282
Marine Mammal Physiology: Requisites for Ocean Living
In order to obtain a sensory threshold, two main experimental methods are usually
applied in marine mammal research, the staircase method/method of limits or the method of constant stimuli (Gescheider 1997; Ehrenstein and Ehrenstein 1999). Using the staircase method,
the intensity of the stimulus is decreased step by step by one unit until the experimental
subject ceases to respond correctly. Whenever the subject answers incorrectly, defined as
a transition point, the stimulus intensity is again increased by one unit. Thresholds are
then calculated by averaging over approximately six transition points. A threshold can
also be approached starting with subthreshold intensity values. However, with respect
to the test subject’s motivation a descending series offers the advantage that the session
starts with suprathreshold intensity values that are easily detectable. The staircase method
allows determining a threshold within one session which is very effective and sometimes
even necessary if, for example, changes in the environment need to be considered (see for
example, Weiffen et al. 2006). However, the thresholds obtained are often not the lowest
as the experimental subjects rapidly learn that if they respond incorrectly, the task will be
easier in the following trial.
Using the method of constant stimuli, a set of stimuli is selected before data collection with stimulus intensities well above and below the putative threshold. This set of
stimuli is repeatedly presented to the experimental subject and its response is noted.
Finally, the threshold can be calculated from the psychometric function that depicts the
performance (in %) of the subject averaged over all presentations as a function of the
stimulus intensity. Threshold estimation with the method of constant stimuli are more
time-consuming than the estimation with a staircase method as several sessions are
needed before data collection can be completed. However, the thresholds are usually
lower in comparison to those obtained with a staircase method as the subject cannot
develop expectancies.
In psychophysical experiments, two experimental paradigms are generally used to
present the stimuli, a go/no go-paradigm or a two-alternative-forced-choice-paradigm. In a
go/no go-experiment, two stimulus categories are defined, the go-stimulus and the no gostimulus. When confronted with the go-stimulus, the experimental subject has to give an
answer by, for example, moving its head to a response target; when confronted with the no
go-stimulus, the subject needs to stay stationary for a predefined time interval. Stimulus
presentation is thus successively, and four responses are possible: a hit (go-stimulus and
go-response), a miss (go-stimulus but no go-response), a correct rejection (no go-stimulus
and no go-response), and a false alarm (no go-stimulus but go-response). As a high hit
rate can be achieved by always answering with a go-response, during analysis, hits and
false alarms need to be considered, and it is required to document a low false alarm rate
to make the data included in the hit rate valid. Using a go/no go-paradigm additionally
requires a perfect control of the time interval as some subjects tend to give a go-response
in the final milliseconds of stimulus presentation of the no go-time interval when stimulus
intensity is close to threshold.
In a two-alternative-forced-choice (2AFC) paradigm, the experimental subject always
has to discriminate between two stimuli which can be presented either simultaneously
or successively. In each trial, the subject is forced to give a predefined response such as
indicating the position of the positive stimulus (S+), which is followed by reinforcement,
whereas choosing the negative stimulus (S−) is not reinforced. In a 2AFC experiment,
side preferences need to be documented and counteracted by an equal number of trials with the S+ on either side (Gellermann 1933). Generally, the 2AFC procedure faster
leads to a threshold in comparison to the go/no go-paradigm as in the latter only the
performance in respect to one stimulus category, usually the go-stimuli, is analyzed to
Marine Mammal Physiology: Requisites for Ocean Living
In order to obtain a sensory threshold, two main experimental methods are usually
applied in marine mammal research, the staircase method/method of limits or the method of constant stimuli (Gescheider 1997; Ehrenstein and Ehrenstein 1999). Using the staircase method,
the intensity of the stimulus is decreased step by step by one unit until the experimental
subject ceases to respond correctly. Whenever the subject answers incorrectly, defined as
a transition point, the stimulus intensity is again increased by one unit. Thresholds are
then calculated by averaging over approximately six transition points. A threshold can
also be approached starting with subthreshold intensity values. However, with respect
to the test subject’s motivation a descending series offers the advantage that the session
starts with suprathreshold intensity values that are easily detectable. The staircase method
allows determining a threshold within one session which is very effective and sometimes
even necessary if, for example, changes in the environment need to be considered (see for
example, Weiffen et al. 2006). However, the thresholds obtained are often not the lowest
as the experimental subjects rapidly learn that if they respond incorrectly, the task will be
easier in the following trial.
Using the method of constant stimuli, a set of stimuli is selected before data collection with stimulus intensities well above and below the putative threshold. This set of
stimuli is repeatedly presented to the experimental subject and its response is noted.
Finally, the threshold can be calculated from the psychometric function that depicts the
performance (in %) of the subject averaged over all presentations as a function of the
stimulus intensity. Threshold estimation with the method of constant stimuli are more
time-consuming than the estimation with a staircase method as several sessions are
needed before data collection can be completed. However, the thresholds are usually
lower in comparison to those obtained with a staircase method as the subject cannot
develop expectancies.
In psychophysical experiments, two experimental paradigms are generally used to
present the stimuli, a go/no go-paradigm or a two-alternative-forced-choice-paradigm. In a
go/no go-experiment, two stimulus categories are defined, the go-stimulus and the no gostimulus. When confronted with the go-stimulus, the experimental subject has to give an
answer by, for example, moving its head to a response target; when confronted with the no
go-stimulus, the subject needs to stay stationary for a predefined time interval. Stimulus
presentation is thus successively, and four responses are possible: a hit (go-stimulus and
go-response), a miss (go-stimulus but no go-response), a correct rejection (no go-stimulus
and no go-response), and a false alarm (no go-stimulus but go-response). As a high hit
rate can be achieved by always answering with a go-response, during analysis, hits and
false alarms need to be considered, and it is required to document a low false alarm rate
to make the data included in the hit rate valid. Using a go/no go-paradigm additionally
requires a perfect control of the time interval as some subjects tend to give a go-response
in the final milliseconds of stimulus presentation of the no go-time interval when stimulus
intensity is close to threshold.
In a two-alternative-forced-choice (2AFC) paradigm, the experimental subject always
has to discriminate between two stimuli which can be presented either simultaneously
or successively. In each trial, the subject is forced to give a predefined response such as
indicating the position of the positive stimulus (S+), which is followed by reinforcement,
whereas choosing the negative stimulus (S−) is not reinforced. In a 2AFC experiment,
side preferences need to be documented and counteracted by an equal number of trials with the S+ on either side (Gellermann 1933). Generally, the 2AFC procedure faster
leads to a threshold in comparison to the go/no go-paradigm as in the latter only the
performance in respect to one stimulus category, usually the go-stimuli, is analyzed to