Ultra-high Field fMRI Reveals Effect of Ketamine on Vocal Processing in Common Marmosets

Abstract

Auditory deficits are a well-known symptom in neuropsychiatric disorders such as schizophrenia. The noncompetitive N-methyl-d-aspartate receptor antagonist ketamine has been used to model sensory and cognitive deficits in nonhuman primates, but its whole-brain effects remain largely unknown. Here we employed ultra-high field functional magnetic resonance imaging at 9.4 T in awake male and female marmoset monkeys (Callithrix jacchus) to compare brain activations to conspecific vocalizations, scrambled vocalizations, and nonvocal sounds following the administration of a subanesthetic dose of ketamine. Our findings reveal a broad suppression of activations across auditory regions following ketamine compared with saline. Additionally, we observed differential effects depending on the type of sound, with notable changes in the mediodorsal thalamus and anterior cingulate cortex, particularly during the processing of vocalizations. These findings suggest a potential overlap between the effects of ketamine and neural disruptions observed in schizophrenia, particularly affecting vocalization processing.

Keywords: auditory processing; fMRI; ketamine; marmoset monkeys; schizophrenia.

Figure 1.

Brain networks activated by all auditory stimuli following saline and ketamine administrations. This figure represents the group functional maps from nine awake marmosets, showing brain regions with significantly greater activations in response to all auditory stimuli (conspecific vocalizations, scrambled vocalizations, and nonvocal sounds) compared with the baseline (no sounds). The top panels display results from the saline administration (control), and the bottom panels show results from the ketamine administration. Activation maps are shown on both lateral and medial views of the marmoset's cortical surfaces, with subcortical responses depicted on coronal slices. A white line delineates the brain regions according to the Paxinos parcellation (Paxinos et al., 2013) of the NIH marmoset brain atlas (Liu et al., 2018). The reported brain regions are identified based on a threshold corresponding to a Z test statistic > 1.96 (p < 0.05, two-sided) and corrected for multiple comparisons using FWE cluster-size correction (α = 0.05) applied across all voxels. Aud, auditory cortex; IC, inferior colliculus; MG, medial geniculate nucleus; Th, thalamus; Th-MD, mediodorsal thalamus.

Figure 2.

Brain regions showing main and interaction effects of ketamine administration. This figure illustrates (A) the regions showing a significant main effect of administration, highlighting areas with differential responses between saline and ketamine administrations across all sound types, and (B) the regions showing a significant interaction effect between administration (saline, ketamine) and condition (vocalizations, scrambled vocalizations, and nonvocal sounds). Activation maps are displayed on both lateral and medial views of the fiducial marmoset cortical surfaces, with subcortical activations presented on coronal slices for panel A. White lines delineate brain regions according to the Paxinos parcellation of the NIH marmoset brain atlas (Paxinos et al., 2013; Liu et al., 2018). In panel A, significant regions are shown at a threshold of Z test statistic > 1.96 (yellow, greater responses for saline) or <−1.96 (blue, greater responses for ketamine; p < 0.05; two-sided), corrected for multiple comparisons using FWE cluster-size correction (α = 0.05). In panel B, significance was assessed using F statistics converted to Z statistics (Z test statistics > 1.65; p < 0.05; one-sided), corrected for multiple comparisons using FWE cluster-size correction (α = 0.05). Aud, auditory cortex; IC, inferior colliculus; MG, medial geniculate nucleus.

Figure 3.

Differential brain activations for ketamine versus saline for each auditory condition. This figure displays group functional maps from nine awake marmosets, illustrating regions where responses were greater following saline administration compared with ketamine administration for each auditory stimulus: conspecific vocalizations A, scrambled vocalizations B, and nonvocal sounds C. The response maps are masked by the significant interaction effects map and are depicted on both lateral and medial views of the fiducial marmoset cortical surfaces, with subcortical activations shown on coronal slices. Brain regions are delineated according to the Paxinos parcellation of the NIH marmoset brain atlas (Paxinos et al., 2013; Liu et al., 2018). The reported responses meet a threshold corresponding to Z test statistics > 1.96 (yellow scale, saline > ketamine) or <−1.96 (blue scale, ketamine > saline; p < 0.05, two-sided), corrected for multiple comparisons using FWE cluster-size correction (α = 0.05). Hyp, hypothalamus; IC, inferior colliculus; MG, medial geniculate nucleus; Th, thalamus; Th-MD, mediodorsal thalamus.

Figure 4.

Beta values obtained for each auditory condition following ketamine and saline administration in selected areas across hemispheres. This figure presents a series of bar plots for the right and left hemispheres, illustrating beta values extracted in each auditory condition under saline and ketamine administration for 21 ROIs: marmoset vocalizations saline (blue), marmoset vocalizations ketamine (light blue), nonvocal sounds saline (red), nonvocal sounds ketamine (light red), scrambled marmoset vocalizations saline (yellow), and scrambled marmoset vocalizations ketamine (light yellow). The plots display mean beta values with error bars indicating the standard error of the mean. Horizontal lines with stars above the subplots indicate significant differences between saline and ketamine administration within each condition (blue, vocal; red, nonvocal; yellow, scrambled), determined by rmANOVA with post hoc correction and corrected for multiple comparisons: *p < 0.05; **p < 0.01; ***p < 0.001. ROIs represents regions known to be involved in auditory processing in marmosets, selected among the primary auditory cortex, adjacent temporal areas, anterior cingulate area 32, subcortical areas MG, IC, and Th-MD, as well as visual area V1 as a control region.

Figure 5.

Brain network showing differential responses to conspecific vocalizations versus scrambled vocalizations following ketamine and saline administrations. This figure illustrates group functional maps from nine awake marmoset monkeys, showing regions with significant response differences between conspecific vocalizations and scrambled vocalizations under saline administration (top panel) and ketamine administration (bottom panel). The maps are masked by the significant interaction effects between administration and condition. Responses are displayed on both lateral and medial views of the fiducial marmoset cortical surfaces, with subcortical activations presented on coronal slices. White lines demarcate brain regions according to the Paxinos parcellation of the NIH marmoset brain atlas (Paxinos et al., 2013; Liu et al., 2018). The reported regions meet a threshold corresponding to Z test statistics > 1.96 (yellow, greater responses for vocalizations) or <−1.96 (blue, greater responses for scrambled vocalizations; p < 0.05; two-sided), corrected for multiple comparisons using FWE cluster-size correction (α = 0.05).

Figure 6.

Brain network showing differential responses to conspecific vocalizations versus nonvocal sounds following ketamine and saline administrations. This figure illustrates group functional maps from nine awake marmoset monkeys, highlighting regions with significant response differences between conspecific vocalizations and nonvocal sounds under saline administration (top panel) and ketamine administration (bottom panel). The maps are masked by the significant interaction effects between administration and condition. Responses are displayed on both lateral and medial views of the fiducial marmoset cortical surfaces, with subcortical activations presented on coronal slices. Brain regions are delineated according to the Paxinos parcellation of the NIH marmoset brain atlas (Paxinos et al., 2013; Liu et al., 2018). The reported regions meet a threshold corresponding to Z test statistics > 1.96 (yellow, greater responses for vocalizations) or <−1.96 (blue, greater responses for nonvocal sounds; p < 0.05, two-sided), corrected for multiple comparisons using FWE cluster-size correction (α = 0.05).