Page 36 - Volume 12, Issue 2 - Spring 2012

Fig. 1. Any sound can be mathematically separated into a slowly varying temporal envelope (ENV) and a rapidly varying temporal fine structure (TFS) component. Each of these temporal components of sound has been suggested to have important perceptual roles. Recent evidence suggests that many listeners with sensorineural hearing loss have a specific deficit in their ability to use TFS cues.
Temporal coding in individual AN fibers
To characterize the effects of SNHL on temporal coding in the auditory periphery, our experiments involve recording spike-train responses from chinchillas with normal hearing and from those with a noise-induced hearing loss (e.g., Kale and Heinz, 2010). Calibrated stimuli are presented via a mon- aural closed-field acoustic system. Glass micropipettes are used to record extracellularly from individual AN fibers as they exit the internal auditory canal, and action potential are timed with 10-μs accuracy. The noise-induced hearing loss results in elevated thresholds of AN
fibers by ~40 dB, along with broad-
ened tuning that provides a good
model of moderate SNHL. The excel-
lent sensitivity and sharp tuning that
benefit normal-hearing listeners are
thought to relate to cochlear amplifi-
cation provided to low-level sounds
by outer-hair-cell (OHC) electro-
motility (Robles and Ruggero, 2001;
Liberman et al., 2002). Inner hair cells
(IHCs) are the sensory transducers
that convert basilar-membrane
motion into neural responses.
Numerous physiological studies of
SNHL support a framework in which
damage to both OHCs and IHCs is
important for the neural coding of
sound (Liberman and Dodds, 1984;
Heinz and Young, 2004).
Furthermore, recent studies suggest
that significant synaptic degeneration
can occur following noise exposure,
even in cases of “temporary” hearing
loss (Kujawa and Liberman, 2009).
coding in AN fiber responses to a variety of stimuli, ranging from simple pure tones, to more complex but narrowband stimuli such as sinusoidally amplitude modulated (SAM) tones and single-formant stimuli, to complex broadband sounds such as noise and speech. The strength of temporal coding to periodic signals is most often quantified using the vector- strength metric, which ranges from 0 to 1 to indicate the degree to which spikes phase lock to a particular phase of the periodic cycle. For non-periodic stimuli, recently developed techniques that use shuffled correlograms (similar to auto- and cross-correlation functions in signal processing) provide met- rics that can be used to quantify temporal coding strength to the TFS and ENV components of the sound (Joris, 2003; Louage et al., 2004).
For both the simple and complex stimuli we have con- sidered in quiet conditions, our results have not shown any decrease in the strength of TFS coding in AN fibers. We interpret these results as suggesting that the fundamental ability of individual AN fibers to encode the rapidly varying TFS components of sound is not degraded with noise- induced hearing loss. Thus, the most straightforward hypothesis based on the perceptual studies does not appear to be correct. This result could be interpreted as suggesting that a peripheral TFS coding deficit does not exist, and is per- haps created at more central levels of the auditory system. However, if we consider the fact that listeners with SNHL often do not actually have so much trouble in quiet situa- tions, but rather struggle primarily in noisy conditions, then it is always possible that our studies of temporal coding in quiet situations have been missing something. In fact, very recent studies in our lab suggest that temporal coding of even pure tones in background noise is degraded following SNHL.
Fig. 2. The temporal structure of sound is transmitted to the central nervous system through 30,000 auditory nerve (AN) fibers, each of which is tuned to a specific characteristic frequency (CF) determined by its location along the basi- lar membrane. Action potentials (or spikes) occur at times that are synchronized (or “phase locked”) to the temporal fine structure and/or the temporal envelope of band-pass portions of the sound that enters the ear. Our experiments and computational modeling have been addressing three hypothesized physiological correlates for perceptual temporal fine-structure (TFS) deficits: (1) degraded TFS coding within individual AN fibers, (2) degraded spatio-temporal (across-fiber) cues in terms of reduced traveling-wave delays, and (3) whether cochlear transformations between acoustic TFS and neural envelope (ENV) complicate interpretations that perceptual TFS deficits are truly based on degraded physiological TFS coding.(Figure modified by Eric Young from Sachs et al. (2002), with permission from Springer and the Biomedical Engineering Society.)
We have characterized temporal
Physiological Correlates of Perceptual Deficits 35