Figure 1. (a) Section through the human head showing auditory canal, middle ear and inner ear. (b) The cochlea is a fluid-filled tube coiled into a three-turn spiral. (c) In cross-section it has three compartments, with the hair cells arranged in rows running the length of the cochlea within the organ of Corti. Sound vibrations cause displacement of hair cell stereocilia relative to the overlying tectorial membrane.

Figure 2. Scanning electron micrograph of the organ of Corti with the tectorial membrane removed. Three rows of outer hair cells and the single row of inner hair cells can be seen, which extend for the length of the cochlea. Each hair cell is capped by a cluster of stereocilia; small movements of these ‘hairs’ are responsible for sensing sound vibrations. Inner hair cells sense sound and transmit this information via the type I spiral ganglion afferents; the outer hair cells transmit information via smaller type II afferent fibres and are thought to serve as the cochlear amplifier. Micrograph kindly provided by Dr. David Furness, Keele University, UK.

Figure 3. The encoding of sound by the cochlea. (a) Sensory transduction. This figure shows the sequence of events leading to the transduction of sound into a train of action potentials. Sound enters the auditory canal, moving the tympanic membrane, which vibrates the three ossicular bones in the middle ear. The stapes transmits this amplified signal to the oval window in the cochlea. The resulting travelling wave displaces the basilar membrane and causes bending of the inner hair stereocilia (IHCs). This movement changes the activity of ion channels, causing depolarization or hyperpolarization of the hair cell membrane potential, which in turn modulates calcium influx and transmitter release onto the type I spiral ganglion cell afferents. The resultant action potentials propagate into the cochlear nucleus via the 8th nerve. (b) Tonotopy. An uncoiled cochlea is shown (see Figure 2b) with two travelling waves represented above. A high-frequency (orange) sound resonates the base of the cochlea and generates action potentials in the spiral ganglion cells projecting to that region. A low-frequency sound (green) activates a region of the cochlea close to the apex and this stimulates a different group of afferents. The action potential response is similar in both fibres (below), but the different cochlear positions of the hair cells that respond to a given frequency of sound allows this information to be encoded into a particular set of afferent fibres. This tonotopic relationship is preserved at many levels of auditory processing. (c) Characteristic or best frequency. The tonotopic relationship means that each afferent fibre responds best (with the lowest threshold) to sound of a given frequency. Each afferent responds to sound of a characteristic frequency. (d) The volume of a sound is thought to be encoded by the number of action potentials. This probably includes action potentials both within and around a particular frequency range; that is, as the volume increases, additional afferents are recruited (volley principle).

Figure 4. Ascending auditory pathway. The three levels of central auditory processing are illustrated. The brainstem auditory pathway consists of the cochlear nucleus and superior olivary complex; the midbrain contains the inferior and superior colliculi and the auditory thalamic relay (the medial geniculate nucleus), which projects to the primary auditory cortex (AI) and the surrounding secondary auditory centres. Abbreviations: CN, cochlear nucleus; DCN, dorsal cochlear nucleus; aVCN, anterioventral cochlear nucleus; DAS, dorsal acoustic stria; IAS, intermediate acoustic stria; SOC, superior olivary complex; LSO, lateral superior olive; MNTB, medial nucleus of the trapezoid body; MSO, medial superior olive; PON, periolivary nuclei; LL, lateral lemniscus; DNLL, dorsal nucleus of the lateral lemniscus; INLL, intermediate nucleus of the lateral lemniscus; VNLL, ventral nucleus of the lateral lemniscus; IC, inferior colliculus; CIC, central nucleus of the inferior colliculus; BIC, brachium of the inferior colliculus; DCx, dorsal cortex of the inferior colliculus; LN, lateral nucleus of the inferior colliculus; SC, superior colliculus; MGN, medial geniculate nucleus; dMGN, mMGN, vMGN, dorsal, medial and ventral MGN, respectively; AI, primary auditory cortex; Py, pyramids; Tz, trepezoid body.

Figure 5. The cochlear nucleus. The cochlear nucleus (CN) is located on the lateral edges of the brainstem, below the cerebellum; this diagram shows the nucleus as viewed from the side. The spiral ganglion axons that carry the input from the cochlea enter the CN via the 8th nerve. The axons maintain their tonotopic relationship and bifurcate into anterior-ventral and rostrodorsal pathways. The dorsal (DCN) and ventral (VCN) cochlear nuclei form the first central processing of the auditory input. The DCN is a laminar structure concerned with the spectral properties of a sound, while the VCN forms the major inputs to the binaural brainstem pathways concerned with sound localization. The locations of the principal cells are indicated by the symbols and by the dashed lines.

Figure 6. Cells of the anterioventral cochlear nucleus have characteristic morphologies and firing properties. Firing properties are depicted here as poststimulus time histograms (PSTH). These represent the rate of firing plotted against time from the start of a sound stimulus (as indicated by the filled bar). (a) The primary afferents or spiral ganglion axons give a short burst of action potentials at the start of a sound and then a sustained discharge for the duration of the sound. This is known as a ‘primary’ response pattern. (b) Bushy cells are so called because they have just few dendrites, which tend to form a bush-like appearance. They receive their primary afferent input from the endbulbs of Held and follow the firing pattern of the primary afferents very closely; hence they are often referred to as having a ‘primary-like’ firing pattern. (c) Stellate or multipolar cells have a more conventional neuronal appearance, they respond to a tone with short bursts of action potentials with distinct, precisely timed pauses, which are called ‘chopper’ responses. (d) Octopus cells possess thick sparsely branched dendrites that tend to originate from one pole of the soma. They fire predominantly at the onset of a stimulus.

Figure 7. Descending auditory pathways. Each level of the auditory pathway possesses both ascending and descending pathways. The descending projection to the hair cells originates in the superior olivary complex of the brainstem (from the same regions involved in sound localization – see below) and forms an important part of the sensitivity control of the cochlear sense organ. These efferent axons form the olivocochlear bundle (OCB) and originate from either medial or lateral divisions of the superior olivary complex (SOC). The lateral division makes presynaptic connections with primary efferent terminals on the inner hair cells, while the medial division makes direct contact with the outer hair cells.

Figure 8. The brainstem binaural auditory pathway. This diagram shows a transverse section of the brainstem at the level of the seventh nerve. The 8th nerve enters the aVCN and excites the bushy cells. The globular bushy cells send a large diameter axon that crosses the brainstem in a tract called the trapezoid body. The terminals of these axons form the calyx of Held giant synapses on the cell bodies of principal cells in the medial nucleus of the trapezoid body (MNTB). The MNTB in turn gives an inhibitory projection to both the medial superior olive (MSO) and the lateral superior olive (LSO). The MSO receives a bilateral excitatory input from the spherical bushy cells and is responsible for interaural time difference computation. The LSO receives an excitatory input from the ipsilateral bushy cells, which is integrated with the inhibitory input from the MNTB as the first stage of interaural level difference computation.

Figure 9. The Jeffress model for interaural time difference computation in the medial superior olive (MSO). Three MSO cells are shown. Each receives excitatory synaptic inputs from both cochlea onto opposite dendrites. The path length of the axons varies along the length of the nucleus, the longer path lengths will give a longer latency response and hence act as a delay line. The cell labelled ‘Right’ receives the longest delay-line response from the right ear and the shortest delay-line response from the left ear. At zero time a sound originating from the far right-hand side is heard, it generates an excitatory postsynaptic potential (EPSP) in all the MSO neurons, but with a range of latencies. About 600 s later the same sound reaches the other ear and triggers EPSPs from the left side. This sound crosses the brainstem and activates the same population of cells. The latency of the left and right EPSPs is the sum of their respective conduction time (+c, for the sound to cross the head) and the delay introduced by the delay-line (+dl). Only those MSO cells that receive coincident left and right EPSPs will be able to generate an action potential as shown in the lower inset. Cell 3 will only receive a coincident input when the sound originates from the right; this is signalled by firing an AP. By the same mechanism cell 1 will only receive coincident EPSPs when the sound originated from the far left. By setting up a complete range of conduction delays, the 180° range of azimuth locations can be specified in this model.

Figure 10. Sound localization in the barn owl. The barn owl uses ITDs to discriminate azimuth location and ILD to detect vertical location. The auditory field of ‘view’ is shown for timing isotherms (ITD, upper graph measured in microseconds), and intensity (ILD, lower graph measured in dB). Combining these two maps of auditory space pinpoints the origin of the sound as indicated in the overlaid map (centre).

Figure 11. Low-threshold potassium currents are responsible for enabling the MNTB neuron to faithfully follow the pattern of activity in its calyx of Held. Dendrotoxin-I (DTx) degrades the transmission of action potentials across the calyx of Held/MNTB synapse. Using whole-cell patch clamp recordings from the MNTB neuron in an in vitro brain slice preparation, the synaptic responses can be evoked by electrical stimulation of the axons in the trapezoid body, as shown in the inset. The postsynaptic responses are shown below, with single stimuli on the left (a, c) and a train of stimuli at 100 Hz shown on the right (b, d). Normally, a single presynaptic action potential generates one postsynaptic action potential (superimposed on the EPSP). In the same recording a 100 Hz presynaptic stimulus generates a train of action potentials in the postsynaptic MNTB neuron, which closely follows the stimulus train. Addition of 100 nmol L^–1 dendrotoxin-I blocks a low threshold potassium current and allows the giant EPSP to generate a train of action potentials in response to a single stimulus. The response to a 100 Hz train is now totally confused with many more action potentials being generated. This evidence and other voltage-clamp data demonstrate that low-threshold potassium currents play an important role in damping excitability, so that the EPSP can generate one precisely timed action potential rather than multiple action potentials that are not coherent with the timing of the auditory input. Reproduced with permission from the Journal of Neuroscience (Brew and Forsythe, 1995).