) and is activated more in the brains of deaf populations for peripheral visual motion detection tasks (Bavelier et al., 2001; Bola et al., 2017). These results spark interest as the MT/MST complex has a similar function to that of the posteromedial (PMLS) and posterolateral (PLLS) lateral suprasylvian areas, which have also been shown to be responsive to moving visual stimuli in the control cat (von Grünau et al., 1987). Furthermore, a correlation between cortical thickness measurements and visual motion detection thresholds has previously been found in the right PT of deaf humans (Shiell et al., 2016). Shiell and colleagues found that in the deaf, the thickness of PT was associated with visual motion detection thresholds, with a thicker PT correlating with a lower motion detection threshold. Feline DZ has been associated functionally with visual motion detection enhancement in deaf cats (Lomber et al., 2010), pointing to a possible translational link between feline DZ and human PT, although more direct investigations may be required to discover the true connection. These results suggest the possibility that regions associated with specific tasks – in this case visual motion detection – are more susceptible to auditory-deprivation-induced crossmodal plasticity in humans. Interestingly, the left PLLS and bilateral DZ were found to be significantly thicker in the present work (Table 3), potentially pointing to measurable overlaps in the cortical response to deafness across human and feline brains.
Differential thinning and thickening of various cortical loci were observed in the deaf feline group relative to healthy controls. The majority of the changes and trends were increases in thickness in sensory areas and decreases in non-sensory ones, with plasticity affecting the left hemisphere slightly more than the right. As many sensory regions process inputs from multiple modalities in the hearing brain, it is postulated that in the absence of auditory input, these regions undergo structural and/or functional modifications related to compensatory crossmodal plasticity that allow them to process extrasensory information more efficiently. These results lend support to the task-retainment theory of crossmodal plasticity, although the true causes of the present MR-measured thickness changes are as of yet unknown. Future work using tissue microscopy and/or electrophysiology could reveal the underlying micro- and macrostructural changes in the cortex following perinatal deafness in the cat, using these noted ROIs as rationalized sites of plasticity for these more targeted studies.
This work was supported by the 10.13039/501100000024Canadian Institutes of Health Research [grant number 159597].
Stephen G. Gordon: Writing – review & editing, Writing – original draft, Investigation, Formal analysis, Conceptualization. Alessandra Sacco: Writing – review & editing, Methodology, Formal analysis. Stephen G. Lomber: Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.