Abstract
Temporal processing of auditory data plays a crucial role in our
proposed model of tinnitus development through stochastic resonance
(SR). The model assumes a physiological mechanism optimizing auditory
information transmission (as quantified by autocorrelation (AC)
analysis) into the brain by adding the optimal amount of neuronal noise
to otherwise subthreshold signals. We hypothesize that this takes place
at the second synapse of the auditory pathway in the dorsal cochlear
nucleus (DCN). We propose that after hearing loss, this neuronal noise
is increased in the affected frequency-band to improve hearing
thresholds at the cost of upward propagation of this added noise, which
finally may be perceived as tinnitus. We already showed the improvement
of hearing thresholds in a large population of patients. Until now, we
did not investigate the differences in hearing thresholds based on the
biological constraints of early auditory temporal processing (phase
locking) that is only possible up to frequencies of 5 kHz. In this
report, we grouped our patient database (N=47986) according to tinnitus
pitch (TP) of below (TP<5kHz) or above (TP>5kHz)
the 5 kHz limit or having no tinnitus (NT) and compared their mean
audiograms. We found that TP<5kHz patients showed
significantly better hearing thresholds than all other patient groups
independent of age. No improvement was seen for TP>5kHz
patients who even showed worse thresholds than NT patients for high
frequencies. These results are further evidence for our SR model of
tinnitus development and the existence of AC analysis at the level of
the DCN.