The effect of Ldopa on fMRI of language processing

1
The effect of L-dopa on fMRI of language
processing

• Namhee Kim , Madalina E. Tivarus, Prem K. Goel,
  David Q. Beversdorf

2
Introduction

• L-dopa, a precursor for dopamine
• Amplify strong signals, and dampen weak signals
• Restrict the semantic network
• Demonstrated in a lexical priming test
  (Kischka,96)
• Under placebo, significant differences in RT,
• non-related vs. related and non-related vs.
  indirectly related
• Under L-dopa
• non-related vs. related
• Decreased indirect semantic priming effect under
  L-dopa was discussed in terms of restriction of
  the semantic network

3
Introduction

• We aimed to examine the effect of L-dopa on
  language processing with fMRI
• Phonological task and Semantic task

4
Methods

• Participants 15 healthy subjects right handed
• Experimental paradigm
• 2 sessions L-dopa and placebo
• In each session, 2 runs
• Semantic (phonological) task (24 sec) with rest
  (30 sec)
• Total time for each run was 4 min and 6 s
• A cue word for 3 seconds followed by 15 words.
• Each word was presented for approximately 1100
  ms, with a 300 ms inter-stimulus interval (a
  blank screen)
• Respond by pressing one of the two buttons YES or
  NO.
• YES if the word in the list was related by
  meaning (or rhyme) to the cue in semantic
  (phonologic) task

5
Methods

• fMRI procedure
• 1.5 T General Electric scanner with a quadrature
  head coil.
• The BOLD contrast functional data were collected
  using a gradient echo EPI pulse sequence
• TR3 s TE40 ms a90 FOV240 mm matrix 6464
• 28 axial slices for whole brain coverage
• 5 mm slice thickness

6
Block Design from Tivarus et al. (2008)
7
Analysis

• Preprocessing
• Realignment (motion correction)
• 6 parameter (rigid body) spatial transformation
• Spatial Normalization to MNI Space
• Template EPI.mnc in SPM
• an average of 13 subjects' mean fMRI images
• Spatial smoothing
• FWHM in mm (8 8 8)
• Independent Component Analysis (ICA)

8
Analysis

• ICA steps
• Perform ICA for each subject
• Choose 25 ICs by doing a preliminary analysis
• Group ICA of fMRI Toolbox (GIFT)
• Select task positive ICs
• correlation with convolved task related HRF with
  threshold 0.33
• Number of selected ICs varies over subjects
• Generate a summary source map for each subject

9
Analysis

• ICA
• Summarize the sources by the maximum value
• Example
• Assign 5 if S1(vok)2, S2(vok) 5 and
  S3(vok)1
• Assign -5 if S1(vok)2, S2(vok) - 5 and
  S3(vok)1
• Negative values are found in the suppressed
  regions for the given task in a task-positive
  source
• Effects
• High sensitivity for the detection of the
  task-involved regions in the brain

10
Analysis I. Group activation map

• Group activation maps of 4 conditions by the task
  and drug types
• phonological/L-dopa, phonological/placebo
• semantic/L-dopa, semantic/placebo
• T-test over the summary maps of subjects for each
  condition as above
• Threshold values
• Voxel level p-value 0.005 (t-value 2.98)
• Cluster level p-value 0.05 (spatial extent
  14-18)

11
Phonologic task activation map by treatments
Left parietal
Red L-dopa BluePlacebo
Anterior LIFG
Left fusiform gyrus
thalamus
Posterior LMTG
12
Semantic task activation map by treatments
Red L-dopa Blue Placebo
Anterior LIFG
Posterior LMTG
Thalamus Medial dorsal
13
Analysis II. Test on the differences in the group
activation maps by drug conditions

• 2 group t-test
• Threshold values
• Voxel level p-value 0.01 (t-value 2.47)
• Cluster level p-value 0.05 (spatial extent
  30-32)

14
Results Test on the differencesL-dopa vs.
Placebo
phonology
thalamus
LSTG LMTG
Red L-dopagtplacebo Blue L-dopaltPlacebo
Frontal lobe
thalamus
Semantics
15
Summary

• Proposed method reveals a similar activation map
  as compared to McDermott et al. (2003) with the
  same task
• Semantic task has activations on
• BA47, BA44/45, BA22/21, BA37, left BA40 and
  cerebellum
• Phonological task activations on
• BA6/BA44, left BA40 (bilateral in McDermott)
• Majority of the group activation maps by the drug
  conditions (L-dopa, placebo) overlapped
• However, the group activation maps in thalamus
  showed difference between drug conditions
• Greater activation in placebo for phonological
  task

16
Summary

• The dopamine neurons primarily project to frontal
  and subcortical regions
• L-dopa is associated with greater frontal
  activation in the semantic task, perhaps relating
  to its effect gating in the semantic network in
  the priming experiment

17
Summary

• However, while the increased temporal activation
  associated with L-dopa could relate to more
  focused signal processing in phonological tasks,
  this seems surprising given the distribution of
  DA projections
• Our results may suggest that these effects in the
  phonological task are related to decreased
  thalamic gating with L-dopa

18
Further research questions
LIFG

• In GLM analysis, the difference in activation in
  thalamus with the same drug conditions was not
  significant
• Thalamus might be indirectly related to the
  language processing resulting in delayed response
• The time course in thalamus needs to be extracted
  and inspected
• Searching a semantic network connecting all the
  ROIs and explaining meaningfully
• Searching ROIs highly connected to the thalamus
• Functional connectivity analysis among all the
  ROIs

LPAR
Thalamus
LMTG
LFUS
19
References

• Calhoun V., D., Adali T., Pearlson, G., D.,
  Pekar, J., J. (2001) A method for making group
  inference from functional MRI data using
  independent component analysis., Human Brain
  Mapping, 14, 140-151
• Kischka, U., Kammer, T., Maier, S., Weisbrod, M.,
  Thimm, M., Spitzer, M. (1996) Dopaminergic
  modulation of semantic network activation.,
  Neuropsychologia, 34, 1107-1113
• McDermott, K., B., Pertersen, S., E., Watson, J.,
  M., Ojemann, J., G. (2003). A procedure for
  identifying regions preferentially activated by
  attention to the semantic and phonological
  relations using functional magnetic resonance
  imaging. Neuropsychologia, 41 293-303
• Tivarus M., E., Hillier, A, Schmalbrock, P.,
  Beversdorf, D., Q. (2007) Functional connectivity
  in an fMRI study of semantic and phonological
  processes and the effect of l-Dopa. Brain
  Language

20
Acknowledgement

• This research was funded by NINDS (K23
  NS43222-Beversdorf) and a pilot grant from NAAR
  (National Alliance for Autism Research-Beversdorf)
  and the OSU Research Investment Fund