Mammals have evolved multiple sensory systems to gather information of external events in parallel streams. The interaction of different sensory input is crucial for forming a collective representation of the environment and facilitating behavioral responses. In cortical regions, multisensory interaction has been commonly observed^1,2. Yet in lower subcortical areas, this has been less extensively investigated and existing studies are mainly focused on the superior colliculus (SC)^3,4, a midbrain structure that is inherently multimodal as it integrates anatomical input from multiple sensory systems. In neighbor to the SC, the inferior colliculus (IC) is a pivotal station in both the ascending and descending pathways of the auditory system. Although evidence has suggested that the IC also receives anatomical input from non-auditory regions, e.g., the retina⁵ and the visual cortex⁶, it is considered fundamentally an auditory only structure. More recently, a few studies found that the IC could represent non-auditory information, such as eye-position or saccade-related signals^7,8. However, it remains unclear how the other sensory input interacts with auditory processing within the IC. In this study, we employed BOLD fMRI to investigate whether visual stimulation can influence auditory processing within the IC.

Animal preparation Adult male rats (n=6, 300-350g, SD strain) were anesthetized with 1.0% isoflurane during fMRI experiment. Auditory stimulation was generated by a magnetic speaker and delivered through custom-made tubes into the right ear of animals⁹. Visual stimulation was generated by a blue (473nm wavelength) DPSS laser and presented to both eyes through a fiber that was placed 1.5cm in front and along the midline of the eyes.

Auditory and visual stimulation First, BOLD responses to the visual stimulation alone were examined by presenting light pulses (1Hz, peak power: ~1.5mW) in a block-design paradigm (20s on and 40s off). Note the visual stimulation here was much stronger than naturalistic situation (<0.2mW). Subsequently, to investigate whether the visual stimulation affects auditory processing in the IC, a broadband noise (sound pressure level: 90dB) was presented in a block-design paradigm (20s on and 50s off) while the light pulse stimulation was presented from 10s before to 10s after every second sound-on period (Figure 1).

fMRI acquisition and analysis All fMRI data was acquired at 7T using GE-EPI (FOV=32×32mm², matrix=64×64, α=56°, TE/TR=20/1000ms, 12 contiguous slices with 1mm thickness). Data were first realigned, co-registered, in-plane smoothed and high-pass filtered before the standard GLM analysis was applied to identify significant BOLD responses (p<0.05, corrected for FWE).

Figure 2 shows the BOLD responses in the visual system, including the SC, lateral geniculate nucleus (LGN) and visual cortex (VC), evoked by the light pulse stimulation. No significant BOLD response was observed in the IC.

Figure 3 compares the noise-evoked IC BOLD responses with and without the presence of light pulse stimulation. During light stimulation, the noise response within the IC became weaker (p<0.05, paired t-test on averaged β value). This indicates that the strong visual stimulation inhibits IC noise sensation.

Our imaging results show that the presence of a strong visual stimulation does not directly induce BOLD signaling in the IC, yet it decreases the subsequent noise-evoked BOLD responses throughout the IC. This clearly indicates that the visual stimulation generates an inhibitory effect on the IC responsiveness to acoustic noise. More importantly, these results reveal that multisensory interaction can occur in the early stage of sensory processing - in a subcortical nucleus that is commonly considered unisensory⁷. Such early interactions can likely improve the accuracy of sensory processing in upstream structures. Note that the visual stimulation employed in this study was much stronger than naturalistic situation, potentially noxious or even causing pain. With its onset preceding that of the noise stimulation, its inhibitory effect on the following noise-evoked IC responses was actually expected, as the brain may be more oriented to resolving this noxious stimulus¹⁰. Interestingly, previous studies showed that input from a different modality can either enhance or reduce the saliency of a noxious/pain stimulus¹⁰. This echoes our present results that indicate a potentially noxious stimulus can also affect the brain responses to other innocuous stimulation. The visual influences on auditory midbrain processing shown in this study can be underlain by multiple neural pathways. For example, the SC likely relays the visual input to the IC. Our recent fMRI study indicates that the VC normally provides facilitatory influence to the IC⁹, yet it is still possible that multisensory activities integrated in the VC can mediate the IC responses. In future studies, we will further investigate how the audiovisual interaction is dependent on the intensity and the relative timing of the two sensory inputs.