Training the ACC with localized EEG-neurofeedback – a pioneer study Sina Radke1,2, Tanja Kellermann3, Lydia Kogler1, Stefanie Schuch4, Herbert Bauer5, Birgit Derntl1,2 1Department
of Psychiatry, Psychosomatics and Psychotherapy, Uniklinik RWTH Aachen, Germany; 2Jülich Aachen Research Alliance (JARA) - Translational Brain Medicine, Jülich/Aachen, Germany; 3 Department of Neurosciences, Medical University of South Carolina, Charleston, SC, USA; 4 Department of Cognitive & Experimental Psychology, RWTH Aachen University, Germany; 5 SCAN-Unit, Faculty of Psychology, University of Vienna, Austria
Background
Sample
Neurofeedback is a method which provides real-time signals and therefore can be used to regulate brain activity ‘online’. Recent methodological advances have made it possible to employ EEG-based feedback resulting from local brain activity (LBA-EEG Neurofeedback, Bauer et al., 2011). The aim of the current study is to train the anterior cingulate cortex (ACC) of healthy individuals within a 10-day period using LBA-EEG Neurofeedback. For that purpose, the effectiveness of the training is evaluated by direct parameters of the training as well as by performance and transfer measures assessed before and after the training period via EEG, fMRI, DTI, resting-state, self-report questionnaires and behavioral performance.
Aimed sample • 20 healthy volunteers: 18-40 yrs, right-handed • 2 groups based on region of training: ACC vs. DLPFC (BA 24/32 vs. BA 46) • 20 units of training within a 10-dayperiod: 2 x 70 trials, i.e., 2 x 8 min, per day
Current sample N = 10 (ACC) and N = 3 (DLPFC) (M age = 24 yrs; 7F, 6M)
Range: 13-20 units of training
Methods & Outcomes
LBA EEG-Neurofeedback
fMRI tasks and evaluation fMRI: 2 Stroop-like tasks (Age- & Emotional Stroop); Resting-State; DTI
Age-Stroop: judge the person‘s age (younger, middle, older; 48 stimuli x 4 = 192 trials)
OLDER
950 900
Day 1
64-Ch
EEG: Cognitive, i.e., Age-Stroop BEM: boundary element method SMS: simultaneous multiple source LORETA: low resolution electromagnetic tomography
OLDER
Task: Make the frame turn green
Congruent Incongruent
All stimuli were presented for 3s with an ITI of 4s (+/-1s). Mean feedback frequencies were analyzed in a 2x4 rm ANOVA (group x time points)
750 700
Day 2
All stimuli were presented for 1s with an ITI of 3s (+/-1s). Mean RTs of correct responses were analyzed in a 2x2 rm ANOVA (congruency x time points)
Emotional Stroop: judge the person‘s emotion (fearful, happy, sad; 30 stimuli x 8 = 240 trials) HAPPY
Day 2-11
1100
Congruent Incongruent
Frequency in %
28,7
27,4
20
EEG: Cognitive, i.e., Age-Stroop Day 11
22,7
*
15
ACC DLPFC
10
900 850 800
Pre
HAPPY GLM: 6 relevant regressors (congruency x stimulus category)
6,7
5,2
5,7
0
Unit 1
Unit 6
Unit 11
Last Unit
Mean feedback frequency of the two groups indicating sig. more frequent feedback in the ACC-group (N = 10) than in the DLPFC-group (N = 3) at all depicted time points.
Post
Mean RTs (with SE) of the ACC-group (N = 10) indicating sig. congruency effects both pre- and post-training
fMRI: brain activity post > pre training
8,6
5
*
950 RTs in ms
Direct training parameter: Frequency
25,1
*
1050
35
25
Post
Mean RTs (with SE) of the ACC-group (N = 10) indicating sig. congruency effects both pre- and post-training
1000
30
*
800
Pre
OLDER
EEG-Neurofeedback Training
(70 stimuli/trials, half of them congruent)
RTs in ms
3T MR scanner TE = 28; TR = 2s; 34 slices; 3.3 mm³ voxel size
OLDER
*
850
fMRI: 2 Stroop-like tasks (Age- & Emotional Stroop); Resting-State; DTI Day 12 Note. * p < .05. FMRI-image thresholded at T = 4.86
Cluster in the mid-orbital gyrus, extending to the ACC (maximum at 6 24 -10; z = 7.77) showing more activity during the Age-Stroop after than prior to the neurofeedback training in the ACC-group (N = 10). A similar effect is currently lacking in the DLPFC-group.
Planning & Outlook Preliminary results suggest the newly developed Age-Stroop task to be successful in not only eliciting behavioral congruency effects, but also in facilitating ACC-based neurofeedback. After data collection is complete, it will be examined whether and how neurofeedback resulting from localized brain activity can alter cognitive functions. To achieve this, pre- and post-training behavioral parameters, event-related potentials, functional MRI as well as resting-state and DTI scans will be compared. Further analyses will focus on group differences in performance, brain activity and connectivity. By monitoring participants‘ mood throughout the training, we aim to capture its impact on subjective affect. In this manner, the feasibility of the training will be evaluated and discussed. These results may aid in optimizing EEG-based local brain activity neurofeedback trainings, which may ultimately lead to further developments with regard to therapeutic interventions for patients with affective disorders, e.g., depression. Importantly, these patients often demonstrate dysfunctions not only in performancemonitoring, but also in emotional processing and social adaptive behavior. Here, determining transfer effects to other, socioemotional contexts will be of particular relevance. Bauer, H., Pllana, A., Sailer, U. (2011). The EEG-based local brain activity (LBA-) feedback training. Activitas Nerv Sup Rediv 53, 107-113. Funded by a START grant from the Medical Faculty of the RWTH Aachen to T.K.
contact:
[email protected]
