Determining the individual tinnitus frequency of tinnitus patients is a standard requirement
in clinical settings and is important for research and treatment alike. However, the
method used to reliably measure this frequency is controversial.
(Dominant) tinnitus frequency was measured in patients (N = 117) suffering from chronic
tinnitus using an iPod-based recursive matching procedure. To get an estimate for
individual variability measurements were performed twice a day on five consecutive
days. For this procedure, the frequency range (1 to 16 kHz) is bisected into two equally
large subintervals of which the patient has to decide being more similar to their
tinnitus. Depending on the outcome, new subintervals are calculated. The smallest
step size was 1/12 octave.
Mean and median across measures were almost identical. The mean standard deviation
turned out to be less than 1/2 octave. In a data driven approach we determined patients
with particularly high variability. These were patients (N = 5) whose standard deviation
(SD) across the ten measures differed more than two SD (about 1/2 octave) from the
mean SD of the whole sample. Inspecting these cases on an individual basis suggested
high variability on the first two days as well as the occurrence of octave confusion,
however with sometimes more than three octaves. Reducing systematically the number
of measurements resulted in slightly more patients with high deviation of the mean
and higher variability across the whole sample.
Taken together, repetitive recursive matching appears to be a promising method to
reliably estimate the individual tinnitus frequency for both research and clinical
purposes alike.
