BEHAVIOR
Schlund
/ STIMULUS
MODIFICATION
CONTROL
/ October 2000

Bedside hearing screenings are routinely conducted by speech and language pathologists for
brain injury survivors during rehabilitation. Cognitive deficits resulting from brain injury, however, may interfere with obtaining estimates of auditory thresholds. Poor comprehension or
attention deficits often compromise patient abilities to follow procedural instructions. This article describes the effects of jointly applying behavioral methods and psychophysical methods to
improve two severely brain-injured survivors’ attending and reporting on auditory test stimuli
presentation. Treatment consisted of stimulus control training that involved differentially reinforcing responding in the presence and absence of an auditory test tone. Subsequent hearing
screenings were conducted with novel auditory test tones and a common titration procedure.
Results showed that prior stimulus control training improved attending and reporting such that
hearing screenings were conducted and estimates of auditory thresholds were obtained.

When Instructions Fail
The Effects of Stimulus Control Training
on Brain Injury Survivors’ Attending
and Reporting During Hearing Screenings
MICHAEL W. SCHLUND
Kennedy Krieger Institute

Diffuse brain injury resulting from stroke or head trauma may produce disorientation, inattention, aphasia, or hearing losses that compromise instruction following. Speech and language therapists routinely conduct brief hearing screenings to assess for hearing loss
before pursuing formal audiological testing, especially when audiological services (e.g., rural subacute skilled nursing facilities) are not
readily available. Instructions are primarily used during hearing
screenings to establish patient responding to auditory test stimuli. A
common problem, however, is that brain injury survivors may be
unable to consistently follow instructions. Consequently, estimates of
auditory thresholds may be difficult to obtain during screenings. This
article describes the application of a stimulus control training proceBEHAVIOR MODIFICATION, Vol. 24 No. 5, October 2000 658-672
© 2000 Sage Publications, Inc.

658

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

Schlund / STIMULUS CONTROL

659

dure that originates in the stimulus control literature to facilitate
patient attending and responding during hearing screenings. For an
overview of stimulus control see Dinsmoor (1995a, 1995b).
The difficulties associated with conducting hearing screenings
with brain injury survivors unable to follow instructions parallels the
difficulties faced by psychologists studying nonhuman animal auditory systems and audiologists conducting hearing screenings with
children. Fortunately, a variety of operant methods are available in the
nonhuman animal literature on stimulus control that may help rehabilitation professionals overcome some of the assessment challenges
presented by brain injury survivors. The methods combine psychophysical methods, such as titration methods (Harrison & Turnock,
1975; Hendricks, 1966; Maurissen, 1988; Mosokowitz & Kitzes,
1966), which are also common methods in clinical audiology, and
principles and methods characteristic of behavioral approaches
(Atkinson, 1958; Bekesey, 1947; Gourevitch & Hack, 1966; Gourevitch,
Hack, & Hawkins, 1960; Harrison, 1984; Harrison, Abelson, & Fisher,
1960; Hunter, 1914; Stebbins, Green, & Miller, 1966). Similarly, the
application of reinforcement principles form the core of numerous
clinical audiology procedures developed for assessing children, such
as conditioned play audiometry (Olsen, 1994), tangible reinforcement
operant conditioning audiology (Wilson & Richardson, 1991), and
visual reinforcement operant conditioning audiology (Thompson,
Thompson, & Vethivelu, 1989). However, unlike nonhuman animal
studies, verbal instructions often remain an essential component of
procedures used to prepare children for testing and maintaining responding during testing. A notable exception is a study by Woolcock and
Alferink (1982) that employed reinforcement and errorless discrimination training procedures.
NONHUMAN ANIMAL MODEL

One characteristic common between nonhuman animal stimulus
control experiments and human hearing screenings is a focus on the
integrity of functional relations between environmental stimuli and
behavior. An experiment by Harrison (1984) provides a model for
developing clinical procedures devoted to improving hearing screen-

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

660

BEHAVIOR MODIFICATION / October 2000

ing performances of brain injury survivors unable to consistently follow instructions. Harrison’s experimental procedure is based on the
go right, go left successive method developed by Hunter (1914) to
study auditory stimulus control. Harrison used a test chamber that
contained two levers mounted on one wall and speakers mounted to
the left side and right side of the two levers. Monkeys were trained to
press both levers under a 45-second variable interval reinforcement
schedule that delivered liquid, on average, every 45 seconds following
a bar press. Stimulus control training used a 45-decibel sound pressure
level (SPL) (wide band noise bursts) stimulus presented randomly
through the left or right speaker. The stimulus remained on until a bar
press occurred. If the auditory stimulus was presented to the left
speaker and a left bar press occurred, the response was reinforced.
Similarly, if the auditory stimulus was presented to the right speaker
and a right bar press occurred, the response was reinforced. All other
bar presses terminated the stimulus and were not reinforced. Results
showed that stimulus control over responding progressed from 50%
accuracy (chance) on each bar to more than 90% accuracy. Thus, two
three- term relations or contingencies were established:
S+1 (left 45-dB stimulus): R1 (left bar press): Reinforcement (liquid)
S+2 (right 45-dB stimulus): R2 (right bar press): Reinforcement
(liquid).

Harrison’s (1984) experiment illustrates a procedure for studying
stimulus control relations with nonhuman animals, and it demonstrates how differential reinforcement may establish and maintain two
three-term contingencies. Similarly, speech therapists are also concerned with stimulus control relations. Hearing screenings investigate
how systematic changes in auditory stimuli along a dimension (e.g.,
hertzes [Hz], decibels [dB]) exert functional changes over behavior
along a dimension (e.g., rate, latency, probability) with the goal of
determining auditory thresholds. Although Harrison’s procedure
employed differential reinforcement as the mechanism to establish
functional relations, speech therapists often rely on instructions as the
mechanism to establish and maintain functional relations. Inconsistent or absent responding during screenings may represent a failure in
instructions as the mechanism for establishing functional relations.

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

Schlund / STIMULUS CONTROL

661

The nonhuman animal stimulus control literature suggests that differential reinforcement may serve as an alternative mechanism under
such circumstances.
STIMULUS CONTROL TRAINING

Harrison’s (1984) experimental procedure may be adapted to facilitate the development of basic functional relations between auditory
test stimuli and behavior when instructions fail. To improve patient
attending to relevant test stimuli and reporting during hearing screenings, elements of Harrison’s procedure may be useful for developing a
stimulus control training procedure for use before screenings. These
elements include using several auditory stimuli, correlating stimuli
with reinforcement, and requiring several distinct responses.
Overall, training would focus on establishing several basic threeterm contingencies. In particular, training would need to focus on
establishing control by stimuli above and below threshold to prepare
for screenings that involve responding to stimuli that lie above and
below threshold. For example, a patient would learn to emit one response
in the presence of an auditory stimulus and a separate response in the
absence of an auditory stimulus. This would require employing two
compound stimuli. One compound stimulus (S+1) would consist of a
light stimulus paired with an audible test stimulus (an auditory tone
that lies well above threshold). A target response (R1) would be reinforced in the presence of the compound stimulus (S+1). The second
compound stimulus (S+2) would consist of a light stimulus paired
with no stimulus tone, which may be viewed as a simple stimulus. A
second, separate response (R2) would be reinforced in the presence of
the compound stimulus (S+2). Modeling Harrison’s procedure, the
stimulus control training procedure would focus on establishing the
following three-term contingencies:
S+1 (light + 70-dB, 4,000-Hz tone): R1 (tone present): Reinforcement
S+2 (light + no tone): R2 (tone absent): Reinforcement.

In one sense, the contingencies form two extremes or anchors of a
psychometric function. That is, S+1 serves as a stimulus that lies well
above threshold and S+2 serves as a stimulus that lies below threshold.

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

662

BEHAVIOR MODIFICATION / October 2000

Such a procedure may serve as a first step toward conducting hearing
screenings that involve presenting many stimuli that lie above and
below threshold.
There are also noteworthy reasons for applying reinforcement and
employing two responses during training and subsequent hearing
screenings. The application of reinforcement has been shown to maintain high levels of attending to relevant environmental stimuli
(Dinsmoor, 1995a) and is common to cognitive and behavioral treatments in rehabilitation (e.g., Parente & Herrmann, 1996). By requiring responding to both the presence and absence of auditory test stimuli, the negative influences of cognitive impairments on responding
may be minimized. Specifically, if responding is required for stimuli
below threshold and responding fails to occur, the trial provides no
meaningful data. Without the required response following the presentation of auditory stimuli (i.e., stimuli that lie below threshold), it may
be difficult for the clinician to ascertain whether stimuli were below
threshold or poor attending compromised detection.
In the present study, two brain injury survivors exhibiting severe
cognitive impairments and hearing loss learned to discriminate the
presence and absence of an auditory stimulus through stimulus control training. Subsequent hearing screenings were conducted using a
titration procedure common to animal psychophysical studies and
clinical audiology. Screening involved presenting a variety of novel
compound stimuli. Based on responding to stimuli, decibel levels
increased or decreased on the following trial. The decibel level associated with relatively equal percentages of each response (i.e., 50% tone
absent and 50% tone present) provided an estimate of the auditory
threshold.

METHOD
PARTICIPANTS

Patient 1 (S1) was a 65-year-old White male who survived a cerebral vascular accident. Axial-computed tomography images of the
brain showed edema of the right cerebral hemisphere with effacement

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

Schlund / STIMULUS CONTROL

663

of the sulci and a mild mass effect compressing the lateral ventricles.
The patient ranged from 5 to 6 on the Rancho Los Amigos cognitive
scale. Because of severe impairments, the patient could not complete a
brief neurocognitive behavioral examination. Evaluation by a speech
and language pathologist noted decreased cognition, verbal apraxia,
and expressive and receptive language deficits. During initial screening, S1 was reasonably attentive and cooperative but did not comprehend verbal or written instructions. Consequently, S1 could not participate in a hearing screening.
Patient 2 (S2) was a 49-year-old White male who survived a severe
closed-head injury. Axial-computed tomography images of the brain
suggest diffuse axonal shearing and injury to the right frontoparietal
region. The patient ranged from 4 to 5 on the Rancho Los Amigos cognitive scale. Because of severe impairments, the patient could not
complete a brief neurocognitive behavioral examination. Evaluation
by a speech and language pathologist noted significant processing
delays, poor following of instruction, disorientation, and severe attentional deficits. During initial screening, S2 was inattentive, unable to
consistently follow verbal or written instructions, withheld responding, and often stated hearing a test tone during and between trials
(false alarms [FAs]—see below).
APPARATUS

A MAICO MA27 audiometer, 14 × 17 × 7 inches, was used to
present a warbling tone (±5%; 5 Hz rate) stimulus to participants
through headphones. A common handheld flashlight provided a light
stimulus.
PROCEDURE

Stimulus control training. All auditory test stimuli were presented
through headphones. The clinician sat facing S1 with the flashlight,
and the operator of the audiometric equipment remained seated
behind S1. A coffee cup and a red and white poker chip were placed on
a small table in front of S1. For S2, the tester sat facing S2 with the
flashlight and with the audiometric equipment hidden from view.

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

664

BEHAVIOR MODIFICATION / October 2000

Figure 1 outlines the discrete trial procedure employed. A trial
began with the presentation of one of two compound stimuli. The
stimuli consisted of a compound light, 70-decibel 4000-Hz-tone stimulus and a compound light, no-tone stimulus. Each compound stimulus was presented for 3 seconds, and stimuli were presented in a random order, with equal frequency. Initially, hand-over-hand assistance
and modeling were required to establish target responses to stimuli.
Responding for S1, shown in Figure 1, consisted of placing a red poker
chip in a cup (S2 waved his arms above his head and nodded “yes”)
following the presentation of the compound light, 70-decibel,
4000-Hz-tone stimulus. Responding for S1 also consisted of placing a
white poker chip in a cup (S2 shook his head “no”) following the presentation of the compound light, no-tone stimulus. Correct responses
were reinforced (S+). All incorrect responses were not reinforced
(S–). Reinforcers consisted of verbal praise, cheers, smiles, and handshakes. Each auditory channel was trained individually.
Hearing screenings. A titration procedure was employed to obtain
estimates of auditory thresholds. Within sessions, the decibel level of
a probe tone (range 5 to 65 dB) decreased 5 decibels on the trial following a participant’s response that indicated detection of the probe
tone, and it increased 5 decibels on the trial following the participant’s
response that indicated nondetection of the probe tone. Figure 2 outlines the screening procedure. The outline for S1 (different from S2’s
by response type—see above) shows that probe stimuli (range 5 to 65
dB) paired with a light were presented for 3 seconds. All target
responses were reinforced (S+). The same procedure was employed
with S2 with the addition of catch trials presented randomly during
testing (p = .50). Catch trials consisted of presenting the compound
light, no-tone stimulus. All correct catch trial responses were
reinforced.
Frequencies of 500, 1000, and 8000 Hz were manipulated between
sessions. Sessions ended if responding appeared stable during the first
10 trials or, during sessions, responding stabilized for approximately 5
trials.

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

Schlund / STIMULUS CONTROL

665

Figure 1. Temporal order of events for a stimulus control training trial.
NOTE: Following stimulus presentation, a correct response (shaded white) was reinforced (S+)
and an incorrect response (shaded gray) was not reinforced (S–).

Figure 2. Temporal order of events for a screening trial with titration.
NOTE: Following stimulus presentation, all target responses were reinforced (S+) and the decibel level was titrated based on responding.

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

666

BEHAVIOR MODIFICATION / October 2000

RESULTS

Stimulus control training. Figure 3 shows responding under titration for S2 at 4000 and 8000 Hz before stimulus control training. Left
auditory channel responding at 4000 Hz shows stability near 50 dB,
whereas right auditory channel responding steadily increased and
then decreased to 20 dB. On catch trials 24 and 25, the compound light,
no-tone stimulus was presented, and FAs occurred. Erratic responding
was also observed at 8000 Hz. On catch trials 6 through 14, numerous
FAs occurred.
During stimulus control training, participants were required to emit
15 consecutive correct responses for each auditory channel before
receiving a hearing screening. Results showed that both S1 and S2 met
the training criterion for each channel within 25 trials.
Hearing screenings. Stimulus control training for S1 successfully
established consistent responding and enabled the speech therapist to
conduct a hearing screening. During the screening, responding under
titration produced stable functions and provided threshold estimates.
Figure 4 shows the titrated dB level by trial for S1 at 500, 1000, 4000,
and 8000 Hz and means taken from the last five titrated trials. In general, responding was maintained under titration, and estimates of
auditory thresholds were obtained. Mean decibel levels suggest hearing losses in the left and right auditory channels at high frequencies
(4000 and 8000 Hz).
Figure 5 shows responding under titration for S2 at 500, 1000,
4000, and 8000 Hz following stimulus control training and means
taken from the last five titrated trials. No false alarms occurred on
catch trials. In general, responding to auditory test stimuli under titration was more consistent after stimulus control training. Data show
orderly declines in decibel levels to thresholds for both auditory channels and across all frequencies tested. Mean dB levels suggest hearing
losses in the right auditory channel and losses at high frequencies
(4000 and 8000 Hz) in the left auditory channel.

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

Schlund / STIMULUS CONTROL

667

Figure 3. Results of responding under titrating decibel levels within sessions for Patient 2
prior to stimulus control training.
NOTE: FA = false alarm. CR = correct rejection.

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

668

BEHAVIOR MODIFICATION / October 2000

Figure 4. Results of responding under titrating decibel levels within sessions for Patient 1
after stimulus control training.
NOTE: Four test conditions (500, 1000, 4000, and 8000 Hz) and the mean decibel level from
each test condition (bottom graph) are presented.

DISCUSSION

A stimulus control training procedure, based on methods used to
study stimulus control with nonhuman animals, was employed to
improve two brain injury survivors’ attending and reporting during

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

Schlund / STIMULUS CONTROL

669

Figure 5. Results of responding under titrating decibel levels within sessions for Patient 2
after stimulus control training.
NOTE: Four test conditions (500, 1000, 4000, and 8000 Hz) and the mean decibel level from
each test condition (bottom graph) are presented.

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

670

BEHAVIOR MODIFICATION / October 2000

hearing screenings. The stimulus control training procedure involved
differential reinforcement of responding in the presence and absence
of an auditory stimulus. Following the establishment of two basic
three-term contingencies, screenings were conducted that involved
titrating dB level within sessions to obtain estimates of auditory
thresholds. Means taken from stable titration functions suggested the
presence of hearing losses for both patients. In general, the intervention was successful in preparing patients, who displayed significant
cognitive impairments that prevented assessment, for a common hearing screening procedure by a speech therapist.
Modeling the stimulus control training procedure after methods
common to nonhuman animal research (Harrison, 1984) facilitated
auditory stimulus control over patient responding. In the present
study, compound stimuli that may be viewed as anchors or endpoints
of a psychometric function were employed to prepare patients for
hearing screenings that involved responding to stimuli above and
below threshold. In essence, patients learned to functionally communicate the perceived presence and perceived absence of auditory test
stimuli to the clinician. The rapid response acquisition during training
and response maintenance during screenings appears clinically significant given the severity of the patients’ deficits. Some clinicians might
question the validity of the auditory thresholds obtained. Such questions, however, are related to the application of the titration procedure.
The purpose of the study was to demonstrate that training could
improve patient attending and reporting such that hearing screenings
could be conducted rather than to demonstrate the validity of thresholds obtained using a titration procedure.
The application of a two-response procedure, which is a modification of the go right, go left successive method, also appears critical to
both training and screening. During stimulus control training, it was
possible to establish two distinct three-term contingencies and assess
the integrity of stimulus control over responding before conducting
the more complex hearing screenings. During screenings, it was possible to observe alternating patterns of subjective detection and subjective nondetection in each titrated function near threshold. Overall,
employing two responses increased our confidence of the integrity of
functional relations as well as the auditory threshold estimates obtained.

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

Schlund / STIMULUS CONTROL

671

The data also provided the rehabilitation team with preliminary information about patients’ potential hearing deficits until formal audiological testing results were obtained.
Future research on the application of stimulus control methods with
brain injury survivors might focus on identifying variables that affect
application, such as type of injury, length of coma, or type of cognitive
impairments. At a broad level, research must also examine the effects
of reinforcement on cognitive and behavioral impairments (e.g.,
Schlund, 1999). In this regard, laboratory research may illuminate the
extent to which survivors are able to report, anticipate, and respond
appropriately to consequences. Given the growing recognition of
the effectiveness of behavioral principles and methods in rehabilitation, basic and clinical research is needed that outlines how behaviorenvironment relations may be differentially affected by brain injury.

REFERENCES
Atkinson, C. J. (1958). The use of the eyelid reflex as an operant in audiometric testing. Journal
of the Experimental Analysis of Behavior, 2, 212.
Bekesey, G. V. (1947). A new audiometer. Acta Otolaryngology, 35, 411-422.
Dinsmoor, J. A. (1995a). Stimulus control: Part I. Behavior Analyst, 18, 51-68.
Dinsmoor, J. A. (1995b). Stimulus control: Part II. Behavior Analyst, 18, 253-269.
Gourevitch, G., & Hack, M. H. (1966). Audibility in the rat. Journal of Comparative and Physiological Psychology, 62, 289-291.
Gourevitch, G., Hack, M. H., & Hawkins, J. E. (1960). Auditory thresholds in the rat measured
by an operant technique. Science, 131, 1046-1047.
Harrison, J. M. (1984). The functional analysis of auditory discrimination. Journal of the Acoustical Society of America, 75, 1848-1854.
Harrison, J. M., Abelson, R., & Fisher, G. (1960). Some relations between the auditory system
and the medulla and auditory-stimulus functions. Journal of the Experimental Analysis of
Behavior, 3, 207-220.
Harrison, J. M., & Turnock, M. T. (1975). Animal psychophysics: Improvements in the tracking
method. Journal of the Experimental Analysis of Behavior, 23, 141-147.
Hendricks, J. (1966). Flicker thresholds as determined by a modified conditioned suppression
procedure. Journal of the Experimental Analysis of Behavior, 9, 501-506.
Hunter, W. S. (1914). Auditory sensitivity of the white rat. Journal of Animal Behavior, 4,
215-222.
Maurissen, J.P.J. (1988). Quantitative sensory assessment in toxicology and occupational medicine: Applications, theory, and critical appraisal. Toxicology Letters, 43, 321-343.
Mosokowitz, H., & Kitzes, L. (1966). A comparison of psychophysical methods using animals.
Journal of the Experimental Analysis of Behavior, 9, 515-519.

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015

672

BEHAVIOR MODIFICATION / October 2000

Olsen, S. O. (1994). Comparative study of the usefulness of play-conditioned audiometry with
wooden rings and electronic video peep show. Scandinavian Audiology, 23, 265-268.
Parente, R., & Herrmann, D. (1996). Retraining cognition: Techniques and applications.
Gaithersburg, MD: Aspen.
Schlund, M. W. (1999). A case study on self awareness: Effects of feedback and review on verbal
self report and remembering following brain injury. Brain Injury, 13, 375-380.
Stebbins, W. C., Green, S., & Miller, F. L. (1966). Auditory sensitivity of the monkey. Science,
153, 1646-1647.
Thompson, M., Thompson, G., & Vethivelu, S. (1989). A comparison of audiometric test methods for 2-year-old children. Journal of Speech and Hearing Disorders, 54, 174-179.
Wilson, W. R., & Richardson, M. A. (1991). Behavioral audiometry. Otolaryngologic Clinics of
North America, 24, 285-297.
Woolcock, J. E., & Alferink, L. A. (1982). An operant tracking procedure in the auditory assessment of profoundly retarded individuals. Journal of Applied Behavior Analysis, 15, 302-307.

Michael Schlund is a licensed psychologist on the faculty at the Kennedy Krieger Institute. He has research interests in learning and decision making deficits in adults and
children with acquired brain injury. He is currently studying the effects of brain injury on
the sensitivity of behavior to contingencies. He is also collaborating with psychiatrists at
Johns Hopkins on fMRI investigations of operant techniques commonly used in clinical
interventions (e.g., different reinforcement, response cost, extinction) and drug treatments for anxiety. As a member of an interdisciplinary neurorehabilitation team, he provides behavior therapy and brain injury education services to patients and families.

Downloaded from bmo.sagepub.com at CALIF STATE POLY UNIV LIBRARY on March 10, 2015