EEG-neurofeedback training of elite singers including fMRI assessments.

1
EEG-neurofeedback training of elite singers
including fMRI assessments.
Conjunct COST B27 and SAN Scientific Meeting,
Swansea, UK, 16-18 September 2006

• Boris Kleber, John Gruzelier, Martin Lotze, Ralf
  Veit , Mike Bensch Niels Birbaumer

Institute of Medical Psychology and Behavioral
Neurobiology, University of Tübingen
2
Aims

• Continuation of a series of studies that
  demonstrated that alpha/theta eeg-biofeedback
  training could enhace the performance of music
  conservatory students (Egner Gruzelier, 2003).

3
Introduction

• In this study we focused on elite classical
  singers (mostly opera).
• Why?
• A homogeneous group may reveal possible effects
  that could get masked otherwise, and
• since singers are their instrument, training
  related changes in performance may be traced more
  easily.

4
Method

• EEG Biofeedback
• Alpha/theta training (a/t acoustic feedback)
• aims at elevating electrocortical theta (5-8 Hz)
  and alpha (8-11 Hz) activity at electrode Pz in
  an eyes closed resting state.
• Sensorimotor-rhythm (SMR visual feedback)
• aims at elevating electrocortical smr (12-15 Hz)
  activity without concurrent rise in theta (5-8
  Hz) activity at electrode C4 in an eyes open
  resting state.
• Nexus-10 with Biotrace Software (MindMedia, NL)

5
Design
3 Groups Assessment 1 Training Assessment 2
1. a/t

• Brain activity related to singing (fMRI).
• Musical performance
• Video and pure audio recordings
• mood and anxiety questionnaires
• heart rate SCL

• Brain activity related to singing (fMRI) and
  feedback training.
• Musical performance
• Video and pure audio recordings
• mood and anxiety questionnaires
• heart rate SCL

10 Sessions á 15 minutes feedback training, 1-2
times the week. No training for the control
group
2. smr
3. control
Analysis of Data
6
fMRI technique

• 1.5 Tesla whole body Scanner (Siemens Vision).
• 66 whole head scans were performed (per block)
  with a Echo planar imaging (EPI) TE 40 ms TA
  3 sec, TR 10 sec,
• Sparse sampling allowed auditory control during
  singing and avoided movement artifacts.
• Data were analyzed with SPM99 using conventional
  preprocessing and fixed effect group statistics
  (plt0.05 False Discovery Rate, FDR).

7
The study is ongoing....

• The data presented here are preliminary data
  representing fMRI pre-/post measurements in
  relation to neurofeedback training.
• 12 subjects selected
• 4 a/t
• 4 smr
• 4 controls

8
Alpha/theta ratio
9
SMR/theta ratio
10
fMRI task
11
fMRI task
Singing task (6 repetitions)
Resting condition (6 repetitions)
Rest Breathing only
12
fMRI sparse sampling
13
fMRI sparse sampling
14
Results

• Postscan minus prescan (control group)
• No differences

15
Results fMRI

• Postscan minus prescan in training groups
• Differential effect for alpha/theta and smr group.

16
Alpha/ Theta Traininga/t (post vs. pre) - minus
- ctrl (post vs. pre)

• Activation in the right medial insula

17
Alpha/ Theta Traininga/t (post vs. pre) - minus
- ctrl (post vs. pre)

• Activation in the right medial insula
• Right temporal pole

18
Alpha/ Theta Traininga/t (post vs. pre) minus
control (post vs. pre)

• Activation in the right medial insula
• Right temporal pole
• Left frontal inferior orbital cortex (BA47)

19
Alpha/ Theta Traininga/t (post vs. pre) minus
control (post vs. pre)

• Activation in the right medial insula
• Right temporal pole
• Left frontal inferior orbital cortex (BA47)
• Pons/Medulla

20
Alpha/ Theta Training

• The Insula
• plays a role in regulating physiological and
  psychological homeostasis (Flynn, Benson,
  Ardila, 1999) and is considered being part of the
  emotional viscerosensory brain (Janig Habler,
  2002).

21
Alpha/ Theta Training

• Temporal pole
• The right temporal pole is correlated with
  attending to ones own emotional experience and
  seem to be involved in the imparting of emotional
  color to subjective experience (Lane, 2000).
• It seems important to consciously and willfully
  self-regulate emotional responses (Mesulam, 1985
  Beauregard et al., 2001).

22
Alpha/ Theta Training

• VLPFC (BA47)
• Left BA47 is involved in semantic processing
  (Fiez, 1997) but also in the passive perception
  of emotional stimuli (visual Blair et al., 1999
  linguistic Wildgruber et al., 2004).
• Activation was found during recognition of
  expressive gestures (Gallagher Frith, 2004)
• Left BA47 might be important for the coding of
  the valence of emotional qualities (Lotze et al.,
  2006)

23
Alpha/ Theta Training

• The Pons
• Is involved in motor control and sensory analysis
  and is important for the level of consciousness
  and for sleep.

24
SMR Trainingsmr (post vs. pre) minus control
(post vs. pre)

• Increased sensorimotor, parietal and auditory
  activation
• Somatosensory Premotor areas (BA3/ 6)

25
SMR Trainingsmr (post vs. pre) minus control
(post vs. pre)

• Increased sensorimotor, parietal and auditory
  activation Somatosensory Premotor areas (BA3/
  6)
• Parietal superior (BA 5/ 40)

26
SMR Trainingsmr (post vs. pre) minus control
(post vs. pre)

• Increased sensorimotor, parietal and auditory
  activation Somatosensory Premotor areas (BA
  3/ 6)
• Parietal superior (BA 5/ 40)
• Right frontal Inferior Operculum

27
SMR Trainingsmr (post vs. pre) minus control
(post vs. pre)

• Increased sensorimotor, parietal and auditory
  activation Somatosensory Premotor areas (BA
  3/ 6)
• Parietal superior (BA 5/ 40)
• Right frontal Inferior Operculum
• Auditory belt area (BA 21 right)

28
SMR Trainingsmr (post vs. pre) minus control
(post vs. pre)

• Increased sensorimotor, parietal and auditory
  activation Somatosensory Premotor areas (BA
  3/ 6)
• Parietal superior (BA 5/ 40)
• Right frontal Inferior Operculum
• Auditory belt area (BA 21 right, BA42 left)

29
SMR Trainingsmr (post vs. pre) minus control
(post vs. pre)

• Increased sensorimotor, parietal and auditory
  activation Somatosensory Premotor areas (BA
  3/ 6)
• Parietal superior (BA 5/ 40)
• Right frontal Inferior Operculum
• Auditory belt area (BA 21 right, BA42 left)
• Cerebellum

30
SMR Training

• Increased activity was found in areas related to
  auditory and motor function
• Areas related to motor function involved primary
  somatosensory (BA3) and premotor cortex (BA6),
  the operculum, the cerebellum and the posterior
  parietal cortex.
• The premotor cortex (PMC) is important for the
  concept, timing and ideation of the movement
  (Lotze et al., 2003)
• The cerebellum is preferentially involved in
  controlling complex movements with involvement of
  sensoric feedback and learned automatic movements
  (Thach et al., 1992)

31
SMR Training

• The inferior frontal operculum belongs to the
  classical perisylvian language system (Jeffries
  et al., 2001) and is involved in phonological
  processing and in motor aspects connected with
  vocal production (Janata Grafton, 2003
  Stanberry, 2005).
• The superior parietal lobe is involved in the
  storage of movement kinematics (e.g., Seitz et
  al., 1997) and is closely connected with the
  posterior SMA and with the PMC (BA6) (Rizzolatti
  et al, 1998).

32
SMR Training

• Areas related to auditory processing (right BA21
  left BA 42)
• The right BA 21 is selectively involved in voice
  perception (Zatorre et al., 2000).
• Left sided auditory areas are usually dominant
  for the perception of temporal changes of an
  auditory signal (e.g. in speech, Schönwiesner et.
  al., 2005) but analytical listening strategies
  can also lead to left hemispheric auditory
  processing (Mazziotta et al., 1982)

33
Conclusions

• Alpha/theta training may lead to an activation of
  brain areas that concentrate on emotion
  modulation.
• This supports the finding that alpha/theta
  training enhanced artistic expression in the
  performance of conservatory music students
  (Egner Gruzelier, 2003).

34
Conclusions

• In contrast, post-smr scans revealed increased
  somatosensory coupling as well as activity in the
  auditory belt area.
• This is interesting, since increased
  sensorimotor or mµ -rhythms are usually
  associated with reduced motor acticity.
• Previous studies have shown that activity in a
  10-Hz mµ band correlated negatively with activity
  in the right postcentral gyrus and posterior
  parietal cortex (BA 5) (Ritter et al., 2003).

35

• Thank you!

36
Thanks to