Content uploaded by Goran B. W. Soderlund
Author content
All content in this area was uploaded by Goran B. W. Soderlund on May 20, 2019
Content may be subject to copyright.
Listen to the noise: noise is beneficial for
cognitive performance in ADHD
Go
¨ran So
¨derlund,
1
Sverker Sikstro
¨m,
2
and Andrew Smart
3
1
Department of Psychology, Stockholm University, Sweden;
2
Lund University Cognitive Science (LUCS), Sweden;
3
New York University, Department of Psychology, USA
Background: Noise is typically conceived of as being detrimental to cognitive performance. However,
given the mechanism of stochastic resonance, a certain amount of noise can benefit performance. We
investigate cognitive performance in noisy environments in relation to a neurocomputational model of
attention deficit hyperactivity disorder (ADHD) and dopamine. The Moderate Brain Arousal model
(MBA; Sikstro
¨m&So
¨derlund, 2007) suggests that dopamine levels modulate how much noise is
required for optimal cognitive performance. We experimentally examine how ADHD and control children
respond to different encoding conditions, providing different levels of environmental stimula-
tion. Methods: Participants carried out self-performed mini tasks (SPT), as a high memory perform-
ance task, and a verbal task (VT), as a low memory task. These tasks were performed in the presence, or
absence, of auditory white noise. Results: Noise exerted a positive effect on cognitive performance for
the ADHD group and deteriorated performance for the control group, indicating that ADHD subjects
need more noise than controls for optimal cognitive performance. Conclusions: The positive effect of
white noise is explained by the phenomenon of stochastic resonance (SR), i.e., the phenomenon that
moderate noise facilitates cognitive performance. The MBA model suggests that noise in the environ-
ment, introduces internal noise into the neural system through the perceptual system. This noise
induces SR in the neurotransmitter systems and makes this noise beneficial for cognitive performance.
In particular, the peak of the SR curve depends on the dopamine level, so that participants with low
dopamine levels (ADHD) require more noise for optimal cognitive performance compared to con-
trols. Keywords: ADHD, stochastic resonance, dopamine, episodic memory, SPT, noise. Abbrevia-
tions: MBA: moderate brain arousal; SR: stochastic resonance; SPT: subject-performed task; VT:
verbal task (VT).
Stochastic resonance is the counterintuitive phe-
nomenon that an optimal amount of noise may un-
der certain circumstances be beneficial for cognitive
performance. The purpose of this study is to examine
the effects of external auditive noise on performance
in an episodic recall task in children with attention
deficit hyperactivity disorder (ADHD). According to
the Moderate Brain Arousal (MBA) model (Sikstro
¨m
&So
¨derlund, 2007), a neurocomputational model of
cognitive performance in ADHD, noise in the en-
vironment introduces internal noise into the neural
system through the perceptual system. This noise is
proposed to compensate for the reduced neural
background activity in ADHD and the hypofunc-
tional dopamine system (Solanto, 2002). The MBA
model predicts that noise enhances memory perfor-
mance for ADHD and attenuates performance for
controls. We will also argue for a link between the
effects of noise, dopamine regulation, and cognitive
performance.
ADHD is a developmental disorder characterized
by behavioral impairments in three domains: in-
attention, impulsivity, and hyperactivity. ADHD is
one of the most commonly diagnosed childhood
psychiatric disorders, affecting approximately 3–7%
(Castellanos & Tannock, 2002) of the childhood
population. A vast literature shows that handling
cognitive flexibility and rigidity during maintenance
of goal-directed behavior is difficult to manage for
ADHD children (Martinussen, Hayden, Hogg-Johnson,
& Tannock, 2005).
It has long been known that cognitive processing is
easily disturbed by noise and other distractors
(Broadbent, 1958). The mechanism behind this ef-
fect, in general terms, is that the distractor removes
attention from the target task. Research on this topic
since 1958 has demonstrated this finding to hold
across a wide variety of target tasks, distractors and
participant populations. Consistent with this, ADHD
children are regarded as more vulnerable to dis-
traction compared to normal controls (Corbett &
Stanczak, 1999) and several studies have demon-
strated results supporting this notion (e.g., Geffner,
Lucker, & Koch, 1996; Higginbotham & Bartling,
1993).
Two recent studies were, however, able to de-
monstrate the counterintuitive finding that under
certain circumstances participants could benefit
from noise and other task-irrelevant sounds pre-
sented concurrently with the target task. Abikoff,
Courtney, Szeibel, and Koplewicz (1996) showed that
children with ADHD were not distracted by back-
ground music, which can be considered as task-
irrelevant noise. Surprisingly, the results further
Conflict of interest statement: No conflicts declared.
Journal of Child Psychology and Psychiatry 48:8 (2007), pp 840–847 doi:10.1111/j.1469-7610.2007.01749.x
2007 The Authors
Journal compilation 2007 Association for Child and Adolescent Mental Health.
Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA
showed a noise-induced improvement in perfor-
mance in the target (arithmetic) task. To the best of
our knowledge Abikoff’s finding has been replicated
just once, by Gerjets, Graw, Heise, Westermann, and
Rothenberger (2002), where noise was induced by
music.
These studies, however, did not provide a satis-
factory theoretical account for why noise was
beneficial for performance. Here we suggest that the
phenomenon known as stochastic resonance can be
used to account for noise-induced improvement in
cognitive performance. Stochastic resonance (SR) is
the phenomenon that detection of a subthreshold
signal is enhanced by addition of noise in a non-
linear system. SR occurs in any system where
detection requires passing of a threshold, so that
the added noise allows for the combined noise and
signal to pass the threshold, permitting detection of
the signal (Moss, Ward, & Sannita, 2004). This
psychophysical phenomenon is present in biological
sensory systems in animals and humans (Russell,
Wilkens, & Moss, 1999). It has been found in sev-
eral modalities; tactile, hearing, and vision (see
Moss et al., 2004 for a review). The effect is not
restricted to sensory processing. Stochastic reson-
ance has been found in cognitive tasks where
auditory noise improved the speed of arithmetic
computations in a normal population (Usher &
Feingold, 2000). Stochastic resonance is usually
quantified by plotting detection, or cognitive per-
formance, as a function of noise intensity. This re-
lation exhibits an inverted U-curve, where
performance peaks at a moderate noise level. That
is, moderate noise is beneficial for performance
whereas too little, or too much, noise attenuates
performance (see Figure 1). After a review of noise
and cognition studies (e.g., Baker & Holding, 1993)
we suggest that, in order to induce the SR effect,
noise has to be continuous (in order not to be
attention-removing) and at a high energy level at all
frequencies, as in white or pink noise.
Stochastic resonance has been shown to be a
ubiquitous natural phenomenon (Moss et al., 2004).
In the brain, stochastic resonance plays an import-
ant role in dopamine signaling (Li, von Oertzen, &
Lindenberger, 2006). Dopamine modulates neural
responses and function by increasing the signal-
to-noise ratio (SNR) through enhanced differenti-
ation between background or efferent firing and
afferent stimulation. Dopamine thus produces a
suppressive influence on spontaneous activity,
explaining its apparent inhibitory actions, and sim-
ultaneously causes an enhanced excitability in re-
sponse to afferent-driven stimulation (J.D. Cohen,
Braver, & Brown, 2002). It has been suggested that
high activity of catecholamine neuromodulators in
prefrontal neurons is associated with a high SNR of
information processing (Kiefer, Ahlegian, & Spitzer,
2005). Thus, too low or too high neuromodulatory
activity results in a low SNR and worse cognitive
performance in such areas as working memory and
inhibitory control. That is, the relation between
cognitive performance and dopamine transmission
shows an inverted U-shaped curve where either too
high, or too low, levels attenuate performance
(Goldman-Rakic, Muly, & Williams, 2000). Conver-
ging evidence indicates that hypo- or dysfunctioning
catecholamine systems in the prefrontal cortex
(PFC), among other areas, are a central neurobiolo-
gical substrate of the cognitive and behavioral defi-
cits associated with ADHD (Arnsten & Li, 2005).
Based on neurocomputational modeling (Sikstro
¨m&
So
¨derlund, 2007), we suggest that dopamine-
deprived neural systems, such as are thought to
occur in ADHD (Solanto, 2002) or in aging (Erixon-
Lindroth et al., 2005), require more noise to induce
SR (see Figure 1).
ADHD is believed to involve a hypofunctional
dopamine system (Solanto, 2002). The MBA model
assumes, consistent with earlier dopamine models
(Li & Sikstro
¨m, 2002; Servan-Schreiber, Printz, &
Cohen, 1990), that the level of dopamine is modu-
lated by the gain parameter in the sigmoid activa-
tion-function. A low dopamine level corresponds to a
low gain, yielding a relatively more linear input–
output relation in neural cells compared to high
dopamine and high gain. The neural system is
influenced by stochastic resonance as the signal
plus noise passes a threshold during generation of
action potentials. Neurocomputational simulations
by Sikstro
¨m and So
¨derlund (2007) showed that low
dopamine levels in ADHD subjects shift performance
on the stochastic resonance curve (inverted U-curve)
to the right, so that ADHD subjects, for a given noise
level, operate on the part of the curve where noise is
beneficial for performance whereas under the same
conditions controls operate on the part of the curve
where performance declines (see Figure 1). Input
noise
performance
Control
ADHD
SPT
ADHD
improve
with noise
VT
control
decline
with noise
Figure 1 ADHD needs more noise for optimal per-
formance compared to control. Note. The figure shows
the stochastic resonance phenomena where perform-
ance on cognitive tests (y-axis) is optimal for moderate
noise levels (x-axis), and attenuated for both too low
and too high noise levels. More noise is required for
optimal performance in low dopamine (ADHD) com-
pared to high dopamine (control) neural systems, where
dopamine modulates the gain in the sigmoid activation-
function. SPT has a higher SNR ratio (left side of the
figure) compared to VT (right side)
Listen to the noise 841
2007 The Authors
Journal compilation 2007 Association for Child and Adolescent Mental Health.
parameters to the model are external noise and
signal, which activate internal neural noise and sig-
nal. Through the SR phenomenon these provide an
output measured by cognitive performance. Thus,
these simulations provide a straightforward predic-
tion of noise-induced improvement in cognitive per-
formance for ADHD. The purpose of this paper is to
explicitly set up an experiment to test this novel
prediction.
Because we were interested in investigating per-
formance for different signal and noise levels (map-
ping to different parts of the stochastic resonance
curve), we used four different encoding conditions.
The conditions were: external auditive noise vs. no
noise and high vs. low memory performance tasks.
Low memory performance is associated with a high
internal noise level whereas high memory perform-
ance is associated with a low internal noise level. The
external auditive noise activates the internal neural
noise and the internal noise influences performance
through the phenomenon of SR.
For the low noise, or high recall performance, we
used a self-performed task. The self-performed task
(SPT) paradigm is known to help focus attention by
means of enactment. SPT yields an efficient encoding
condition that requires few conscious strategies (R.L.
Cohen, 1981). Participants are presented with verbal
commands, simple verb–noun sentences such as
‘roll the ball’ or ‘break the match’. While these com-
mands are presented, participants are asked to
perform the action indicated by each command. At
the subsequent memory test, participants are
instructed to remember as many of the verbal com-
mands presented as possible. For the high noise, or
low recall performance, we used a verbal task (VT)
that includes the same type of verbal commands,
and the same study time, as in the SPT condition
except that they are presented to the participant
without any instructions to perform any actions.
Results from experiments using this paradigm are
very stable; memory performance after enacted
phrases (SPT) is consistently superior to the ones
without enactment (VT) and is generally referred
to as the SPT effect (see Nilsson, 2000, p. 137 for a
review).
The MBA model predicts that cognitive perform-
ance in ADHD children benefits from noisy environ-
ments because the dopamine system modulates the
SR phenomenon. It suggests that the stochastic
resonance curve is right shifted in ADHD due to
lower gain or lower dopamine. The MBA model pre-
dicts that for a given cognitive task ADHD children
require more external noise or stimulation, com-
pared to control children, in order to reach optimal
(i.e., moderate) brain arousal level. However, in the
high noise condition (VT task) performance will be
near the peak for ADHD children whereas controls
will operate on the part of the SR curve where there is
too much noise for optimal performance. That is,
noise will attenuate performance for controls but not
for ADHD children. In the low noise condition (SPT
task) ADHD children will operate on the part of the
SR curve where noise is beneficial for performance,
whereas controls operate near the peak. That is, in
the SPT condition noise will increase performance for
ADHD children but not for controls. Each participant
is exposed to four conditions: white auditory noise
and a control condition without noise during SPT
and VT encoding.
Method
Participants
Forty-two children, aged 9.4–13.7 years, participated in
the study. The ADHD group consisted of 21 boys and no
girls. This group was diagnosed by pediatricians (in
hospitals or local neuro-teams) according to the guide-
lines of DSM-IV (APA, 1994). Fifteen of the children
were diagnosed ADHD-combined type (ADHD-C) and
six as predominantly inattentive (ADHD-I). Diagnoses
were given 1–4 years prior to the experiment and the
children were 6–11 years old at the time of diagnosis
(M ¼8.1 yrs). An interview based on Conner’s rating
scale for teachers confirmed, in all cases, the diagnostic
distinction between ADHD-C and ADHD-I at the time of
the experiment.
Although most of the participants (14) did not use
medication, a smaller group (7) of the ADHD children
used methylphenidate, supplied for one month or lon-
ger (see Table 1). The medicated children comprised the
ADHD-C group in six cases; only one child in the
ADHD-I group was given medication. The medication
was administered in the morning; three of the children
also got an additional dose during the day. For ethical
and practical reasons, the medicated children remained
on medication and the test was conducted in the
morning during a normal school day. The participants
used no other types of medication. To control for poss-
ible confounding effects, the medicated and non-
medicated groups were analyzed separately. We focus
on the non-medicated participants. Further co-dia-
gnoses such as conduct disorder and mental retarda-
tion were used as exclusion criteria. The ADHD children
attended either regular school in small separate groups
(10 children) or schools for children with special needs
(11 children).
The control group was matched to the ADHD group
on the basis of four inclusion criteria; district (controls
were chosen from the same area as the experimental
Table 1 Participant characteristics
School performance
Medicated N Age (SD) (1) (2) (3) (M)
ADHD (no med.) – 14 11.2 (1.2) 2 9 3 (2.1)
ADHD (tot) 7 21 11.2 (1.1) 3 13 5 (2.1)
Control – 21 11.2 (1.1) 3 12 6 (2.1)
Note. Only boys were tested. School performance was judged
by teachers as: 1 ¼below average, 2 ¼average, or 3 ¼above
average. Medication was Methylphenidate or equivalent.
842 Go
¨ran So
¨derlund, Sverker Sikstro
¨m, and Andrew Smart
2007 The Authors
Journal compilation 2007 Association for Child and Adolescent Mental Health.
group), gender (boys), age (months), and school per-
formance (teacher ratings). Teachers made a judgment
of school performance on three levels; average, above
average and below average, based on what is expected
for the age according to the curriculum (see Table 1).
Teachers’ school performance ratings corresponded
well with the earlier WISC scores obtained at the time of
the ADHD diagnosis. IQs below 80 were excluded.
Teacher interviews confirmed that all control children
were well within the normal range on the Conner rating
scale and intelligence was within a normal range. The
study was conducted at participants’ schools following
permission from parents, headmasters, and approval of
the local ethics committee at the Department of Psy-
chology, Stockholm University.
Design
The design was a 2 ·2·2, where type of encoding
(subject performed task vs. verbal task) and noise (no
noise versus noise) were the within-subject manipula-
tions and the between-group variable was ADHD versus
Control.
Materials
The to-be-remembered (TBR) items consisted of 96
sentences divided into 8 separate lists with 12 verb–
noun sentences in each list. Each sentence consisted of
a unique verb and a unique noun (e.g., ‘roll the ball’).
The sentences were placed in random order. All to-be-
remembered sentences were recorded on a CD. In the
no noise conditions the sentences were read in silence
and in the noise conditions they were read in the
presence of white noise. The equivalent continuous
sound level of the white noise and the speech signal was
81 and 80 dB (A), respectively. Thus, the signal-to-
noise ratio was )1 dB. The noise level was set in
accordance with earlier studies where an effect of SR on
cognition (arithmetic) was obtained for a normal popu-
lation (Usher & Feingold, 2000) and on working memory
for Alzheimer patients (Belleville, Rouleau, Van der
Linden, & Collette, 2003). Recordings were made in a
sound studio.
Procedure
Participants were tested individually before lunch. The
experiment lasted for about 45 minutes. First, two
training sentences were presented. There were 4 condi-
tions; SPT, SPT + noise, VT, and VT + noise. SPT/VT
conditions comprised every second list and noise or no
noise was used during every second SPT/VT encoding
condition. The encoding conditions (SPT/VT, no noise/
+noise) were counterbalanced across participants so
that each list was present in every condition equally
many times. List-order (1–8) and condition-order (SPT/
VT and no noise/+noise) were also counterbalanced.
Participants sat at a table screened off from a part of the
table where the to-be-remembered objects were placed.
The items in the SPT conditions required one or two
physical objects. These objects were given to particip-
ants at the time of presentation of the sentence (spoken
as commands to the participants) and were then hidden
behind a screen after the actions had been performed.
The rate of presentation was the same for all conditions
and controlled by the recording on the CD. A new sen-
tence was read every 9th second. Time taken to present
each list of 12 sentences was approximately 1 minute
and 40 seconds. Directly after presentation of the last
item in a list, participants performed a free-recall test in
which they spoke out loud as many sentences as poss-
ible, in any order. Recall time was measured and the
maximum allowed time was 2 minutes.
Results
Recall performance
A2·2·2 mixed ANOVA was conducted with one
between-subject factor, Group (ADHD vs. Control)
and two within-subjects factors, encoding condition
(SPT vs. VT) and noise (no noise vs. +noise). Con-
sistent with earlier SPT studies, strict scoring was
used for the nouns (exact matches were required)
and lenient scoring was used for the verbs (exact
matches not required).
There was a main effect of encoding, where SPT
outperformed VT (F(1,33) ¼45.85, p¼.000, eta
2
¼
.58). The interaction between group and noise was
also significant (see Figure 2, F(1,33) ¼5.73, p¼
.023, eta
2
¼.15). No other main effects or inter-
actions were found. When medicated children were
included in the assessment the interaction between
group and noise became stronger (F(1,40) ¼8.41,
p¼.006, eta
2
¼.17).
Table 2 shows means and standard deviations for
the proportion of correctly recalled items divided into
group, medication, noise level, and encoding condi-
tion. Consistent with our hypothesis, noise
enhanced performance for the ADHD group (M ¼.44
vs. .46) and impaired performance for the control
group (M ¼.47 vs. .43). Paired-sample t-tests were
conducted to test the predictions of SNR within
tasks. In the SPT conditions, consistent with the
prediction ADHD participants performed better with,
compared to without, noise, when all ADHD parti-
Control
ADHD
0.0
0.1
0.2
0.3
0.4
0.5
0.6
No Noise
Noise
Free Recall, Percent Correct
Interaction p = .023
Figure 2 Percentage correct answers in free recall as a
function of noise and group
Listen to the noise 843
2007 The Authors
Journal compilation 2007 Association for Child and Adolescent Mental Health.
cipants were included in the analysis (t(20) ¼2.56,
p¼.01, one-tailed); however, when only the
non-medicated children were included the result did
not reach significance but indicated a trend (t(13) ¼
1.59, p¼.07, one tailed). In the control group, noise
did not significantly influence SPT performance.
However, consistent with the prediction, the control
group performed significantly lower in VT + noise
compared to the VT condition (t(20) ¼)2.47, p¼
.01, one-tailed) whereas noise did not influence the
ADHD group in the VT noise condition. In summary,
the tests of our directed hypotheses were significant.
These tests are more precise tests of our predictions
than the three-way interaction between group, noise,
and encoding which was not significant (F(1,33) ¼
.49, p¼.49, eta
2
¼.02).
Discussion
The most intriguing result in the present study is the
positive effect of white noise on performance for the
ADHD children. This noise effect was present in both
the non-medicated and medicated children. This
supports the MBA (Moderate Brain Arousal) model
(Sikstro
¨m&So
¨derlund, 2007), suggesting that the
endogenous (neural) noise level in children with
ADHD is sub-optimal. MBA accounts for the noise-
enhancing phenomenon by stochastic resonance
(SR). The model suggests that noise in the environ-
ment introduces internal noise into the neural sys-
tem through the perceptual system. Of particular
importance, the MBA model suggests that the peak
of the SR curve depends on the dopamine level, so
that participants with low dopamine levels (ADHD)
require more noise for optimal cognitive performance
compared to controls.
Three ADHD models – cognitive-energetic (Ser-
geant, 2000), delay aversion (Sonuga-Barke, 2002b),
and optimal stimulation (Zentall & Zentall, 1983) –
argue that state factors have to be taken into account
when explaining deficits seen in ADHD. These state
factors could be conceptualized as arousal and
activation regulation and deficiencies that lead to
impairments in allocation of cognitive resources.
However, in contrast to MBA, none of these models
use stochastic resonance modulated by dopamine as
an explanatory framework to account for cognitive
performance in ADHD.
The cognitive-energetic model focuses on energetic
levels. For example, ISIs have been found to alter the
participants’ energetic state, where both over- and
under-arousal could be induced by the applied event
rate (Sergeant, 2000, 2005). As confirmatory evid-
ence, methylphenidate has been found to have the
same effect as an increased event rate, where both
are seen as state-regulating factors (van der Meere,
Gunning, & Stemerdink, 1999). Furthermore, ADHD
children show reduced P300 amplitudes to cues and
distractors (Banaschewski et al., 2003). Energetic
level can also be manipulated through cognitive load,
signal intensity and novelty (Sergeant, 2005). The
MBA model is consistent with this but also points
out the possibility of increasing energetic level
irrespective of task and improving cognitive
performance by the use of noise.
The optimal stimulation model (Zentall & Zentall,
1983) is a homeostatic model, suggesting that there
is an optimal level of stimulation toward which
organisms strive. It is argued that hyperactivity
stems from low levels of arousal and serves to
maintain an optimal arousal level. Hyperactivity,
impulsivity, and a short attention span should be
seen as a form of self-stimulation to achieve an
optimal arousal level. Behaviors supporting this view
are reward-seeking and stimulation-seeking behav-
iors often seen in ADHD (Zentall & Zentall, 1983).
More recent research has found that in the presence
of highly appealing toys ADHD children spent half as
much time attending to, and recalled less of, the
content in TV programs (Lorch et al., 2000). The MBA
model is consistent with the proposed need of
external stimulation in ADHD but elaborates on the
conditions when this stimulation will be beneficial.
In the delay aversion model attention is allocated
toward environmental stimulation that speeds up
the perceived passage of time. Intolerance of waiting
is manifested as a tendency to select an immediate
reward rather than a larger delayed reward (Sonuga-
Barke, 2002b). Altered reward processes in ADHD
(Sonuga-Barke, 2003) could be explained as a ceiling
effect due to an excessive phasic DA response to
novel stimuli. Delay aversion is found in over-sensi-
tivity to inter-stimulus intervals (Sonuga-Barke,
2002a), an increase in activity and inattention dur-
ing delay periods (van der Meere et al., 1999), and
avoidance of delay. This over-sensitivity to external
stimulation is suggested to be caused by an over-
Table 2 Proportion of items correctly recalled across encoding conditions and groups (SPT–VT, Noise–No noise, ADHD–Control,
Medicated–Non-medicated)
Group N
Type of encoding
SPT (SD) SPT+noise (SD) VT (SD) VT+noise (SD)
ADHD 21 .47 (.12) .52 (.12) .40 (.12) .41 (.10)
ADHD-non-medicated 14 .47 (.11) .50 (.11) .40 (.15) .42 (.11)
ADHD-medicated 7 .49 (.14) .55 (.13) .39 (.06) .39 (.06)
Control 21 .52 (.14) .50 (.13) .42 (14) .35 (.13)
844 Go
¨ran So
¨derlund, Sverker Sikstro
¨m, and Andrew Smart
2007 The Authors
Journal compilation 2007 Association for Child and Adolescent Mental Health.
active alerting system in ADHD that makes behavi-
oral responses maladaptive to external demands
(Nigg & Casey, 2005). This view is complementary to
the MBA model, where prolonged ISIs produce in-
sufficient phasic responses generating too little do-
pamine, and resulting in a dysfunctional arousal
state (Sikstro
¨m&So
¨derlund, 2007).
The beneficial effects of noise in cognitive perfor-
mance for ADHD have not been considered earlier,
nor have these effects been systematically tested, in
the literature. Surprisingly few experiments have
explored the possibilities of stimulating participants
with noise and there are no theories about positive
effects of noise in the literature apart from SR ex-
periments referred to in the introduction. Most ex-
periments since Broadbent’s days deal with the
negative effects of noise and distraction. We know of
only two ADHD studies using noise stimulation;
however, neither of these invoked the concept of
stochastic resonance as an explanatory framework
nor are they theory driven, rather they refer to gen-
eral appeal or arousal. Abikoff et al. (1996) attrib-
uted the enhancing effect to increased level of
general appeal counteracting boredom, and Gerjets
and colleagues (Gerjets et al., 2002) to optimal
stimulation in line with the early optimal stimulation
theory (Zentall & Zentall, 1983).
However, research has shown enhancing effects of
white noise on non-clinical groups (90 dB) on simp-
ler, short-term memory tasks such as anagrams
(Baker & Holding, 1993) whereas speech noise was
detrimental. These noise effects also interacted with
other variables such as gender and time of the day
(Holding & Baker, 1987), which makes these results
equivocal. In simple addition tasks white noise
(80 dB) improved performance, in both elderly and
younger participants, as compared to a no-noise
condition (Harrison & Kelly, 1989). More recent ex-
periments providing white noise found no effect on
cognition in digit-span recall in comparison with
irrelevant speech, which attenuated performance
(Belleville et al., 2003; Rouleau & Belleville, 1996). In
Belleville et al.’s (2003) experiment a small, but non-
significant, increment was seen among older and
Alzheimer patients as compared with young particip-
ants using white noise (75 dB). Furthermore, extra
noise required in old age to induce SR was modeled
by Li and colleagues (2006). White noise also im-
proved performance in monkeys in a delayed task
experiment, whereas Mozart’s piano music was
found detrimental (Carlson, Rama, Artchakov, &
Linnankoski, 1997). In experiments were ecolo-
gically relevant noise was studied, effects on episodic
and semantic memory showed that both road traffic
noise (62 dB followed by 78 dB sequences) and
meaningful irrelevant speech were detrimental for
memory performance. Episodic memory was found
particularly vulnerable to noise and irrelevant
speech was most detrimental for memory perform-
ance. Under some conditions road traffic noise did
not interfere with memory recall at all (Boman, En-
marker, & Hygge, 2005). For example, in Zentall and
Shaw’s (1980) experiment high levels of speech noise
(69 dB) were detrimental for ADHD whereas low
levels (64 dB) were beneficial for cognitive perform-
ance. However, fan noise where the main energy is
below 1000 Hz did not have a positive effect on
ADHD children. Noise and signal levels were also
lower as in the present experiment (50dBHL, SNR
+10 dB) (Geffner et al., 1996). Stimulus levels in the
present experiment were placed according to earlier
studies that have found SR in cognitive tests (Usher
& Feingold, 2000).
In summary, the literature review above suggests
that noise has to be continuous (i.e., not attention-
removing) and at a high energy level at all fre-
quencies, for example white or pink noise, to induce
the SR effect. Furthermore, beneficial noise levels
may vary between groups, i.e., ADHD subjects, the
elderly and people with Alzheimer’s require more
noise to induce SR. In a follow-up experiment we will
manipulate noise levels under the hypothesis that
they have to be estimated on an individual basis, see
arguments below.
As the first paper studying the stochastic reson-
ance phenomenon in ADHD, there are limitations in
the current study that should be investigated in fu-
ture studies. For example, our study investigated
only two noise levels and two encoding conditions,
thus it would be interesting to include more levels so
that the entire stochastic resonance curve can be
mapped out. Further studies should also measure
individual dopamine levels, and study how these
levels correlate to symptom severity, intellectual
capacity and the required noise level. Medicated
participants responded as strongly to noise as non-
medicated participants. However, medication is a
confounding variable, and it would be interesting to
look for interaction between noise and medication
during noise exposure.
There are several clinical implications of the MBA
model. For example, it can be used to understand
shortcomings in cognitive functioning for patient
groups where changes in the dopamine system have
been identified. While noise affects children selecti-
vely, it can be used as a complement or as an alter-
native to medication in ADHD. Moreover, reinforced
cognitive processing by noise could have applied
implications for clinical groups, as well as for normal
populations. The MBA model may be used to create
appropriate and adaptive environments for ADHD
children, especially in school settings. White noise
can be replaced with more pleasant auditive stimu-
lation such as music or other pleasant sounds.
Klingberg and colleagues have attained remark-
able results with Robomemo, a computer game that
trains working memory (Klingberg et al., 2005). In
this context, the MBA model can serve as a tool for
tailoring individually adapted treatments for ADHD
children. Computerized training programs are
Listen to the noise 845
2007 The Authors
Journal compilation 2007 Association for Child and Adolescent Mental Health.
particularly interesting because crucial variables
can be manipulated easily and precisely. This pro-
vides us with the hope of creating long-term changes
as an alternative to short-term medications. Further
research will exploit the effect of white noise and
stochastic resonance in the context of learning and
ADHD.
Correspondence to
Go
¨ran So
¨derlund, Department of Psychology,
Stockholm University, 106 91 Stockholm, Sweden;
Tel: ++46 8 16 38 76; Fax: ++46 8 15 93 42; Email:
gsd@psychology.su.se
References
Abikoff, H., Courtney, M.E., Szeibel, P.J., & Koplewicz,
H.S. (1996). The effects of auditory stimulation on the
arithmetic performance of children with ADHD and
nondisabled children. Journal of Learning Disabil-
ities,29, 238–246.
APA. (1994). Diagnostic and statistical manual of mental
disorders (4th edn). Washington DC: American Psy-
chiatric Association.
Arnsten, A.F., & Li, B.M. (2005). Neurobiology of
executive functions: Catecholamine influences on
prefrontal cortical functions. Biological Psychiatry,
57, 1377–1384.
Baker, M.A., & Holding, D.H. (1993). The effects of noise
and speech on cognitive task performance. The
Journal of General Psychology,120, 339–355.
Banaschewski, T., Brandeis, D., Heinrich, H., Albrecht,
B., Brunner, E., & Rothenberger, A. (2003). Associ-
ation of ADHD and conduct disorder – brain electrical
evidence for the existence of a distinct subtype.
Journal of Child Psychology and Psychiatry,44,
356–376.
Belleville, S., Rouleau, N., Van der Linden, M., &
Collette, F. (2003). Effect of manipulation and irrele-
vant noise on working memory capacity of patients
with Alzheimer’s dementia. Neuropsychology,17, 69–
81.
Boman, E., Enmarker, I., & Hygge, S. (2005). Strength
of noise effects on memory as a function of noise
source and age. Noise Health,7, 11–26.
Broadbent, D.E. (1958). The effects of noise on beha-
viour. In. Elmsford, NY: Pergamon Press, Inc.
Carlson, S., Rama, P., Artchakov, D., & Linnankoski, I.
(1997). Effects of music and white noise on working
memory performance in monkeys. Neuroreport,8,
2853–2856.
Castellanos, F.X., & Tannock, R. (2002). Neuroscience
of attention-deficit/hyperactivity disorder: The
search for endophenotypes. Nature Reviews Neuro-
science,3, 617–628.
Cohen, J.D., Braver, T.S., & Brown, J.W. (2002).
Computational perspectives on dopamine function
in prefrontal cortex. Current Opinion in Neurobiology,
12, 223–229.
Cohen, R.L. (1981). On the generality of some memory
laws. Scandinavian Journal of Psychology,22, 267–
281.
Corbett, B., & Stanczak, D.E. (1999). Neuropsycholo-
gical performance of adults evidencing attention-
deficit hyperactivity disorder. Archives of Clinical
Neuropsychology,14, 373–387.
Erixon-Lindroth, N., Farde, L., Wahlin, T.B., Sovago, J.,
Halldin, C., & Backman, L. (2005). The role of the
striatal dopamine transporter in cognitive aging.
Psychiatry Research,138, 1–12.
Geffner, D., Lucker, J.R., & Koch, W. (1996). Evaluation
of auditory discrimination in children with ADD and
without ADD. Child Psychiatry and Human Develop-
ment,26, 169–180.
Gerjets, P., Graw, T., Heise, E., Westermann, R., &
Rothenberger, A. (2002). Deficits of action control and
specific goal intentions in hyperkinetic disorder. II:
Empirical results [Handlungskontrolldefizite und
sto
¨rungsspezifische Zielintentionen bei der Hyper-
kinetischen Sto
¨rung: II: Empirische Befunde]. Zeits-
chrift fu
¨r Klinische Psychologie und Psychotherapie:
Forschung und Praxis,31, 99–109.
Goldman-Rakic, P.S., Muly, E.C., 3rd., & Williams, G.V.
(2000). D (1) receptors in prefrontal cells and circuits.
Brain Research.Brain Research Reviews,31, 295–
301.
Harrison, D.W., & Kelly, P.L. (1989). Age differences in
cardiovascular and cognitive performance under
noise conditions. Perception and Motor Skills,69,
547–554.
Higginbotham, P., & Bartling, C. (1993). The effects of
sensory distractions on short-term recall of children
with attention deficit-hyperactivity disorder versus
normally achieving children. Bulletin of the Psycho-
nomic Society,31, 507–510.
Holding, D.H., & Baker, M.A. (1987). Toward mean-
ingful noise research. The Journal of General Psycho-
logy,114, 395–410.
Kiefer, M., Ahlegian, M., & Spitzer, M. (2005). Working
memory capacity, indirect semantic priming, and
stroop interference: Pattern of interindividual pre-
frontal performance differences in healthy volunteers.
Neuropsychology,19, 332–344.
Klingberg, T., Fernell, E., Olesen, P.J., Johnson, M.,
Gustafsson, P., Dahlstrom, K., et al. (2005). Com-
puterized training of working memory in children with
ADHD – a randomized, controlled trial. Journal of the
American Academy of Child and Adolescent Psychi-
atry,44, 177–186.
Li, S.C., & Sikstro
¨m, S. (2002). Integrative neuro-
computational perspectives on cognitive aging,
neuromodulation, and representation. Neuroscience
and Biobehavioral Reviews,26, 795–808.
Li, S.C., von Oertzen, T., & Lindenberger, U. (2006). A
neurocomputational model of stochastic resonance
and aging. Neurocomputing,69, 1553–1560.
Lorch, E.P., Milich, R., Sanchez, R.P., van den Broek,
P., Baer, S., Hooks, K., et al. (2000). Comprehension
of televised stories in boys with attention deficit/
hyperactivity disorder and nonreferred boys. Journal
of Abnormal Psychology,109, 321–330.
Martinussen, R., Hayden, J., Hogg-Johnson, S., &
Tannock, R. (2005). A meta-analysis of working
memory impairments in children with attention-
deficit/hyperactivity disorder. Journal of the Amer-
ican Academy of Child and Adolescent Psychiatry,44,
377–384.
846 Go
¨ran So
¨derlund, Sverker Sikstro
¨m, and Andrew Smart
2007 The Authors
Journal compilation 2007 Association for Child and Adolescent Mental Health.
Moss, F., Ward, L.M., & Sannita, W.G. (2004). Stochas-
tic resonance and sensory information processing: A
tutorial and review of application. Clinical Neurophy-
siology,115, 267–281.
Nigg, J.T., & Casey, B.J. (2005). An integrative theory of
attention-deficit/ hyperactivity disorder based on the
cognitive and affective neurosciences. Development
and Psychopathology,17, 785–806.
Nilsson, L.G. (2000). Remembering actions and words.
In E. Tulving & F.I.M. Craik (Eds.), The Oxford
handbook of memory (pp. 137–148). Oxford: Oxford
University Press.
Rouleau, N., & Belleville, S. (1996). Irrelevant speech
effect in aging: An assessment of inhibitory processes
in working memory. The Journals of Gerontology.
Series B, Psychological Sciences and Social Sciences,
51, P356–363.
Russell, D.F., Wilkens, L.A., & Moss, F. (1999). Use of
behavioural stochastic resonance by paddle fish for
feeding. Nature,402, 291–294.
Sergeant, J.A. (2000). The cognitive-energetic model: An
empirical approach to attention-deficit hyperactivity
disorder. Neuroscience and Biobehavioral Reviews,
24, 7–12.
Sergeant, J.A. (2005). Modeling attention-deficit/
hyperactivity disorder: A critical appraisal of the
cognitive-energetic model. Biological Psychiatry,57,
1248–1255.
Servan-Schreiber, D., Printz, H., & Cohen, J.D. (1990).
A network model of catecholamine effects: Gain,
signal-to-noise ratio, and behavior. Science,249,
892–895.
Sikstro
¨m, S., & So
¨derlund, G.B.W. (2007). Stimulus
dependent dopamine release in ADHD. Manuscript
under revision.
Solanto, M.V. (2002). Dopamine dysfunction in AD/HD:
Integrating clinical and basic neuroscience research.
Behavioral Brain Research,130, 65–71.
Sonuga-Barke, E.J. (2002a). Interval length and time-
use by children with AD/HD: A comparison of four
models. Journal of Abnormal Child Psychology,30,
257–264.
Sonuga-Barke, E.J. (2002b). Psychological hetero-
geneity in AD/HD – a dual pathway model of behaviour
and cognition. Behavioral Brain Research,130, 29–36.
Sonuga-Barke, E.J. (2003). The dual pathway model of
AD/HD: An elaboration of neuro-developmental char-
acteristics. Neuroscience and Biobehavioral Reviews,
27, 593–604.
Usher, M., & Feingold, M. (2000). Stochastic resonance
in the speed of memory retrieval. Biological Cyber-
netics,83, L11–16.
van der Meere, J., Gunning, B., & Stemerdink, N.
(1999). The effect of methylphenidate and clonidine
on response inhibition and state regulation in chil-
dren with ADHD. Journal of Child Psychology and
Psychiatry,40, 291–298.
Zentall, S.S., & Shaw, J.H. (1980). Effects of classroom
noise on performance and activity of second-grade
hyperactive and control children. Journal of Educa-
tional Psychology,72, 830–840.
Zentall, S.S., & Zentall, T.R. (1983). Optimal stimula-
tion: A model of disordered activity and performance
in normal and deviant children. Psychological Bul-
letin,94, 446–471.
Manuscript accepted 24 January 2007
Listen to the noise 847
2007 The Authors
Journal compilation 2007 Association for Child and Adolescent Mental Health.
