@inproceedings {Kato1544_2016,
	year = {2016},
	author = {Kato, Yoko and Klump, Georg Martin},
	title = {Effects of Masker Bandwidth, Mistuning and Binaural Cues on the Representation of Components in a Harmonic Tone Complex in the Gerbil Inferior Colliculus},
	booktitle = {Assoc. Res. Otolaryng. MidWinter Meeting (ARO)},
	URL = {http://c.ymcdn.com/sites/www.aro.org/resource/resmgr/Abstract_Archives/UPDATED_2016_ARO_Abstract_Bo.pdf},
	abstract = {Harmonicity is an important cue to group components of
sounds from a specific source in the acoustic environment.
If a component of a harmonic sound is slightly shifted in
frequency, that tone becomes distinct from other harmonics
components in a complex with a specific fundamental
frequency. In a human psychophysical study, Klein-Hennig
and colleagues (2012) determined masked thresholds for
detecting a single component of a tone complex in relation
to its mistuning and binaural phase by asking the subjects
to report the presence of the target component presented at
different sound levels. Unexpectedly, when adding masker
harmonics outside the auditory filter tuned to the target
frequency the masked threshold increased. This is in contrast
to the expectation, that a higher number of components would
make it easier to separate the complex from the target.
To investigate the mechanisms underlying this observation,
we recorded single- and multi-unit responses from inferior
colliculus neurons of anesthetized gerbils that have been
demonstrated to have a high sensitivity for detecting
mistuning of components in a harmonic complex (e.g.,
Klinge-Strahl et al. 2013). Complex tone maskers were
centered at the target frequency and either had 8 or 32
harmonic components equally distributed to both sides of
the target. Maskers with two different phase relation between
components were applied: random phase and sine phase
maskers. Target tones were either tuned or mistuned with
respect to the masker fundamental. Maskers were always
played with the same phase to both ears. Targets of different
levels were played either with the same phase or with a
phase difference of 180 degrees between the two ears. Spike
rate and temporal patterns of spike sequences (analysis
with van Rossum [2001] metrics) were used to determine
neuronal response thresholds. Neuronal sensitivity (da) was
determined by subjecting the response measures to an ROC
analysis comparing responses to masker plus target and
masker alone.
In general, an increased target level resulted in a higher da for
both the rate response and the temporal spike pattern. Similar
to the human psychophysical data (Klein-Hennig et al. 2012),
the sensitivity for both response measures was considerably
reduced in a fraction of the neurons if the bandwidth of the
harmonic complex masker was increased. Also the phase
relation of the masker components affected the response.
However, the impact of mistuning and target phase was only
small. The relation between the neurophysiological and the
psychophysical data will be discussed.
Funding
This study was funded by the DFG (TRR 31 and Cluster of
Excellence “Hearing4all”)}
}