Transcript Slide 1

COGNITIVE TRAINING AND ADHD:
Can training each day keep impairments
at bay?
Rosemary Tannock, PhD
Erskine Visiting Fellow, University of Canterbury(Feb-April, 2015)
Senior Scientist, Neurosciences & Mental Health Research Program
The Hospital for Sick Children; &
Professor Emerita (Special Education, Psychiatry), University of Toronto
[email protected]
CFPC CoI Templates: Slide 1
Disclosure: potential conflicts of interest
• Faculty: Rosemary Tannock, PhD
• Relationships with commercial interests:
• Research Grants: Federal grants (IES-USA; CIHR)
• Research Support: Cogmed; Purdue Pharma
• Honoraria: Shire; Janssen-Ortho
• Consulting Fees: Biomed Central (publisher) Editors Advisory Group
• Other:
• Royalties: Springer, as Co-Editor of book (Behavioral Neuroscience
of ADHD and its Treatment, 2011)
• Member DSM-5 Workgroup on ADHD, & liaison member to
Neurodevelopmental Disorders workgroup (for Learning Disorders)
• Member International Steering Committee for WHO International
Classification of Functioning (ICF)-Core Set for ADHD
• Affiliate member WHO ICD-11 Specific Learning Disorders
subcommittee
Scope of this talk
• Cognitive training has wide range of meaning
• Application of self-monitoring & self-reinforcement
techniques to enhance functioning
(e.g., Abikoff et al., 1988, J Abnorm Child Psychol, 16:411-432)
• Application of EEG feedback for self-monitoring &
sustaining attention – Neurofeedback
(e.g., Hurt et al., 2014, Child Adolesc Psychiatric Clin N Am, 23:465486)
• Intensive, adaptive practice of specific cognitive process
(e.g., Klingberg, 2010, Trends Cog Sci 14: 317-324)
• This talk focuses on computerized cognitive training
designed to target WM: specifically on a software
program called “COGMED”
Neuroplasticity
Scientific dogma until 1970’s…
“In the adult [brain] centres, the nerve paths are
something fixed, ended, immutable. Everything may
die, nothing may be regenerated.”
Santiago Ramon v Cajal (1913). Spanish physician,
neuroanatomist & Nobel Laureate
By contrast, scientists now recognize that the human
brain (even in adulthood) shows remarkable
neuroplasticity
IS WORKING MEMORY CAPACITY A
MALLEABLE FACTOR?
CAN IT BE IMPROVED BY
COMPUTERIZED TRAINING?
By improving a person’s WM, beneficial
effects should also be expected in various
related abilities utilizing WM, including realworld behavior (beyond the Laboratory)
6
COMPUTERIZED
COGNITIVE TRAINING
Cogmed™ [abbrev.
CWMT]
Jungle Memory™
Cognifit™
BrainTrain: Captain’s Log™
Lumosity™
7
Features of Cogmed WM Training
• Intensive & adaptive training
• Adaptive: automatically, continuously adjusted
in difficulty relative to individual’s WM capacity
• Extensive repetition, practice, feedback –
designed to enhance the development &
efficiency of underlying neural substrates (for
WM)
• Underlying assumption: improvements in WM
will generalize or transfer to other tasks or
activities that rely on the same neural networks
or require WM (Klingberg, 2010)
Computerized Working Memory Training
(Cogmed JM/RM/QM)
 45 min training/day
 5 days/week,5 weeks
 Adaptive algorithm
◦ individually-based
 Reinforcement
◦ Immediate performance-based
feedback;
◦ internal reinforcement activities
◦ external reinforcement for completing
pre-specified # sessions
 Weekly monitoring calls from
licensed provider, using uploaded
tracking data
Cogmed/Pearson
http://www.cogmed.com/rm
9
What does the training entail?
• EQUIPMENT
• Software (license per person)$
• Computer (per person) linked to internet$
• [Headphones for group administration] $
• COACHING/Supervision$
• Weekly telephone call from a trained & licensed ‘coach’
to give feedback on performance, give advice about
training activities, answer questions
• For youth - Daily supervision of training –
parent, school-aide, volunteer, (often by members of
research study)
The beginning…a startling finding
suggesting that WM might be mutable –
like a muscle!
But a very small sample ( n ~ 7 per
group)
Double-blind but not randomized
• In 2002, Torkel Klingberg, a Swedish researcher
challenged the prevailing notion that WM capacity
is fixed - he reported that 5 weeks of playing
specific memory-based computer games (every
day for about 30-45 mins), not only boosted WM,
of children with ADHD but also intellectual ability!
Promising findings
from first randomized controlled trial!
Klingberg et al. (2005) JAACAP
(Klingberg et al, 2005)
Adaptive training:
n=20 ‘ADHD’
Non-adaptive training:
n=24’ ADHD’
Repeat this letters in the same
order they are given....
”2 8 4 7 2 9”
ES:
.93/.92
ES:
.73/?
Moreover, training effects appeared to
transfer to other cognitive functions!
Parent rating of ADHD symptoms
(Conners) no evidence of change in teacher ratings
Controlled for whether training done at home or school
Strong claims for WM training
Including those by Cogmed
The spread of Cogmed
~58 published studies; 90+ ongoing studies
•
Healthy young adults
Older/elderly adults
• Adults who have had a stroke
• Individuals with brain injury
•
•
Children with: WM deficits; ADHD; cochlear
implants; cancer treated with radiation; Down
syndrome, low IQ, born prematurely…
Cogmed is now operating in > 1000 schools
world-wide (& is available in NZ)
So, does WM training
(cogmed) enhance WM
capacity?
Cogmed WM Training: Reviewing the Reviews
Shinaver*, Entwistle*, Söderqvist*. APPLIED NEUROPSYCHOLOGY: CHILD,
3: 163–172, 2014 (*Pearson Assessment)
• “Cogmed has indeed demonstrated reliable immediate
improvements in WM capacity in samples of typically
developing children…& adults….children with ADHD…with
poor WM…cochlear implants…very low birth
weight…adolescents at extremely low birth weight….pediatric
cancer survivors…& brain injured adults”
• “In a little more than a decade, there is evidence that
suggests that Cogmed has a significant impact upon
visual-spatial and verbal WM, and these effects
generalize to improved sustained attention up to 6
months”
• “In conclusion, we find that there is a consensus in showing
Cogmed working memory training is sold as a tool for
improving cognitive abilities, such as attention and reasoning.
At present, this program is marketed to schools as a means
of improving underperforming students’ scholastic
performance, and is also available at clinical practices as a
treatment for ADHD. We review research conducted with
Cogmed software and highlight several concerns regarding
methodology and replicability of findings. We conclude that
the claims made by Cogmed are largely unsubstantiated,
and recommend that future research place greater emphasis
on developing theoretically motivated accounts of working
memory training.
Conclusions from recent Meta-Analysis
“Collectively, meta-analytic results indicate that claims
regarding the academic, behavioral, and cognitive benefits
associated with extant cognitive training programs are
unsupported in ADHD.”
“The methodological limitations of the current evidence base,
however, leave open the possibility that cognitive training
techniques designed to improve empirically documented
executive function deficits may benefit children with ADHD”.
Optimal research designs for studies of
WM training
• Randomized control design
• Active control group – to control for
• Test-retest effects & Placebo effects;
• Expectations of trainers/testers/participants
• Motivation effects; engagement in training
• Measurement
• Objective measures (if subjective - blinded informants)
• Latent changes (more than one measure/construct)
• Transfer of training effects (far transfer)
• Duration of changes (sustained effects)
• Mechanisms of change
From: Shipstead, Redick, Engle (2012) Psychol Bull [advance online publication]
24
Published RCTs of WM training in ADHD
Study
Cogmed
N
Contro
lN
Design/lo
cation
Control
Total training
(min/weeks)
Beck 2010
27
24
RCT/H
Waitlist
750/6 weeks
Green et al 2012
12
14
RCT/H
NonAdaptive
615-1000/5
weeks
Klingberg et al 2005 20
24
RCT/ H -S
NonAdaptive
1000/5 weeks
Van DonganBoomsma 2014
27
24
RCT/H
NonAdaptive
375/5 weeks
Chacko et al 2014
44
41
RCT/H
Egeland et al.,
2014
Non750-1125/5 wk
Adaptive
0.57
[99%CI .29 - .82]
VERBAL WM: SMD =
38
RCT/S= 0.47
Waitlist
VISUAL37WM: SMD
[95%CI750-1125/5-7
.23 - .70]
w
CORTESE ET AL.,920150 META ANALYSIS
23
RCT/H
Waitlist
750/5 weeks
Gropper et al 2014
39
Gray et al., 2012
36
24
TOTAL
243
211
RCT/S
Adaptive
750/5 weeks
4 RCTs from Tannock Lab
• Study 1: Does WM training enhance WM & other aspects of
cognitive, academic, behavioral functioning in adolescents
with severe Learning Disabilities & comorbid ADHD?
(Gray et al., 2012, JCPP)
• Study 2. Does WM training enhance WM & other aspects of
cognitive, academic, behavioral functioning in college
students with ADHD and/or Learning Disabilities?
(Gropper et al., 2014., JAD)
• Study 3. Working Memory Training in Post-Secondary
Students with Attention-Deficit/Hyperactivity Disorder:
Pilot Study of the Effects of Training Session Length
(Mawjee et al., in press., JAD)
• Study 4. Does WM training have specific or non-specific
effects on WM in college students with ADHD? (Mawjee et al,
under review)
Two randomized controlled trials
Gray et al 2012; Gropper et al, 2014
 Treatment-resistant sample ‣ University students with
of secondary-school
students with severe LD
with comorbid ADHD,
 Attend semi-residential
school funded by Ministry
of Education (Ontario)
 Sample size = 60
ADHD and/or LD
‣ Registered with student
disability services (Most taking
reduced course load, poor time
management, lower academic
grades)
 Mean age:14.3 yrs (1.2yr)
‣ Sample size = 62
 All medicated & receiving
‣ Mean age = 27.9 yrs(7.1yr)
intense academic
remediation
Study 1:
High-school LD/ADHD
‣ None medicated
Study 2:
University ADHD/LD
Intervention: treatment arms
• Comparison of two
active, computerized
intervention arms
• Working Memory
Training (Cogmed RM)
• 20 x 45-min sessions
• Academic Training
(Academy of Math)
• 20 x 45-min sessions
• During school day
Study 1:
High-school LD/ADHD
• Comparison of active vs
inactive intervention arm
• Working Memory
Training (Cogmed QM)
• 25 x 45-min sessions
• Wait-List Control with
some phone-calls
• Own schedule (at
home/residence)
Study 2:
University ADHD/LD
Measures
• Criterion
• WISC-IV Digit SPAN F/B
• CANTAB Spatial Span F
• Near Transfer
• CANTAB Spatial WM
• D2 Test of Attention
• Far Transfer
• Academics
• WRAT-4 Progress Monitoring
•
• Criterion
• WAIS –IV Digit Span F/B/S
• CANTAB Spatial Span F
• Near Transfer
• CANTAB Spatial WM
• Paced Auditory Serial Addition
• Ruff 2& 7 Selective Attention
• Far Transfer
•
• Nelson Denny Reading Test
Behavior (Parent/teacher)
• SWAN, IOWA
• Working Memory Rating Scale
Study 1:
High-school ADHD/LD
Academics
• WJ-III Achievement
•
Behavior (self-rating)
ASRS, Cognitive Failures
Study 2:
University ADHD/LD
Training Compliance
 Good overall
• Generally good overall
 Attrition at post-test:16%
• Attrition by post-test: 10%
(n=8) of 60 participants,
equally distributed across
treatment arms
 4 unable to cope with academic
load & this study
 3 moved & left school midprogram
 1 due to computer problems
◦ No difference between completers
& non-completers
 ANALYSIS: Intent-to-Treat
 No Follow-up
Study 1:
High-school LD/ADHD
(n=6) of 62 participants, equally
distributed across treatment
arms
• ANALYSIS: Intent-to-Treat
• But, attrition by follow-up: 30%
from WM group; 30% from waitlist group
• primarily conflict with exam
schedules
• Thus follow-up analysis based
on As-Treated
Study 2:
University ADHD/LD
Compliance outcomes
• WM training group:
70% obtained WM
Improvement Index > 17
• (Mean Improvement
score = 18.85, SD = 6.3)
• AOM group:
57% mastered >10
skills
• WM training Group
97% obtained WM
Improvement Index >17
(M=25.72, SD=6.54)
• 92% completed the
required 25 sessions
• Mean # skills mastered =
19.81, SD=14.14.
Study 1:
•High-school LD/ADHD
Study 2:
University ADHD/LD
Overview of Results
• Criterion
• WISC-IV Digit SPAN F/B
• CANTAB Spatial Span F
• Criterion
• WAIS –IV Digit Span F/B/S
• CANTAB Spatial Span F
• Far Transfer
• Academics
• Far Transfer
• Academics
• Near Transfer
• Near Transfer
• CANTAB Spatial WM
• CANTAB Spatial WM
• Paced Auditory Serial Addition
• Working Memory Rating Scale
• Ruff 2& 7 Selective Attention
• D2 Test of Attention
• Nelson Denny Reading Test
• WRAT-4 Progress Monitoring
• Math – trend for Math Training
•
Behavior (Parent/teacher)
• SWAN, IOWA
Study 1:
High-school LD/ADHD
• WJ-III Achievement
•
Behavior (self-rating)
ASRS, Cognitive Failures
Study 2:
University ADHD/LD
Training effects on Criterion Measure
DS-Backwards
Intent-to-Treat Analysis: Ancova post-test, covarying pretest score
ES: Cohen’s d = .55
33
Standardized Level Equivalent scores (LE)
Training effects on WRAT-PM Math
p = .08
Study 2: Maintenance of gains in criterion measures at 2-month follow-up
Effect size ŋ2 = .08
Exp Group 28% greater
improvement than controls
Analysis: As Treated
Repeated measures:
Exp(WM): n=23
Wait-List : n=16
Effect size ŋ2 = .22
Exp Group 47% greater
improvement than controls
Revised manuscript resubmitted to PlosOne Jan 2015
So our next question was…
Does WM training improve WM when
controlling for participant’s motivation,
engagement & expectancy?
(using independent intervention &
research teams)
Methods: an RCT
• Participants: 97 post-secondary students with ADHD
• Treatment arms:
• standard-length adaptive Cogmed WM training: 45-
min/session 5 days /week, 5 weeks,
• shortened-length adaptive version: 15 min/session, 5
days/week, 5 weeks
• waitlist control group provided with weekly telephone
advice about ADHD, WM, Disability Services etc
• All three groups received weekly telephone calls from
trained, CMWT coaches independent from research team
• Procedures.
• Measures taken before, 3 weeks after training period; those in the two
CWMT groups were also assessed 3 months post-training.
Hypothetical patterns of findings
Outcome measures
• Current symptom
validation
ASRS: self- interview, Q
other-Q
• Compliance
# sessions completed &
# weeks to complete
# coach calls completed
Cogmed Training Index
Attrition
• Criterion
WAIS–IV Digit Span F/B/S
• CANTAB Spatial Span F/B
• WRAML FW F/B
•
• Near Transfer
CANTAB Spatial WM
• Kahneman’s Addition Task
•
• Far Transfer
Symptoms: ASRS
• Function: BDEFS, CFQ
•
Results: examples
An unexpected
pattern of findings
(ceiling effects? Nonspecific training
effects?…)
These findings are
based on analysis at
the group level..what
about at the
individual level?
Brinley Plots:
A visual approach to
displaying data at both
group & individual leve
calculation of RCI is in
progress..
But perhaps our neural measures will
provide further insight into effects of
CWMT on WM
Neural measure of training effects on WM
ADHD impaired in encoding stage of WM
Kim et al., (2014) Clin Neurophysiol
• ADHD (n=30) lower scores on behavioural working memory
tasks compared to CTL (n-25), suggesting impaired
behavioural WM performance
• Smaller P3 amplitude in ADHD group compared to CTL
group: both load conditions, at parietal–occipital sites.
.
Group differences in maintenance stage
(preliminary findings)
• ADHD (n= 136, ~50% F); CTL (n=41, ~50% F)
• ADHD poorer WM on standardized tests of WM
• ADHD vs CTL
• Behavioral results: delayed match-to-sample task
• ADHD tended to perform more poorly (more errors)
than CTL
• Neural results:
• ADHD lower alpha power, particularly for high load,
compared to CTL
Preliminary data: training effects on
WM during maintenance phase
• Neural pathways associated with working memory do seem
to show treatment changes for the high intensity standardlength training. These effects were only present during the
high load condition at parietal sites.
• No evidence for behavioral task differences were found.
Alpha power
Waitlist
Pre
Post
Shortened-length
Standard-length
Interpretation & Discussion of results
• “Half-empty”
• “Half-full”
• No robust evidence of
• Robust evidence that
transfer of treatment
gains to untrained
WM activities or daily
functioning
computerized cognitive
training enhances WM –
as measured on
standardized
neuropsychological
tests-& perhaps neurally
Or does the program simply result in
learning how to do the tests better!
Caveat
• Not yet an evidence-based intervention
for ADHD - But premature to discard
cognitive training
• Most of the published research studies have
methodological weaknesses (including ours)
• Need for better studies & more sensitive
measures (e.g., measure concurrent behavior &
WM)
• Advances in the training paradigm may yield
better outcomes
Spencer-Smith M, Klingberg T (2015)
Benefits of a Working Memory Training Program
for Inattention in Daily Life: A Systematic Review
and Meta-Analysis. PLoS ONE 10(3): e0119522.
Acknowledgements
Collaborating Organizations
 Cogmed America Inc
◦ provision of research licenses
 JVS-Toronto
◦ licensed Cogmed providers
 Ontario Provincial
Demonstration Schools
 University Disability Services
Drs. H Gottlieb
& RChairs
Kronitz (JVS)
• Canada
Research
(RT)
Program
Desiree Smith,
Denise
Murnaghan (OPDS)
• Provincial Centre of
Peter Chaban, Min-Na
Excellence
forMarisa
Child Catapang
& Youth
Hockenberry,
Mental
(HSC) Health at CHEO (RT)

Dr Marc Lewis
(OISE)of Health
• Canadian
Institutes
Best
Research
GraduateBanting
Students&(OISE)
Award
(RG)
 Rachel Gropper, Steven
Woltering, Zhongxu Liu,
Sarah Gray,
Christine Popovich
◦ York University, University of
Toronto
 The Hospital for Sick Children

Dr Torkel Klingberg (consultant)
 Ontario Institute for Studies in
Education/University of Toronto
Funding
Research
Team
TIME FOR ME TO STOP!
ANY QUESTIONS?