Editing Absolute threshold of hearing (section)

== Psychophysical methods for measuring thresholds ==

Measurement of the absolute hearing threshold provides some basic information about our [[auditory system]].<ref name="Gelfand, 1990"/> The tools used to collect such information are called psychophysical methods. Through these, the [[perception]] of a physical stimulus (sound) and our psychological response to the sound is measured.<ref name="Hirsh, 1952">Hirsh I J.,1952. "The Measurement of Hearing". United States of America: McGraw-Hill.</ref>

Several psychophysical methods can measure absolute threshold. These vary, but certain aspects are identical. Firstly, the test defines the stimulus and specifies the manner in which the subject should respond. The test presents the sound to the listener and manipulates the stimulus level in a predetermined pattern. The absolute threshold is defined statistically, often as an average of all obtained hearing thresholds.<ref name="Gelfand, 1990"/>

Some procedures use a series of trials, with each trial using the 'single-interval "yes"/"no" paradigm'. This means that sound may be present or absent in the single interval, and the listener has to say whether they thought the stimulus was there. When the interval does not contain a stimulus, it is called a "catch trial".<ref name="Gelfand, 1990"/>

=== Classical methods ===
Classical methods date back to the 19th century and were first described by [[Gustav Theodor Fechner]] in his work ''Elements of Psychophysics''.<ref name="Hirsh, 1952"/> Three methods are traditionally used for testing a subject's perception of a stimulus: the method of limits, the method of constant stimuli, and the method of adjustment.<ref name="Gelfand, 1990"/>

[[Image:Method of limits.png|thumb|Series of descending and ascending runs in Method of Limits]]
; '''Method of limits''': In the method of limits, the tester controls the level of the stimuli. Single-interval ''yes/no'' paradigm' is used, but there are no catch trials.

: The trial uses several series of descending and ascending runs.

: The trial starts with the descending run, where a stimulus is presented at a level well above the expected threshold. When the subject responds correctly to the stimulus, the level of intensity of the sound is decreased by a specific amount and presented again. The same pattern is repeated until the subject stops responding to the stimuli, at which point the descending run is finished.

: In the ascending run, which comes after, the stimulus is first presented well below the threshold and then gradually increased in two [[decibel]] (dB) steps until the subject responds. As there are no clear margins to 'hearing' and 'not hearing', the threshold for each run is determined as the midpoint between the last audible and first inaudible level.

: The subject's absolute hearing threshold is calculated as the mean of all obtained thresholds in both ascending and descending runs.

: There are several issues related to the method of limits. First is anticipation, which is caused by the subject's awareness that the turn-points determine a change in response. Anticipation produces better ascending thresholds and worse descending thresholds.

: [[Habituation]] creates completely opposite effect, and occurs when the subject becomes accustomed to responding either "yes" in the descending runs and/or "no" in the ascending runs. For this reason, thresholds are raised in ascending runs and improved in descending runs.

: Another problem may be related to step size. Too large a step compromises accuracy of the measurement as the actual threshold may be just between two stimulus levels.

: Finally, since the tone is always present, "yes" is always the correct answer.<ref name="Gelfand, 1990"/>

; '''Method of constant stimuli''': In the method of constant stimuli, the tester sets the level of stimuli and presents them at completely random order. [[Image:Method of Constant Stimuli.png|thumb|Subject responding "yes"/"no" after each presentation]]

: Thus, there are no ascending or descending trials.

: The subject responds "yes"/"no" after each presentation.

: The stimuli are presented many times at each level and the threshold is defined as the stimulus level at which the subject scored 50% correct. "Catch" trials may be included in this method.

: Method of constant stimuli has several advantages over the method of limits. Firstly, the random order of stimuli means that the correct answer cannot be predicted by the listener. Secondarily, as the tone may be absent (catch trial), "yes" is not always the correct answer. Finally, catch trials help to detect the amount of a listener's guessing.
: The main disadvantage lies in the large number of trials needed to obtain the data, and therefore time required to complete the test.<ref name="Gelfand, 1990"/>

; '''Method of adjustment''': Method of adjustment shares some features with the method of limits, but differs in others. There are descending and ascending runs and the listener knows that the stimulus is always present. [[Image:Method of Adjustment.png|thumb|The subject reduces or increase the level of the tone]]

: However, unlike in the method of limits, here the stimulus is controlled by the listener. The subject reduces the level of the tone until it cannot be detected anymore, or increases until it can be heard again.

: The stimulus level is varied continuously via a dial and the stimulus level is measured by the tester at the end. The threshold is the mean of the just audible and just inaudible levels.

: Also this method can produce several biases. To avoid giving cues about the actual stimulus level, the dial must be unlabeled. Apart from the already mentioned anticipation and habituation, stimulus persistence (preservation) could influence the result from the method of adjustment.

: In the descending runs, the subject may continue to reduce the level of the sound as if the sound was still audible, even though the stimulus is already well below the actual hearing threshold.

: In contrast, in the ascending runs, the subject may have persistence of the absence of the stimulus until the hearing threshold is passed by certain amount.<ref name="Hirsh & Watson, 1996">Hirsh I J., Watson C S., 1996. Auditory Psychophysics and Perception. Annu. Rev. Psychol. 47: 461–84. Available to download from: http://arjournals.annualreviews.org/doi/pdf/10.1146/annurev.psych.47.1.461 . Accessed 1 March 2007.</ref>

=== Modified classical methods ===
{{main article|Pure tone audiometry}}

=== Forced-choice methods ===
Two intervals are presented to a listener, one with a tone and one without a tone. The listener must decide which interval had the tone in it. The number of intervals can be increased, but this may cause problems for the listener who has to remember which interval contained the tone.<ref name="Gelfand, 1990"/><ref name="Miller et al., 2002">Miller et al., 2002. "Nonparametric relationships between single-interval and two-interval forced-choice tasks in the theory of signal detectability". Journal of Mathematical Psychology archive. 46:4;383–417. Available from: http://portal.acm.org/citation.cfm?id=634580. Accessed 1 March 2007.</ref>

=== Adaptive methods ===
Unlike the classical methods, where the pattern for changing the stimuli is preset, in adaptive methods the subject's response to the previous stimuli determines the level at which a subsequent stimulus is presented.<ref name="Levitt, 1971">{{cite journal |author=Levitt H. |date=1971 |title=Transformed up-down methods in psychoacoustics |journal=J. Acoust. Soc. Am. |volume=49 |issue=2 |pages=467–477 |doi=10.1121/1.1912375 |pmid=5541744 |url=http://scitation.aip.org/getabs/servlet/GetabsServlet?prog=normal&id=JASMAN00004900002B000467000001&idtype=cvips&gifs=yes |access-date=1 March 2007|url-access=subscription }}</ref>

==== Staircase (up-down) methods ====
[[Image:Simple Up-Down Method.png|thumb|Series of descending and ascending trials runs and turning points]]

The simple ''1-down-1-up method'' consists of a series of descending and ascending trial runs and turning points (reversals). The stimulus level is increased if the subject does not respond and decreased when a response occurs. Similar to the method of limits, the stimuli are adjusted in predetermined steps. After obtaining from six to eight reversals, the first one is discarded and the threshold is defined as the average of the midpoints of the remaining runs. Experiments have shown that this method provides only 50% accuracy.<ref name="Levitt, 1971"/> To produce more accurate results, this simple method can be further modified by increasing the size of steps in the descending runs, e.g. ''2-down-1-up method'', ''3-down-1-up methods''.<ref name="Gelfand, 1990"/>

====Bekesy's tracking method====
[[Image:Bekesy's Tracking Method.png|thumb|The threshold being tracked by the listener]]

Bekesy's method contains some aspects of classical methods and staircase methods. The level of the stimulus is automatically varied at a fixed rate. The subject is asked to press a button when the stimulus is detectable. Once the button is pressed, the level is automatically decreased by the motor-driven [[attenuator (electronics)|attenuator]] and increased when the button is not pushed. The threshold is thus tracked by the listeners, and calculated as the mean of the midpoints of the runs as recorded by the automat.<ref name="Gelfand, 1990"/>