1Disclosure: no potential conflicts of interest. Funding EU, SNF, University grants.

OHBM 2014 Electrical Neuroimaging course 2014

EEG & ERP - baselines

Daniel Brandeis

1) Department of Child and Adolescent Psychiatry, 2) Center for Integrative Human Physiology, 3) Neuroscience Center Zurich, University of Zurich

and ETH Zurich, Switzerland, 4) Department of Child and Adolescent Psychiatry and P sychotherapy,

Central Institute of Mental Health, Medical Faculty Mannheim/ Heidelberg University, Germany

4th Human Brain Mapping, June 8-12, 2014, Hamburg, Germ any

2

Baseline

•Baseline vs reference

•Baseline state

•Baseline condition

•Baseline correction, filtering

•Informed baseline choice

3

+

++

EEG sources, map zero, baseline

Scherg

Brandeis & Michel 2009

Baseline (t) ≠ reference (x,y,z)but implict zero baseline at recording reference

4

Baseline - reference

Mismatch negativity: Gottselig et al 2004

Michel et al 2009

Baseline (t) ≠ reference (x,y,z)but implict zero baseline at recording reference

Nasion

EarsMastoids

5

EEG, Microstates

From Michel et al: Electrical Neuroimaging

EEG Baseline: amplifier technical zero + filter settings

6

EEG: envelopes

3sec1-1

+100 µV

-100 µV

slow

3 sec1-1

+100 µV

-100 µV

fast(er)

Brookes et al 2011; Hipp et al 2012

Event-related EEG changes (ERD, ERS): baseline as reference interval

7

Maps and waveshapes: displaying the same data

ERP Baseline: technical zero + filters; not prestimulus activity

8

ms (Milliseconds)

+50 µV

-50 µV

5001000

Event-related potentials (ERP) –phase locked averaging of EEG

(Microvolt,1/1’000’000

Volt)

1. Event

2. Event

... Event

30. Event

Average

---- ERP: time course of information processingERP Baseline: noise mean zero, prestimulus mean ?

9

Baseline

Lehmann et al 1982, 1984; Koenig & Gianotti 2009

Isolate … Distort

VEP >P300

+CNV

-Baseline (CNV)

=VEP-correction>CNV- inversion

10

Deconvolution

Woldorff et al 1993

Lütkenhöner 2011

Deconvolving / separating overlapping activity requires jitter between events

11

H

Typical ERP Attention Test

cued Continuous Performance Test (CPT O-X, A-X) Regular letter sequence : press for O-X sequence (10%)

fixed SOA 1.65s – systematic time estimation & preparation

O S

O X

: Attention, preparation & Go-NoGo- Test

Ignore

Attend & Prepare - CNV

Inhibit

Attend & Prepare - CNV

Execute

Rosvold et al 1956, van Leeuwen et al 1998, Doehnert et al 2013…

12

Cue

Distractor

+7.0 µV

-7.0 µV

0.1 0.2 0.4 1.5 seconds

0.0

2.0

4.0

6.0

0.1 0.4 0.9 1.4 1.9

Strength in µV

P1 N1

P300

G:\MM6\MM_P1\Tn148.DAT

P300

Attention - baseline

Attention: Preparation before 0 s, Filter after 0.1 s, Resource after 0.4s …

CNV

13

Baseline –preparatory CNV

McLoughlin et al 2011

*

*

14

CNV relevant preparatory prestimulus activity

Reduced CNV in ADHD - Impaired preparation is familal

-- Controls -- Nonaffected Sibs -- ADHD

Albrecht et al. 2012

15

Baseline – correction vs distortion

Subtraction assumes constant, sustained, irrelevant baseline activity

16

Baseline - affects variance

17

[Hbdeoxy] [Hboxy]

O2

fMRI Activation … O2

fMRI – EEG (t)

0 5 10 15 20s 0 .5 1.0 1.5 2.0s

EEG/MEG Activation … +/- Polarization

Activation - SynchronizationSummation- Cancellation

Localise – EstimateIndirect - Direct

Slow - Fast

Normalisation

Statistical Parametric MapImage time-series

Parameter estimates

General Linear ModelRealignment Smoothing

Design matrix

Anatomicalreference

Spatial filter

StatisticalInference

RFT

p <0.05

fMRI: Methods (SPM)

No baseline correction

19EEG-course-jan14 HS 2013

Baseline

•Baseline - technical zero, filter, zero mean

•Baseline - state, condition

•Baseline „correction“ - often inadequate

•Baseline - analysis & informed choice