BOLD fMRI study of ultrahigh frequency encoding in the inferior colliculus
Introduction
Many vertebrates have evolved the ability to communicate with ultrahigh frequency (UHF) vocalizations to decrease the likelihood of detection by predators (Brudzynski, 2009, Brudzynski, 2013, Gadziola et al., 2012, Portfors, 2007, Takahashi et al., 2010). For example, male mice emit vocalizations with frequencies up to 110 kHz when encountering female mice, and female mice produce frequency-modulated upsweeps above 60 kHz when their pups are removed from the nest (Portfors, 2007). Rats emit short chirping 50 kHz vocalizations during juvenile play, male agonistic behaviors and sexual behaviors (Brudzynski, 2009, Portfors, 2007). Recently, there is increasing interest in studying the neural mechanisms underlying the learning, production and representation of such vocalizations (Arriaga and Jarvis, 2013, Woolley and Portfors, 2013) because it may improve our understanding of complex sound (e.g., speeches) processing in humans (Rauschecker and Scott, 2009).
One important question is how the auditory system encodes UHFs, in order to process vocalizations that contain them (Romanski and Averbeck, 2009, Woolley and Portfors, 2013). In the auditory system, the majority of neurons are sensitive to only a narrow spectrum of sounds centered about a characteristic frequency (CF). The topographic alignment of the CFs (i.e., the tonotopy) is a fundamental organizing principle maintained from the cochlea to the auditory cortex (Ehret and Fischer, 1991, Malmierca, 2003). For example, in the primary auditory nucleus in the midbrain, called the inferior colliculus (IC), which is a compulsory relay for all ascending auditory projections (Malmierca, 2003) and a region that efficiently encodes vocalizations (Holmstrom et al., 2010), the tonotopic organization is a basic feature in its central nucleus (CNIC), where the CFs of neurons run from low to high along the dorsolateral-ventromedial dimension (Cheung et al., 2012a, Cheung et al., 2012b, De Martino et al., 2013, Malmierca et al., 2008, Ress and Chandrasekaran, 2013, Schreiner and Langner, 1997, Yu et al., 2005). In the cortical regions of the IC, the frequency organization is usually non-tonotopic, with neurons showing complex frequency tuning properties, such as broadband or multipeak receptive fields (Duque et al., 2012, Hernandez et al., 2005, Malmierca et al., 2011, Stebbings et al., 2014). Given the importance of UHFs in social communications, they are expected to occupy a substantial area in the tonotopic map of the CNIC, or the frequency receptive fields of neurons in non-tonotopic IC regions. But surprisingly, there is limited representation of UHFs in the auditory system of many species. In mouse IC, neurons with CF higher than 60 kHz are rarely found and occupy a very small tonotopic area compared to those with lower CFs (Stiebler and Ehret, 1985, Yu et al., 2005). Similarly, in rat IC, representations of frequencies higher than 40 kHz are seldom delineated (Cheung et al., 2012a, Cheung et al., 2012b, Malmierca et al., 2008). This mismatch immediately poses a basic question how UHFs are encoded in the auditory system of these species.
Current evidence of UHF encoding is found in electrophysiological studies that probed neuronal responses to behaviorally relevant sound in the auditory system. In the mustached bat, neurons in the auditory cortex as well as the IC were found to display excitatory or inhibitory responses when two tones with different frequencies were presented simultaneously (two-tone paradigm) with one of them in the frequency range of bat biosonar or vocalizations (Leroy and Wenstrup, 2000, Mittmann and Wenstrup, 1995, Suga et al., 1979). This neuron property, i.e., responding best to the combination of two different frequency components, was called combination sensitivity (Leroy and Wenstrup, 2000, Portfors and Felix, 2005, Portfors and Wenstrup, 2002). Combination sensitivity was believed to allow neural representation of spectral and temporal information conveyed in echolocation signals or social vocalizations, and was suggested to be a common mechanism for processing complex sound (Portfors, 2004, Portfors and Wenstrup, 2002). Combination sensitive neurons were also found in mouse IC (Felix and Portfors, 2007, Portfors and Felix, 2005). In both bat and mouse, the facilitatory or inhibitory frequencies were often found to be in the UHF range (Portfors and Felix, 2005, Portfors and Wenstrup, 2002), indicating that combination sensitivity might be a neural mechanism underlying UHF encoding shared by different mammals.
Recently, it was observed that neurons in the central auditory system could also respond to the two-tone paradigm by exploiting cochlear nonlinearities. When simultaneously stimulated with two different frequencies (f1 and f2, f2 > f1), the cochlear characterized by mechanical nonlinearity will generate distortion tones with frequencies at e.g. f2–f1 (quadratic) (Robles et al., 1991). The distortion products can propagate along the cochlea to their CF site on the basilar membrane and then excite neurons with the same CF at higher levels of the auditory pathway, including the cochlear nucleus (Faulstich and Kossl, 1999), the IC (Abel and Kossl, 2009, McAlpine, 2004) and the auditory cortex (Purcell et al., 2007). It has been shown that at least in some cases, cochlear distortion products were responsible for the neuronal activities that were thought to arise from combination sensitivity (McAlpine, 2004). It was also noticed that some IC neurons could respond to the cochlear distortion of UHF tones (Portfors and Roberts, 2014, Portfors et al., 2009), suggesting that cochlear distortion could also be a mechanism underlying UHF encoding.
By demonstrating the neuronal responses in multiple auditory nuclei that arise from combination sensitivity and cochlear distortion, the above-mentioned studies provided important clues for understanding UHF encoding. However, it remains unclear how prevalent these mechanisms are in an auditory nucleus, since relatively small numbers of neurons were usually recorded in those electrophysiological studies. Furthermore, the previous studies were limited by the large variation of frequency values used in the two-tone paradigm. Since different neurons display different tuning properties, one of the two frequencies was usually fixed at the CF or limited in the frequency range of a vocalization that the neuron was sensitive to (Portfors and Felix, 2005, Portfors et al., 2009, Portfors and Wenstrup, 2002). This strategy could effectively reduce the large number of combinations to test, but also induced large variation in the frequency values tested in a study. With such limitations, it was not addressed whether UHFs are encoded in a generally organized way in the auditory system as like the tonotopic encoding of lower frequencies is, or only particular UHFs can be encoded by some specialized neurons. Moreover, it remains unknown how the two mechanisms are combined by the auditory system in processing UHFs, e.g. whether they are adopted by spatially different neuron populations or the same set of neurons at different temporal stages.
Blood-oxygen-level-dependent (BOLD) functional magnetic resonance imaging (fMRI) (Ogawa et al., 1990) with large field of view is a non-invasive technique that can measure the hemodynamic responses as neural correlates (Logothetis et al., 2001) in multiple brain structures with relatively high spatial and temporal resolutions. Recent development in BOLD fMRI has been demonstrated for studying the rodent auditory system (Cheung et al., 2012a, Cheung et al., 2012b, Gao et al., 2014, Lau et al., 2013, Lau et al., 2015, Yu et al., 2009, Zhang et al., 2013), along with previous auditory fMRI studies conducted on humans (Barton et al., 2012, De Martino et al., 2013, Sigalovsky and Melcher, 2006), primates (Baumann et al., 2011, Kayser et al., 2007, Tanji et al., 2010) and songbirds (Boumans et al., 2007, Van Meir et al., 2005, Voss et al., 2007).
In this study, BOLD fMRI was applied to investigate the responses to UHF sounds in rat auditory system. Rat UHF vocalizations, single UHF pure tones, and pairs of UHF tones interleaved with pure tones equivalent to their quadratic distortion products were presented to the animals. The UHF values were systematically varied between 45 kHz and 75 kHz with 10 kHz steps, in order to examine the organization of UHF representation that may potentially exist. The findings were mainly focused on the IC and lateral lemniscus (LL), a subcollicular auditory nucleus that projects to the IC (Malmierca, 2003).
Section snippets
Animal preparation
All animal experiments were approved by the local animal research ethics committee. Animals were prepared for fMRI experiments as described in our previous studies (Chan et al., 2010, Cheung et al., 2012a, Cheung et al., 2012b, Gao et al., 2014, Lau et al., 2011, Lau et al., 2013, Lau et al., 2015, Zhang et al., 2013, Zhou et al., 2012, Zhou et al., 2014). Briefly, male Sprague-Dawley rats (300–350 g, N = 10) were used in this study. They were initially anesthetized with 3% isoflurane and then
Results
Fig. 3C displays the activation (β value) map of the vocalizations from a representative animal, overlaid on the T2W anatomical images. Rat brain templates that correspond to the center of each slice are also presented to identify major auditory structures (Paxinos and Watson, 2005). The left IC and LL (contralateral to the side of stimulation) were activated robustly by the vocalizations. Fig. 3D presents the map computed from the averaged fMRI images of the ten animals. Activated structures
Discussion
The present fMRI results demonstrated that UHFs are represented in the auditory system with different mechanisms from tonotopy. The strong BOLD responses in the IC, LL and MGB to UHF vocalizations demonstrated that the auditory system is highly sensitive to UHFs. However, BOLD responses to single UHF tones, which would be naturally expected from this sensitivity, were not observed. Together with previous electrophysiological studies (Kelly et al., 1991, Malmierca et al., 2008), this indicated
Conclusion
The present fMRI study demonstrated the robust sensitivity of the auditory system to UHF vocalizations, despite limited representation of single UHF pure tones in tonotopy. More importantly, with BOLD responses to pairs of UHF pure tones, the present study indicated that combination sensitivity and cochlear distortion are employed to represent UHFs by large, but spatially distinctive neuron populations, in the ventromedial and dorsolateral side of the IC, respectively. The sophisticated
Abbreviations
- UHF
ultrahigh frequency
- CF
characteristic frequency
- IC
inferior colliculus
- LL
lateral lemniscus
- MGB
medial geniculate body
- BOLD
blood-oxygen-level-dependent
- fMRI
functional MRI
- T2W
T2 weighted
- GLM
general linear model
Acknowledgments
The authors would like to thank Professor Dan H. Sanes from New York University for his insightful comments on the manuscript. This work was supported in part by Hong Kong Research Grants Council (HKU7837/11M), Croucher Foundation for Senior Researcher Fellowship (EXW) and Lam Woo Foundation for Endowed Professorship (EXW).
References (82)
- et al.
Mouse vocal communication system: are ultrasounds learned or innate?
Brain Lang.
(2013) Ethotransmission: communication of emotional states through ultrasonic vocalization in rats
Curr. Opin. Neurobiol.
(2013)- et al.
Functional MRI of postnatal visual development in normal and hypoxic-ischemic-injured superior colliculi
NeuroImage
(2010) - et al.
BOLD fMRI investigation of the rat auditory pathway and tonotopic organization
NeuroImage
(2012) - et al.
High fidelity tonotopic mapping using swept source functional magnetic resonance imaging
NeuroImage
(2012) - et al.
Neuronal activity and tonotopy in the auditory system visualized by c-fos gene expression
Brain Res.
(1991) - et al.
Excitatory, inhibitory and facilitatory frequency response areas in the inferior colliculus of hearing impaired mice
Hear. Res.
(2007) - et al.
The inferior colliculus is involved in deviant sound detection as revealed by BOLD fMRI
NeuroImage
(2014) - et al.
The inferior colliculus of the rat: a quantitative analysis of monaural frequency response areas
Neuroscience
(2005) - et al.
Sound frequency and binaural response properties of single neurons in rat inferior colliculus
Hear. Res.
(1991)
BOLD responses in the superior colliculus and lateral geniculate nucleus of the rat viewing an apparent motion stimulus
NeuroImage
Long-term, passive exposure to non-traumatic acoustic noise induces neural adaptation in the adult rat medial geniculate body and auditory cortex
Neuroimage
The structure and physiology of the rat auditory system: an overview
Int. Rev. Neurobiol.
Computer-assisted 3-D reconstructions of Golgi-impregnated neurons in the cortical regions of the inferior colliculus of rat
Hear. Res.
Combination-sensitive neurons in the inferior colliculus
Hear. Res.
Auditory thalamus responses to guinea-pig vocalizations: a comparison between rat and guinea-pig
Hear. Res.
The dominant role of inhibition in creating response selectivities for communication calls in the brainstem auditory system
Hear. Res.
Spectral integration in the inferior colliculus of the CBA/CaJ mouse
Neuroscience
Mismatch of structural and functional tonotopy for natural sounds in the auditory midbrain
Neuroscience
Excitatory and facilitatory frequency response areas in the inferior colliculus of the mustached bat
Hear. Res.
Over-representation of species-specific vocalizations in the awake mouse inferior colliculus
Neuroscience
Effects of sound level on fMRI activation in human brainstem, thalamic and cortical centers
Hear. Res.
The auditory corticocollicular system: molecular and circuit-level considerations
Hear. Res.
How challenges in auditory fMRI led to general advancements for the field
NeuroImage
Effect of sound intensity on tonotopic fMRI maps in the unanesthetized monkey
NeuroImage
Spatiotemporal properties of the BOLD response in the songbirds' auditory circuit during a variety of listening tasks
NeuroImage
Representation of the purr call in the guinea pig primary auditory cortex
Hear. Res.
Conserved mechanisms of vocalization coding in mammalian and songbird auditory midbrain
Hear. Res.
Functional magnetic resonance imaging of sound pressure level encoding in the rat central auditory system
NeuroImage
Brain resting-state functional MRI connectivity: morphological foundation and plasticity
NeuroImage
Sensitive response to low-frequency cochlear distortion products in the auditory midbrain
J. Neurophysiol.
Orthogonal acoustic dimensions define auditory field maps in human cortex
Proc. Natl. Acad. Sci. U. S. A.
Spectral determination of responses to species-specific calls in the dorsal nucleus of the lateral lemniscus
J. Neurophysiol.
Orthogonal representation of sound dimensions in the primate midbrain
Nat. Neurosci.
Neural representation of spectral and temporal features of song in the auditory forebrain of zebra finches as revealed by functional MRI
Eur. J. Neurosci.
Communication of adult rats by ultrasonic vocalization: biological, sociobiological, and neuroscience approaches
ILAR J.
Spatial organization of frequency preference and selectivity in the human inferior colliculus
Nat. Commun.
Topographic distribution, frequency, and intensity dependence of stimulus-specific adaptation in the inferior colliculus of the rat
J. Neurosci.
Neuronal response to cochlear distortion products in the anteroventral cochlear nucleus of the gerbil
J. Acoust. Soc. Am.
Social vocalizations of big brown bats vary with behavioral context
PLoS One
“Sparse” temporal sampling in auditory fMRI
Hum. Brain Mapp.
Cited by (13)
BOLD-fMRI in the mouse auditory pathway
2018, NeuroImageProcessing of Ultrasonic Vocalizations in the Auditory Midbrain of Mice
2018, Handbook of Behavioral NeuroscienceCitation Excerpt :Further support for the use of distortion products in encoding ultrasonic stimuli comes from fMRI studies in rats. A strong BOLD fMRI signal has been observed in tonotopic regions in the IC of rats equal to the difference between two high-frequency pure tones (e.g., 20 kHz regions activated by 45 and 65 kHz pure tones), suggesting that neurons tuned to low frequencies can be evoked by distortion products (Gao et al., 2015). Thus, indirect evidence in the IC of both mice and rats suggests that low-frequency distortion products are created and used for encoding USVs.