Partial hearing loss can cause neurons in the auditory and audiovisual cortices to increase their responsiveness to visual stimuli; however, behavioral studies in hearing-impaired humans and rats have found that the perceptual ability to accurately judge the relative timing of auditory and visual stimuli is largely unaffected. To investigate the neurophysiological basis of how audiovisual temporal acuity may be preserved in the presence of hearing loss-induced crossmodal plasticity, we exposed adult rats to loud noise and two weeks later performed
Following sensory deprivation, such as vision or hearing loss, the brain has the capacity to undergo extensive reorganization, which is often characterized by an increased responsiveness of neurons in the deprived sensory cortex to the spared senses (i.e., crossmodal plasticity) (for review, see [
Despite the high prevalence of
In normal-hearing subjects, there is clear evidence of several behavioral advantages afforded by the brain’s natural ability to integrate auditory and visual information, including improved detection, localization, and identification of the stimuli. In addition, psychophysical testing has revealed that auditory and visual stimuli presented within ~100 ms offset from each other can be bound into a unified percept, with subjects showing difficulty in accurately judging whether the auditory or visual stimulus was presented first. Ultimately, because neuroimaging studies in humans have shown that synchronized activity in the multisensory cortex underlies audiovisual temporal acuity [
To date, no studies have investigated whether changes in the temporal precision of audiovisual processing occur at the neuronal level following adult-onset hearing loss or if these crossmodal effects differ across the neighboring regions of the multisensory cortex that normally integrate audiovisual stimuli. Thus, in the present study, we used
The present study included two experimental series—each using a separate group of adult male Sprague-Dawley rats (
Using appetitive operant conditioning, rats were trained on a two-alternative forced-choice paradigm that assessed their ability to perform audiovisual temporal-order judgments (TOJ;
Behavioral training and testing were conducted in a standard modular test chamber (ENV-008CT; Med Associates Inc., St. Albans, VT) that was housed within a sound-attenuating box (29” W by 23.5” H by 23.5” D; Med Associates Inc.). The front wall of the behavioral chamber included a nose poke as well as a left and right feeder trough, each fitted with an infrared detector to monitor the rat’s performance. The test chamber was illuminated by a house light on the back wall. Real-time processing hardware (RZ6 and BH-32, Tucker Davis Technologies, Alachua, FL) were interfaced with the test chamber. Custom behavioral protocols running in MATLAB (EPsych Toolbox,
The auditory stimulus was a 50 ms noise burst (75 dB SPL; 1-32 kHz) presented from a speaker (FT28D, Fostex, Tokyo) mounted on the ceiling of the behavioral chamber near the front wall. The intensity of the auditory stimulus was calibrated using custom MATLAB software with a 1/4-inch microphone (2530, Larson Davis, Depew, NY) and preamplifier (2221; Larson Davis). The visual stimulus was a 50 ms light flash (27 lux) from an LED (ENV-229M; Med Associates Inc.) located above the center nose poke. An LED light meter (model LT45, Extech Instruments, Nashua, NH) was used to determine the intensity of the visual stimulus.
Prior to commencing behavioral training, rats were weighed daily and maintained on a food-restricted diet until they neared 85% of their free-feeding body mass. Over the course of several stages of training, rats learned to associate a given audiovisual stimulus condition with a specific feeder (i.e., TOJ task: auditory-
Pharmacological silencing of the V2L cortex disrupts audiovisual temporal acuity in rats. (a) An overview of both the TOJ task and the SJ task that were used to screen rats for their audiovisual temporal acuity. Across several stages, rats were trained to select the right or left feeder trough depending on the stimulus condition presented (i.e., TOJ task: auditory-
Behavioral tasks
V2L injection sites
TOJ psychophysical curves
SJ psychophysical curves
Once rats had successfully completed all stages of behavioral training, they were prepared for chronic implantation of bilateral guide cannulae into the V2L cortex, as this would ultimately allow for the local microinfusion of artificial cerebrospinal fluid (aCSF) or muscimol prior to behavioral test sessions. In preparation for surgery, the rats were anaesthetized with isoflurane (induction: 4%; maintenance: 2%) and body temperature was maintained at 37°C using a homeothermic heating pad (507220F; Harvard Apparatus). A subcutaneous injection of meloxicam (1 mg/kg) was administered before surgery and as needed postsurgery for pain management. Once a surgical plane of anesthesia was achieved, rats were placed in a stereotaxic frame with blunt ear bars, a midline incision was made in the scalp, and the dorsal aspect of the skull was cleaned with a scalpel blade. In an effort to improve postsurgical recovery, we elected to have the guide cannulae enter into the cortex on a dorsal-medial-to-ventral-lateral approach, as this left the temporalis muscles intact. After small burr holes were drilled into the skull, stainless steel guide cannulae (26 gauge, 3 mm in length) were bilaterally implanted to target the V2L cortex using the following coordinates: 6 mm caudal to bregma, 5.6 mm lateral to the midline, and a 10° angle (Figure
The rats returned to daily behavioral training after they had fully recovered from surgery, and once their performance again achieved >80% accuracy, experimental test sessions were introduced in which novel SOAs were presented (described below). Ultimately, each rat performed two experimental test sessions following the local microinfusion of either aCSF, which served as the control condition, or muscimol, a potent agonist of GABA-A receptors, which was used to silence the neuronal activity within the V2L cortex.
Microinjections were performed in awake animals using infusion cannulae that extended 1.2 mm beyond the length of the chronically implanted guide cannulae. On a testing day, a given rat received a bilateral infusion of either aCSF (0.5
During the TOJ test sessions, 7 SOAs were randomly delivered (i.e., 0, ±40, ±100, and ±200 ms) and rats performed a minimum of 10 trials at each of the novel SOAs. For the SJ test sessions, 5 SOAs were randomly delivered (i.e., 0, 10, 40, 100, and 200 ms) and rats were presented with at least 18 trials at each of the novel SOAs. For both behavioral tasks, 70% of the trials presented consisted of training stimuli (i.e., TOJ task: ±200 ms SOA and SJ task: 0 and 200 ms SOA), while the remaining 30% of the trials was made up of the random presentation of the novel SOAs. This distribution of trials has been previously shown to reduce the potential of developing a side bias [
For each of the TOJ test sessions, performance across all 7 SOAs was measured as the proportion of trials in which the rat perceived the stimuli as visual first (i.e., it responded to the right feeder trough; Figure
In a separate group of rats (
Prior to the
Under ketamine (80 mg/kg; IP) and xylazine (5 mg/kg; IP) anesthesia, rats were bilaterally exposed to broadband noise (0.8–20 kHz) for two hours at 120 dB SPL and body temperature was maintained at ~37°C using a homeothermic heating pad. This noise exposure was selected because it has been shown to be effective at inducing changes in the auditory cortex [
Following the final hearing assessment, each rat was maintained under ketamine/xylazine anesthesia and fixed in a stereotaxic frame with blunt ear bars. Anesthetic depth was assessed by the absence of a pedal withdrawal reflex, and supplemental doses of ketamine/xylazine were administered IM as needed. An incision was made along the midline of the skull, and the dorsal aspect of the skull was cleaned with a scalpel blade. The left temporalis muscle was reflected and removed using a blunt dissection technique in order to provide access to the temporal bone overlying the left auditory and multisensory cortices. A stereotaxic micromanipulator was used to make a mark on the skull 6 mm caudal of bregma, which represents the approximate stereotaxic coordinates of the lateral extrastriate visual (V2L) cortex [
In each animal, two recording penetrations were performed which encompassed the majority of the audiovisual cortex. At each of the recording locations (described in detail below), a small slit was made in the dura and a 32-channel linear electrode array was slowly inserted perpendicular to the cortical surface (Figure
Recording site reconstruction and audiovisual-evoked CSD analysis in the multisensory zone of the V2L cortex (V2L-Mz). (a) The schematic shows a reconstruction of all recording penetrations in normal-hearing control rats (grey;
In each rat, laminar recordings were completed in two brain regions: (1) the multisensory zone of the lateral extrastriate visual cortex (V2L-Mz; corresponding to the 2 mm ventral mark made on the skull using our measurements) and (2) the auditory zone of the lateral extrastriate visual cortex (V2L-Az; 2.5 mm ventral). Consistent with previous studies demonstrating that higher-order sensory cortices occur at the intersection of the primary sensory cortices [
At each of the recording locations, auditory, visual, and combined audiovisual stimuli were presented using a TDT RZ6 processing module (100 kHz sampling rate) and custom MATLAB software. Auditory stimuli consisted of 50 ms noise bursts (1-32 kHz) from a speaker (MF1, TDT) positioned 10 cm from the right pinna on a 30° angle from the midline. The intensity of the auditory stimulus was customized for each rat, such that it was presented 40 dB above the rat’s click threshold (control:
The CSD analysis provides a spatial profile of ionic flow and a measure of the total current density that enters or leaves the extracellular matrix through the cell membrane [
The CSD analysis reveals the flow of ions into and out of the neural tissue across the cortical thickness. Current sinks represent the flow of positive ions into the neural tissue from the extracellular space, which is reflective of events such as active excitatory synaptic populations and axonal depolarization [
Consistent with Schormans et al. [
Audiovisual-evoked CSD profiles within the multisensory zone of the V2L cortex in response to 3 different SOAs. Representative CSD profiles (a, c, and e) and extracted CSD waveforms (b, d, and f) at SOAs of (a, b) 0 ms, (c, d) 30 ms, and (e, f) 50 ms in response to a combined audiovisual stimulus. CSD waveforms were extracted from the electrode showing the largest amplitude from each of the individual sinks (denoted by the dashed lines on the CSD images for the supragranular (sSk, red), granular (gSk, green), infragranular upper (iSk-upper, blue), and infragranular lower (iSk-lower, black) responses; sinks are positive, whereas sources are negative. In each of the plots, the horizontal black bar denotes the presentation of the visual stimulus (50 ms LED flash at 15 lux) and the grey horizontal bar shows the timing of the auditory stimulus (50 ms noise burst at 40 dB above click threshold). The black arrow within the CSD waveforms on panels (b,f) shows the location of the visual response, demonstrating that the visual response changes from occurring second at an SOA of 0 ms to occurring first at an SOA of 50 ms.
To examine the overall strength of postsynaptic currents in each of the cortical areas, the average rectified CSD (AVREC) measure was applied to the CSD analysis [
Multisensory interactions were quantified by comparing the response of the combined audiovisual stimulus to that of the unimodal stimulus that evoked the largest response in each experiment [
At the conclusion of both of the experimental series, the rats were injected with sodium pentobarbital (100 mg/kg; IP) in preparation for exsanguination via transcardial perfusion with 4% paraformaldehyde. Brains were serially sectioned (50
Statistical analyses were conducted on the data using various procedures, including repeated-measures analysis of variance (ANOVA), one-way ANOVA, or paired/unpaired
In Experiment 1, we investigated the contribution of the V2L cortex to (1) the perception of simultaneity during a TOJ task and (2) synchrony perception during an SJ task, by locally infusing the GABA-A receptor agonist, muscimol, prior to behavioral testing and ultimately comparing the performance results to those following the control condition (i.e., aCSF infusion). During the TOJ test sessions, the proportion of trials that were perceived as visual first was determined for all 7 SOAs, ranging from -200 ms (i.e., auditory-first) to +200 ms (i.e., visual-first). A two-way repeated-measures ANOVA revealed a significant interaction of infusion by SOA (
During the SJ test sessions, the proportion of trials that were perceived as synchronous (i.e., the rat responded to the right feeder trough) was determined for all 5 SOAs ranging from 0 ms (i.e., synchronous) to 200 ms (i.e., asynchronous; visual stimulus presented 200 ms before the auditory stimulus). Overall, a two-way repeated-measures ANOVA revealed the main effects of infusion and SOA (
The collective results of Experiment 1 show for the first time that the V2L cortex is directly involved in the perceived timing of audiovisual stimuli. Moreover, the fact that these results confirm the importance of the V2L cortex in TOJ task performance was interesting given that our previous studies on hearing-impaired rats showed a preservation of audiovisual temporal perception despite extensive crossmodal reorganization in the V2L cortex in the weeks following noise-induced hearing loss. In considering this apparent paradox, we conducted Experiment 2 in which
Consistent with [
For all electrophysiological experiments, the intensity of the auditory stimulus (50 ms noise burst, 1-32 kHz) was adjusted for each rat in order to control for potential differences in hearing sensitivity among rats. To account for each rat’s noise-induced hearing loss, the auditory stimulus was presented 40 dB SPL above its click threshold. As such, the auditory stimulus that was presented during the electrophysiological experiments to the noise-exposed rats was greater in comparison to the controls (noise exposed:
Using the analysis of CSD sink amplitudes, we investigated whether noise-induced crossmodal plasticity within the multisensory zone of the lateral extrastriate visual cortex (V2L-Mz) altered audiovisual temporal processing across the cortical layers. Within V2L-Mz—a region previously shown to exhibit increased visual responsiveness following exposure to loud noise [
A loss of the characteristic audiovisual temporal profile was observed across the majority of layers of the multisensory zone of the V2L cortex in noise-exposed rats. Averaged CSD waveforms from the (a) supragranular, (b) granular, (c) infragranular upper, and (d) infragranular lower layers within the V2L-Mz in response to audiovisual stimuli presented at SOAs of 0, 30, and 50 ms. The black horizontal bar denotes the presentation of the visual stimulus, and the grey horizontal bar shows the timing of the auditory stimulus. The dark lines represent the group mean, and the shading represents the SEM for the noise-exposed rats (blue;
Supragranular layer
Granular layer
Infragranular (upper) layer
Infragranular (lower) layer
As shown in Figure
In addition to examining the effects of noise-induced hearing loss within distinct cortical layers, AVREC waveforms were computed in order to provide additional information about the temporal pattern of the overall strength of the postsynaptic currents [
Noise-induced hearing loss enhanced the audiovisual-evoked AVREC amplitudes at select SOAs within the multisensory zone of the V2L cortex. (a) AVREC waveforms in response to audiovisual stimuli presented at SOAs of 0, 10, 30, and 50 ms (from left to right) for noise-exposed rats (blue;
AVREC waveforms
AVREC peak amplitude
AVREC peak latency
Similar to V2L-Mz, it has been previously demonstrated that there is increased visual responsiveness within the auditory zone of the V2L cortex following noise-induced hearing loss [
A decrease in audiovisual-evoked CSD amplitudes was generally observed within the auditory zone of the V2L cortex in noise-exposed rats. Averaged CSD waveforms from the (a) supragranular, (b) granular, (c) infragranular upper, and (d) infragranular lower layers within the auditory zone of the V2L cortex in response to audiovisual stimuli presented at SOAs of 0, 30, and 50 ms. The black horizontal bar denotes the presentation of the visual stimulus, and the grey horizontal bar shows the timing of the auditory stimulus. The dark lines represent the group mean, and the shading represents the SEM for the noise-exposed rats (blue;
Supragranular layer
Granular layer
Infragranular (upper) layer
Infragranular (lower) layer
While the multisensory zone of V2L demonstrated an overall increase in CSD sink amplitude, an opposite pattern emerged in the more ventrally located auditory zone of the V2L cortex (V2L-Az). As shown in Figure
To further examine the consequences of a partial hearing loss on the auditory zone of the V2L cortex, the overall strength of the postsynaptic currents was examined by computing AVREC waveforms for each of the groups. To do so, AVREC peak amplitude and latency were extracted from the waveforms in response to audiovisual stimuli across a range of SOAs. Overall, a two-way repeated-measures ANOVA revealed a main effect of the group (
Audiovisual-evoked AVREC amplitude and latency within the auditory zone of the V2L cortex following noise-induced hearing loss. (a) Audiovisual-evoked AVREC waveforms within the auditory zone of the V2L cortex at SOAs of 0, 10, 30, and 50 ms for noise-exposed rats (blue;
AVREC waveforms
AVREC peak amplitude
AVREC peak latency
To further examine changes in audiovisual processing following noise-induced hearing loss, the magnitude of response interaction was calculated for the granular sink and AVREC peak amplitudes by comparing audiovisual-evoked amplitudes to the unimodal stimulus that produced the largest amplitude. More specifically, the magnitude of response interaction for both the granular sink data and AVREC data was calculated for each group at all temporal offsets ranging from 0 ms (synchronous) to 50 ms (visual leading) within both V2L-Mz and V2L-Az. Consistent with the neuronal response profile observed in the superior colliculus [
For the granular sink dataset, an initial three-way repeated-measures ANOVA found a significant interaction of the area by SOA by group (
The magnitude of multisensory response interactions varied across the regions of the V2L cortex before and after noise exposure. To assess how hearing loss affected the sensitivity of neurons in the multisensory and auditory zones of the V2L cortex to the relative timing of the auditory and visual stimuli, the magnitude of the multisensory response interaction was calculated by comparing the amplitude of the granular sink in response to the combined audiovisual stimulus to that of the separately presented unimodal stimulus that evoked the largest response. Overall, a differential effect was observed between the noise-exposed rats (
Additional support for a functional transition in the cortical region showing the greatest degree of audiovisual response interaction was evident from analyses of the AVREC data collected from the multisensory and auditory zones of the V2L cortex in noise-exposed rats versus controls. As shown in Figure
Compensatory plasticity in the auditory zone of the V2L cortex preserves audiovisual temporal processing following moderate hearing loss. Using the AVREC amplitude as a measure of the overall strength of postsynaptic currents in a given cortical region, the magnitude of the multisensory response interactions was then calculated at each SOA to determine how noise-induced hearing loss affected the sensitivity of neurons in the multisensory and auditory zones of the V2L cortex to the relative timing of the auditory and visual stimuli. Ultimately, the temporal profile observed in V2L-Mz of control rats (a), in which there was a significant increase in the magnitude of the multisensory response interaction at SOAs of 30 and 40 ms, was consistent with the temporal profile that emerged within V2L-Az of noise-exposed rats (d). Following two-way repeated-measures ANOVAs, paired sample
Multisensory zone of V2L in controls
Multisensory zone of V2L in noise exposed
Auditory zone of V2L in controls
Auditory zone of V2L in noise exposed
Sensory profile in controls
Sensory profile in noise exposed
Following moderate hearing loss, neurons in the auditory cortex as well as the higher-order audiovisual cortex maintain a residual capacity for sound processing, while also now demonstrating crossmodal plasticity, a phenomenon characterized by an increased responsiveness to visual stimuli [
Previous studies on normal-hearing rats have reported that the V2L cortex, which is wedged between the primary visual cortex (V1) and the dorsal auditory cortex (AuD), shows several hallmarks of cortical multisensory processing consistent with other mammals [
Based on these electrophysiological findings, it would be reasonable to suspect that the V2L cortex plays a role in perceptual tasks that require audiovisual temporal acuity, such as those in which the rats must judge the temporal order of auditory and visual stimuli (TOJ task) or whether the auditory and visual stimuli were presented synchronously or not (SJ task). To investigate this possibility, we chronically implanted cannulae into the V2L cortex of normal-hearing rats that had been trained to perform the TOJ or SJ task and then microinfused muscimol (or aCSF) prior to behavioral testing to determine the effect of pharmacological silencing of the V2L cortex on audiovisual temporal acuity. Ultimately, this novel experimental series revealed that the inactivation of the V2L cortex (1) caused a shift in the perception of simultaneity during the TOJ task, such that the light flash now had to be presented much earlier before the noise burst for the two stimuli to be perceived as having occurred simultaneously, and (2) caused a lengthened epoch of time over which the physically asynchronous auditory and visual stimuli were perceived to have occurred at the same moment in time (i.e., the temporal binding window increased on the right side of physical synchrony) (Figure
Our previous studies on noise-exposed rats found a significant reduction in the auditory-evoked activity in V2L-Mz (despite increasing the noise burst intensity to control for their elevated hearing thresholds) and a concomitant increase in visual responsiveness in the neighboring region, V2L-Az [
At present, the cellular mechanisms underlying the functional shift in multisensory convergence across the neighboring cortical regions remain elusive. With respect to hearing loss-induced crossmodal plasticity in general, it has been postulated that cortical reorganization may emerge via (1) altered multisensory processing in subcortical loci that ultimately manifests as cortical plasticity [
To date, the vast majority of studies that have investigated the behavioral consequences of hearing loss-induced crossmodal plasticity have focused on humans and laboratory animals with profound hearing loss. Given the improved processing of peripheral visual stimuli and visual motion [
The present study is aimed at advancing our understanding of the nature and extent of sensory reorganization that occurs following moderate hearing loss in adulthood, with an emphasis on how this highly prevalent form of sensory deprivation impacts audiovisual temporal processing at the neuronal level. Using a rat model of noise exposure and layer-specific electrophysiological recordings of postsynaptic potentials in neighboring regions within the lateral extrastriate visual (V2L) cortex, we have shown for the first time that adult-onset hearing loss does not result in a loss of temporally precise audiovisual processing but rather a shift in the cortical region displaying this capacity for temporal sensitivity. Indeed, although the neurons in the multisensory zone of the V2L cortex of noise-exposed rats no longer showed the canonical enhancement of multisensory responses when the visual stimulus preceded the auditory stimulus by ~30 ms, this temporal profile emerged in the neighboring cortical region, the once predominantly auditory zone of V2L. Future studies are needed to uncover the cellular mechanisms associated with this compensatory plasticity and whether the transition in the functional boundary of the audiovisual cortex is indeed the neural substrate for the preservation of audiovisual temporal perception reported in hearing-impaired subjects.
Auditory brainstem response
Point of subjective simultaneity
Artificial cerebrospinal fluid
Repeated-measures ANOVA
Analysis of variance
Standard error of the mean
Dorsal auditory cortex
Synchrony judgment
Average rectified current source density
Stimulus onset asynchrony
Current source density
Supragranular sink
Decibels, sound pressure level
Temporal binding window
Granular sink
Temporal order judgment
Intramuscular
Primary visual cortex
Intraperitoneal
Lateral extrastriate visual cortex
Infragranular sink-lower
Auditory zone of the lateral extrastriate visual cortex
Infragranular sink-upper
Multisensory zone of the lateral extrastriate visual cortex
Just noticeable difference
Local field potential.
The electrophysiological data used to support the findings of this study are available from the corresponding author upon request.
The authors declare that there is no conflict of interest regarding the publication of this paper.
We would like to acknowledge Daniel Stolzberg (New York University) for his assistance with the current source density analysis, as well as Albert Vo, Aly Balbaa, and Katharine Pacoli for their assistance with behavioral training. This work was supported by the Natural Science and Engineering Research Council of Canada (NSERC Discovery Grant 435819-2013 RGPIN to BLA) and the Canadian Institutes of Health Research (CIHR 137098 to BLA).