A controlled comparison of EEG- neurofeedback and attention training for children with ADHD. Tony Steffert, Annie Frick, Martin Batty, Malcolm Hawken,

AA controlled comparison of EEG- controlled comparison of EEG-neurofeedback and attention training neurofeedback and attention training

for children with ADHDfor children with ADHD..

Tony Steffert, Annie Frick, Martin Batty, Malcolm Hawken, Claire Sturge, Katherine Loveday, John

Gruzelier.

Limitations of previous neurofeedback studies

• Rossiter & LeVaque (1995)– Combined neurofeedback with additional interventions– Failed to control for patient-therapist contact

• Linden et al (1996)– Small sample size: difficult to generalise– Failed to control for patient-therapist contact

• Fuchs et al (2003), Monastra et al (2002)– Failure to randomly allocate treatment.– Failed to control for patient-therapist contact

• This has limited the acceptance of neurofeedback as a viable alternative

Project Aim

• To examine the efficacy of neurofeedback in a randomised control trial, assessing its effectiveness in comparison to another computerised intervention. Groups will be stratified according to medication status (medicated vs. non-medicated)

Procedure

Mid session

test

NeurofeedbackNo medication.

Attention task No medication.

Neurofeedback+ Medication.

Attention task+ Medication

Pre-Test

And

QEEG

Time 1

Post-Test

And

QEEG

Time 2

Group 1

Group 2

Group 3

Group 4

30 Sessions

N =80

Subjects randomly allocated

Assessment Measures

• Test of Variables of Attention (TOVA)• Attention Network Test (ANT)• Suffolk Reading Scale (SRS)• Wide Ranging Aptitude Test – Maths (WRAT)• Eysenck Personality Questionnaire (EPQ)

Assessment Measures

• Strengths and Difficulties Questionnaire (SDQ)– Parent Ratings– Teacher Ratings

• Parenting Alliance Measure (PAM)

EEG Assessment Measures

• All assessments carried out following a medication ‘washout period’ of 48 hours

• 19-channel qEEG assessment on the Mitsar– Eyes Open,

– Eyes Closed,

– Listening

– Drawing

Selection Criteria

• No co-morbid psychiatric or neurological conditions

• Subjects receive a DSM IV subtype

(to inform the training protocol)

• Any subjects with significantly elevated beta (compared to Kropotov database) excluded

Control - Captain’s Log Computer Attention Training

• A ‘mental gym’

• Increases focused attention and short-term memory, while reducing impulsivity using specifically designed tasks

• Used in previous research with ADHD children (Sandford, 1994; Slate et al, 1998)

• Also successfully used with psychotic patients (Bell et al, 2001; Burda et al, 1994) and in cognitive rehabilitation programmes with brain injured patients (e.g., Stathopoulou & Lubar, 2001)

Training Protocol

• All subjects receive 30 sessions of Neurofeedback or Captain’s Log training

• Participants in Captain’s Log group have 5 minute eyes open baseline then 6 x 5 minute games per session (each game at 2 levels)

• Participants in Neurofeedback group have 5 minute eyes open baseline followed by 6 x 5 minute NF training per session

• Training protocol in Neurofeedback group dependent upon subtype. All training at CZ (Beauregard, 2005, Lubar and Lubar,1984)

Neurofeedback training protocol• ADHDin 30 sessions inhibit θ/enhance β1

(enhance attention)

• ADHDhyp 30 sessions inhibit θ/enhance SMR (reduce hyperactive behaviour)

• ADHDcom children 2 protocols: Sessions 1-15 inhibit θ/enhance SMR Sessions 16-30 inhibit θ/enhance β1• Rationale: inhibit motor activity to reduce hyperactive behaviour,

while increasing attention..

Assessment Measures

• Attentional Network Task (ANT) Posner et al (2002, 2004)

• Measures 3 separate attentional networks (alerting, orienting, executive control), based on earlier functional studies (e.g. Posner & Peterson, 1990)

ANT

• Alerting – maintaining an alert state; Calculated by Mean RT no cue – Mean RT double cue.

• Orienting – selecting information;

Mean RT centre cue – Mean RT spatial cue.

• Executive Attention – resolution of conflict;

Mean RT incongruent – Mean RT congruent.

Results

• All data from 23 subjects (11 NF, 12 CL)

• Data analysed using 2 (GROUP) x 3 (TIME) mixed ANOVA. Medication status and subtype not included as factors in the preliminary analysis

• Any violations in sphericity corrected using Greenhouse-Geisser adjustment

Omission Errors

0

5

10

15

20

25

30

35

40

Pre Mid Post

Time of Testing

Nu

mb

er

of

Err

ors

(%

)

Neurofeedback

Captain's Log

Pre test, t = ns. Mid test, t = -2.55 (d.f =18), p= .024

Post test, t = -2.47 (d.f =18), p= .024

TOVA Omission Errors

ANT Accuracy

75

80

85

90

95

100

Pre Post

Time of Testing

Ac

cu

rac

y (

%)

Neurofeedback

Captain's Log

Pre test t=ns. Post test, t=2.34 (d.f=18), p=.017 (2-tailed)

Strengths and Difficulties (SDQ) Parent Ratings

Both groups improved equally on the

Stress and

Hyperactivity subscales.

Interpretation

• We demonstrated some advantages for NF in improving attention on both computerised measures– TOVA omissions– Attention Network Task accuracy.

• This was in comparison with training on a Computerised Attention test, a control comparison that has not previously been reported on.

• Notwithstanding both groups improved on parent ratings of hyperactivity and stress.

Neurofeedback Session Analysis

• Future work will include analysis of the neurofeedback EEG data, both within and between sessions

• Neurofeedback learners will be distinguished from non-learners

• This differentiation will enable us to determine whether learned modulation of the EEG correlates with improved performance on the various cognitive and behavioural measures

Methodological Issues

• Threshold settings

•  Reward system

•  Fixed protocols

•  5 minute trials to long

•  Filter settings

Dr Catherine Loveday The University of Westminster.

Drs Linda and Michael Thompson.

Dr Sturge and colleagues, Northwick Park Hospital.

Dr David Vernon.

ADHD Support Group, Harrow.

Silvio Adriovati, Christoph Moehlbrink, Andrea Oskis, Silvie Rainer, Keri Thornton, Wallis Levin.

All the parents and children who have taken part in the study.

Thanks to: