Experimental,Design,for,

Brain,Imaging,I,

Susan&Bookheimer&

fMRI,experimental,design:,A,basic,plan,

Define&mental&process&

to&examine&

Define&tasks&to&manipulate&

that&process&

Measure&fMRI&data&

during&tasks&

Compare&fMRI&data&

between&tasks&

Replace&�fMRI&data�&with&�RT�&

and&you&have&cogniBve&

psychology!&

Conceptual,and,methodological,

aspects,of,experimental,design,

•  There&are&two&aspects&of&fMRI&design&that&are&important&to&disBnguish&

• Conceptual&design&•  How&do&we&design&tasks&to&properly&measure&the&processes&of&interest?&

•  The&issues&here&are&very&similar&to&those&in&cogniBve&psychology&

• Methodological&design&•  How&can&we&construct&a&task&paradigm&to&opBmize&our&ability&to&measure&the&effects&of&interest,&within&the&specific&constraints&of&the&fMRI&scanning&environment?&

IV’s,and,contrasts:,basics,•  There&are&(almost&always)&two&or&more&condiBons&in&acBvaBon&imaging&

• We&make&a&series&of&assumpBons&about&the&cogni&ve&and&the&neural&processes&involved,&and&their&relaBon&to&each&other,&in&every&experiment&

•  The&logic&involved&and&choosing&tasks&and&contrasBng&them,&and&the&problems&of&assumpBons&in&these&choices,&spans&all&experimental&designs&

•  In&this&context,&makes&no&difference&whether&we&use&event&related&or&blocked&designs,&eg.&&�Null�&events&in&ER&designs&oOen&=&�rest�&in&block&designs&

•  Challenge:&Know&your&assumpBons;&choose&the&design&that&best&answers&your&quesBon,&minimizes&your&assumpBons,&and&allows&you&to&test&your&assumpBons&

General,Design,approaches,

•  SubtracBonUbased&designs&•  Factorial&Designs&• Parametric&Designs&

•  SelecBve&AYenBon&Designs&• AdaptaBon&Designs&• ConjuncBon&approaches&• Weird&stuff&

General,Design,approaches,

• SubtracBonUbased&designs&• MulBUlevel&hierarchical&subtracBon&

•  Simple&subtracBon&and&direcBonality&

• Common&baselines&and&parallel&comparisons&

•  Tailored&baselines&

Experimental,Example:,

semantic,processing,,

Experimental Question: Language&sBmuli&can&enter&the&brain&through&various&modaliBes:&auditory,&printed&word,&meaningful&pictures.& What areas of the brain are essential for language (semantic) processing, independent of input modality, and unrelated to other factors such as motor output? What specific processes take place during language performance? -

&“HOUSE”&HOUSE

The,subtraction,method,

•  Acquire&data&under&two&condiBons&

•  These-condi&ons-puta&vely-differ-only-in-the-cogni&ve-process-of-interest-

•  Compare&brain&images&

acquired&during&those&

condiBons&

•  Regions-of-difference-reflect-ac&va&on-due-to-the-�subtracted�-process-of-interest- Petersen et al., 1988

Hierarchical,subtraction,

example'from'Petersen,'1991,

•  Rest Control&

• Auditory&words&or&Visual&Word&(passive&presentaBon)&

• Reading&or&repeaBng&words&vs&passive&words:&motor&areas&&

• GeneraBng&words&vs.&repeaBng:&semanBc&(language)&areas&

- }

Semantic

Motor

Sensory

Design: Subtract out sensory and motor aspects of language from higher order semantic tasks using hierarchically based subtractions

- } - }

The,pure,insertion,assumption,

•  SubtracBon&requires&a&strong&assumpBon&of&

�pure&inserBon�&

•  Inser&on(of(a(single(cogni&ve(process(does(not(affect(any(of(the(other(processes((no(interac&ons)(

• �Controlled�&variables&do&not&change&as&you&add&processes&(move&“up”&the&hierarchy)&•  Eg,&Neural&processes&involved&in&reading&words&out&loud&are&idenBcal&to&those&in&reading&silently,&plus&motor&

•  &PI&must&hold&at&both&neural&and&cogniBve&levels&

Pure,insertion,in,a,hierarchical,model,

S& M&

“rest”& Sensory:&View&

words&

Motor:&Read&

words&

SemanBc:&generate&

words&

L&

In&hierarchical&subtracBon&models,&you&pile&up&the&pure&inserBon&assumpBons&with&each&addiBonal&

hierarchical&level!&&

Failure,of,pure,insertion,in,a,hierarchical,

model,

S&V1&

M&M1&

“rest”&Sensory:&View&

words&

Motor:&Read&

words&

SemanBc:&generate&

words&

L&

=&M&

M&

MotorUsensory&

U1& 1&

Motor:&Read&

words&

=&U1& 1&L&

M&SemanBcU&Motor&

•  One&level&of&hierarchy&&&•  Test&for&violaBon&of&addiBvity&assumpBon&

•  Allows&you&to&see&common&areas&acBve&for&A&and&B&

•  Assumes&A&and&B&have&similar&psychometric&properBes&

(ie,&level&of&difficulty,&variaBon,&and&distribuBon&in&the&

populaBon)&

•  Need&addiBonal&approach&to&see&unique&areas&

Ex B Ex A

Control

Common Baseline Is(there(evidence(for(pure(inser&on:((Does(reading(words(aloud(add(motor(areas(without(changing(silent(reading(areas?((

Oral Reading

Silent reading

Rest

&&&&&Read&&

�HOUSE�&

Name&

Simple,subtraction,assumptions,

• Make&assumpBons&about&&• What&your&tasks&are&doingU&do&they&tap&into&the&processes&of&interest&

•  How&they&differ&(what&variables&are&shared,&what&are&unique)&

• OOen&assume&(mistakenly)&that&differences&

are&due&to&increases&in&one&condiBonU&that&

which&is&the&�higher&order�&task&or&the&

experiment&(vs.&control)&task&From Morcom and Fletcher, NeuroImage, 2006

•  EG:&Seeing&words&vs.&hearing&words&•  Alone,&see&no&common&areas&

•  Good&adjunct&to&common&baseline&

•  Use&common&baseline&as&mask&to&reduce&errors&and&

increase&power&in&likely&areas&

•  Assumes-similar-psychometric-proper&es-of-A-and-B-

Ex B Ex A

Ex A Ex B >

>

Assuming,no,hierarchy:,Parallel,Comparisons,,

,

See Words

Speak Words

Speak Words

See Words

>

>

Word-order-change-

Word-change-

DapreYo&et&al:&SemanBc&vs.&SyntacBc&Processing&

The&city&is&west&of&the&school&

&West&of&the&school&

is&the&city&

The&city&is&west&of&the&school&

&The&town&is&west&of&

the&school&

Common&Baseline&comparison&

DapreYo&and&Bookheimer,&Neuron,&1999&

47&45&

DapreYo&et&al:&SemanBc&vs.&SyntacBc&Processing&

Parallel&comparison&

Common,baseline,and,

Parallel,comparisons,

Sem& Syn&Other&

lang,&task&

“rest”& SemanBc&decision&

SyntacBc&decision&

Direct&comparisons&

Sem&

Other&lang,&task&

Syn&

But,one,task,could,be,harder…..,

Sem&Syn&

Other&lang,&task&

“rest”& SemanBc&decision&

vs.&rest&

SyntacBc&decision&vs.&

rest&

Direct&comparisons&

Sem&Other&lang,&task&

Syn&

UUUAssumes&similar&psychometric&properBes&of&tasks&

Other&lang,&task&

•  Example&study&(ThomsponUSchill,&PNAS&1997):&Do&frontal&areas&implicated&in&semanBc&processing&really&involve&semanBcs,&or&are&they&instead&important&for&response&selecBon&(independent&of&task)&

• Use&3&different&tasks:&generaBon,&classificaBon,&and&comparison;&each&has&its&own&control,&each&as&different&levels&of&selecBon&demand&

Tailored,baseline,

Do(visual(language(and(auditory(language(s&muli(share(seman&c(representa&ons?(U  Want(to(look(only(at(seman&c(areas,(not(sensory(areasA(

tailor(control(tasks(for(each(experimental(tasks,(each(control(task((matched(for(unwanted(variables(

ThompsonUSchill&et&al&PNAS&1997&

Tailored,baseline,

• Assumes&baseline&tasks&control&for&(sensory&input)&equally&

•  &&reading&a&word&controls&for&generaBng&an&associate&to&a&word,&as&well&as&matching&a&word&controls&&for&matching&a&property&&

Tailored,baseline,

•  Assumes&similar&psychometric&properBes&of&both&experimental&and&both&control&tasks&

•  The&distribuBon&of&behavior&and&its&neural&responses&(magnitude,&RT,&sd,&etc)&for&high&selecBon&generaBon&compared&to&reading&:&is&the&same&as&that&for&high&selecBon&classificaBon&compared&to&the&control&

•  Need&very&extensive&behavioral&tesBng&&

Factorial,design,

•  A&factorial&design&involves&mulBple&concurrent&subtracBons&

•  Allows&for&tesBng&of&interacBons&between&components&

•  SBll&requires&pure&inserBon&assumpBon&and&task&decomposiBon&

•  But&addiBvity&can&be&tested&for&the&specific&factors&that&are&manipulated&

Directed,Attention,Models,

•  All&sBmuli&idenBcal&in&all&condiBons&

•  Direct&aYenBon&towards&different&features&•  Implicit&or&explicit&

•  Assumes&process&is&modified&by&directed&aYenBon&

•  Assumes&passive&processing&does&not&capture&your&variable&of&interest&&

EG,Corbetta,et,al,

Selective,attention,

•  In&every&condiBon,&all&three&variables&change&•  Told&to&respond&to&a&shape,&color&or&movement&

change&in&different&blocks&

•  SelecBvely&acBvates&form,&color,&moBon&centers&

•  Assumes&your&process&of&interest&is&modified&by&

directed&aYenBon&(oOen&true,&not&always)&

•  Can&be&done&explicitly&or&implicitly&(You&can&change&a&

factor&that&does&not&change&instrucBons)&

•  Read&words;&some&may&be&high,&others&low&imagery&

Parametric,designs,

•  Employs&conBnuous&variaBon&in&a&sBmulus/task&parameter&•  E.g.,&working&memory&load,&sBmulus&contrast&

•  Inference:&•  ModulaBon&of&acBvity&reflects&sensiBvity&to&the&modulated&parameter&

Boynton,et,al.,,1996, Assumptions,of,parametric,designs,

•  Pros:&you&don’t&have&to&design&a&control&conidBonU&no&subtracBon&

•  AssumpBon&of&pure&modulaBon&•  Each&level&of&the&task&differs&quanBtaBvely&in&the&level&of&engagement&of&the&process&of&interest,&rather&than&qualitaBvely&

•  Assumes&you&can&define&the&magnitude&differences&across&levels&(usually&assumes&equality,&but&not&necessarily&

•  Failures:&•  Response&is&a&step&funcBon&•  There&are&different&processes&engaged&at&different&levels&

Cohen,et,al.,,1996,

0&and&1Uback&do&not&require&WM;&2&does;&step&funcBon&with&a&qualitaBve&change&in&process&

Priming/adaptation,designs,

• PresentaBon&of&an&item&mulBple&Bmes&leads&to&

changes&in&acBvity&•  Usually&decreased&acBvity&upon&repeBBon&

•  Inference:&•  Regions&showing&decreased&acBvity&are&sensiBve&to&(i.e.&represent)&whatever&sBmulus&features&were&repeated&

• Requires&version&of&pure&modulaBon&assumpBon&•  Assumes&that&processing&of&specific&features&is&reduced&but&that&

the&task&is&otherwise&qualitaBvely&the&same&

Can,adaptation,fMRI,characterize,

,neural,representations?,,

,

•  A&voxel&containing&neurons&that&respond&to&all&poliBcians,&irrespecBve&of&party&

•  A&voxel&containing&some&specifically&DemocraBc&

neurons,&and&other&specifically&Republican&

neurons.&&

Two&sBmuli:&can&neurons&tell&the&

difference?&

From&R.&Raizada&

Responses,to,individual,stimuli,

do,not,show,whether,neurons,can,tell,the,

difference,

•  Different&sets&of&neurons&are&acBve&

within&the&voxel,&

but&overall&fMRI&

responses&are&

indisBnguishable&&

From&R.&Raizada&

Neural,adaptation,to,repeated,stimuli,does,show,the,difference:,

What,counts,as,repetition,for,neurons,in,a,voxel?,

It�s a politician" Same neurons, adapting:"It�s a politician again"

It�s a Republican" Different, fresh neurons: It�s a Democrat"

From&R.&Raizada&

Adaptation,in,bilingual,subjects,

,Do(different(language(share(seman&c(representa&ons(across(languages(in(bilingual(subjects?((Chee(et(al(

Chee,et,al,2003,

Main&effect&for&meaning&(adaptaBon)&in&LIFG,&not&LOcc&

Conjunction,analysis,(Price,&,Friston,,1997),

•  Perform&several&parallel&subtracBons&

•  Each&of&which&isolates&only&the&process&of&interest&•  Find&regions&that&show&common&acBvaBon&across&all&of&these&

Ex A - Ctl Ex B - Ctl Ex C - Ctl

A AND B AND C

from&Price&&&Friston,&1997&

BTLA],all,tasks,involving,accessing,phonology,

Problems,with,conjunction,analysis,,

(Caplan,&,Moo,,2003),

• Many&assumpBons&about&what&processes&are&involved&

•  Implicit&processing&•  Subjects&may&engage&processes&that&are&not&necessary&for&the&taskU&does&not&measure&magnitude&differences&

•  InteracBons&between&processing&stages&•  ConjuncBon&only&gets&rid&of&interacBons&if&they&do&not&acBvate&the&same&regions&to&the&same&degree&across&tasks&

• We&use&this&approach&for&finding&consistent&but&lowUlevel&acBvaBons&in&clinical&mapping&&

&

Factor]determined,component,classi^ication:,

Badre,,Poldrack,et,al,2005,

IFG,dissociations,

Badre,&Poldrack&etc&2005&

2]group,designs,

•  Build&on&any&of&the&prior&designs&•  AddiBonal&between&group&comparisons&

•  Hypothesis&sounds&something&like:&

•  “&the(differences(between(experimental(and(control(task(in(my(pa&ent(group(differs(from(that(difference(in(controls”(

•  Assumes&baseline&task&performance&is&equal&

• What&if&one&group&cannot&perform&the&task&well?&

•  Eg:&dyslexia:&give&a&nonUword&reading&task&

Counterbalancing,

• With&more&than&2&condiBonsU&essenBal&

•  EG:&Low,&medium&and&high&stress&condiBons&•  HabituaBon&•  Order&effects&&eg&High&carryUover&

• Complete&counterbalancing&(recruit&in&groups&of&N!&where&N&is&the&total&number&of&condiBons)&•  1&2&3&&&&132&&&&231&&&&213&&&312&&&&312&

•  LaBn&Square&(recruit&in&groups&of&N&condiBons)&•  123&&231&&312&•  Each&condiBon&in&each&serial&order&•  &assumes-no-taskDtask-order-interac&ons-

Summary,

•  No&design&is&perfect&•  &Use&that&which&is&most&consistent&with&your&specific&

research&quesBon&

• MulBple&�baseline�&condiBons&help&interpretaBon&

•  Behavioral&data&is&extremely&helpful&in&supporBng&

many&assumpBons&and&prevenBng&criBcal&errors&

•  Beware&of&your&assumpBons!&all&designs&make&

assumpBons&that&are&not&fully&verifiable;&know&

them!&

•  Freely&admit&your&design&limitaBons&