Mechanisms of Forward Masking

1
Mechanisms of Forward Masking Magdalena
Wojtczak and Neal F. Viemeister, Department of
Psychology, University of Minnesota

• Prediction based on adaptation

• Purpose
• Two mechanisms have been proposed to account for
  forward masking
• adaptation of neural response after the offset
  of a forward masker
• persistence of neural activity extending beyond
  the offset of a forward masker.
• The purpose of the presented experiments was to
  provide additional insight into the forward
  masking effect with the aim of distinguishing
  between the two suggested underlying mechanisms.

Adaptation would lead to an attenuation (a) of
the responses to P and sM.
(2)
(3)
where
Intensity of internal noise (IIN) was estimated
assuming that the criterion ratio for absolute
threshold is the same as it is above threshold
(IINIabs/k). Eqs. (1) and (2) produce
equivalent predictions, thus both models of
forward masking account for the data equally
well. Similar to Widin and Viemeister (1980) and
Oxenham (2001), this experiment does not allow
for distinguishing between the two explanations
for forward masking.
Experiment 2 Effect of forward masking on
central masking
11
Experiment 1 Simulation of persistence Growth of
simultaneous masking was measured in the presence
of a fixed-level forward masker. A schematic
illustration of the signal interval of a 3-IFC
task is shown below. Methods Stimuli fM
100-10000 Hz, 150 ms, 30 dB SPL/Hz sM
100-10000 Hz, 20 ms, -18 - 30 dB SPL/Hz P
100-10000 Hz, 10 ms (5-ms ramps), level varied
adaptively d 5 and 75 ms Subjects 2 subjects
with normal hearing
Central masking involves interaction of the
stimuli presented to the left and right ear.
Forward masking is not observed across ears and
thus, is likely peripheral to the site of central
masking. Shore (1998) showed that
forward-masking characteristics measured in some
units of the VCN are in close agreement with the
characteristics measured psychophysically.
Central masking may result from an interaction of
inputs from lower processing stages at the level
of MSO (Zwislocki, 1972). These observations
support the assumption that the mechanism
underlying forward masking is peripheral to that
underlying central masking. Thus, we can study
the effect of a forward masker by examining its
effect on central masking. The experiment
involved three steps Step 1 Forward masking was
measured monaurally for three masker-probe
delays 5, 25, and 55 ms. Step 2 Detection of
the probe (presented to the left ear) was
measured in the presence of a fixed-level central
masker (presented to the right ear). Step 3 Step
2 was repeated with the probe and the central
masker presented after a monaural fixed-level
forward masker, for three masker-probe delays 5,
25, and 55 ms. A schematic illustration of one
trial in Step 3 is shown below
Fig. 2. Thresholds in central masking measured
with and without a monaural forward masker.
Experiment 3 Centering of binaural image in the
presence of forward masker
The effect of a forward masker was examined by
measuring the level of the probe presented in the
unmasked ear that produced a centered image when
gated synchronously with the contralateral
forward-masked probe. A method of adjustment was
used and a cue indicating the centered image was
provided. The stimuli and the subjects were the
same as in Experiment 2. The level of the probe
in the ear with the forward masker was set to 70
dB SPL.
Methods
Stimuli fM 100-10000 Hz, 150 ms, 20 and 30 dB
SPL/Hz cM 4 kHz, 10 ms (5-ms ramps), 70 and 80
dB SPL P 4 kHz, 10 ms (5-ms ramps), level
varied adaptively d 5, 25, and 55 ms Subjects 3
subjects with normal hearing
Results
Results
Figure 1 shows data from two listeners, for
delays (d) of 5 ms (circles) and 75 ms
(triangles). The yellow hexagon represents
absolute threshold for probe detection.
For all three masker-probe delays, levels of the
4-kHz tone in the unmasked ear necessary for a
centered image were lower than 70 dB
SPL. Attenuation of the response to the 4-kHz
probe presented after fM was computed for results
from three experiments (forward masking, central
masking and image centering) and is shown in Fig.
3.
Results
Figure 2 shows individual data for two spectrum
levels of the forward masker, 20 and 30 dB SPL
(shown in the two rows of panels). It is
important to note that the probe cannot be
detected monaurally since the threshold in
central masking is always higher than the
absolute threshold. Thus, the task is not
equivalent to a monaural detection task with a
contralateral stimulus. Subjects detected the
probe using binaural cues, such as a slight
increase in loudness or a slight movement in the
intracranial position of the probe (it was a 3IFC
task with feedback so the subjects could chose
the interval that sounded different). In central
masking, a higher level of the probe was needed
for threshold in the presence of fM. A likely
explanation is that fM has the effect of
attenuating the response to the probe. Assuming
that a criterion ratio is required for threshold
in central masking
References Oxenham, A. J. (2001). Forward
masking Adaptation or integration, JASA 109,
732-741. Shore, S. E. (1998). Influence of
centrifugal pathways of ventral cochlear nucleus
neurons, JASA 104, 378-389. Widin, G. P., and
Viemeister, N. F. (1980). Masker interaction in
pure-tone forward masking, JASA 68,
475-479. Zwislocki, J. J. (1972). A theory of
central auditory masking and its partial
validation, JASA 52, 644-659.
Fig. 3. Estimated attenuation.
. Thus,
(4)
without fM
, after fM
Fig. 1. Growth of masking in the presence of a
forward masker
It is possible that the adaptation equally
affects both the termination of the masker and
the probe and that forward masking is purely the
result of persistence. In such a case for
threshold in forward masking
(5)

• Prediction based on persistence of excitation

Conclusion Two experiments involving
interaction of inputs from both ears (central
masking and image centering) revealed that the
response to a stimulus presented after a forward
masker is attenuated compared with the unmasked
response. The estimated attenuation is in
relatively good agreement with the amount of
forward masking. This result suggests that
adaptation is a predominant mechanism of forward
masking.
In this case both, persistence and attenuation
would influence central masking and
Predictions shown by solid lines in Fig. 1 were
obtained assuming that threshold is observed when
a constant criterion ratio is reached
(6)
Eq. (6) leads to estimates of a in dB that
decrease with delay d. Thus, the assumption of
adaptation equally affecting fM and P cannot hold
since adaptation of the termination of the masker
cannot depend on the masker-probe delay. If a
were assumed to be constant for all delays d than
decreasing persistence would have to lead to an
increased IP(fM) with increasing delay, which was
not observed in the data. This indicates that
persistence alone cannot account for forward
masking.
(1)
Funded by NIH/NIDCD grant DC00683