Hindawi

Behavioural Neurology
Volume 2019, Article ID 5682050, 16 pages
https://doi.org/10.1155/2019/5682050

Review Article
Behavioral and Cognitive Impacts of Mindfulness-Based
Interventions on Adults with Attention-Deficit Hyperactivity
Disorder: A Systematic Review

Hélène Poissant ,1 Adrianna Mendrek,2 Nadine Talbot ,3 Bassam Khoury,4

and Jennifer Nolan1
1
Universite du Quebec à Montréal, Education and Pedagogy Department, Montréal, Quebec, Canada H3C 3P8
2
Bishop’s University, Psychology Department, Sherbrooke, Quebec, Canada J1M 1Z7
3
Universite du Quebec à Trois-Rivières, Sciences Education Department, Trois-Rivières, Canada G9A 5H7
4
McGill University, Educational and Counselling Psychology, Montréal, Quebec, Canada H3A 1Y2

Correspondence should be addressed to Hélène Poissant; [email protected]

Received 10 August 2018; Revised 18 December 2018; Accepted 20 January 2019; Published 4 April 2019

Academic Editor: Jesus Pastor

Copyright © 2019 Hélène Poissant et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Mindfulness-based interventions (MBIs) are becoming increasingly popular as treatments for physical and psychological problems.
Recently, several studies have suggested that MBIs may also be effective in reducing symptoms of attention-deficit hyperactivity
disorder (ADHD). Most studies have examined the effectiveness in children, but there are now a sufficient number of individual
treatment trials to consider a systematic review in adults. Majority of existing systematic reviews and meta-analyses only
consider ADHD symptoms as an outcome, and most of them do not fully report potential biases of included studies, thus
limiting considerably their conclusions. This is an important facet because some studies could be found ineligible to be included
in future analysis due to their low quality. In this systematic review, we followed the PRISMA/PICO criteria and we thoroughly
assessed the risks of bias for each of the selected studies according to Cochrane guidelines. We searched the available literature
concerning MBIs in adult participants with ADHD using PsycINFO, PubMed, Scopus, and ERIC databases. In total, 13 studies
conducted with 753 adults (mean age of 35.1 years) were identified as eligible. Potential moderators such as participants’ age,
ADHD subtypes, medication status, comorbidity, intervention length, mindfulness techniques, homework amount, and training
of therapists were carefully described. Aside from measuring the symptoms of ADHD, outcome measures were categorized into
executive/cognitive functioning, emotional disturbances, quality of life, mindfulness, and grade point average at school.
According to presented descriptive results, all the studies (100%) showed improvement of ADHD symptoms. In addition,
mindfulness meditation training improves some aspects of executive function and emotion dysregulation. Although these are
promising findings to support treatment efficacy of MBIs for ADHD, various biases such as absence of randomization and lack
of a control group may affect the actual clinical value and implications of the studies. Moreover, the relatively low quality of
selection and performance criteria in several studies, as well as relatively high attrition bias across studies, call for caution before
considering conducting further analysis.

1. Introduction (body scan, sitting meditation, and mindful yoga) that are
thought to help participants cultivate nonjudgmental, mind-
Mindfulness-based interventions (MBIs) have gained in pop- ful awareness of present-moment experience. Recently, sev-
ularity over the past decade. Clinical trials provide evidence eral studies have suggested that MBIs may also be effective
of their effectiveness in the treatment of depression, anxiety, in reducing symptoms of attention-deficit hyperactivity dis-
addictions, and other mental health problems. Most MBIs order (ADHD). Most studies have examined the effectiveness
involve three somatically focused meditative techniques in children. There are now a sufficient number of individual
2 Behavioural Neurology

treatment trials to consider a systematic review in adults. The present systematic review attempts to provide more
To our knowledge, only Cairncross and Miller [1] have evidence for the use of MBIs in adults with ADHD by
conducted a meta-analysis (six studies in children and four means of investigation into several variables and character-
in adults) on ADHD to measure the impact of MBIs on istics that may moderate the effectiveness of MBIs. It is
symptoms of ADHD. The authors only considered publi- important for us to report these elements so that the relation-
cation bias in their study. The present systematic review ship between these characteristics and intervention effective-
goes further by exploring in detail seven types of biases ness can be better understood. Moreover, incorporation of an
according to Cochrane guidelines. This is an important exhaustive analysis of biases with an assessment of the quality
aspect of the present review because it allows identifying of studies is presented here as an essential element of the
studies with poor quality. systematic review.
This systematic review examines if MBIs are effective
treatments of attention-deficit hyperactivity disorder (ADHD)
in adults. ADHD is characterized by marked behavioral 2. Methods
symptoms such as inattention and/or hyperactivity and act-
2.1. Eligibility Criteria. In order to conduct the systematic
ing impulsively. The prevalence of the disorder is about
review, we used the criteria from PRISMA-P [8]. These cri-
3-4% in adults, and it is higher in males than in females.
teria allow following a stepwise methodology in conducting
ADHD often exists simultaneously with other conditions,
and reporting the outcomes of the systematic review. A data
such as anxiety, depression, and personality disorders. The
sheet based on the PRISMA-P protocol was designed and
most common treatment of ADHD consists of administra-
comprised information extracted from each selected study
tion of psychostimulant medications. However, the pharma-
based on (1) research design (including RCT, N-RCT, the
cotherapy is not always effective and is associated with
presence of a waiting list, pre-posttest, follow-up, and base-
various side effects. Thus, MBIs represent a much-welcomed
line), (2) characteristics of participants (including number,
addition to available treatments or a stand-alone therapy.
age, gender, diagnosis with subtype, comorbidity, and medi-
Comprehension of the mechanisms mediating the effec-
cation), (3) characteristics of the intervention (including
tiveness of MBIs in ADHD at both the behavioral and neuro-
type, description, length, identity of therapists, and their
nal levels has greatly improved. Thus, three large neural
experience), and (4) characteristics of outcomes (ADHD
networks have been implicated both in ADHD and in mind-
symptoms, executive functioning, emotional disturbance,
fulness meditation: the default mode network, salience net-
mindfulness, and quality of life).
work, and central executive network [2–7]. Because of
We limited inclusion to peer-reviewed empirical published
space constraints, we concentrate in the present review only
studies that examined effects of meditation or mindfulness-
on the behavioral level, considering the effects of MBIs on
based interventions (MBIs) on symptoms of ADHD. All stud-
hyperactivity, emotion dysregulation, deficits in attention
ies were published in English. We excluded books, reviews,
and executive function (EF), and other problems. Most of
meta-analyses, qualitative, psychometric, or single-case stud-
these dysfunctional behaviors seem to improve following
ies, and duplicates.
MBIs. The improvement of these behaviors in turn contrib-
utes to the general well-being of adults with ADHD.
However, due to the heterogeneous presentation of 2.2. Information Sources and Search. We consulted Psy-
ADHD, it is important to document the interindividual cINFO, PubMed, Scopus, and ERIC databases from the first
differences of individuals with ADHD. These differences available date until June 2018 (+reference lists of previous
between individuals with ADHD may affect the success of reviews). By the end of the search, the selected studies cover
MBIs in reducing symptoms. For example, there is more a period from 2002 to 2018. The search terms “ADHD
evidence to suggest that MBIs are effective in reducing AND meditation OR mindfulness” and associative terms
inattention, so perhaps the intervention is more helpful for (e.g., impulsivity, inattention, and hyperactivity) were con-
individuals with ADHD with a predominantly inattentive sidered for inclusion. Impact of MBIs was extended to
type [1]. Adults being more often characterized as inatten- cognition, EF, and brain structure alterations. We included
tive (instead of hyperactive) may react in a different man- randomized and nonrandomized control trials (RCT and
ner to MBIs compared to children. Other variables could N-RCT, respectively), pre-posttest (within-group) studies,
also affect whether MBTs are effective in improving func- clinical trials, prospective or follow-up studies, and single-
tioning in individuals with ADHD. For example, it is or double-blind studies. Studies were excluded if they (1)
unclear how factors such as the length of intervention, were conducted with children, (2) did not include a mind-
mindfulness techniques, amount of homework, homework fulness- or meditation-based treatment, (3) did not include
compliance, and training of the therapist affect the out- a group of ADHD or ADD or hyperactivity disorders, (4) did
come of therapy [1]. These are important elements that not examine treatment effects, (5) did not report clinical
should be taken into account. Moreover, besides symptoms outcomes, and (6) described solely mindfulness or medita-
of inattention and hyperactivity, other indicators of efficacy tion instructions. We included different forms of MBIs as
of MBTs had been reported, namely, EFs, emotional distur- long as the intervention contained significant elements of
bance, quality of life, and academic performance. We con- mindfulness. We also excluded “gray literature,” reports,
sider these additional elements useful in supporting the and unpublished studies. By the end of the selection pro-
portrayal of ADHD. cess, we included thirteen studies conducted with adults,
Behavioural Neurology 3

young adults, and college students. The full electronic pro- Attrition bias (4) is based on the recommendation by
cess search strategy for our databases is described below. Higgins et al. [10] to rate studies with above 20% attrition
of participants as high risk. An attrition rate lower than
2.3. Risk of Bias in Individual Studies. We looked for risk of 20% where the groups (i.e., treatment and control) are equiv-
bias in each individual study. Thus, we designed and adapted alent yields a low risk. Thus, studies were rated high risk if the
a classification to report potential bias for each individual attrition rate was greater than 20% and the authors did not
study using the Cochrane Collaboration [9] recommenda- use any analyses to compensate for the missing data. Studies
tions. Biases included (1) “sequence generation” (e.g., Is the were rated unclear risk if the authors did not explicitly pro-
allocation sequence acceptably generated? “YES” if explicitly vide attrition rates or the computation of the attrition was
mentioned that the “patients were randomly allocated”), (2) not possible using the provided data.
“allocation concealment” (e.g., Is the allocation acceptably Reporting bias refers to reporting the outcome data par-
concealed? “YES” if participants and researcher could not tially or omitting to report scales (or subscales) that may lead
foresee assignment because of an explicit mention of a to a bias. We decided to rate the study low risk if all the scales
method to conceal allocation), (3) “blinding of participants, and subscales were reported. We rated the study unclear risk
personnel, and outcome assessors” (e.g., Is knowledge of when subscales were not fully reported or it was unclear
the allocated treatment plenty prevented during the study? whether omitting to report the subscales led to a bias.
“YES” in case of blinding or if the authors judged improba- Other biases included a researcher’s allegiance and fund-
bly that the outcome measurement was influenced by no ing source. These biases addressed the authors’ role in the
blinding, and (4) “selective outcome reporting” (e.g., Are study development or implementation, as well as acknowl-
partial outcome data adequately addressed? “YES” if explic- edgement of any conflict of interest. Studies were rated high
itly mentioned that nonmissing outcome data or reasons for risk when authors were actively involved in delivering inter-
missing outcome data had little impact on outcomes). The ventions, evaluating participants, or conducting any other
authors’ judgments involved answering specific questions aspects of the study. Studies were rated low risk when authors
for each query and providing a detailed entry addressing were not involved in conducting the study. Studies were rated
the sources of bias. In all cases, an answer “YES” indicates unclear when the authors did not report their involvement
a low risk of bias, an answer “NO” indicates a high risk of and information from the paper did not suggest their
bias, and an answer “unclear” indicates an uncertain risk of involvement. Regarding funding, studies were rated high risk
bias (p. 196, [9]). when sources of funding can cause a conflict of interest, low
Selection bias comprises random “sequence generation” when the study was not funded or when the sources of fund-
(1) and “allocation concealment” (2) (p. 196, [9]). According ing were disclosed with a nonconflict of interest statement,
to the criteria proposed by the Cochrane Collaboration [10], and unclear when the sources of funding were not reported.
a study was rated low risk on random sequence generation if Discussions about the judgment ratings were provided in
the method used to allocate sequence produces equivalent an iterative way until consensus about the ratings was
groups. If random sequence generation was not described reached between judges (H.P., B.K., and A.M.). Prior to these
in sufficient detail but the study was described as random- discussions, the rating coauthors familiarized themselves
ized and the groups were equivalent, we rated the risk with a series of articles and a document containing specific
unclear. The studies, which have not fulfilled either condi- instructions and examples of rating the studies from the
tion, were rated high risk. Regarding allocation concealment, Cochrane Collaboration’s tools.
we rated a study low risk when its method to conceal the
allocation sequence could not be predicted in advance of 2.4. Summary Measures and Additional Analyses. The results
or during intervention. If allocation concealment was not of the present systematic review are first presented in a narra-
described with sufficient accuracy to allow an appreciation tive manner. Tables 1 and 2 give an overview of a PICO
of whether it could be foreseen but participants would not description of each individual study considering two separate
necessarily identify the group to which they belong (i.e., types of research design: within and between subjects (see
treatment or control), the risk was rated unclear. The Tables 1 and 2, respectively). For the bias analysis, apprecia-
remaining studies, which have not fulfilled either condition, tion of the quality of each study was converted into numeric
were rated high risk. variables with quality scores ranging from 0 (high risk) and 1
Performance bias and detection bias (3) reflect the blind- (unclear risk) to 2 (low risk) on each of the seven bias evalu-
ing of participants and personnel (e.g., facilitators or trainers) ation measures (the maximum score of quality for each study
and reflect the blinding of assessors to the condition. We is 14).
considered studies low risk if they described the blinding
of outcome assessors and used only self-report subjective 3. Results
measures and/or objective measure (e.g., neuropsychological
measures). Unclear risk corresponded to self-report mea- 3.1. Study Selection. The final literature search resulted in 720
sures but nonblinding of assessors. For the nature of inter- studies: PsycINFO (n = 225), PubMed (n = 460), Scopus
ventions in this study (MBIs), it is not habitual, nor always (n = 23), and ERIC (n = 12). An Endnote file was first cre-
advantageous, to blind personnel or participants. Therefore, ated, and abstracts of all articles were saved on an electronic
we expected studies to have an elevated risk for performance file for further examination. The search was conducted in two
and/or detection bias. consecutive sessions from October 2016 to January 2017 and
4 Behavioural Neurology

Table 1

(a) Characteristics of single-group studies: research design and participants

Age
1st author % % ADHD
Research design participant N % medication status % comorbid disorders
(date) males subtypes
(y/o)
19-44, C = 75, 37.5 (MDD/Sx),
Hesslinger x = 31 9, I = 12 5, 37.5 (MPH),
Pre-post 15 62.2 25 (social phobia),
(2002) [14] s=9 0 H = 12 5 12.5 (other)
25 (insomnia)
Baseline (T1), 19–63, 74.7 (atomoxetine+central 71.1 (at least one
Morgensterns C = 86 6,
posttreatment (T2), x = 37 4, 98 31.6 stimulants), 88.2 comorbid DSM-IV
(2016) [42] I = 13 4
3-month follow-up (T3) s = 10 4 (psychoactive drugs) diagnosis)
Zylowska x = 48 5, C = 50, 83 (mood), 33 (AD),
Pre (T1) to post (T2) 24 38 I = 42, H = 8 63
(2008) [11] s = 10.9 33 (ODD), 92 (any)
(1) Statistics: n.r. = nonreported; y/o = years old; x = mean; s = standard deviation. (2) ADHD subtypes: ADHD = attention-deficit and hyperactivity disorders;
I = inattentive; H = hyperactive; C = combined. (3) Comorbid disorders: AD/Sx = anxiety disorder or symptoms; ODD = oppositional defiant disorder;
MDD/Sx = major depressive disorder or symptoms. (4) MPH = methylphenidate.

(b) Characteristics of single-group studies: intervention

1st author (date) Intervention Intervention length Therapist Informants

Hesslinger ST-based DBT+mindfulness 2 h/week,
Psychotherapists trained in DBT Self, objective
(2002) [14] components 13 weeks = 26 h
Two clinical psychologists who are trained
DBT (elements of acceptance,
Morgensterns 2 h/week, in CBT+DBT experienced from previous
mindfulness, functional behavioral Self
(2016) [42] 14 weeks = 28 h study phases (T.H.) or had clinical supervision
analysis, psychoeducation)
from the experienced group leader
2.5 h/week,
Zylowska Experienced mindfulness instructor
MAP+psychoeducation 8 weeks = 20 h+daily Self, objective
(2008) [11] (D.W.)+ADHD researchers (L.Z. & S.S.)
at-home practice
Intervention: CBT = cognitive behavioral therapy; DBT = dialectical behavior therapy; MAP = mindful awareness program; ST = skills training.

(c) Characteristics of single-group studies: measures of outcome

1st author Cognitive/executive Emotional Quality Academic

ADHD symptoms Mindfulness
(date) function disturbance of life performance
Fluency, Stroop,
Hesslinger
ADHD-CL, SCL-16 DSS, KLT, BDI n.r. n.r. n.r.
(2002) [14]
d2-Test, WMS-R
CADHDSC_SRF: ADHD
AAQ-9,
Morgensterns symptoms (functional
n.r. BDI, BAI, PSS AAQoL, n.r. MAAS
(2016) [42] impairment,
KSQ
aggression irritability)
Zylowska ADHD Rating Scale IV ANT, TMT,
BAI, BDI n.r. n.r. n.r.
(2008) [11] (adults), DSM-IV DST, VOC
(1) ADHD symptoms: ADHD-CL = Attention-Deficit Hyperactivity Disorder Checklist; CADHDSC_SRF = Current ADHD Symptom Scale-Self-Report Form;
DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; SCL-16 = Symptom Check List. (2) Cognitive/executive function:
ANT = Attention Network Task; d2-Test = selective attention; DSS = Digit Symbol Subtest; DST = Digit Span Test (WAIS-R); KLT = Konzentrations-
Leistungs-Test; TMT = Trail Making Test; VOC=vocabulary subtest (WAIS-R); WMS-R = Wechsler Memory Scale-R (mental control, digit span, visual
memory span). (3) Emotional disturbance: BAI = Beck Anxiety Inventory; BDI = Beck Depression Inventory; PSS = Perceived Stress Scale. (4) Quality of
life: AAQ-9 = Adult Quality of Life Questionnaire; AAQoL = Adult ADHD Quality of Life Questionnaire; KSQ = Karolinska Sleep Questionnaire. (5)
Mindfulness: MAAS = Mindful Attention Awareness Scale.

later updated from April 2018 to July 2018 resulting in the After reviewing the abstracts of the 542 remaining studies,
incorporation of two new articles (with the help of N.T.). 458 studies were classified as irrelevant (erratum, theoretical
The first step consisted of the elimination of 178 duplicates paper, qualitative study, no ADHD group, program develop-
(+1 erratum). The final study selection was based on eligibil- ment, case study, study protocol, no treatment, and no quan-
ity assessment from two independent reviewers (H.P. and titative outcome). From the remaining 84 studies, 56 were
A.M.). Disagreements were resolved through discussions. reviews or meta-analyses; therefore, they were eliminated
 Table 2

(a) Characteristics of between-group studies (age and percentages only for ADHD groups with treatment (Tx)): research design and participants

N of ADHD,
HC, Tx, and % ADHD % medication
1st author (date) Research design Age (y/o) % males % comorbid disorders
WL/no/other subtypes status
Tx
Behavioural Neurology

None (3 months 14 (AD), 57 (Sx), 4 (OC), none

18-65, x = 40,
Bachmann (2018) [13] RCT, pre/post 40, 0, 21, 19 38 C = 81, I = 19 before and during (Schizo, BD, SD, AU, SUI/SI,
s = 10 58
the study) ND, Somato)
18-45, x = 31 2,
Bueno (2015) [44] N-RCT, pre/post s=7 5 43, 17, 29, 31 54.5 n.r. 69.7 (MPH) n.r.

ADHD-treated vs. ADHD-WL x = 36 6, C = 73 5, 61.22 (MPH),

Cole (2016) [40] at baseline, post, 3-month and 62, 0, 49, 13 54 46.94 (MDD, BD, AD, SD, BPD)
s = 10 02 I = 22 5, H = 5 24.49 (other)
6-month follow-up, end of treatment
25.6 (BPD), 30.8 (other PD),
x = 33 8, C = 69 2, 43.6 (MPH), 17.9 (social anxiety), 10.3
Edel (2017) [32] N-RCT, WL, pre-post 91, 0, 39, 52 59
s = 10 1 I = 30 8 38.5 (other) (MDD/Sx), 5.1 (dysthymia),
2.6 (SD)
C = 5 9,
x = 21 2, 29.4 (MPH), 41.2
Fleming (2015) [35] RCT, baseline, post, 3-month follow-up 33, 0, 17, 16 58.8 I = 88 2 + 5 9 AD, MDD/Sx (% n.r.)
s = 1 67 (other), 29.4 (none)
(4 (Sx))
RCT, ADHD-treated vs.
19-24, x = 20 2, C = 6 3, I = 28.6 (MPH), 42.8
Gu (2018) [48] ADHD-WL pre-post, 54, 0, 28, 26 57.1 n.r.
s = 1 03 93 3 (other), 28.6 (none)
3-month follow-up
CAARS-INV:
18-65, x = 36 5, I = 5 2, H = 5 8 46 (MPH), 15
Hepark (2019) [45] RCT, WL 103, 0, 55, 48 38 CAARS-SR: n.r.
s = 10 (other), 40 (none)
I = 4 3, H = 5 1
RCT, ADHD-MBCT+TAU vs.
18+, x = 39 7, C = 50, I = 38, 38 (MDD/Sx), 2 (BD), 13 (AD),
Janssen (2019) [50] ADHD-TAU, baseline, post, 120, 0, 60, 60 47 60
s = 11 1 H=8 70 (Somato), 2 (ED), 2 (dysthymia)
3-month and 6-month follow-up
18-50,
C = 27 3,
Mitchell (2017) [25] WL & treatment group, pre-post x = 40 55, 20, 0, 11, 9 45.5 I = 72 7 54.5 (“stimulants”) 54.5
s = 6 83
19-53, x = 39 5, 38 (MPH), 24
Schoenberg (2014) [12] RCT, WL, pre-post 50, 0, 26, 24 37.5 n.r. n.r.
s=9 5 (other), 38 (none)
(1) Statistics and design: n.r. = nonreported; y/o = years old; x = mean; s = standard deviation; HC = healthy control; N-RCT = nonrandomized control trial; RCT = randomized control trial; Tx = treatment;
WL = waiting list. (2) ADHD subtypes: ADHD = Attention-Deficit and Hyperactivity Disorders; I = inattentive; H = hyperactive; C = combined. (3) Comorbid disorders: AD/Sx = anxiety disorder or symptoms;
AU = autism; BD = bipolar disorder; BPD = borderline personality disorder; ED = eating disorder; MDD/Sx = major depressive disorder or symptoms; ND = neurological disorders; OC = obsessive compulsive
disorders. (4) PD = any personality disorder; Somato = somatoform disorder; SD = substance dependence; SUI\SI+suicidality: self-injurious behavior. MPH = methylphenidate.
5
 6
(b) Characteristics of between-group studies: intervention

1st author (date) Intervention Intervention length Therapist Informants

2.5 h/week, 8 weeks = 20 h+daily
Bachmann (2018) [13] MAP or PE n.r. Self, objective
home practice
2.5 h/week, 8 weeks = 20 h+daily
Bueno (2015) [44] MAP or no intervention Highly experienced practitioners Self, objective
home practice
DBT (+elements of mindfulness) 2 h individual psychotherapy+group/week,
Nurses, psychologists, and psychiatrists,
Cole (2016) [40] or CBT modules 12-month period = 96 h+homework Self
trained in DBT+CBT
(impulsivity/hyperactivity, attention) assignments
MBT (+mindfulness component of Experienced psychologist working with
Edel (2017) [32] 2 h/week, 13 weeks = 26 h Self, expert-rated scale
DBT) or ST (DBT-oriented skills training) ADHD+5 y experience in DBT/MBT
Group leader (A.P.F.), coleader (L.R.M.),
1.5 h group/week, 8 weeks = 12 h+7 graduate students in clinical psychology
DBT (+elements of mindfulness)
Fleming (2015) [35] 10 min individual coaching/week+90 min with DBT training & intervention, Self, objective
or skills handouts
group (1st week follow-up) psychologist with experience in ADHD
students
Group leader, psychiatrist specializing in
1 h individual/week,
ADHD+8 y experience, MBCT trainers,
Gu (2018) [48] MBCT 6 weeks = 6 h+30 min Self, objective
psychologist with experience in ADHD
self-practice/day workbook psychoeducation
students
Psychiatrist specializing in ADHD (10 y),
mindfulness teacher (S.H.) & nurse
12-week meditation exercises built up
Hepark (2019) [45] MBDT (+PE) specialist, Association of Self, investigator
gradually+home practice 30 min/day
Mindfulness-Based Teachers, 150 h
education (MBSR)/MBCT
2.5 h group/week, 8 weeks = 20 h+6 h Mindfulness teachers at different levels
Janssen (2019) [50] MBDT (+PE) Self, objective
silent day+home practice 30 min/day of competence
Mitchell (2017) [25] MAP 2.5 h/week, 8 weeks = 20 h+home practice Ph.D. clinical psychology Self, objective
3 h/week, 12 weeks = 36 h+30-45 min Psychiatrist specializing in ADHD
Schoenberg (2014) [12] MBCT Self, objective
self-practice/day with 9 y training in MBCT
Intervention: CBT = cognitive behavioral therapy; DBT = dialectical behavior therapy; MAP = mindful awareness program; MBT/MBCT = mindfulness-based training; PE = psychoeducation; ST = skills training.

(c) Characteristics of between-group studies: measure of outcome

1st author (date) ADHD symptoms Cognitive/executive function Emotional disturbance Quality of life Academic performance Mindfulness
Bachmann (2018) [13] CAARS-SR/OR n-back n.r. n.r. n.r. n.r.
Bueno (2015) [44] ASRS ANT, CPT-2 BDI, STAI-T, PANAS-X AAQoL n.r. n.r.
Cole (2016) [40] ASRS v1.1, BIS-11 n.r. BDI-II, BHS, STAXI WHOQoL-BREF, QFS n.r. KIMS
Edel (2017) [32] DSM-IV-(SR/OR) n.r. WRI GSES n.r. MAAS
Fleming (2015) [35] BAARS-IV, BADDS CPT-2 BAI, BDI-II AAQoL GPA FFMQ
Gu (2017) [48] CAARS-S:SV ANT BAI; BDI-II VAS GPA MAAS
Behavioural Neurology
 Behavioural Neurology

Table 2: Continued.

1st author (date) ADHD symptoms Cognitive/executive function Emotional disturbance Quality of life Academic performance Mindfulness
Hepark (2019) [45] CAARS-INV, CAARS-SR BRIEF-ASR BDI-II-NL, STAI OQ 45.2 n.r. KIMS
Janssen (2019) [50] CAARS-INV:SV, CAARS-S: SV BRIEF-A n.r. OQ 45.2, MHC-SF n.r. FFMQ-SF, SCF-SF
DEFS, BRIEF-A, ANT, CPT,
Mitchell (2017) [25] Current ADHD Symptom Scale DERS, DTS n.r. n.r. n.r.
DST, TMT, WAIS-R
Schoenberg (2014) [12] CAARS-S:SV CPT-X n.r. OQ 45.2 n.r. KIMS
(2) ADHD symptoms: ASRS (v1.1) = Adult ADHD Self-Report Scale; BAARS-IV = Barkley Adult ADHD Rating Scale-IV; BADDS = Brown ADD Rating Scales; BIS-11 = 11th version of the Barratt Impulsiveness
Scale; CAARS-S:SV = Conners’ Adult ADHD Self-Rating Scale; CAARS-SV = Conners’ Adult ADHD Rating Scales-Screening Version; CAARS-SR = self-report version of the Conners’ Adult ADHD Rating Scale;
CAARS-INV = investigator rating version of the Conners’ Adult ADHD Rating Scale; DSM-IV-(SR/OR) = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (self-rating/other-rating). (2)
Cognitive/executive function: ANT = Attention Network Task; BRIEF-A = Behavior Rating Inventory of Executive Functioning-Adult Version; BRIEF-ASR = Behavior Rating Inventory of Executive
Function-Adult Self-Report version; CPT-2 = the Conners’ Continuous Performance Test-2nd edition; CPT-X = visual Continuous Performance Task; DEFS = Deficits in Executive Functioning Scale;
DST = Digit Span Test (WAIS-R); TMT = Trail Making Test; WAIS-R = Wechsler Adult Intelligence Scale-Revised. (3) Emotional disturbance: BAI = Beck Anxiety Inventory; BDI-II/BDI-II-NL = Beck
Depression Inventory (2nd edition); BHS = Beck Hopelessness Scale; DERS = Difficulties in Emotion Regulation Scale; PANAS-X = Affect Schedule-Expanded form; STAI = State-Trait Anxiety Inventory;
STAXI = State-Trait Anger Expression Inventory; WRI = Wender–Reimherr Interview. (4) Quality of life: AAQoL = Adult ADHD Quality of Life Questionnaire; GSES = Generalized Self-Efficacy Scale;
MHC-SF = Positive Mental Health Short Form; OQ 45.2 = Outcome Questionnaire; QFS = questionnaire of social functioning; VAS = Visual Analog Scale (personal health status); WHOQoL-BREF = World
Health Organization Quality of Life. (5) Mindfulness: MAAS = Mindful Attention Awareness Scale; FFMQ = Five Facet Mindfulness Questionnaire; KIMS = Kentucky Inventory of Mindfulness Skills;
SCF-SF = Self-Compassion Short Form. (6) Academic performance: GPA = Grade Point Average.
7
8 Behavioural Neurology

Literature search produced

720 papers:
PubMed: 460
PsycINFO : 225
Scopus : 23
ERIC : 12

178 papers are

duplicate
records and 542 papers are
removed recorded once
1 paper is
an erratum

458 papers are 84 papers are

irrelevant relevant

56 papers are 28 papers are

reviews or meta- empirical
analyses studies

13 papers are
with adults and
15 papers are young adults
with children and retained for
further meta-
analysis

Figure 1: Flow chart of the eligibility criteria.

(but we examined their reference lists). Among the 28 Bachmann et al. [13], the ESs were not provided by the
remaining empirical studies, 15 were conducted with chil- authors. Therefore, for consistency, we calculated the ESs
dren, adolescents, and/or their parents; therefore, they were using Comprehensive Meta-Analysis (CMA) software. Since
excluded. Thirteen studies conducted with adults, young Bachmann et al. [13] did not find evidence for a significant
adults, and college students corresponded to all the selection main effect of type of treatment (MAP vs. the psychoeduca-
criteria; therefore, they were included. A detailed illustration tion comparison group), we did not calculate any ES for this
of the study selection process is found in Figure 1. The sys- study (see below).
tematic review of the thirteen studies selected is presented In an early pilot study, Hesslinger et al. [14] evaluated a
qualitatively (narrative review). training based on dialectical behavior therapy (DBT) to suit
the special needs of adult patients with ADHD. The overall
3.2. Study Characteristics and Results of Individual Studies. treatment goal was that patients would “control ADHD
Each of the following study characteristics (e.g., study size, rather than being controlled by ADHD.” Prior to and follow-
PICOS, and follow-up period) and results of individual ing group therapy, symptoms were assessed using self-rating
study are presented in Tables 1 and 2. Outcomes in terms scales of ADHD-CL (from DSM-IV), a short version of the
of benefits are presented in a narrative manner for each SCL-16 to assess nervousness, memory deficits, carelessness,
study in the form of a simple summary data for each excitability, emotional outburst, self-reproach, difficulties
intervention group. For briefness, we choose not to list the to start, inferiority complex, sleep disturbances, concentra-
effect size (ES) for every outcome (e.g., all tests and subtests tion deficits, feeling of tension, embarrassment, exertion,
and T1-T2-T3-T4, totalizing 470 ESs for all articles). Rather, restlessness, worthlessness, thinking something is wrong
a summary of effect sizes (ESs) is given for each study with a with comprehension, and the BDI [15]) to assess depressive
main focus on symptom outcomes and posttest (T2) or symptoms (see the appendix for the complete names of tests).
follow-up. Most ESs were retrieved directly from the articles. In addition, neuropsychological testing was performed at
However, for Zylowska et al. [11], Schoenberg et al. [12], and baseline and following treatment, including a verbal and
Behavioural Neurology 9

letter fluency test, the Stroop test indicating mental speed and comprising symptom domains such as attentional difficul-
inhibitory EFs, the digit symbol subtest to evaluate divided ties, hyperactivity/restlessness, (hot) temper, affective lability,
attention, a test of continuous attention, the d2-Test measur- emotional overreactivity, disorganization, and impulsivity.
ing selective attention, and tests measuring mental control, General linear models with repeated measures revealed that
digit span, and visual memory span indicating short-term both programs resulted in a similar reduction of ADHD
memory, working memory (WM), and general attentional symptoms. The effect sizes were in the small-to-medium
capacities [16]. The DBT treatment resulted in mild to range (ESs ranging from 0.06 to 0.49). However, some degree
moderate improvements on all the measured symptoms of decrease in ADHD symptoms (30%) was more prominent
and even greater improvements in neurocognitive function for the MBT participants since 30.8% of them showed
(ESs ranging from 0.99 to 2.22). improvement compared to 11.5% of the ST participants.
Subsequently, Zylowska et al. [11] enrolled adults and Fleming et al. [35] conducted an RCT evaluating DBT
adolescents with ADHD in the mindful awareness pro- training adapted for college students with ADHD random-
gram (MAP) [17, 18] adapted to meet the challenges of ized to receive either DBT or skills handouts. ADHD symp-
ADHD symptoms, including a psychoeducational compo- toms, EF, and related outcomes were assessed at baseline,
nent. Self-report scales of ADHD, depression, and anxiety posttreatment, and 3-month follow-up. Authors used the
symptoms and several cognitive tests were administered to BAARS-IV (based on DSM-5 criteria) to assess ADHD
participants during pre- and postintervention sessions. symptoms. Self-report of current symptoms was used as a
ADHD symptoms were assessed via the ADHD Rating Scale primary outcome measure. EF was assessed with the BADDS,
IV [19] that measures the severity of symptoms. Self-reports a self-report questionnaire [36] that yields scores on cate-
of anxiety and depression were assessed using the BAI [20] gories of EF: organization and prioritization, focused and
and BDI. Attention was assessed using the ANT [21] measur- sustained attention, regulation of alertness and sustained
ing three aspects of attention: alerting (maintaining a vigilant effort, affect modulation, and WM. Anxiety and depressive
state of preparedness), orienting (selecting a stimulus among symptoms were assessed with self-report measures of BAI
multiple inputs), and conflict (prioritizing among competing and BDI-II [36]. The CPT-2 [37] provides assessment of
tasks). Authors also used the Stroop test and measure of sustained attention, inhibition, and response variability
attentional conflict with the ANT [22]; the Trail Making Test [38, 39]. Overall, participants receiving DBT group skills
[23], which assesses set-shifting and inhibition; the Digit training showed greater treatment response rates (59-65%
Span Test, which measures WM; and the vocabulary subtest vs. 19-25%) and higher clinical recovery rates (53-59% vs.
(WAIS-R) [24]. Improvements were found in depression and 6-13%) on ADHD symptoms (ES = 0 84 at follow-up) and
anxiety as well as improvements in ADHD self-reported EF (ES = 0 81 at follow-up).
symptoms (ESs ranging from 0.50 to 0.93) and measures In a similar approach, Cole et al. [40] addressed training
of attentional conflict and set-shifting after the training skills by means of cognitive behavioral therapy (CBT) or
(ES = 0 93 and 0.43, respectively). DBT. They assessed the benefits of the program to reduce
Later on, Mitchell et al. [25] tested the impact of MAP for residual symptoms. Patients with ADHD who were poor
adults with ADHD on symptoms, EF, and emotion dysregu- responders to medication were enrolled in a one-year pro-
lation. Adults were stratified by ADHD medication status gram where they received individual therapy, associated with
and then randomized into a group-based mindfulness group therapy with different modules that included mindful-
treatment or waitlist group. The authors observed large ness (along with emotion regulation, interpersonal effective-
effect sizes in improvement of self-reported and clinician ness and distress tolerance, impulsivity/hyperactivity, and
ratings of ADHD symptoms (ESs ranging from 1.35 to attention). Each subject was assessed at baseline, at 3 and 6
3.14) and EF (ESs ranging from 1.45 to 2.67) as well as months, and at the end of the treatment for ADHD severity
self-reported emotion regulation (ESs ranging from 1.27 to with the ASRS v1.1, for depression with BDI-II, for hopeless-
1.63), for the treatment group relative to the waitlist group. ness with the BHS, for anger experience, expression, and con-
EF self-report scales included the DEFS [26] and the trol with STAXI [41], and for impulsivity with BIS-11. The
BRIEF-A [27], which consists of nine scales: Inhibit, Shift, ADHD patients were compared with ADHD patients on a
Emotional Control, Self-Monitor, Initiate, WM, Plan/Orga- waiting list. Overall, the treatment was associated with signif-
nize, Task Monitor, and Organization of Materials. Emotion icant improvements in almost all dimensions. The most sig-
dysregulation was assessed by the DERS [28] and the DTS nificant changes were observed with large to moderate
[29]. The DERS assesses how often emotionally dysregulated effect sizes for depression (ES = −0 84) followed by ADHD
behavior occurs. Additional EF tasks were also administered: severity (ES = −0 63) and hopelessness (ES = −0 52).
the ANT, the CPT [30] to measure response inhibition, the Morgensterns et al. [42] also used DBT for adults with
Digit Span Test [31] to measure WM, and the Trail Making ADHD in an outpatient psychiatric context. The treatment
Test to assess attentional set-shifting and inhibition. uses elements such as acceptance, mindfulness, functional
Edel et al. [32] recruited adults with ADHD and nonran- behavioral analysis, and psychoeducation to target problems
domly assigned them to mindfulness-based training (MBT, common in ADHD. Self-rating scales were administered at
including elements of DBT) or a skills training group (ST). baseline before the first session (T1), posttreatment (T2),
The WRI [33] and scales covering the inattention and hyper- and 3-month follow-up (T3). Self-rating of current ADHD
activity/impulsivity symptoms (DSM-IV, [34]) were used for symptoms was measured by the Current ADHD Symptom
pre- and postassessment. The WRI is an expert-rated scale Scale-Self-Report Form [43] that contains three parts: (1)
10 Behavioural Neurology

the symptoms for ADHD, (2) impairment in major life areas, and hyperactivity/impulsivity scores) were assessed by a
and (3) symptoms of irritability and aggressiveness. More- clinician with the CAARS-INV [46] as well as with the
over, participants completed self-rating questionnaires for self-report of the CAARS-SR [46]. EF was assessed using
assessing symptoms of psychiatric comorbidity: the BDI the BRIEF. The BDI-II was used to assess the presence of
and the BAI. The main results indicated that approximately depression symptoms. The Dutch version of the STAI [47]
80% of the participants attended at least two-thirds of the ses- was administered. The findings indicate that MBCT resulted
sions. ADHD symptoms (ES = 0 22) and functional impair- in a significant reduction of ADHD symptoms as assessed
ment (ES = 0 15) in everyday life were reduced. The results by the investigator or self-reported (Cohen’s d = 0 78 and
were stable at 3-month follow-up. Variables such as age, 0.64, respectively). Significant improvements in EF were
comorbidity, ADHD medication status, and IQ level did also found (Cohen’s d = 0 93). However, no improvements
not predict outcomes. were observed for depressive and anxiety symptoms.
A study from Bueno et al. [44] addresses the impact of Schoenberg et al. [12] looked at the effects of MBCT on
MAP on affective problems and impaired attention. Adults neurophysiological correlates (event-related potentials (ERPs))
with ADHD and healthy controls underwent MAP sessions of performance monitoring in adults with ADHD. Half of
while similar patients and controls did not undergo the inter- patients were randomly allocated to MBCT, and the other half
vention. The authors evaluated MAP-induced changes in to a waitlist control. Inattention and hyperactivity-impulsivity
mood and attention using several measures: (1) the ASRS ADHD symptoms, psychological distress, and social function-
for symptom assessment, (2) the BDI for attitudes related ing were assessed. Clinical scales (the CAARS-S:SV) were
to depression, (3) the STAI to describe how people feel at administered pre- and post-MCBT (or waiting list (WL)). Par-
a particular moment, and (4) the PANAS-X to assess feel- ticipants also completed a standard visual continuous perfor-
ings or moods. Combinations of these ratings yield to mance task (CPT-X). Examining results for CAARS-S:SV
“higher-order affective levels” (positive affect and negative indicated reduced inattention, hyperactivity/impulsivity,
affect) and “lower-order affective levels” (fear, sadness, guilt, and global ADHD index symptoms pre to post-MBCT (ESs
etc.). Attention was evaluated using the ANT and the CPT-2, ranged between 0.49 and 0.93). As expected, the main effect
before and after intervention. The authors found that MAP of treatment was evident for CPT-X repeated-measures
enhanced sustained attention (ANT) and detectability on ANCOVAs comparing accuracy score data indicated that
the CPT-2 and improved the mood of patients and healthy the number of false alarms (FA) significantly decreased pre
controls with overall medium effect sizes (g > 0 5) to large to post in the MBCT group alongside a significant slowing
effect sizes (g > 0 8). Because of mixed results regarding the in reaction times.
enhancement of attentional performance (not all attentional Gu et al. [48] conducted a clinical trial to assess MBCT
measures were found significant), the authors call for more efficacy in the treatment of ADHD in college students.
studies that address the efficacy of mindfulness meditation Undergraduates with ADHD between ages 19 and 24 were
for ADHD in terms of its impact on EF. randomized either to receive MBCT or to be put on a waitlist.
In a recent study, Bachmann et al. [13] evaluated the ADHD symptoms, neuropsychological performance, and
impact of MAP on neurocognitive performance in adults related outcomes were assessed pre- (T1) and posttreat-
with ADHD. The authors performed a RCT to investigate ment (T2), as well as at the 3-month follow-up (T3). Clinical
WM with an n-back task during fMRI before and after an assessment was conducted with the CAARS-S:SV to assess
8-week mindfulness intervention. ADHD symptoms were the extent of ADHD symptoms. Anxiety and depressive
assessed using the self- and observer-rated Conners Adult symptoms were measured with the BAI and the BDI-II. In
ADHD Rating Scales (CAARS). The researchers found a addition, academic performance was collected (participants’
significant decrease in ADHD symptoms and significant GPA) using an official transcript. The authors tested the par-
improvement in task performance in both the MAP and ticipants’ neuropsychological performance (MAAS) and
the psychoeducation comparison group post- versus prein- attentional networks with the ANT [49]. At follow-up, results
tervention but did not find evidence for a significant main revealed that participants receiving MBCT showed greater
effect of treatment or a significant interaction effect on any treatment response rates (57%-71% vs. 23%-31%) and symp-
ADHD symptoms (self- and observer-rated) nor on task per- tom reduction (ES = 1 26). Participants also experienced less
formance (WM). Results also revealed significant increased anxiety and depression (ES = 0 75 and 0.53, respectively)
brain activation after MAP in the bilateral inferior parietal than those on the waitlist. Moreover, MBCT participants
lobule, right posterior insula, and right precuneus. A decrease showed greater improvement on most neuropsychological
in self-rated “inattention/memory problems” after MAP performance and attentional scores (ES for MAAS = 1 30,
compared to baseline was associated with stronger activation ES for ANT subscales ranging from 0.19 to 1.19).
in parts of the left putamen, globus pallidus, and thalamus. In a recent study, [50] investigated the efficacy of MBCT
Hepark et al. [45] also looked at the efficacy of an +treatment as usual (TAU) versus TAU only in reducing core
adaptation of mindfulness-based cognitive therapy (MBCT) symptoms in adults with ADHD. Participants were ran-
on core ADHD symptoms and EF. Adults with ADHD were domly assigned to MBCT+TAU, an 8-weekly group therapy
randomly allocated to MBCT or waitlist. Outcome measures including meditation exercises, psychoeducation, and group
included investigator-rated ADHD symptoms, self-reported discussions, or TAU only, including pharmacotherapy and/
ADHD symptoms, EF, depressive and anxiety symptoms, or psychoeducation. Outcomes were ADHD symptoms
and patient functioning. Symptoms (total ADHD, inattention, rated by blinded clinicians (CAARS-INV) and self-reported
Behavioural Neurology 11

(CAARS-S), EF, mindfulness skills, self-compassion, positive disturbance, quality of life, mindfulness, and grade point
mental health, and general functioning (see details in average at school.
Table 2(c)). Outcomes were assessed at baseline, posttreat-
ment (T1), and 3- and 6-month follow-up (T2 and T3, respec- (1) ADHD Symptoms. Prior to and following treatment,
tively). In MBCT+TAU patients, a significant reduction of researchers used different self-rating scales to assess symp-
clinician-rated ADHD symptoms (CAARS-INV) was found toms of ADHD. Among the most frequently used self-
at posttreatment (T1) (ES = 0 41) and was maintained at the report scales were the following:
6-month follow-up. MBCT+TAU patients compared with
TAU patients also reported significant improvements in (1) The Conners’ Adult ADHD Rating Scale
self-reported ADHD symptoms (ES = 0 37, 0.71, and 0.79 at (CAARS-SR/CAARS-S:SV) from Conners et al. [46,
T1, T2, and T3, respectively), mindfulness skills, self-com- 51] ([12, 13, 45, 48, 50], n = 5)
passion, and positive mental health up to the 6-month
(2) The Adult ADHD Self-Report Scale (ASRS v1.1)
follow-up. Patients in MBCT+TAU reported improvement
from Kessler et al. (2005) ([40, 44], n = 2)
in executive functioning (EF) but only at the 6-month
follow-up. A significant group x time interaction showed that (3) The Attention-Deficit Hyperactivity Disorder Check-
EF further improved over time in MBCT+TAU compared list (ADHD-CL) from DSM-IV (1994) ([14, 32],
with TAU resulting in an effect size of d = 0 49 at the n = 2)
6-month follow-up. The authors concluded that MBCT
might be a valuable treatment option alongside TAU for The complete list of scales is available in Tables 1(c)
adult ADHD aimed at alleviating symptoms. and 2(c).

3.3. Synthesis of Studies. For convenience, we divided the (2) Executive/Cognitive Functioning. As additional measures
above studies (n = 13) according to the two main research of outcomes, the most frequently used tests for executive/-
designs: within-group (Tables 1(a)–1(c)) and between- cognitive functioning were as follows:
group (Tables 2(a)–2(c)). Three studies used a within-group
design (with two or more time points) while the remaining (a) Objective tasks
studies used different between-group designs (N-RCT, RCT).
All studies, except one [44], did not include a healthy control (1) Various versions of Attention Network Test
group. Follow-up evaluations varied from none to three or (ANT) [21, 22, 49] ([11, 25, 44, 48], n = 4)
six months.
(2) Various versions of the Conners’ Continuous
3.3.1. Participants. The sum of participants with ADHD was Performance Test (CPT/CPT-2/CPT-X) from
753 with a mean age of 35.1 years (18-65 y/o). About half of Conners [30, 37] ([12, 25, 35, 44], n = 4)
participants were males (47.7%). The combined and inatten- (3) The TMT (n = 2)
tive subtypes of ADHD were the most predominant. Most
participants were on medication, with psychostimulants like (4) The Stroop test indicating mental speed and
methylphenidate (MPH) being the most frequently reported. inhibitory EFs ([11, 14], n = 2)
Comorbidity was present in all studies, with major depressive (5) The digit span, vocabulary, memory scale, and
disorders and mood disorders being frequently reported. A digit symbol subtests from the Wechsler Adult
summary of participants’ characteristics is presented in Intelligence Scale-Revised (WAIS-R) [24, 31]
Tables 1(a) and 2(a). ([11, 14, 25], n = 2)
3.3.2. Intervention. Our definition of MBI intervention (6) n-back ([13], n = 1)
included mindfulness and/or meditation as a principal or a
partial component of the intervention. That included various (b) Subjective questionnaire
adaptations of (1) dialectical behavior therapy (DBT), (2)
mindful awareness program (MAP), and (3) mindfulness- (1) The Behavior Rating Inventory of Executive
based/cognitive training (MBT/MBCT). The duration of Functioning-Adult Version (BRIEF-A) from
treatment varied considerably across studies (from six to Roth et al. [27] ([25, 45, 50], n = 3)
96 hours, mode value of 20 hours). The presence or absence
of homework also accounted for variability between studies. The complete list of outcome measures is available in
Therapists included clinical psychologists, psychology gradu- Tables 1(c) and 2(c).
ate students, mindfulness instructors, practitioners, group
leaders, ADHD researchers, nurses, and psychiatrists. A (3) Emotion Disturbance. Self-reports of anxiety, depression,
summary of intervention characteristics is presented in and other emotional disturbances were often assessed using
Tables 1(b) and 2(b). the following:

3.3.3. Outcomes. Besides the measures of ADHD symptoms (1) The Beck Depression Inventory (BDI/BDI-II/B-
(inattention and hyperactivity), outcome measures can be DI-II-NL) from Beck et al. [15] ([11, 14, 35, 40, 42,
categorized into executive/cognitive functioning, emotional 44, 45, 48], n = 8)
12 Behavioural Neurology

Random sequence generation (selection bias)

Allocation concealment (selection bias)

Blinding of participants and personnel (performance bias)

Blinding of outcome assessment (detection bias)

Incomplete outcome data (attrition bias)

Selective outcome reporting (reporting bias)

Other biases (research allegiance, funding, confounds)

0 0.25 0.5 0.75 1

Low risk
Unclear risk
High risk

Figure 2: Methodological quality graph: a review of the authors’ judgements about each methodological quality item presented as percentages
across all included studies.

(2) The Beck Anxiety Inventory (BAI) from Beck et al. the Cochrane Collaboration. As alluded earlier, this analysis
[20] ([11, 35, 42, 48], n = 4) of bias gives us a useful complement of information besides
the sole calculation of effect sizes. Slight differences in inter-
(3) The State-Trait Anxiety Inventory (STAI) from Van pretation of bias were discussed and solved between the
der Ploeg [47] ([44, 45], n = 2) judges (HP, BK, and AM). Assessment of risk of bias was
(4) The Outcome Questionnaire (OQ 45.2) from Lambert rated high (red), uncertain (yellow), or low (green) for each
et al. (1996) ([12, 45], n = 2) individual study and category (see Figure 2), then compiled
into percentages of studies that fall into high, uncertain, or
The complete list of outcome measures is available in low risk on each category of bias (see Figure 3).
Tables 1(c) and 2(c).
3.5. Selection Bias. Only three studies were rated low risk on
(4) Mindfulness. Mindfulness was assessed using the follow- that criterion [12, 13, 50]. Four studies (31%) were rated
ing self-report questionnaires: unclear risk because of insufficient data on sequence genera-
tion. The remaining six (46%) studies were rated high risk.
(1) The Mindful Attention Awareness Scale (MAAS) Regarding allocation concealment, the same three studies
from Brown and Ryan (2003) ([32, 42, 48], n = 3) were judged as low risk because the occultation of the alloca-
(2) The Kentucky Inventory of Mindfulness (KIMS) tion sequence could not be predicted. Only one study was
from Baer et al. (2004) ([12, 40, 45], n = 3) judged as unclear risk because participants would not neces-
sarily identify the group to which they belong (i.e., treatment
At two occasions, authors used the Five Facet Mindful- or control). The remaining nine (69%) studies did not fulfill
ness Questionnaire (FFMQ) from Baer et al. (2006) ([35, either conditions and were considered high risk.
50], n = 2).
3.6. Performance and Detection Bias. Not surprisingly, the
(5) Quality of Life and Others. Self-report questionnaires or vast majority of studies (92%) in this review were rated high
semistructured interviews (QFS) were used in most studies risk for performance bias (blinding of participants and of
to evaluate the level of functioning and quality of life: personnel). Only [50] satisfied this criterion according to
our interrater judgment. Detection bias reflects the blinding
(1) The ADHD Quality of Life Questionnaire (AAQoL) of assessors to the treatment condition. Overall, eleven stud-
from Brod et al. (2006) ([35, 42, 44], n = 3) ies (85%) were judged as low risk of detection bias (blinding
of assessors), one study as high risk, and one as unclear risk as
Other less frequently used measures, such as OQ 45.2 it included self-report measures but nonblinded assessors.
(n = 2), are listed in Tables 1(c) and 2(c).
3.7. Attrition Bias. According to the application of the 20%
3.4. Risk of Bias within and across Studies. After characteriz- cut-off criteria, six studies (46%) in this review were consid-
ing each study according to PICO, we then evaluated each ered low risk of attrition bias. Three studies (23%) were rated
study (n = 13) on the seven categories of bias established by high risk (attrition rate > 20%, high differential attrition, and
Behavioural Neurology 13

Other biases (research allegiance,

Blinding of outcome assessment
Random sequence generation

personnel (performance bias)

Blinding of participants and

Selective outcome reporting

Incomplete outcome data
Allocation concealment

funding, confounds)
(reporting bias)
(detection bias)
(selection bias)

(selection bias)

(attrition bias)

Mean /2
Bachmann et al. (2018) 2 2 0 2 1 2 2 1.57
Bueno et al. (2015) 0 0 0 2 2 2 2 1.14
Cole et al. (2016) 0 0 0 2 2 1 2 1.00
Edel et al. (2017) 0 0 0 0 1 2 2 0.71
Fleming et al. (2015) 1 1 0 2 2 2 1 1.29
Gu et al. (2018) 1 0 0 2 2 2 2 1.29
Hepark et al. (2019) 1 0 0 2 1 2 2 1.14
Hesslinger et al. (2002) 0 0 0 2 1 2 1 0.86
Jansenn et al. (2019) 2 2 2 2 2 2 2 2.00
Mitchell et al. (2017) 1 0 0 1 2 2 1 1.00
Morgensterns et al. (2016) 0 0 0 2 0 2 1 0.71
Schoenberg et al. (2014) 2 2 0 2 0 1 2 1.29
Zylowska et al. (2008) 0 0 0 2 0 2 1 0.71

High risk
Unclear risk
Low risk

Figure 3: Methodological quality summary: a review of the authors’ judgments about each methodological quality item for each
included study.

no imputation of missing data). The remaining three studies our interrater judgment of quality, Bachmann et al. [13]
had no sufficient description of attrition (or impossibility to and Janssen et al. [50] were the most robust and valid studies
compute the attrition rate) and thus were judged as having with 1.57/2 and 2/2 overall quality scores, respectively.
unclear risk. In sum, the majority of studies (but [12, 13, 50]) were
considered high risk on selection biases (random sequence
3.8. Reporting Bias. Most of the studies (11/13) were rated generation, allocation concealment), and all but [50] had a
low risk since the outcome data were reported on all used performance bias (blinding of participants and personnel).
scales and subscales. Two other studies were rated unclear See Figure 3 and supplementary materials.
risk as subscales were not fully reported.

3.9. Other Biases and Limitations. Overall, eight studies 4. Discussion

(62%) were rated low risk, while the five remaining studies
were judged as unclear risk. More precisely, we could iden- 4.1. Summary of Evidence. In this systematic review, we
tify an author’s role in the study development and/or imple- assessed cognitive and behavioral effects observed in 13
mentation (e.g., delivery of the intervention) in only three studies using MBIs to alleviate ADHD symptoms and to
(23%) out of the 13 studies. In four of the studies (31%), improve executive function and emotion dysregulation
the authors identified a funding source. Other studies were among adults with ADHD. All the studies (100%) showed
not funded or did not report the funding source. Other lim- improvement of ADHD symptoms following an MBI.
itations that were reported by the authors of each paper were Researchers have also found a significant improvement on
included as an additional source of information (in narrative cognitive task performance in post- versus preintervention
form) but were not rated (see supplementary materials or with treatment as usual (TAU). For most patients, reduc-
(available here)). tion of ADHD symptoms was maintained at posttreatment
After conversion of the high, uncertain, and low risk (3- to 6-month follow-up). In studies addressing other out-
scores into numeric variables (0, 1, and 2), we found stud- comes, patients reported significant improvements in mind-
ies’ quality mean scores ranging between 0.71 and 2 (with fulness skills, self-compassion, and positive mental health up
2 being the highest quality) for each study. According to to the 6-month follow-up.
14 Behavioural Neurology

However, we also found the quality of studies to be vari- Appendix

able with a tendency for more recent studies to have less
biases. Notably, [50] was given a perfect score according to
Full Names of Most Frequently Used Outcome
our application of the Cochrane Collaboration standards. A Measures (See Also Tables 1(c)–2(c))
vast majority of studies were judged as high risk on the per-
formance bias (blinding of participants and personnel), and ADHD symptoms measures:
several had issues with the selection biases (random sequence
(1) ADHD-CL = Attention-Deficit Hyperactivity Disor-
generation, allocation concealment). As mentioned earlier, it
der Checklist
is not habitual, nor always advantageous, to blind personnel
or participants in this type of intervention, so the elevated (2) ASRS (v1.1) = Adult ADHD Self-Report Scale
risk for performance and/or detection bias may be inevitable.
Attrition bias was found to have high or unclear risk in more (3) BAARS-IV = Barkley Adult ADHD Rating Scale–IV
than a half of the studies. The reason for dropout of partici- (4) BADDS = Brown ADD Rating Scales
pants was not always clearly specified in those studies, so it
is difficult to decide if it might be related to adverse effects (5) BIS-11 = 11th version of the Barratt Impulsiveness
or to some discomfort with treatment or instead to some Scale
incidental reasons. (6) CAARS-INV = investigator rating version of the
Despite the above limitations, most studies (except Conners’ Adult ADHD Rating Scale; CAARS-S: SV =
one) scored well on the detection bias, meaning that the Conners’ Adult ADHD Self-rating Scale; CAARS-
trainers were not involved in the assessment of the partici- SV = Conners’ Adult ADHD Rating Scales–Screening
pants and therefore could not interfere with the outcomes. Version; CAARS- SR = self-report version of the
Moreover, most studies (except for the two studies being Conners’ Adult ADHD Rating Scale
unclear) were found free of suggestion of selective reporting
(reporting bias). (7) CADHDSC_SRF = Current ADHD Symptom Scale-
In sum, most studies show that mindfulness training or Self-Report Form
structured programs with mindfulness components appear (8) DSM-IV (SR/OR) = Diagnostic and Statistical Man-
useful for patients who respond partially or not at all to drug ual of Mental Disorders, 4th Edition (Self-rating/
therapy. Indeed, group skills training may be efficacious, Other-rating)
acceptable, and feasible for treating ADHD among college
students and adult patients. Mindfulness meditation training (9) DSM-IV = Diagnostic and Statistical Manual of Men-
seems to improve ADHD behavioral symptoms (inattention, tal Disorders, 4th Edition; SCL-16 = Symptom Check
hyperactivity, and impulsivity) and some facets of EF and List
emotion dysregulation. Although these are promising find-
ings to support treatment efficacy of MBIs for ADHD, vari- Cognitive/EF measures:
ous biases such as the absence of randomization and lack of
a control group may affect the importance of outcomes. (1) ANT = Attention Network Task
Other factors such as those documented in the present study (2) BRIEF-A = Behavior Rating Inventory of Executive
(see Tables 1 and 2) may also impact on the outcomes. For Functioning-Adult Version
example, the amount of home exercise, type of monitoring
of participants’ progress, or absence from sessions may also (3) BRIEF-ASR = Behavior Rating Inventory of Execu-
affect the outcomes. tive Function-Adult Self-Report version
(4) CPT-2 = The Conners’ Continuous Performance
5. Conclusions, Future Research, Test-2nd edition
and Limitations
(5) CPT-X = visual Continuous Performance Task
The aim of this systematic review was to look for symp- (6) d2-Test = selective attention
toms and additional indicators of improvement of ADHD
mediated through mindfulness interventions. Each study (7) DEFS = Deficits in Executive Functioning Scale
measured many outcomes, namely, executive functions,
emotional disturbance, quality of life, and academic perfor- (8) DSS = Digit Symbol Subtest
mance. Some outcomes were considered to be important (9) DST = Digit Span Test (WAIS-R)
(e.g., symptoms), while others were surrogate outcomes
(e.g., attention test). Despite its comprehensiveness, this (10) KLT = Konzentrations-Leistungs-Test
review was not without limitations, mostly because of hetero- (11) TMT = Trail Making Test
geneity of available studies. Although all studies included a
mindfulness-based intervention for ADHD, there was a sub- (12) VOC = Vocabulary subtest (WAIS-R)
stantial variability among them, e.g., difference in sample size (13) WAIS-R = Wechsler Adult Intelligence Scale-Revised
and duration of intervention. Future studies and potential
meta-analysis should consider these factors. (14) WMS-R = Wechsler Memory Scale-R
Behavioural Neurology 15

Emotional disturbance measures: [5] H. McCarthy, N. Skokauskas, A. Mulligan et al., “Attention

network hypoconnectivity with default and affective network
(1) BAI = Beck Anxiety Inventory hyperconnectivity in adults diagnosed with attention-defi-
cit/hyperactivity disorder in childhood,” JAMA Psychiatry,
(2) BDI = Beck Depression Inventory; BDI-II = Beck vol. 70, no. 12, pp. 1329–1337, 2013.
Depression Inventory (2nd edition) [6] L. J. Norman, C. O. Carlisi, A. Christakou et al., “Shared and
(3) BHS = Beck Hopelessness Scale disorder-specific task-positive and default mode network dys-
functions during sustained attention in paediatric attention--
(4) DERS = Difficulties in Emotion Regulation Scale deficit/hyperactivity disorder and obsessive/compulsive
disorder,” NeuroImage: Clinical, vol. 15, pp. 181–193, 2017.
(5) PANAS-X = Affect Schedule-Expanded form [7] J. Sidlauskaite, E. Sonuga-Barke, H. Roeyers, and R. Wiersema,
(6) PSS = Perceived Stress Scale “State-to-state switching in ADHD: default mode network
abnormalities,” International Journal of Psychophysiology,
(7) STAI = State-Trait Anxiety Inventory vol. 108, p. 29, 2016.
[8] L. Shamseer, D. Moher, M. Clarke et al., “Preferred reporting
(8) STAXI = State-Trait Anger Expression Inventory
items for systematic review and meta-analysis protocols
(9) WRI = Wender–Reimherr Interview (PRISMA-P) 2015: elaboration and explanation,” BMJ,
vol. 349, article g7647, 2015.
Conflicts of Interest [9] J. P. Higgins and S. Green, “Guide to the contents of a
Cochrane protocol and review,” in Cochrane Handbook for
The authors declare that there is no conflict of interest Systematic Reviews of Interventions: Cochrane Book Series,
regarding the publication of this paper. pp. 51–79, Wiley-Blackwell, 2008.
[10] J. P. T. Higgins, D. G. Altman, P. C. Gotzsche et al., “The
Authors’ Contributions Cochrane Collaboration’s tool for assessing risk of bias in ran-
domised trials,” BMJ, vol. 343, article d5928, 2011.
H.P. conducted the literature search, data analysis, and [11] L. Zylowska, D. L. Ackerman, M. H. Yang et al., “Mindfulness
data interpretation and wrote the manuscript. A.M. inter- meditation training in adults and adolescents with ADHD: A
rated the selection of articles and of analysis of bias and cow- feasibility study,” Journal of Attention Disorders, vol. 11,
rote the manuscript. N.T. conducted the literature search no. 6, pp. 737–746, 2008.
(update), created the figures, and collected the data. B.K. con- [12] P. L. A. Schoenberg, S. Hepark, C. C. Kan, H. P. Barendregt,
ducted the analysis of bias (interrating) and data interpreta- J. K. Buitelaar, and A. E. M. Speckens, “Effects of mindfulness-
tion reviewed the manuscript. J.N. conducted the literature based cognitive therapy on neurophysiological correlates of
performance monitoring in adult attention-deficit/hyperac-
search (phase 1) and is responsible for the endnote creation.
tivity disorder,” Neurophysiologie Clinique, vol. 125, no. 7,
pp. 1407–1416, 2014.
Acknowledgments [13] K. Bachmann, A. P. Lam, P. Sörös et al., “Effects of mindful-
ness and psychoeducation on working memory in adult
The project was funded by the Fonds de recherche du
ADHD: a randomised, controlled fMRI study,” Behaviour
Québec-Société et culture (FRQSC). Research and Therapy, vol. 106, pp. 47–56, 2018.
[14] B. Hesslinger, L. Tebartz van Elst, E. Nyberg et al., “Psycho-
Supplementary Materials therapy of attention deficit hyperactivity disorder in adults:
a pilot study using a structured skills training program,”
Full details of analysis of bias for each of individual studies European Archives of Psychiatry and Clinical Neuroscience,
(n = 13). (Supplementary Materials) vol. 252, no. 4, pp. 177–184, 2002.
[15] A. T. Beck, R. A. Steer, and G. K. Brown, BDI-II Manual,
References Psychological Corporation, San Antonio, TX, USA, 1996.
[1] M. Cairncross and C. J. Miller, “The effectiveness of [16] D. Wechsler, WMS-R: Wechsler Memory Scale-Revised, Psy-
mindfulness-based therapies for ADHD: a meta-analytic chological Corporation, 1987.
review,” Journal of Attention Disorders, pp. 1–17, 2016. [17] J. Kabat-Zinn, Full Catastrophe Living: Using the Wisdom of
[2] S. Cortese, C. Kelly, C. Chabernaud et al., “Toward systems Your Body and Mind to Face Stress, Pain and Illness, Delacorte,
neuroscience of ADHD: a meta-analysis of 55 fMRI studies,” New York, NY, USA, 1990.
American Journal of Psychiatry, vol. 169, no. 10, pp. 1038– [18] Z. V. Segal, J. M. G. Williams, and J. D. Teasdale, Mindfulness-
1055, 2012. Based Cognitive Therapy for Depression, Guilford Press, New
[3] W. Francx, M. Oldehinkel, J. Oosterlaan et al., “The executive York, NY, USA, 2013.
control network and symptomatic improvement in attention-- [19] G. J. DuPaul, The ADHD Rating Scale: Normative Data, Reli-
deficit/hyperactivity disorder,” Cortex, vol. 73, pp. 62–72, ability, and Validity, Unpublished manuscript, University of
2015. Massachusetts Medical Center, Worcester, UK, 1990.
[4] E. B. Liddle, C. Hollis, M. J. Batty et al., “Task-related default [20] A. T. Beck, N. Epstein, G. Brown, and R. A. Steer, “An inven-
mode network modulation and inhibitory control in ADHD: tory for measuring clinical anxiety: psychometric properties,”
effects of motivation and methylphenidate,” Journal of Child Journal of Consulting and Clinical Psychology, vol. 56, no. 6,
Psychology and Psychiatry, vol. 52, no. 7, pp. 761–771, 2011. pp. 893–897, 1988.
16 Behavioural Neurology

[21] J. Fan and M. Posner, “Human attentional networks,” Psychia- disorder: do we need neuropsychologically impaired sub-
trische Praxis, vol. 31, pp. 210–214, 2004. types?,” Biological Psychiatry, vol. 57, no. 11, pp. 1224–1230,
[22] C. J. Golden, Stroop Color and Word Test: A Manual for 2005.
Clinical and Experimental Use, Stoelting, Chicago, IL, [40] P. Cole, S. Weibel, R. Nicastro et al., “CBT/DBT skills training
USA, 1978. for adults with attention deficit hyperactivity disorder
[23] R. M. Reitan, Manual for Administration of Neuropsychologi- (ADHD),” Psychiatria Danubina, vol. 28, pp. 103–107, 2016.
cal Test Batteries for Adults and Children, Neuropsychology [41] C. D. Spielberger, “State-Trait Anxiety Inventory,” in Corsini
Laboratory, Tucson, AZ, USA, 1979. Encyclopedia of Psychology, vol. 1, p. 1698, John Wiley & Sons,
[24] D. Wechsler, Wechsler Adult Intelligence Scale–Revised, The Inc., 2010.
Psychological Corporation, New York, NY, USA, 1981. [42] E. Morgensterns, J. Alfredsson, and T. Hirvikoski, “Structured
[25] J. T. Mitchell, E. M. McIntyre, J. S. English, M. F. Dennis, skills training for adults with ADHD in an outpatient psychi-
J. C. Beckham, and S. H. Kollins, “A pilot trial of mindful- atric context: an open feasibility trial,” ADHD Attention Deficit
ness meditation training for ADHD in adulthood: impact and Hyperactivity Disorders, vol. 8, no. 2, pp. 101–111, 2016.
on core symptoms, executive functioning, and emotion dys- [43] R. A. Barkley and K. R. Murphy, Attention-Deficit Hyperactiv-
regulation,” Journal of Attention Disorders, vol. 21, no. 13, ity Disorder: A Clinical Workbook, Guilford Press, 1998.
pp. 1105–1120, 2017.
[44] V. F. Bueno, E. H. Kozasa, M. A. da Silva, T. M. Alves, M. R.
[26] R. A. Barkley, Barkley Adult ADHD Rating Scale-IV (BAAR- Louzã, and S. Pompéia, “Mindfulness meditation improves
S-IV), Guilford Press, 2011. mood, quality of life, and attention in adults with attention
[27] R. M. Roth, P. K. Isquith, and G. A. Gioia, BRIEF-A: Behavior deficit hyperactivity disorder,” BioMed Research International,
Rating Inventory of Executive Functions-Adult Version, Psy- vol. 2015, Article ID 962857, 14 pages, 2015.
chological Assessment Resources, Lutz, FL, USA, 2005.
[45] S. Hepark, L. Janssen, A. de Vries et al., “The efficacy of
[28] K. L. Gratz and L. Roemer, “Multidimensional assessment of adapted MBCT on core symptoms and executive functioning
emotion regulation and dysregulation: development, factor in adults with ADHD: a preliminary randomized controlled
structure, and initial validation of the difficulties in emotion trial,” Journal of Attention Disorders, vol. 23, no. 4, pp. 351–
regulation scale,” Journal of Psychopathology and Behavioral 362, 2019.
Assessment, vol. 26, no. 1, pp. 41–54, 2004.
[46] C. K. Conners, D. Erhardt, and E. Sparrow, CAARS Adult
[29] J. S. Simons and R. M. Gaher, “The distress tolerance scale: ADHD Rating Scales-Technical Manual, Multi-Health Sys-
development and validation of a self-report measure,” Motiva- tems, Toronto, Ontario, Canada, 1999.
tion and Emotion, vol. 29, no. 2, pp. 83–102, 2005.
[47] H. M. Van der Ploeg, Handleiding bij de Zelf-Beoordelings
[30] C. K. Conners, The Conners’ Continuous Performance Test, Vragenlijst [Manual for the State-Trait Anxiety Inventory],
Multi-Health Systems, Toronto, ON, Canada, 1995. Harcout Assessment B.V, Amsterdam, Netherlands, 2000.
[31] D. Wechsler, Wechsler Adult Intelligence Scale–Fourth Edition
[48] Y. Gu, G. Xu, and Y. Zhu, “A randomized controlled trial of
(WAIS-IV), The Psychological Corporation, San Antonio, TX,
mindfulness-based cognitive therapy for college students with
USA, 2008.
ADHD,” Journal of Attention Disorders, vol. 22, no. 4, pp. 388–
[32] M.-A. Edel, T. Hölter, K. Wassink, and G. Juckel, “A compar- 399, 2018.
ison of mindfulness-based group training and skills group
[49] J. Fan, B. D. McCandliss, T. Sommer, A. Raz, and M. I. Posner,
training in adults with ADHD: an open study,” Journal of
“Testing the efficiency and independence of attentional net-
Attention Disorders, vol. 21, no. 6, pp. 533–539, 2017.
works,” Journal of Cognitive Neuroscience, vol. 14, no. 3,
[33] M. Rösler, W. Retz, P. Retz-Junginger et al., “ADHS-diagnose pp. 340–347, 2002.
bei erwachsenen,” Der Nervenarzt, vol. 79, no. 3, pp. 320–327,
2008. [50] L. Janssen, C. C. Kan, P. J. Carpentier et al., “Mindfulness-
based cognitive therapy v. treatment as usual in adults with
[34] American Psychiatric Association, Diagnostic and Statistical
ADHD: a multicentre, single-blind, randomised controlled
Manual of Mental Disorders, American Psychiatric Associa-
trial,” Psychological Medicine, vol. 49, no. 1, pp. 55–65, 2019.
tion, Washington, DC, USA, 4th edition, 1994.
[51] C. K. Conners, D. Erhardt, and E. Sparrow, Conners’ Adult
[35] A. P. Fleming, R. J. McMahon, L. R. Moran, A. P. Peterson, and
ADHD Self-Rating Scale (CAARS-S:SV), Psychological Assess-
A. Dreessen, “Pilot randomized controlled trial of dialectical
ments Australia, Jannali, NSW, Australia, 2008.
behavior therapy group skills training for ADHD among col-
lege students,” Journal of Attention Disorders, vol. 19, no. 3,
pp. 260–271, 2015.
[36] T. E. Brown, Brown Attention Deficit Disorder Scales, The
Psychological Corporation, San Antonio, TX, USA, 1996.
[37] C. K. Conners, Conners’ Continuous Performance Test II
(CPT-II): Computer Program for Windows Technical Guide
and Software Manual, Multi-Health Systems, North Tona-
wanda, NY, USA, 2000.
[38] K. R. Murphy, R. A. Barkley, and T. Bush, “Executive function-
ing and olfactory identification in young adults with attention
deficit-hyperactivity disorder,” Neuropsychology, vol. 15, no. 2,
pp. 211–220, 2001.
[39] J. T. Nigg, E. G. Willcutt, A. E. Doyle, and E. J. S. Sonuga-
Barke, “Causal heterogeneity in attention-deficit/hyperactivity