SPM 2002 Experimental Design Daniel Glaser Institute of

SPM 2002 Experimental Design Daniel Glaser Institute of Cognitive Neuroscience, UCL Slides from: Rik Henson, Christian Buchel, Karl Friston, Chris Frith, Cathy Price, Ray Dolan The Wellcome Department of Cognitive Neurology, UCL London UK http//: www. fil. ion. ucl. ac. uk/spm

Data transformations Kernel Design matrix Smoothing General linear model Image time-series Realignment Statistical parametric map (SPM) Statistical inference Normalisation Gaussian field theory p <0. 05 Template Parameter estimates

Stimuli SPM{T} + isoluminant stimuli (even) + V 5 speed isochromatic stimuli (odd)

A taxonomy of design • Categorical designs Subtraction Conjunction - Pure insertion and cognitive subtraction - Differential event-related f. MRI - Testing multiple hypotheses • Parametric designs Linear Nonlinear - Cognitive components and dimensions - Polynomial expansions and neurometric functions - Nonlinear event-related f. MRI • Factorial designs Categorical Parametric - Additive factors and pure insertion - Psychopharmacological designs - Adaptation, modulation and dual-task inference - Linear and nonlinear interactions - Psychophysiological Interactions

A categorical analysis Experimental design Word generation G Word repetition R RGRGRG G - R = Intrinsic word generation

BOLD EPI f. MRI at 2 T, TR 3. 2 sec. Words presented every 16 secs; (i) studied words or (ii) new words Differential event-related f. MRI Parahippocampal responses to words SPM{F} testing for evoked responses SPM{F} testing for differences studied words new words Peri-stimulus time {secs}

Cognitive Conjunctions Components Object Recognition Phonological Retrieval Visual Processing R P V Tasks Object naming Colour naming R, P, V, Rx. P P, V Object viewing Colour viewing R, V V Object recognition responses (R) The Conjunction [ R, V - V ] & [ R, V, P, Rx. P - V, P ] R, Rx. P & R = R viewing naming

A parametric analysis The [nonlinear] effect of time assessed with the SPM{T}

Nonlinear parametric responses: Inverted ‘U’ response to increasing word presentation rate in the DLPFC SPM{F} 42 wpm Regression and design matrix

Neurometric functions: Extrastriate responses to stimuli of increasing duration: Stimulus-specific or contextual effects? SPM{F} Adaptation of neuronal responses time{ms} Attentional modulation 0 exposure duration {ms} 800

Hemodynamic responses to words in the left peri-auditory region • • • BOLD EPI T 2* at 2 Tesla TR = 1. 7 s 34 s epochs of aurally presented words at 0, 15, 30, 60 and 90 words per minute 1200 volume time-series State-related responses Event-related 60 wpm 32 secs Peri-stimulus time SPM{F}

Context-sensitive responses time stimulus n-1 stimulus n+1 interaction between stimuli response n

Volterra series - a general nonlinear input-output model input u(t) Simulated input kernel estimates response y(t) nonlinear saturation

Nonlinear hemodynamic responses kernel coefficients - h SPM{F} p < 0. 001 SPM{F} testing H 0: kernel coefficients = h = 0

Nonlinear hemodynamic responses - implications Hemodynamic response interference in terms of the effect of a prior stimulus nonlinear saturation

Rate-dependent responses: Processing one stimulus in the context of others ERPs - single words 42 wpm Integrated response 1400 ms 42 wpm 600 ms 1400 ms 700 ms nonlinear interactions 84 wpm ? Rate {wpm}

Nonlinear BOLD responses left right BOLD Linear r. CBF responses r. CBF Peri-auditory responses in PET and f. MRI Word presentation rate

A factorial analysis Time x condition interactions (i. e. adaptation) assessed with the SPM{T}

Interactions and context-sensitive effects A 2 x 2 layout Context 1 (no naming) 1 2 Context 2 (naming) 3 4 without A & with A (e. g. . recognition) interaction effect (A x Context) A A task 1 2 Context 1 3 4 Context 2

Interaction effects in the left inferotemporal region A PET study of object naming V R P Rx. P Conjunction (name object - shape) & (view object - shape) = (R + Rx. P) & R = R Interaction (name object - shape) - (view object - shape) = (R + Rx. P) - R = Rx. P Object-specific activations adjusted r. CBF Components Visual processing Object recognition Phonological retrieval Interaction Context: no naming

Psychopharmacological studies • PET, 6 subjects, 6 conditions • 3 x 2 factorial design, Buspirone x Memory left parahippocampal responses sub- suprapreacutepre post- acute post

Dissociating the direct and modulatory effects of context Context-sensitive effect = S x N response A direct effect of context = C time C frequency response time S frequency

Responses in the right posterior superior temporal region to increasing rates (15 - 90 wpm) of word production r. CBF {ml/dl/min} 84 77 whilst repeating words and producing new ones

Interactions between set and event-related responses: Attentional modulation of V 5 responses attention to motion attention to colour

Non-linear parametric interactions Increasing word presentation rate in a subject and patient with a CVA Differential responses in the left hippocampus SPM{F}

Psychophysiological interactions Context X source target Set stimuli Modulation of stimulus-specific responses source target Context-sensitive connectivity source target

Psychophysiological interactions in the right inferotemporal region: Modulation of face-selective responses by PPC SPM{Z} Face stimuli PPC IT adjusted r. CBF Faces Objects medial parietal activity

Psychophysiological interactions: SPM{T} V 5 activity Attentional modulation of V 2 - V 5 contribution Attention V 2 V 5 activity time attention no attention V 2 activity