Dynamic Causal Modelling for ERPERFs Practical session Marta

Dynamic Causal Modelling for ERP/ERFs Practical session Marta Garrido and Stefan Kiebel Thanks to James Kilner and Karl Friston

Outline • Hands on : application to the Mismatch Negativity (MMN) • Demo • Results

DCM for Evoked Responses functional connectivity vs. effective connectivity causal architecture of interactions estimated by perturbing the system and measuring the response The aim of DCM is to estimate and make inferences about the coupling among brain areas, and how that coupling is influences by changes in the experimental contex. differences in the evoked responses changes in effective connectivity

mode 1 Data acquisition and processing Oddball paradigm standards deviants mode 2 pseudo-random auditory sequence time 80% standard tones – 500 Hz 20% deviant tones – 550 Hz mode 3 preprocessing raw data 128 EEG scalp electrodes • convert to matlab file data reduction to • filter principal spatial • epoch modes • down sample • artifact correction • average (explaining most ERPs / ERFs of the variance) time (ms)

The Mismatch Negativity (MMN) is the ERP component elicited by deviations within a structured auditory sequence peaking at about 100 – 200 ms after change onset. 4 b standards deviants MMN 3 a 2 VEOG 1 m. V HEOG 0 -1 -2 -3 -4 -100 -50 c 0 50 100 150 ms 200 250 300 350 400

What are the mechanisms underlying the generation of the MMN? DCM specification a plausible model… Forward - F Backward - B Both - FB Opitz et al. , 2002 3 STG IFG 5 STG 4 l. IFG l. A 1 1 A 1 2 input Doeller et al. , 2003 modulation of effective connectivity l. STG r. IFG r. A 1 r. STG

Matlab spm eeg choose time window choose data number of svd components choose polhemus file sources or nodes in your graph DCM. AF DCM. AB DCM. AL specify extrinsic connections to driving input DCM. C Intrinsic connections modulatory effect from DCM. B compare models estimate the model visualise output

Demo

Results

S 2 A 1 S 4 A 1 S 5 IFG S 6 IFG STG STG STG A 1 A 1 A 1 Forward Backward Intrinsic input S 2 i A 1 input S 4 i A 1 input S 5 i IFG input S 6 i IFG STG STG STG A 1 A 1 A 1 input modulation of effective connectivity

results Bayesian Model Comparison model space 4 x 10 F - negative free energy -2. 55 S 5 i -2. 6 S 4 -2. 65 -2. 75 -2. 8 S 2 i -2. 85 -2. 9 S 4 i S 5 -2. 7 S 2 S 6 i

Thank you