Auditory Streaming and Prediction in Tinnitus Sufferers

The aim of this study was to determine whether auditory streaming (segregation of incoming sounds into separate sound sources) and the ability to anticipate future auditory events are affected in tinnitus sufferers compared with nontinnitus controls with matched levels of hearing. It was hypothesized that tinnitus would result in abnormal electroencephalography responses to tone deviants and tone omissions compared to controls for frequencies near the pitch of tinnitus, and this should correspond with increased levels of cortical γ and θ oscillatory rhythms.

Sixteen individuals with tinnitus (10 men and 6 women; age, 53.44; SD, 12.92 years) and 14 control participants (8 men and 6 women; age, 50.25; SD, 18.54 years) took part in the study. A modified version of the ABA streaming paradigm, with repeating triplet pattern of two frequencies (A and B) presented as A-B-A, was used to examine deviant-related prediction error. Omission-related prediction errors were examined using a modified version of a tone-omission paradigm. Regions of interest were frontocentral, left frontal, right frontal, and temporal lobes.

A larger N1c waveform was elicited in the absence of any tone deviation within the left primary auditory cortex of tinnitus participants. No differences were present between groups for omissions. The only difference in oscillatory band activity between the two groups in this study was in response to tones 7 semitones different from tinnitus pitch, with significantly lower β-2 band activity present for the tinnitus group, correlating most with activity within the right inferior occipital gyrus.

The findings from this study imply that cortical-level auditory stream segregation is altered among individuals with tinnitus.