Building a meaningful sound object

Abstract

This first lecture began with an update on recent data concerning the mechanical response of the cochlea to sound, as measured by interferometry and coherent optical tomography techniques. These data provide major objections to certain ideas often accepted by models of sound stimulation of the cochlea. They show that the reticular lamina, the surface of the cochlea's sensory neuroepithelium, plays a key role in sound detection. One such study (by Tianying Ren, Wenxuan He and David Kemp) [1] shows that the vibration of the reticular lamina is amplified in vivo by a factor of 30 compared with that of the basilar membrane.

Unlike the basilar membrane, the reticular lamina undergoes active amplification of its vibration over the entire region from the base of the cochlea to the peak of the cochlear wave. This vibration is thus dominated by the " active " contribution made by the outer hair cells (OHCs), and reflects very little of the vibration of the basilar membrane, the mechanical support of the auditory organ. A second study (by the groups of Alfred Nuttal and Anders Fridberger) [2], which we have discussed in greater detail, highlights a possible role for the reticular lamina (and not the basilar membrane) in the envelope extraction of the cochlear vibration waveform. This envelope extraction - an essential step in the coding of sounds for speech perception, as shown by Robert Shannon's work - could therefore be the result of a mechanical mechanism within the cochlea, and not only of further processing in the auditory pathways, as proposed so far.

In a second part of the lecture, we discussed the transition from the periphery to the higher-level relays of auditory processing, and more specifically the formation of so-called " auditory streams ", in which sound information is separated into different channels each containing distinct sound cues that are processed separately. This occurs in particular when listening to a sound from multiple sources (two voices, or two distinct objects producing different but overlapping sounds) : the textbook case of the cocktail party.

In this context, we have discussed a recent study by John Middlebrooks' group [3] based on electrophysiological recordings made on neurons in different structures of the auditory pathways in rats and cats. These analyses clearly show how the activity of neurons in response to two sounds coming from different directions changes, as one moves up the auditory pathway, from a response that encodes a mixture of the two sounds to one in which distinct sounds have distinct (or segregated) representations : this segregation, still very weak in the central nucleus of the inferior colliculus, begins to appear in the nucleus of the brachium, then increases in the medial geniculate body of the thalamus, and becomes very clear in the primary auditory cortex.

Another, very different, but equally essential aspect was discussed : how the plasticity of the auditory system is adjusted over the course of life by the various experiences and behavioral needs. We are interested in the results of a study conducted by Robert Froemke's group [4], which analyzes the role of oxytocin (a pituitary hormone involved in parturition and lactation) in the development of a well-characterized behavior in mice, that of maternal rescue of nestlings lost from the nest, in response to the perception of their distress calls. Through a combination of electrophysiological and behavioral experiments, the authors demonstrate a mechanism by which oxytocin, when combined with the presentation of distress vocal cues (reproduced using loudspeakers), induces this rescue behavior in virgin mice from an initial " naïve " state in which the mice do not respond to distress calls. Oxytocin rapidly alters the state of the brain, inducing a transition from a disordered response of inhibitory neurons (parvalbumin and somatostatin) in the auditory cortex to a much more robust and synchronized response characterized by a fine balance between the excitatory and inhibitory activities of these neurons. This transition involves type 2 (OXTR-2) oxytocin receptors, which are expressed by the inhibitory neurons of the auditory cortex (parvalbumin and somatostatin). Surprisingly, OXTR-2 receptors are present only in the left auditory cortex, and virtually not in the right. There is thus a cortical lateralization of a receptor playing a critical role in the detection of distress calls. All in all, this is a fine example of a general neuromodulation mechanism for social behavior.

Finally, we mention a study from the same group [5], which highlights another example of attention modulation of auditory perception during a behavioral task performed by mice. This modulation also involves the inhibitory neurons of the primary auditory cortex, this time activated by the cholinergic synapses formed by the axonal terminals of Meynert's nucleus basalis (or magnocellular nucleus basalis) in the basal proencephalon. These different examples show that sound perception is largely modulated by context and associated information to form physiologically meaningful percepts or sound objects, a modulation that plays an essential role in behavioral and cognitive tasks.