Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Stopping and response control : application of a novel tracking task Morein-Zamir, Sharon 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2006-200544.pdf [ 7.43MB ]
Metadata
JSON: 831-1.0092957.json
JSON-LD: 831-1.0092957-ld.json
RDF/XML (Pretty): 831-1.0092957-rdf.xml
RDF/JSON: 831-1.0092957-rdf.json
Turtle: 831-1.0092957-turtle.txt
N-Triples: 831-1.0092957-rdf-ntriples.txt
Original Record: 831-1.0092957-source.json
Full Text
831-1.0092957-fulltext.txt
Citation
831-1.0092957.ris

Full Text

S T O P P I N G A N D R E S P O N S E C O N T R O L : A P P L I C A T I O N O F A N O V E L T R A C K I N G T A S K by S H A R O N M O R E I N - Z A M I R B A . , B e n - G u r i o n Un ivers i t y , 1997 M . A . , The Un i ve rs i t y o f B r i t i s h C o l u m b i a , 2002 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E M T S F O R T H E D E G R E E O F D O C T O R O F P H I L O S O P H Y in T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Psychology) T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A June 2006 © Sharon Morein-Zamir, 2006 "3 Abstract Stopping a p lanned or ongo ing act ion is one o f the central methods for e x a m i n i n g response control and inh ib i t ion . Th is research area has been dominated b y the appl icat ion o f the "s top -s ignal p a r a d i g m " w h i c h measures the inh ib i t ion o f a planned act ion (Chapter 1). Chapter 2 describes two experiments on the development and in i t ia l testing o f m y nove l t rack ing task w h i c h measures the inh ib i t ion o f an ongoing act ion. O v e r the next three chapters this t rack ing task is used to examine three fundamental assumptions o f stopping research. The first assumption is that a l l forms o f stopping are mediated by a c o m m o n m e c h a n i s m . The invest igat ion reported in Chapter 3 supports this idea as the c lassic stop-s ignal task and the nove l t racking task are shown to tap into the same inhib i tory mechanism. The second assumption is that stopping is governed by different constraints than those govern ing response in i t iat ion. Chapter 4 reports two experiments that argue against this pos i t ion , as stopping and response in i t iat ion are found to be in f luenced s imi la r l y by manipulat ions to st imulus-response compat ib i l i t y . The third assumption is that stopping generalizes to other measures o f response modi f i ca t ion . Chapter 5 reports three experiments that support this th i rd assumpt ion, w i t h expectancy effects h a v i n g a comparable inf luence on response inh ib i t ion and accelerat ion. These data also prov ide converg ing evidence for a rejection o f the second assumption that stopping is unique. Chapter 6 concludes w i t h a summary and integration o f the m a i n findings, an explorat ion o f the pract ical and theoretical impl icat ions o f the work , and f ina l l y , a considerat ion o f some outstanding questions and future direct ions in the f ie ld o f response control and inh ib i t ion . Ill Table o f Contents Abstract i i Table o f Contents i i i L i s t o f Tables v L i s t o f F igures v i A c k n o w l e d g e m e n t s v i i C o - A u t h o r s h i p Statement v i i i C H A P T E R 1 : An Introduction to Response Control and Stopping 1 What is Stopping? 1 A n a l y s i s o f the Stopp ing Literature T o Date 7 Genera l O v e r v i e w o f Studies 12 References 16 C H A P T E R 2: Measuring Online Volitional Response Control with a Continuous Tracking Task 25 Current M e t h o d o l o g y 27 Exp loratory D a t a C o l l e c t i o n 37 Exper iment 1 37 M e t h o d s 37 Resul ts and D i s c u s s i o n 38 Exper iment 2 41 M e t h o d s 42 Results and D i s c u s s i o n 44 Conc lus ions 45 References 48 C H A P T E R 3: Inhibiting prepared verus ongoing responses 51 Methods 54 Results and D i s c u s s i o n 57 Conc lus ions 64 References 66 C H A P T E R 4: Compatibility effects in stopping and response initiation in the continuous tracking task 71 Exper iment 1 76 M e t h o d s 78 Results 82 D i s c u s s i o n 84 Exper iment 2 86 M e t h o d s 86 Results 87 D i s c u s s i o n 88 C o m p a r i s o n between Exper iment 1 and Exper iment 2 88 i v Genera l D i s c u s s i o n 89 References 95 CHAPTER 5: Comparing stopping with another measure of response control: the role of predictability 103 Exper iment 1 109 M e t h o d s 112 Resul ts 118 D i s c u s s i o n 123 Exper iment 2 126 M e t h o d s 127 Resul ts 128 D i s c u s s i o n 131 Exper iment 3 133 M e t h o d s 133 Resul ts -. 134 D i s c u s s i o n 137 Genera l D i s c u s s i o n 137 References 147 CHAPTER 6: General Discussion 155 S u m m a r y 155 Conc lus ions and Impl icta ions o f the Present Results 158 Outstanding Quest ions and Future Di rect ions 163 C l o s i n g Remards 168 References 169 APPENDIX: Copy of UBC Research Ethics Board's Certificate of Approval 173 V L i s t o f Tables Tab le 2 . 1 . Descr ip t i ve statistics for ind iv idua l participants i n Exper iment 1 38 Table 2 .2 . Corre lat ions between dependent measures in Exper iment 1 40 Table 3 . 1 . Corre lat ions between the dependent measures 60 Table 3.2. Resul ts o f mean stop s ignal react ion t ime ( S S R T ) 62 Tab le 4 . 1 . Resul ts o f Exper iments 1 and 2 84 Tab le 5 .1 . Results o f Exper iments 1, 2 and 3 120 VI L i s t o f F igures F igure 1.1. F requency o f publ icat ions over the per iod o f 1960 to 2005 4 F igure 2 .1 , A depict ion o f the typ ica l series o f events in the t rack ing task 31 F igure 2.2. A force over t ime prof i le y ie lded over the course o f a tr ial 32 F igure 2 .3 . Stop s ignal react ion t ime data in Exper iment 1 39 F igure 3 . 1 . The sequence o f events in a stop tr ial dur ing the stop s ignal task 56 F igure 3 .2 . A response force prof i le o f a stop tr ial in the t rack ing task 58 F igure 4 . 1 . A depict ion o f the procedure o f Exper iment 1 80 F igure 5 .1 . A portrayal o f the sequence o f events in a tr ial in Exper iment 1 114 F igure 5.2. I l lustrat ion o f the force prof i le o f two prototypical tr ial 117 F igure 5.3. The results o f Exper iment 1 123 F igure 5.4. The results o f Exper iment 2 130 F igure 5.5. The results o f Exper iment 3 135 A c k n o w l e d g e m e n t s I a m indebted to m a n y people whose support made it possib le , or at least m u c h easier, for me to conduct the research and wr i t ing invo lved i n this thesis. F i rs t , I w o u l d l ike to thank m y supervisor , A l a n K i n g s t o n e , w h o a l l owed me as m u c h f reedom as I wanted and as much support as I needed. I am also grateful to many faculty members for be ing m y col laborators and guides: to R o m e o C h u a and Ian Franks f r o m the S c h o o l o f H u m a n K i n e t i c s , and to L a w r e n c e W a r d , Charlotte Johnston, and T o d d H a n d y f r o m the P s y c h o l o g y Department. L i k e w i s e , I thank the past and present members o f the B r a i n and At tent ion Research lab for be ing great col leagues, and in part icular to the people who helped me navigate through these and other research projects: Stephanie H o , Carey H u h , C h r i s t a - L y n n D o n o v a n , M e e n a Jassal , K e n L i , Jennifer Q u a n , R a c h e l l e S m i t h , Salvador Soto -Faraco , M a r i a Ste l la , and Chr ist ine Y e u n g . In addi t ion , I w i s h to acknowledge the f inanc ia l support I received f r o m the Un i ve rs i t y o f B r i t i s h C o l u m b i a i n c l u d i n g the L i T z e F o n g F e l l o w s h i p , the Pat r i ck D a v i d C a m p b e l l F e l l o w s h i p , the Wal te r C . K o e r n e r F e l l o w s h i p and the Un i ve rs i t y Graduate F e l l o w s h i p s . A very special and personal thank -you to P a u l N a g e l k e r k e w h o prov ided so m u c h support in so many ways , and who came through on the most out landish requests w i t h on ly a bit o f head shaking and many a c o m i c a l remark. F i n a l l y , I w o u l d l ike to express m y gratitude to m y partner M i c h a e l and the rest o f m y f a m i l y and fr iends, for their love , convent ional and unconvent ional support, and their incredible reserves o f patience. V l l l Co-authorsh ip Statement I a m the pr imary author on a l l the P h D w o r k presented in this thesis. The w o r k grew f r o m discussions and col laborat ions w i th the f o l l o w i n g ind iv idua ls : D r . A l a n K ings tone , Dr . R o m e o C h u a , D r . Ian Franks and P a u l Nage lkerke . 1 C H A P T E R 1: A n Introduction to Response Cont ro l and Stopp ing What is Stopping? Cont ro l o f behav ior is a h igher -order cogni t ive funct ion essential for self -regulat ion, that is , a person must be able to purposeful ly direct and execute actions to real ize her or his intended goals. One class o f intentional enactments, p r o v i d i n g a means to measure and investigate se l f regulat ion and control is act ion suppression, also k n o w n as act ion inh ib i t ion and control . A c t i o n inh ib i t ion and control purports to uncover h o w ind iv idua ls suppress or alter their actions and behaviors w h e n those are no longer appropriate. Suppress ing no- longer -appropr iate actions is c ruc ia l in everyday behavior and defic ient in numerous populat ions such as the very y o u n g or o l d , and those w i th disorders such as attention def ic i t hyperact iv i ty disorder ( A D H D ) and schizophrenia . One o f the simplest and most direct ways to investigate vo l i t iona l act ion suppression is by e x a m i n i n g h o w people intent ional ly stop an act ion in response to an external event. W h e n cons ider ing stopping, it is quite easy to think o f everyday instances that exempl i f y its behav iora l relevance: stopping on the p l a y i n g field when the referee b lows the whist le , or stopping in m i d - a c t i o n w h e n the mus ic conductor prov ides an external stop s ignal . E v e n a mundane occurrence such as stopping when w a l k i n g towards a busy road, exempl i f ies the funct ional importance o f stopping. W h e n invest igat ing stopping in the contro l led c i rcumstances o f the laboratory, one parad igm has been part icular ly dominant : the stop s ignal task. The stop s ignal , or countermanding, task has been developed spec i f i ca l l y to examine response inh ib i t ion , i.e., the abi l i ty to intent ional ly control or stop one's actions ( Logan , 1994; L o g a n & 2 C o w a n , 1984; see also L a p p i n & E r i k s e n 1966; O i l m a n , 1973). The procedure o f the stop s ignal task requires participants to per fo rm a s imple or choice react ion t ime ( R T ) task i n response to a v i sua l st imulus (go signal) . F o r example , they m a y be asked to press a right key in response to an ' x ' and a left key in response to an ' o ' and to press as fast as possib le . O n a subset o f trials a subsequent auditory or v i sua l st imulus (stop signal) is presented f o l l o w i n g the go s ignal , ind icat ing that the or ig ina l response is to be cancel led or countermanded. That is , participants are to stop and w i t h h o l d the p lanned act ion and not press the response key . Importantly , the interval or stop s ignal delay, between the go s ignal and the stop s ignal is var ied. W h e n the stop s ignal appears immedia te l y after the go s ignal , participants can general ly stop their response (s ignal - inh ib i t trials). H o w e v e r , when the delay is longer, the l i k e l i h o o d o f car ry ing out a response increases despite the presentation o f the stop s ignal (s ignal - respond tr ials) , as part icipants cannot refrain f r o m execut ing the p lanned act ion. The inh ib i t ion funct ion denotes the proport ion o f s igna l -respond trials as a funct ion o f stop s ignal delay. U s i n g the distr ibut ion o f go R T s together w i t h the inh ib i t ion funct ion and a mathematical mode l that assumes a race between stop and go processes, researchers estimate the unobservable latency to the stop s ignal , or S S R T (stop s ignal react ion t ime) ( B a n d , van der M o l e n & L o g a n , 2 0 0 3 ; L o g a n , C o w a n & D a v i s , 1984). H e n c e , the task and mode l offer an estimate o f the latency to stop a p lanned, unexecuted act ion. The stop s ignal task is also accompanied by an expl ic i t theoretical m o d e l o f response control ( L o g a n & C o w a n , 1984). A c t s o f contro l are def ined as the interactions in a h ierarchica l scheme between an executive or supervisory system that forms intentions and commands , and subordinate systems that interpret the commands and 3 execute them. S ignals to stop or m o d i f y on ly some parameters o f an act ion have p r i v i leged access to the subordinate systems. In contrast, signals that require act ion in i t iat ion or a suf f ic ient ly different response, necessitate s igni f icant process ing and do not have pr iv i leged access to the system. Th is d ist inct ion is consistent w i th models o f control that assume an executive system manag ing subordinate systems to real ize intended actions in a hierarchal manner (e.g., N o r m a n & Sha l l i ce , 1986). The stop process can inhibi t lower leve l processes direct ly , or m o d i f y parameters passed f r o m h i g h - l e v e l processes to l o w - l e v e l ones. Hence , stopping can operate outside the hierarchy o f higher and lower level processes. The theory posits that the stop s ignal task can measure the d i f f i cu l ty and latency o f an act o f control . M o r e o v e r , the stop s ignal task a l lows for the study o f both act ion and cognit ive control w i th in the same parad igm. The stop s ignal task and accompany ing f ramework offer several characteristics that make them especia l ly appeal ing to study stopping in part icular and act ion control and suppression i n general . M o s t important ly , the parad igm prov ides an exp l i c i t latency for the stopping process a l l o w i n g for a measure o f reaction t ime ( R T ) c o m m o n l y used in cogni t ive psycho logy . M o r e o v e r , dissociat ions between the stop and go processes can be invest igated direct ly w i t h i n the same task. In addi t ion, the stop s ignal parad igm is fa i r ly easy to administer and demonstrates considerable adaptabi l i ty to different settings and populat ions (see be low) . L i k e w i s e , the task is be l ieved to isolate response control and inh ib i t ion ef fect ive ly , us ing a clear instance o f act ion suppression. In v i e w o f these characterist ics, it is not surpr is ing that stopping us ing the stop s ignal has been studied extensively by psychologists to many different ends and is cont inuously ga in ing in popular i ty (see F igure 1.1). F i rst and foremost, stopping has been 4 used as a means to investigate response control or how one controls one 's actions. In this capacity s topping is considered a general index o f response contro l . F o r example , the stop s ignal task has been used to explore the point o f no return in movement contro l (e.g., D e Jong , C o l e s , L o g a n & Gratton, 1990; M c G a r r y & Franks , 1997; O s m a n , K o m b l u m & M e y e r , 1986). L i k e w i s e , because o f its generality, it can be used to investigate many product ion systems i n c l u d i n g manual responses (e.g., L o g a n , 1981) , saccadic or smooth pursuit eye movement contro l (e.g., Hanes & S c h a l l , 1995; K o r n y l o , K i l l , Saenz & K r a u s l i s , 2003) as w e l l as speech product ion and control (e.g. L a d e f o g e d , Si lverstein & P a p c u n , 1973). F i n a l l y , as noted above, stopping has led to a fo rmal i zed theory and a m o d e l o f response control (see be low ; L o g a n , 1994; L o g a n & C o w a n , 1984). Figure 1.1. Frequency of publications over the period of 1960 to 2005, using or reviewing the stop signal task. The references were identified by a literature search using PsychlNFO and Pubmed MedLine, as well as by examining references in relevant published papers (Morein-Zamir & Kingstone, in preparation). 70 r — . . . . . . . - . . . 1 60 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 Year 5 S topp ing , or response inh ib i t ion , is also considered to be one o f the executive funct ions, w h i c h are a f a m i l y o f cogni t ive functions ( inc lud ing also w o r k i n g memory , set shi f t ing and planning) be l ieved to be mediated p r imar i l y by the frontal lobes (e.g., A r o n , R o b b i n s & P o l d r a c k , 2 0 0 4 ; Burgess , 1997; P h i l l i p s , 1997). In accordance w i t h this f ramework , s topping in the stop s ignal task has been e m p l o y e d to ident i fy areas and processes in the brain speci f ic to response inh ib i t ion , u s i n g techniques as var ied as single ce l l record ing as w e l l as event related potentials and funct ional magnet ic resonance i m a g i n g in humans ( E R P and f M R I , respect ively ; B a n d & van B o x t e l , 1999; D e Jong et a l . , 1990, 1995; R u b i a , R u s s e l l , Overmeyer , B r a m m e r , B u l l m o r e , Sharma, S i m m o n s , W i l l i a m s et a l . , 2 0 0 1 ; S c h a l l , Hanes & Tay lo r , 2000) . L i k e w i s e , s topping has been used to measure response inh ib i t ion as part o f test batteries a imed at gauging a broad array o f cogni t ive abi l i t ies (e.g., Turner, D o w s o n , R o b b i n s & Sahakian , 2004) . Invest igating stopping also provides a means to detect impa i red response contro l , as w e l l as to measure funct ional i ty w i t h i n and across populat ions. It has been appl ied to measure stopping across the l i fe span (Bedard , N i c h o l s , B a r b o s a , Schachar , L o g a n & Tannock , 2 0 0 2 ; K o k , 1999; K r a m e r , H u m p h r e y , L a r i s h , L o g a n & Strayer, 1994; W i l l i a m s , Ponesse, Schachar , L o g a n & Tannock , 1999) and i n patients w i t h Park inson Disease, conduct disorder, anxiety disorder, aut ism and patients w i t h damage to the frontal lobes (e.g., G a u g g e l , R ieger & Feghoff , 2 0 0 4 ; N i g g , 2 0 0 1 ; Ooster laan, L o g a n & Sergeant, 1998; R ieger & G a u g g e l , 1999). In addi t ion, it has been adapted to rat (Eagle & R o b b i n s , 2003) and pr imate (Hanes, Patterson & S c h a l l , 1998) an imal models . F i n a l l y , s topping research has been conducted to assess vo l i t iona l se l f - regulat ion, be l ieved to be defic ient in disorders such as schizophrenia and A D H D (e.g., B a d c o c k , 6 M i c h i e , Johnson & C o m b r i n c k , 2 0 0 2 ; Ooster laan et a l . , 1998). In fact, the consistent f ind ing o f impai red stopping in A D H D across many studies has led to the dominance o f theories pos i t ing impai red response inh ib i t ion as the p r imary def ic i t in this d isorder (e.g., B a r k l e y , 1997; N i g g , 2001) . M o r e o v e r , stopping has been used to measure impai red self-regulat ion as a result o f the administrat ion o f fore ign substances ranging f r o m drugs, such as cocaine and a l c o h o l , to medicat ions, such as methylphenidate and m o d a f i n i l (e.g., F i l l m o r e , R u s h , K e l l y & H a y s , 2 0 0 1 ; M u l v i h i l l , S k i l l i n g & V o g e l - S p r o t t , 1997; Turner, C l a r k , P o m a r o l - C l o t e t , M c K e n n a , Robb ins & Sahakian , 2004) . In sum, the stop signal parad igm has proven to be an effect ive tool in addressing a mult i tude o f questions in diverse areas o f research i n c l u d i n g act ion and cognit ive control . It has offered r ich in format ion and has p layed an instrumental role in the research o f response inh ib i t ion (for rev iews see L o g a n , 1994; L o g a n & C o w a n , 1984; Ooster laan, et a l . , 1998). A t present, current theor iz ing us ing the stop s ignal task has suggested that stopping possesses three key attributes. F i rst , a single mechanism under l ies a var iety o f stopping behaviors ( B a n d & v o n B o x t e l , 1999; L o g a n & C o w a n , 1984, but see also D e Jong et a l . , 1995). A second characteristic o f stopping assumed by current theor iz ing is that stopping is implemented by processes that are distinct f r o m the processes governing the go response ( L o g a n & C o w a n , 1984). U n l i k e new responses, w h i c h require s igni f icant process ing , the interruption or mod i f i ca t ion o f a response appears to have p r i v i leged access to the response system ( L o g a n & C o w a n , 1984). E m p i r i c a l support for this not ion can be seen in that un l ike the go task, stopping appears to be insensit ive to s ignal predictabi l i ty ( L o g a n & B u r k e l l , 1986; Ramautar , K o k & R i d d e r i n k h o f , 2004) , as w e l l as to practice ( L o g a n & B u r k e l l , 1986; W i l l i a m s et a l . , 1999) and demonstrates unique developmental trends (Bedard et a l . , 2 0 0 2 ; R idder inkhof , B a n d & L o g a n , 1999; W i l l i a m s et a l . , 1999). In addi t ion , several dissociat ions between stopping and go ing are found in populat ions demonstrat ing defic ient stopping such as A D H D (Ooster laan et a l . , 1998). The th i rd attribute is that stopping is a generic instance o f act ion or response contro l , genera l i z ing to other cases o f response control . A c t i o n contro l is inherent in many o f our behaviors , but is usual ly subtle, often leading to a sl ight change in an ongo ing act ion. E x a m p l e s o f response control in the present context inc lude error correct ion, and the m o d i f i c a t i o n o f force or d i rect ion o f an ongo ing movement or other instances where an ex ist ing response is purposeful ly adjusted (e.g., D a y & L y o n , 2 0 0 0 ; H e n r y & Har r i son , 1961; M e g a w , 1972; P isse la , A r z i & Rossett i , 1998; V i n c e & W e l f o r d , 1967; for a rev iew see L o g a n & C o w a n , 1984). The advantage o f studying stopping over these other instances o f response adjustments has been that it is a c lear and extreme behavior that possesses an established protocol o f h o w it is to be measured (Logan , 1994). In sum, i m p l i c i t in many studies is the assumption that stopping engages, in some sense, a general response control mechan ism that is governed by different constraints than those regulat ing the go response (Logan , 1994; L o g a n & C o w a n , 1984; N i g g , 2001) . Th is v i e w , together w i t h the formal m o d e l , prov ides a theoretical f ramework w i t h i n w h i c h m u c h o f the literature on stopping presently exists. A n a l y s i s o f the Stopping Literature to Date Despi te the profound importance and usefulness o f the stop s ignal task it suffers f r o m numerous l imitat ions . F i rst , because stopping is an est imation o f a behav ior that is not d i rect ly observed ( in the case o f a successful inh ib i t ion , no key has been pressed), the 8 computat ion o f S S R T is heav i l y dependent on the mathematical race m o d e l and its assumptions. The most important assumption is the assumption for stochastic independence between the stop and go processes, (although see B a n d et a l . , 2003) as it a l lows for the computat ion o f S S R T . That is , on any g iven t r ia l , k n o w i n g the latency o f one w i l l not prov ide us w i t h in format ion about the latency o f the other. I f this were not the case, then the est imation o f S S R T w o u l d be ser iously c o m p r o m i s e d . In fact, m u c h research has been dedicated to va l idat ing the race m o d e l as a c o m p e l l i n g method to assess response inh ib i t ion (e.g., B a n d et a l . , 2 0 0 3 ; L o g a n , 1994; L o g a n et a l . , 1984). H o w e v e r , there are studies demonstrat ing that the under ly ing assumptions, such as the assumption for independence between the stop and go processes, do not a lways ho ld (e.g., C o l o n i u s , Ozy r t & A rnd t , 2 0 0 1 ; Ozyr t , C o l o n i u s & A r n d t , 2003) . In fact, even i f the race mode l holds true in the standard stop s ignal task, it appears to be compromised when more c o m p l e x tasks such as the stop-change or go -s top -go tasks are employed (Logan & B u r k e l l , 1986; M c G a r r y , C h u a & F ranks , 2003) . The stop-change task is ident ical to the stop-s ignal task but the s ignal indicates that the participant must stop and then engage in a different task (e.g., L o g a n & B u r k e l l , 1986). The go -s top -go task is also an extension o f the c lassic task, where f o l l o w i n g a stop s ignal a second s ignal m a y indicate that the or ig ina l go task is to be resumed ( M c G a r r y , C h u a & Franks , 2003) . Furthermore, at present, there is no conc lus ive method o f va l idat ing the m o d e l on real data since the key test o f compar ing observed stop-respond trials to predicted stop-respond trials generated f r o m the race m o d e l has been proven to be questionable ( B a n d et a l . , 2003) . The i m p l i c a t i o n o f these points when they are taken together is that the S S R T measure may not be v a l i d in c i rcumstances where the race model does not h o l d , i nc lud ing the instances 9 where more c o m p l e x behaviors are examined. M o r e o v e r , there is no establ ished w a y to test w h e n the mode l does and does not h o l d true. A n o t h e r l imi ta t ion o f the stop s ignal task is that no i n d i v i d u a l t r ia l measurements are avai lable , as the S S R T estimations can on ly be computed across a series o f tr ials. Thus , questions necessitating analyses that require measurements on every tr ial cannot be addressed ( M o r e i n - Z a m i r & M e i r a n , 2003) . These include the examinat ion o f sequential effects on stopping across tr ials, and measurements o f var iance (see M o r e i n - Z a m i r & M e i r a n , 2003 for an extensive discussion) . A l t h o u g h it was proposed that stopping var iab i l i ty cou ld be estimated f r o m the inh ib i t ion funct ions that relate proport ion o f unsuccessful inh ib i t ion to stop signal delay, recent evidence suggests that such methods are unrel iable at best ( B a n d , et a l , 2003) . Obta in ing a rel iable measurement o f stopping var iab i l i t y is part icu lar ly important as clear predict ions exist for increased stopping var iab i l i t y in disorders such as A D H D , and their support or negat ion w o u l d have clear theoretical impl icat ions ( N i g g , 2 0 0 1 ; Tannock , 2003) . H e n c e , al though S S R T is one o f the key strengths o f the stop s ignal task, the fact that it is an indirect est imation consist ing o f a single value is a weakness. A n o t h e r l imi tat ion o f the stop s ignal task, repeatedly stated in the literature, is that the task has often not been successful in invest igat ing the nature o f the under l y ing cogni t ive mechanisms o f stopping (e.g., B a n d et a l . , 2 0 0 3 ; L o g a n , 1994; van der W i l d e n b e r g , van der M o l e n & L o g a n , 2002) . T o address this concern , several researchers have examined the interaction between reactive inh ib i t ion (non- intended inh ib i t ion result ing f r o m the execut ion o f some cognit ive process, for example inh ib i t ion result ing f r o m the employment o f selective attention; L o g a n , 1994) and stopping (e.g., K r a m e r et 10 a l . , 1994; R i d d e r i n k h o f et a l . , 1999; van der W i l d e n b e r g et a l . , 2 0 0 2 ; Verb ruggen , L ie fooghe & Vand ie rendonck , 2004, 2005) . The reasoning assumes that s lowed stopping in tasks engaging reactive inh ib i t ion is evidence that both forms o f inh ib i t ion share a c o m m o n mechan ism. Nevertheless , this f o r m o f evidence suffers f r o m numerous drawbacks such as inconsistent results (e.g., van der W i l d e n b e r g et a l . , 2002) , and a h igh l y restricted range o f studies as to date on ly one study examined whether stopping inf luences reactive inh ib i t ion ( R i d d e r i n k h o f et a l . , 1999)). In addi t ion , there are plausible alternative explanations relat ing the observed results to changes in response force rather than to interactions between stopping and reactive inh ib i t ion (van den W i l d n e b e r g , van B o x t e l & van der M o l e n , 2003) . F i n a l l y , longer S S R T s in some condit ions may indeed be evidence for an interaction between the inh ib i t ion generated by the go task w i th the stop process, but it m a y also be evidence for a direct interaction between the stop and go processes suggesting that the theoretical reasoning under l y ing this l ine o f research m a y be f lawed ( M o r e i n - Z a m i r & K ings tone , in preparation). T a k e n together these problems suggest that attempts at c la r i f y ing the under ly ing cognit ive mechanisms o f stopping by e x a m i n i n g interactions between stopping and var ious forms o f cogni t ive inhib i t ions has resulted in on ly l im i ted success. One o f the m a i n points o f concern when e x a m i n i n g the stop s ignal litertuare is that although three key attributes o f stopping have been c lear ly stated in the literature, they have not been r igorously tested. Hence , they can be considered assumptions rather than establ ished characteristics o f stopping. A l t h o u g h the first assumpt ion exp l i c i t l y assumes that stopping as measured by the task is representative o f a l l forms o f stopping, such as stopping an act ion after response execut ion ( Logan , 1994; L o g a n & C o w a n , 11 1984), this has yet to be established. The stop signal task examines only one type of stopping, that of stopping an action before response execution. The stop signal task cannot gauge stopping a response during true execution as in this case the go process will have already won the race, thus it can only measure countermanding a response either at the beginning or in the middle of a series of discrete responses. The three studies to date examining stopping before or during a series of responses have yielded ambiguous results. A n early study indicated that it was harder to stop before compared to during a response (Ladefoged et al., 1973). In contrast, equivalent stopping performance was found before and in the middle of typewriting a word or sentence and it was concluded that they were stopped in much the same way (Logan, 1982). To complete the set of possible outcomes, Osman and colleagues (1990) found that it was harder to stop a sequence of key presses once it had begun. Evidence supporting the second assumption regarding the distinctiveness of stopping from go response measures is also unclear. On the one hand, evidence from domains such as the study of A D H D suggest that inhibiting responses is at the core of the deficit and therefore certainly distinct from other types of behaviors (Barkley, 1997; Oosterlaan et al., 1998). On the other hand, the relatively idiosyncratic nature of the stop signal task, and the findings of deficient go RTs in A D H D (Nigg, 2001; Tannock, 2003), would suggest that the status of stopping as distinct from going is currently unsettled. Furthermore, task idiosyncrasies together with the fact that while go RTs are observed, SSRTs are only estimated may account for some of the other dissociations observed in the literature, such as the resistance of SSRTs to signal predictability (Morein-Zamir & Kingstone, in preparation). 12 The third assumption regarding stopping being representative of other instances of response modification has also not been rigorously tested. There is indeed some evidence supporting the notion that stopping is a good representative as it demonstrates similar reaction times to some forms of response adjustment (e.g., Kudo & Ohtsuki, 1998; Pisella et al., 1998). However, there is also evidence challenging this notion, as differences can be found between various types of response adjustments, limiting the generalizability of any one measure (e.g., Day & Lyon, 2000; Vince & Welford, 1967). Furthermore, it has been explicitly acknowledged that the relationship between stopping and other, subtler forms of control might not be simple and straightfoiward (Logan, 1994). In the least, boundary conditions for the validity of this assumption should be established. Finally, as the research is based on only one basic kind of task, theorizing lacks the ability to rely on converging evidence from other tasks purporting to measure the same processes. Some of the richness observed in normal and abnormal stopping behaviors is lost when only a single instance of stopping is examined. On the same note, many questions, not easily addressed by the stop signal task, are either unasked or remain unanswered (e.g., does stopping before and after response execution employ similar mechanisms?). This in turn, has shaped the current views and theories of response inhibition. General Overview of Experiments The aim of the studies comprising this dissertation is to address some of the shortcomings discussed above. Specifically, an alternative task was developed and utilized to test the three key assumptions directly whilst providing converging evidence to 13 the stop signal task. The novel task required participants to stop an ongoing action after it was initiated. This task allowed for a direct observation of stopping a controlled action on each trial where a stop signal was presented. Somewhat similar reasoning has previously been used by Ladefoged and colleagues and Henry and Harrison (Henry & Harrison, 1963; Ladefoged et al., 1973). However, in those cases, response inhibition was not the primary area of interest (see also Logan, 1982). Moreover, due to technical limitations those prior studies did not provide useful paradigms with which to study stopping. Initially, Morein-Zamir and Meiran (2003) presented a continuous tracking task where the time to initiate a stopping of the tracking response was used as a primary measure of response inhibition. Participants tracked a moving visual target presented on the screen by controlling a response marker with the aid of the computer mouse. Hence, on each trial participants first engaged in tracking and at some later point responded to a signal to stop their action. The time from signal onset to the onset of deceleration of the tracking movement was taken to be a measurement of stopping initiation. One advantage of this task is that it is not dependent on the race model or its assumptions. Likewise, it provides stopping reaction times on each trial, leading to a wealth of information about motor control behaviors and allowing for analyses previously not viable with the stop signal task (Morein-Zamir & Meiran, 2003). This study demonstrated the usefulness of an alternative approach by testing and supporting the independence assumption so central to the stop signal literature. The studies in the present dissertation aimed to develop and utilize an improved tracking task. The thesis follows a manuscript-based format, with the following four chapters presenting four independent studies. The purpose of these studies is to a) provide 14 converging evidence for earlier results that utilized the stop signal task, b) critically examine the key assumptions in ways previously not viable and c) suggest new avenues of research that could, in turn, utilize the strengths and weaknesses o f both tasks to gain a deeper understanding of stopping in particular, and response control in general. To this end, Chapter 2 describes the development and initial testing of the present tracking task. Chapter 3 introduces a study testing the first assumption that stopping as measured in the stop signal indeed generalizes to other forms of stopping (Logan, 1994; Logan & Cowan, 1984). Specifically, the study examined whether stopping before response execution (as is the case in the stop signal task) and stopping after response execution (as is the case in the tracking task) tap into similar mechanisms. If different stopping behaviors engage different mechanisms, then stopping is not unitary, negating the first key assumption. Moreover, i f this turns out to be the case, stopping and the stop signal task in particular, could not be used as a general index of response control. This would severely limit the scope of the stop signal literature and the underlying theoretical framework, requiring the reinterpretation of many of the existing studies. Chapter 4 presents a study investigating the second assumption of whether stopping is vulnerable to some of the same constraints, or rules, governing response initiation (Logan, 1994; Logan & Cowan, 1984). Although previously, response inhibition has been examined indirectly by manipulating the properties of the go process (e.g., by manipulating reactive inhibition: Verbruggen et al. 2004, 2005) the present study attempts to explore the stopping process by manipulating it directly. Stopping and response initiation are compared directly by examining their susceptibility to the effects of stimulus-response (S-R) compatibility. S-R compatibility merely states that the more 15 similar the stimulus and the response, the better the observed performance (e.g., Hommel & Prinz, 1997). The S-R compatibility effect was chosen as it is typically very robust and is commonly found for many different types of responses (Hommel & Prinz, 1997). Such a study wi l l help delineate the extent to which stopping is ruled by different constraints than going, in addition to exploring a new S-R compatibility effect. Chapter 5 presents a study explicitly testing the third assumption that stopping generalizes to other measures of response adjustment. Thus, stopping is compared to another response modification, that of accelerating the controlled tracking response. Likewise, this study provides an additional test of the second assumption. To this end, this study tests one of the most striking results in the stop signal literature - that unlike go RTs, SSRTs do not seem to be influenced by predictability as manipulated by signal frequency (Logan, 1981; Logan & Burkell , 1986). If the third assumption does not hold in this circumstance, then the validity of stopping as an index of response control would be in question. Chapter 6 presents a summary and integration of the main findings, discusses their implications, and offers both insights into current theorizing as well as a discussion of future issues' that could be addressed using the present research methodology. 16 References Aron, A. R., Fletcher, P. C , Bullmore, E. T., Sahakian, B. J., & Robbins, T. W. (2003b). Stop-signal inhibition disrupted by damage to right inferior frontal gyrus in humans. Nature Neuroscience, 6, 115-116. Aron, A. R., Robbins, T. W., & Poldrack, R. A . (2004). Inhibition and the left inferior frontal cortex. Trends in Cognitive Sciences, 8, 170-177. Badcock, J. C , Michie, P. T., Johnson, L. , & Combrinck, J. (2002). Acts of control in schizophrenia: dissociating components of inhibition. Psychological Medicine, 32, 287-297. Band, G. P. H. , & van Boxtel, G. J. M . (1999). Inhibitory motor control in stop paradigms: review and reinterpretation of neural mechanisms. Acta Psychologica, 707,179-211. Band, G. P. PL, van der Molen, M . W., & Logan, G. D. (2003). Horse-race model simulations of the stop-signal procedure. Acta Psychologica, 112, 105-142. Barkley, R. A. (1997a). Behavioral inhibition, sustained attention, and executive functions: constructing a unifying theory of A D H D . Psychological Bulletin, 121, 65-94. Bedard, A. , Nichols, S., Barbosa, J. A. , Schachar, R., Logan, G. D., & Tannock, R. (2002). The development of selective inhibitory control across the life span. Developmental Neuropsychology, 21, 93-111. 17 Burgess, P. W . (1997). Theory and methodology in executive function research. In P. Rabbitt (Ed.), Methodology offrontal and executive function (pp. 81-116). Hove, U K : Psychology Press. Colonius, H . , Ozyrt, J., & Arndt, P. A . (2001). Countermanding saccades with auditory stop signals: testing the race model. Vision Research, 41, 1951-1968. Curtis, C. E . , Cole, M . W. , Rao, V . Y . , & D'Esposito, M . (2004). Canceling planned action: an fMRI study of countermanding saccades. Cerebral Cortex. Day, B . L . , & Lyon, L . N . (2000). Voluntary modification of automatic arm movements evoked by motion of a visual target. Experimental Brain Research, 130, 159-168. De Jong, R., Coles, M . G . H . , & Logan, G . D . (1995). Strategies and mechanisms in nonselective and selective inhibitory motor control. Journal of Experimental Psychology: Human Perception and Performance, 21(3), 498-511. De Jong, R., Coles, M . G . H . , Logan, G . D . , & Gratton, G . (1990). In search of the point of no return: the control of response processes. Journal of Experimental Psychology: Human Perception and Performance, 16, 164-182. Dimitrov, M . , Nakic, M . , Elpern-Waxman, J., Granetz, J., O'Grady, J., Phipps, M . , et al. (2003). Inhibitory attentional control in patients with frontal lobe damage. Brain and Cognition, 52, 258-270. Eagle, D . M . , & Robbins, T. W . (2003). Inhibitory control in rats performing a stop-signal reaction-time task: effects of lesions of the medial striatum and d-amphetamine. Behavioral Neuroscience, 117, 1302-1317. Fillmore, M . T., Rush, C. R., Kelly, T. FL, & Hays, L. (2001). Triazolam impairs inhibitory control of behavior in humans. Experimental and Clinical Psychopharmacology, 9, 363-371. Friedman, N . , & Miyake, A. (2004). The relations among inhibition and interference control functions: a latent-variable analysis. Journal of Experimental Psychology: General, 133, 101-135. Gauggel, S., Rieger, M . , & Feghoff, T. A. (2004). Inhibition of ongoing responses in patients with Parkinson's disease. Journal of Neurology, Neurosurgery and Ps, 75, 539-545. Hanes, D. P., Patterson, W. F., & Schall, J. D. (1998). Role of frontal eye fields in countermanding saccades: visual, movement, and fixation activity. Journal of Neurophysiology, 79, 817-834. Hanes, D. P., & Schall, J. D. (1995). Countermanding saccades in macaque. Visual Neuroscience, 12, 929-937. Henry, F. M . , & Harrison, J. S. (1961). Refractoriness of a fast movement. Perceptual and Motor Skills, 75,351-354. Hommel, B. (1997). Interactions between stimulus-stimulus congruence and stimulus-response compatibility. Psychological Research, 59, 248-260. Kok, A. (1999). Varieties of inhibition: manifestations in cognition, event-related potentials and aging. Acta Psychologica, 101, 129-158. Kornylo, D., Di l l , N . , Saenz, M . , & Krauslis, R. J. (2003). Canceling of pursuit and saccadic eye movements in humans and monkeys. Journal of Neurophysiology, 59,2984-2999. Kramer, A . F. , Humphrey, D . G. , Larish, J. F. , Logan, G . D . , & Strayer, D . L . (1994). Aging and inhibition: Beyond a unitary view o f inhibitory processing in attention. Psychology and Aging, 9(4), 491-512. Kudo, K . / & Oshtuski, T. (1998). Functional modification of agonist-antagonist electromyographic activity for rapid movement inhibition. Experimental Brain Research, 122, 23-30. Ladefoged, P., Silverstein, R., & Papcun, G . (1973). Interruptility of Speech. The Journal of the Acoustical Society of America, 54, 1105-1108. Lappin, J. S., & Eriksen, C. W . (1966). Use of a delayed signal to stop a visual reaction-time response. Journal of Experimental Psychology, 72(6), 805-811. Logan, G . D . (1981). Attention, automaticity, and the ability to stop a speeded choice response. In J. Long & A . D. Baddeley (Eds.), Attention and performance IX. Hillsdate, N J : Lawrence Erlbaum. Logan, G . D . (1982). On the ability to inhibit complex movements: a stop-signal study of typewriting. Journal of Experimental Psychology: Human Perception and Performance, 5(6), 778-792. Logan, G . D . (1994). On the ability to inhibit thought and action: A users' guide to the stop signal paradigm. In D . Dagenbach & T. H . Carr (Eds.), Inhibitory processes in attention, memory, and language (pp. 214-249). San Diego: Academic Press. Logan, G . D . , & Burkell , J. (1986). Dependence and independence in responding to double stimulation: a comparison of stop, change and dual-task paradigms. Journal of Experimental Psychology: Hitman Perception and Performance, 12(4), 549-563. Logan, G. D., & Cowan, W. B. (1984). On the ability to inhibit thought and action: A theory of an act of control. Psychological Review(9l), 295-327. Logan, G. D., Cowan, W. B., & Davis, K. A . (1984). On the ability to inhibit simple and choice reaction time responses: a model and a method. Journal of Experimental Psychology: Human Perception and Performance, 10(2), 276-291. Logan, G. D., & Irwin, D. E. (2000). Don't look! Don't touch! Inhibitory control of eye and hand movements. Psychonomic Bulletin & Review, 7(1), 107-112. McGarry, T., Chua, R., & Franks, I. M . (2003). Stopping and restarting an unfolding action at various times. Quarterly Journal of Experimental Psychology, 56a. McGarry, T., & Franks, I. M . (1997). A horse race between independent processes: Evidence for a phantom point of no return in the preparation of a speeded motor response. Journal of Experimental Psychology: Human Perception and Performance, 23(5), 1533-1542. Megaw, E. D. (1972). Directional errors and their correction in a discrete tracking task. Ergonomics, 15(6), 633-643. Miyake, A. , Friedman, N . , Emerson, M . J., Witzki, A . H. , Howerter, A. , & Wager, T. D. (2000). The Unity and Diversity of Executive Functions and Their Contributions to Complex "Frontal Lobe" Tasks: A Latent Variable Analysis. Cognitive Psychology, 41,49-100. Morein-Zamir, S., & Kingstone, A . F. (in preparation). Examining the stop signal: a critical review. 21 Morein-Zamir, S., & Meiran, N . (2003). Individual stopping times and cognitive control: Converging evidence from a new stop signal paradigm. Quarterly Journal of Experimental Psychology, 56, 469-490. Mulvihill, L. A. , Skilling, T. A. , & Vogel-Sprott, M . (1997). Alcohol and the ability to inhibit behavior in men and women. Journal of Studies on Alcohol, 58, 600-605. Nigg, J. T. (2001). Is A D H D a disinhibitory disorder? Psychological Bulletin, 127, 571-598. Norman, D., & Shallice, T. (1986). Attention to action: willed and automatic control of behaviour. In R. J. Davidson, G. E. Schwartz & D. Shapiro (Eds.), Consciousness and self-regulation (Vol. 4). New York: Plenum Press. Oilman, R. T. (1973). Simple reactions with random countermanding of the "go" signal. In S. Kornblum (Ed.), Attention and Performance IV (pp. 571-581). New York: Academic Press. Oosterlaan, J., Logan, G. D., & Sergeant, J. A . (1998). Response inhibition in AD/HD, CD, comorbid AD/HD + CD, anxious, and control children: a meta-analysis of studies with the stop task. Journal of Child Psychology and Psychiatry, 39, 411-425. Osman, A. , Kornblum, S., & Meyer, D. E. (1986). The point of no return in choice reaction time: controlled and ballistic stages of response preparation. Journal of Experimental Psychology: Human Perception and Performance, 12(3), 243-258. Osman, A. , Kornblum, S., & Meyer, D. E. (1990). Does motor programming necessitate response execution? Journal of Experimental Psychology: Human Perception and Performance, 1990(16), 183-198. Ozyrt, J., Colonius, H., & Arndt, P. A. (2003). Countermanding saccades: evidence against independent processing of go and stop signals. Perception and Psychophysics, 65, 420-428. Phillips, L. H. (1997). Do "frontal tests" measure executive function? Issues of assessment and evidence from fluency tests. In P. Rabbitt (Ed.), Methodology of frontal and executive function (pp. 191-213). Hove, UK: Psychology Press. Pisella, L., Arzi, M., & Yves, R. (1998). The timing of colour and location processing in the motor context. Experimental Brain Research, 121(3), 270-276. Ramautar, J. R., Kok, A., & Ridderinkhof, R. (2004). Effects of stop-signal probability in the stop-signal paradigm: the N2/P3 complex further validated. Brain and Cognition, 56, 234-252. Ridderinkhof, K. R., Band, G. P. H., & Logan, G. D. (1999). A study of adaptive behavior: effects of age and irrelevant information on the ability to inhibit one's actions. Acta Psychologica, 101,315-331. Rieger, M., & Gauggel, S. (1999). Inhibitory after-effects in the stop signal paradigm. British Journal of Psychology, 90,509-518. Rubia, K., Russell, T., Overmeyer, S., Brammer, M. J., Bullmore, E..T., Sharma, T., et al. (2001b). Mapping motor inhibition: conjunctive brain activations across different versions of go/no-go and stop tasks. Neurolmage, 13, 250-261. Schall, J. D., Hanes, H. P., & Taylor, T. L. (2000). Neural control of behavior: countermanding eye movements. Psychological Research, 63, 299-307. Tannock , R. (2003). Neurophsycho logy o f attention disorders. In S . J . Sega lowi tz & I. R a p i n (Eds. ) , Handbook of Neuropsychology (2nd E d i t i o n ed. , V o l . 8, Part II, pp. 7 5 3 - 7 8 4 ) . A m s t e r d a m : E lsev ier . Turner, D . C , C l a r k , L. , P o m a r o l - C l o t e t , E. , M c K e n n a , P. , R o b b i n s , T. W . , & Sahakian , B . J . (2004a). M o d a f i n i l improves cogni t ion and attentional set shi f t ing in patients w i t h chronic schizophrenia . Neuropsychopharmacology(29), 1 3 6 3 - 1 3 7 3 . Turner, D . C , D o w s o n , J . , R o b b i n s , T. W . , & Sahakian , B . J . (2004b) . M o d a f i n i l improves cogni t ion and response inh ib i t ion in adult A D H D . Biological Psychiatry, 55, 1031 - 1 0 4 0 . V i n c e , M . A . , & W e l f o r d , A . T. (1967). T i m e taken to change the speed o f a response. Nature(2\3), 5 3 2 - 5 3 3 . van den W i l d e n b e r g , W . , van B o x t e l , G . J . M . , & van der M o l e n , M . W . (2003) . The duration o f response inh ib i t ion in the stop-s ignal parad igm varies w i th response force. Acta Psychological 14), 115 -129. van den W i l d e n b e r g , W . , van der M o l e n , M . W . , & L o g a n , G . D . (2002) . R e d u c e d response readiness delays stop s ignal inh ib i t ion . Acta Psychologica, 11, 155 -169. Verbruggen , F., L i e f o o g h e , B . , & Vandierendonck . (2004). The interaction between stop s ignal inh ib i t ion and distractor interference in the f lanker and Stroop task. Acta Psychologica, 116,21-31. Verbruggen , F., L i e f o o g h e , B . , & Vand ie rendonck , A . (2005). O n the dif ference between response inh ib i t ion and negative p r i m i n g : evidence f r o m s imple and selective stopping. Psychological Research, 69, 2 6 2 - 2 7 1 . 24 W i l l i a m s , B . R., Ponesse, J . S . , Schachar , R. J . , L o g a n , G . D . , & Tannock , R. (1999). Deve lopment o f inhib i tory control across the l i fe span. Developmental Psychology, 35(1), 2 0 5 - 2 1 3 . 25 t C H A P T E R 2: M e a s u r i n g On l ine V o l i t i o n a l Response Cont ro l w i th a Cont inuous T r a c k i n g T a s k 1 N u m e r o u s tasks have been e m p l o y e d to measure cogni t ive inf luences on motor contro l , i n c l u d i n g error correct ion and the adjustment o f an ongo ing response such as to accelerate, decelerate or change direct ion o f a movement (e.g., H e n r y & Har r i son , 1961; M e g a w , 1972; V i n c e & W e l f o r d , 1967; for a rev iew see L o g a n & C o w a n , 1984). One o f the most notable measures has been that o f response inh ib i t ion , or s topping ( L a p p i n & E r i k s e n , 1966; L o g a n , 1994). The advantage o f studying stopping over other forms o f control has been that it is a clear and extreme fo rm o f control w h i c h is easy to observe and relat ively easy to measure ( L ogan , 1994). S topping has t radi t ional ly been studied us ing the stop s ignal task where participants are asked to press a key i n response to a "go s i g n a l " , w h i c h indicates that a response is to be executed. O c c a s i o n a l l y the go s ignal is f o l l o w e d by a "stop s i g n a l " , w h i c h indicates that the key press is to be stopped and the intended act ion wi thhe ld . A l t h o u g h no overt behavior is observed on a successful stop tr ia l , the use o f a racehorse mode l together w i th the task a l lows for the mathematical est imation o f the latency o f stopping or stop s ignal react ion t ime ( S S R T , L o g a n & C o w a n , 1984). S S R T serves as a direct measure o f the control exerted to prevent the execut ion o f a vo l i t iona l goal -d i rected act ion. T o date, the stop s ignal task has been used in a w ide variety o f settings in order to examine questions ranging f r o m the bal l is t ic elements o f movement to the inhib i tory underpinnings o f disorders such as attention ' A version of this chapter is in press: Morein-Zamir, S., Chua, R., Franks, I., Nagelkerke, P., & Kingstone, A. (in press). Measuring online volitional response control with a continuous tracking task. Behavior Research Methods. 26 def ic i t hyperact iv i ty disorder ( A D H D ) (De Jong , C o l e s , L o g a n & Grat ton , 1990; N i g g , 2 0 0 1 ; Ooster laan, L o g a n & Sergeant, 1998; O s m a n , K o r n b l u m & M e y e r , 1986). The purpose o f the present study is to introduce and explore an alternative task for measur ing stopping, o f fer ing several advantages over the tradit ional stop s ignal task ( M o r e i n - Z a m i r & M e i r a n , 2003) . F o r instance, the task a l lows for the explorat ion o f addit ional instances o f contro l , such as response adjustment and s topping after response execut ion, w h i c h have been largely neglected w i th the increasing popular i ty o f the stop s ignal task. In addi t ion , the new task is not dependent on a mathemat ica l m o d e l such as the race m o d e l or its assumptions, as is the stop s ignal task, and so is resistant to instances where the m o d e l m a y not ho ld (e.g., L o g a n & B u r k e l l , 1986). L i k e w i s e , the alternative task can prov ide a measurement o f stopping R T on each t r ia l , not possible w i t h the stop s ignal task, p r o v i d i n g for more eff ic ient data co l lec t ion and a broader range o f analyses. Spec i f i ca l l y , questions necessitating measurements on every tr ial can be addressed, i n c l u d i n g the examinat ion o f sequential effects and measurements o f variance. Such measures are important as they a l l o w for the testabil ity o f predict ions that to date cou ld not be conf i rmed (e.g., B a n d , van der M o l e n & L o g a n , 2 0 0 3 ; Tannock , 2003) . The standard vers ion o f the t racking task requires part ic ipants to stop an ongoing act ion after it has been init iated (see Ladefoged , S i lverste in & P a p c u n , 1973; L o g a n , 1982 for a s imi la r rationale). O n each tr ia l , participants engage in t rack ing a m o v i n g v isual target presented on the screen and at some later point they respond to a s ignal to stop their act ion. Stopping performance is measured by e x a m i n i n g the response adjustment onset latency to the stop s ignal (Henry & Har r i son , 1961). In the current task, the target moves a long a c i rcu lar trajectory at a constant angular speed and participants 27 control the speed o f a corresponding response marker by press ing on a force sensor. The greater the pressing force, the faster the response marker rotation. Part ic ipants per form the t rack ing task for a var iable amount o f t ime before a stop s ignal is presented. Onset o f response adjustment, i.e., force reduct ion causing rotational decelerat ion, is taken as the measure o f contro l . The current study introduces the new task and its development in detai l . Data col lected over two experiments explore the performance characterist ics o f the task. In the second experiment, onset o f increasing force leading to marker accelerat ion is introduced as an addit ional measure o f response adjustment, and compared to onset o f stopping. Current M e t h o d o l o g y Apparatus A l t h o u g h the development o f the t racking task has been a dynamic process, the core log ic and mot ivat ion def in ing the task and the dependent measures remained constant. The hardware on w h i c h the task was implemented changed over the course o f experiments conducted. The in i t ia l development was per formed w i t h part icipants pressing on a 7.6 x 10.2 x .9 c m a l u m i n u m custom mach ined load c e l l instrumented w i th 4 strain gauges (Omega mode l 350 O h m ) . A d d i t i o n a l data were co l lected w i th responses executed us ing a telegraph key instrumented w i th 2 strain gauges to measure downward force appl ied to the finger-rest. In both cases the analog signals f r o m the strain gauges were a m p l i f i e d (Nor thwood Instruments, mode l I A - 1 0 2 - 5 0 0 ) . The resul t ing force-vol tage s ignal was filtered ( K r o n e - H i t e 3750 analog filter set at 50 H z lowpass) and then sampled at the ±1.25 V o l t range at 1000 H z by an analog to d ig i ta l (A/D) converter (Techmar 28 Labmaster ) instal led in the data co l lec t ion computer. Th is computer had an Intel Pent ium- I , 233 M H z processor and 16 M B R A M running M S W i n d o w s '98 Second E d i t i o n operating system in D O S mode w h i c h contro l led st imulus presentation and data co l lec t ion v ia custom computer software (written in B o r l a n d Turbo Pasca l 6.0). S t i m u l i were presented to participants on a Zen i th 14" F lat Screen moni to r ( Z C M - 1 4 9 0 ) w i th a refresh rate o f 60 H z running in standard V G A v ideo mode (resolution 640 x 480 pixels) and the experimenter moni tored performance on a monochrome moni tor dr iven by a Hercu les compat ib le monochrome v ideo card. A d d i t i o n a l data co l lec t ion was performed " w i th part ic ipants ' response force measured on the selector button o f a m o d i f i e d older m o d e l M a c i n t o s h computer mouse. The computer mouse was m o d i f i e d to accept a force sensor (F lex i force A 2 0 1 - 1 ) under the selector key . The sensor used a pressure-sensit ive ink that changes its e lectr ical resistance w i t h compress ive force, w i th the result ing analog voltage sampled by an A 7 D converter (P ico A D C - 2 1 2 ) at 1000 H z . Th is apparatus was connected to the paral le l port o f a 3 2 0 C D T Tosh iba laptop, w h i c h contro l led st imulus d isp lay and response co l lec t ion . The same d isp lay was presented on the laptop moni tor as w e l l as a s laved moni tor , w i t h performance summary data be ing presented at the end o f each practice tr ia l and at the end o f each exper imental b lock . In al l cases, performance data were stored and analyzed after data co l lect ion . General Procedure In the development o f the new task, many different parameters and their speci f ic values had to be def ined. T o ensure a satisfactory task, several p i lot versions were developed and tested. In some o f the experiments, several o f the parameters were 29 changed or manipulated and the task was further m o d i f i e d . The f o l l o w i n g is a descr ipt ion o f the in i t ia l task, detai l ing also h o w part icular parameters were chosen. In i t ia l ly , several parameters were determined upon inspect ion o f one o f the earlier response inh ib i t ion tasks, employed by S la t te r -Hammel (1960). A d d i t i o n a l guidel ines were adopted f r o m a t rack ing task us ing a s imi la r rationale ( M o r e i n - Z a m i r & M e i r a n , 2003) . In this p r io r task, participants per formed a v isua l t rack ing task, but the apparatus and measurement o f dependent measures di f fered considerably . Part ic ipants tracked a v isual target m o v i n g in a series o f l inear trajectories, w i t h the a id o f a computer mouse. Dece lerat ion onset latency, or S S R T , was calculated us ing an accumulat ive regression a lgor i thm calculated f r o m the spatial locat ion o f the mouse marker over t ime. In this task although stopping cou ld be measured on each t r ia l , motor noise and a poor sampl ing rate (50 H z ) caused the a lgor i thm to fa i l on a substantial port ion o f the trials ( 2 2 % and 3 3 % o f trials in Exper iments 1 and 2, respect ively) . T h e present task possessed several important new characterist ics. The task was more straightforward and the t racking requirements were constrained. Part ic ipants contro l led the rotational speed o f the response marker but not its d i rect ion , thus restr ict ing the noise result ing f r o m motor performance. In addi t ion , the dependent measure was computed f r o m the force appl ied over t ime. Th i s force contro l led the response marker w i th an increased temporal resolut ion o f 1000 H z . The m a i n dependent measure was conceptual ly s imi la r to S S R T in the prev ious t rack ing task and consisted o f the latency to init iate a stop, i.e., to initiate a reduct ion i n appl ied force on the sensor caus ing angular deceleration o f the t rack ing response. H o w e v e r , instead o f comput ing S S R T f r o m the pos i t ion o f the computer mouse and hence f r o m the pos i t ion o f the 30 response marker on the screen, S S R T was now computed f r o m the fo rce -over - t ime prof i les a l l o w i n g for increased re l iabi l i ty . Th is alternate dependent measure enabled the measurement o f S S R T on almost every tr ial . The presented vers ion o f the task proved to be cha l leng ing and engaging yet after a few minutes o f pract ice, performance was stable for the vast major i ty o f participants. Th is transpired to be an important guidel ine to determining task characterist ics. T r ia ls were considered stable i f dur ing the S S R T latency interval the t rack ing force d i d not deviate more than 100 grams and m a x i m u m tracking error was less than 15 degrees. Consequent ly , performance y ie lded force prof i les that were suf f ic ient ly steady so as to a l low a clear demarcat ion o f S S R T on over 9 8 % o f a l l tr ials. In the current t rack ing task, on each tr ial the target marker m o v e d in a predetermined w a y , rotating around the screen at a f i xed speed (see F igure 2.1). B o t h target and response images were smal l c i rcular markers. Part ic ipants were asked to track and advance the response marker so as to over lap the target marker . T o do this, participants pressed on a static force sensor to advance their response marker . The more force appl ied , the faster the rotational rate o f the response marker . O n a typ ica l stop t r ia l , after a var iable per iod o f t rack ing , a s ignal appeared ind icat ing that part icipants were to stop t racking. They d i d this by terminat ing the appl ied pressure on the force sensor without l i f t ing their f inger and as a consequence, the response marker also stopped m o v i n g . S topp ing was stressed as releasing the exerted force, without l i f t ing the f inger f r o m the key to ensure m a x i m a l s imi lar i ty between conceptual and motor stopping ( M o r e i n - Z a m i r , N a g e l k e r k e , C h u a , F ranks & K ings tone , 2004) . S S R T was determined as the point in t ime where the appl ied force on the sensor began a drop towards the in i t ia l basel ine pre-press force. W h e n t rack ing was stable, w i t h no large changes in t racking force, this point was quite dist inct, as can be seen in F igure 2.2. Figure 2.1. A depiction of the typical series of events in the tracking task Target and response markers at the 3" position (0.2-0.6 sec) Target marker moves and tracking is initiated Participant tracks target by moving the response marker (5-15 sec) Following 3 sec of adequate tracking (better than 15°), target stops and participant stops response marker 32 Figure 2.2. A force over time profile y ielded over the course of a trial. F igure 2.2A portrays a stop trial, whi le Figure 2.2B portrays an accelerate trial (see Exper iment 2 for further details of Figure 2.2B). Figure 2.2A. 900 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Time (seconds) Figure .2B. 900 800 -, . </> 700 -E re i _ 600 ; 3 500 -u o 400 -LL 300 -c 200 -o re 100 -H 0 --100 -A Trigger Start O Stimulus Stop O S S R T (261 ms) V 1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Time (seconds) The st imulus and response markers first appeared stationary at the 3 o ' c l o c k pos i t ion and then rotated c l o c k w i s e around a 200 p i x e l (8.1 cm) radius c i rcu lar path. The 33 size o f this pathway was constrained by the height o f the screen. C i r c u l a r s t imul i were chosen, as they do not have a direct ional component and thus w o u l d appear the same no matter where a long the pathway they were presented. Several colors o f the c i rc les were tested, inc lud ing a red target and a green response marker but bright blue and green were eventual ly selected as hav ing little meaning in combinat ion . The sizes o f the s t imul i were chosen so that the response marker fit inside the target dur ing perfect t rack ing performance. A radius o f 5-7 p ixe ls for the green response c i rc le and a radius o f 8 -10 p ixe ls for the b lue target r ing were chosen. Th is entailed a radius o f .3 c m for the green response c i rc le and .3 - .405 c m for the blue target r ing . The size was suf f ic ient ly large to a l l o w for easy v i s ib i l i t y but was smal l enough to necessitate part icipants to f ixate on a part icular area o f the screen. Di f ferent rotational speeds o f the target were also examined. A st imulus rotation rate o f 1 H z was tested and deemed too fast for adequate t racking performance, and so a rotation rate o f .5 H z was chosen. A vert ical d o w n w a r d force o f 3 .924 N e w t o n , equivalent to a 4 0 0 gram mass resting on the sensor, gave a response rotation rate o f .5 H z , match ing the st imulus rotation speed. There was a l inear force-rotat ion rate relat ionship o f .125 H z for every .981 N e w t o n (the force exerted by a 100 gram mass at 1 gravity) . The equipment was cal ibrated us ing standard g ram masses, and the units o f force are reported in equivalent gram mass units for s impl ic i t y . The m i n i m a l threshold o f 100 grams (.981 N e w t o n ) y ie lded a .125 H z rate and the m a x i m u m threshold o f 2000 grams (19.62 N e w t o n ) resulted in a response rotation rate o f 2.5 H z . The opt imal t rack ing force was reduced f r o m an in i t ia l 600 grams to 400 grams to reduce part icipant fatigue. A t 4 0 0 grams the 34 force was sustainable for durations up to an hour and yet the apparatus was not overly sensitive. Analysis of Dependent Measures A response usually consisted of a time interval with zero applied force on the selector key, followed by a rapid onset peaking at 600 grams or more (often 1000 grams, applied to 'catch up' with the target), followed by tracking at the target value of 400 grams (see Figure 2.2). After the presentation of a stop signal there was a marked force release where tracking force decreased from approximately 400 grams back to zero grams. If an acceleration signal was introduced instead of a stop signal (see below), participants increased the applied force well above 400 grams for several seconds. The dependent measures were SSRT and A C C E L R T for the onset of force reduction and force increase leading to stopping and accelerating, respectively. In both cases, performance-adjustment reaction times (RTs) were measured as the latency from signal onset to the onset of the response adjustment. The onset and offset points for changes in applied force were found by analyzing the change in the response force profile over the time course of the trial (Figure 2.2). First the force profile for each trial was digitally filtered at 10 Hz. The value of 10 Hz was chosen because movements faster than 100 ms are not considered voluntary and therefore not of interest in the present context. A standard procedure (see below) was used to detect an adjustment on the force profile. Specifically, the procedure could be used to find any of three possible points: a) following a stop signal, force reduction latency leading to deceleration; b) following an acceleration signal, force increase latency leading to acceleration; and c) following target movement at the beginning of a trial, the force onset 35 latency in i t iat ing the t racking movement 2 . In addi t ion, a s imi la r procedure c o u l d be used to examine the latency at w h i c h the force decrease or increase terminated. In the case o f a force increase leading to accelerat ion, this value was not re l iable , as it was heav i l y dependent on the f inal speed o f the response marker f o l l o w i n g force increase, w h i c h was under the discret ion o f the part icipant and var ied across trials. Thus , response adjustment onset and not offset t imes were the m a i n dependent measures. Nevertheless , the a lgor i thm for f ind ing force change offset is also descr ibed be low. T o find the onset o f force reduct ion, a computer a lgor i thm searched for the first point f o l l o w i n g the stop s ignal at w h i c h the measured force dropped b e l o w 200 grams ( m i d - w a y between the ideal 4 0 0 grams go ing to 0 grams). The rate -of -change, or slope, o f this point was calculated based on the value difference o f the two points immediate ly surrounding it d i v ided by the t ime interval between the points. The a lgor i thm then searched for the first p receding data point where the slope was 1 0 % o f this value. That point in t ime was marked as the onset o f force reduct ion and the latency was calculated as the di f ference between this point in t ime and the s ignal onset t ime marker . The a lgor i thm also w o r k e d fo rward in t ime f r o m the 200 gram threshold to find the point where the slope was 1 0 % o f this value. The difference between this t ime and the target s ignal p rov ided the latency where the appl ied force reduct ion and consequent marker decelerat ion was essential ly complete (also ca l led F I N R T ) . The ca lcu lat ion o f onset latency for a force increase was almost ident ica l . The onset o f a force increase was determined based on the point whose slope was 1 0 % o f the slope measured at the 600 gram threshold. 2 Tracking onset did not prove to be an interesting measure in the two experiments reported, as many anticipations were observed. Additional studies not reported here have been conducted that have demonstrated the reliability of the algorithm at detecting force onset. In essence, the a lgor i thm was used to calculate the force di f ferent ia l o f the tr ial . The slope for each data point was calculated as the difference between the values before and the value after, d i v ided by the t ime interval , y i e l d i n g the change in force in units o f grams/second. Th is resulted in a posit ive value when the force increased and a negative value w h e n force decreased. Peak onset slopes typ ica l l y measured 1000 grams/second. Th is a lgor i thm was appl ied to a l l trials to find force onset and offset, f o l l o w e d by v isual inspect ion o f every t r ia l . T r ia ls were spec i f ica l l y screened for S S R T and F I N R T t imes, in addi t ion to t rack ing stabi l i ty as indicated by the force var iab i l i ty dur ing the S S R T interval (between signal onset and force change onset). O n l y trials w i t h stable t rack ing data dur ing the S S R T interval w i th clear force change onsets and offsets were inc luded for data analysis . A s the final vers ion o f the task led to very stable t rack ing at the moment o f st imulus presentation, less than 2 % o f trials were exc luded (see results sections). Tr ia ls w i t h S S R T values larger than 3.5 standard deviat ions were also exc luded f r o m further analyses. F i n a l l y , S S R T latencies shorter than 50 ms (mi l l i seconds) were considered anticipations and also not inc luded in the analyses. A n addi t ional measure o f interest was t rack ing performance on each tr ia l . Th is was measured by phase error (PE) . Phase error was calculated as the absolute angular pos i t ion error between the st imulus and response markers d isp layed on the computer screen a long the c i rcu lar trajectory, and ranged f rom 0 to 180 degrees. Thus , a smal ler value connoted better performance. The P E value on each tr ia l constituted the average angle for 1 second before s ignal onset sampled at 1000 H z . 37 Exp lo ratory Data C o l l e c t i o n In the f o l l o w i n g sect ion, we present some in i t ia l results generated us ing the t racking task. The f o l l o w i n g two experiments and their corresponding data a l l owed for the in i t ia l explorat ion and ver i f i cat ion o f the task. Exper iment 1 The purpose o f the first experiment was to test the new task and obtain a body o f data for w h i c h descr ipt ive statistics cou ld be obtained and ver i f ied . T h e data analysis a lgor i thm was tested and examined on a sample o f observations. Th i s exper iment also il lustrates the performance on the task and the types o f descr ipt ive statistics that can be gathered. Methods Participants. Ten undergraduates f r o m the P s y c h o l o g y Department at the Un ive rs i t y o f B r i t i s h C o l u m b i a part ic ipated in exchange for course credit (mean age was 19.3 years, £ 0 = 0 . 7 ) . A l l but one were right handed, two were male , and a l l had normal or cor rected - to -normal v i s ion . A f t e r be ing in formed about the nature o f the experiments, part icipants s igned an in formed consent fo rm in accordance w i t h the ethical guidel ines o f the Un i ve rs i t y o f B r i t i s h C o l u m b i a . Procedure and Design. U s i n g the methods descr ibed above, part icipants completed one session o f approx imately 100 exper imental tr ials (see Tab le 2.1 be low) , w i th a stop s ignal on each tr ial . The trials began w i t h stationary target and response s t imul i at the " 3 o ' c l o c k " pos i t ion o f the imaginary c i rc le . A f t e r a random delay o f 2 0 0 -600 ms the target marker began c i r c l i ng c l o c k w i s e at 0.5 H z . A f t e r a r a n d o m t rack ing 38 duration o f 5 -15 seconds, the program ' a r m e d ' the stop s ignal . F o l l o w i n g this a rming t ime, the program moni tored t racking performance such that when the P E was less than 15 degrees for 3000 ms , the stop s ignal was tr iggered. The stop s ignal was the stopping o f the target st imulus , ind icat ing that participants were to stop as fast as poss ib le . The inter-tr ial duration was 2 seconds. F o l l o w i n g 10 practice tr ials, the exper imental tr ials began w i t h se l f - terminat ing breaks every 20 trials. Table 2.1. Descript ive stat ist ics for individual participants in Exper iment 1, inc luding number of trials (No. of trials), means, standard deviat ions and medians for stop s ignal reaction time (SSRT), f inal RT (FINRT) and phase error (PE). S S R T F I N R T P E P N o . N o . o f T r ia ls M SD Mdn M SD Mdn M SD Mdn 1 59 300 56 292 483 69 468 4.0 1.9 4.0 2 119 257 49 254 389 47 379 3.8 2.2 3.7 ,3 100 312 51 307 540 86 540.5 2.9 1.8 2.8 4 98 356 54 358 633 60 631.5 2.6 1.8 2.5 5 98 279 42 278 598 79 586.5 4.1 2.4 3.9 6 98 308 61 299 555 78 547 2.4 1.9 1.9 7 120 287 36 291 538 58 539.5 2.5 2.1 2.1 8 100 266 44 260 489 64 484.5 3.1 1.3 3.1 9 100 291 56 288 612 114 617.5 1.9 1.3 1.6 10 94 338 60 334 604 72 594 3.1 2.0 2.7 99 299 50 295 544 72 539 3.0 1.9 2.77 Note. P no. = Part ic ipant number ; M= M e a n ; SD = standard deviat ions; Mdn = median . Results and Discussion U s i n g the results o f Exper iment 1, the analysis a lgor i thm was developed and over 9 9 % o f the trials y ie lded v a l i d S S R T values. T r ia ls on w h i c h S S R T c o u l d not be computed re l iab ly , or resulted in values larger than 590 ms or smal ler than 50 ms (0 .9%) , were subsequently dropped f r o m a l l analyses. Table 2.1 prov ides descr ipt ive statistics for 39 S S R T , F I N R T and P E for each participant. S S R T s values were in c lose agreement to those prev ious ly found in the stop s ignal procedure (e.g., L o g a n , 1994; L o g a n & C o w a n , 1984) and those found in the mouse - t rack ing task ( M o r e i n - Z a m i r & M e i r a n , 2003) . The shape o f the S S R T distr ibut ion across participants proved to be s imi la r to those o f most reaction t ime funct ions. The box and whisker plot graphs (see F igure 2.3) support this conc lus ion . Some var iab i l i t y in the var ious performance indices can be observed between part ic ipants, w i t h participant 4 be ing slowest overal l to stop, and part ic ipant 9 be ing best overal l in t rack ing performance. Figure 2.3. Stop signal reaction time (SSRT) data in Experiment 1 for each individual participant using box and whisker plots. 600 550 500 450 400 350 | 300 250 200 150 100 ±1.96*Std. Dev. r ~ l ±1.00*Std. Dev. • Mean 4 5 6 PARTICIPANT 10 Corre lat ions between S S R T , final R T and P E across part icipants indicated that S S R T and F I N R T were s igni f icant ly correlated (r=.69, f (8)=2.7,p<.05) . A l t h o u g h both S S R T and F I N R T correlated negat ively w i t h P E , neither corre lat ion reached s igni f icance 40 in the current sample ( - .32 and - . 4 7 , respect ively) . Corre lat ions were further computed between each o f the three measures w i th in each part icipant (see Tab le 2.2). Th is analysis exp lored whether participants were l i k e l y to tradeoff t rack ing and stopping performance. M e a n correlat ions were computed on transformed F isher Z values and then re -transformed (Rosenthal , 1991). The resul t ing mean correlat ion within each participant between S S R T and P E was . 0 1 , and between F I N R T and P E it was . 0 3 , suggesting that even on ind i v idua l trials there was no correlat ion between t racking and stopping performance (see M o r e i n - Z a m i r & M e i r a n , 2 0 0 3 , for a s imi la r conc lus ion) . There was a strong, posit ive correlat ion between S S R T and F I N R T o f .74. H e n c e , P E does not seem to correlate consistently w i th either o f the stopping measures, w h i l e they s ign i f icant ly correlate between themselves. Corre lat ions between each o f the dependent measures and tr ial number were also computed , us ing the above procedures, to examine the effects o f practice across t ime in the exper imental tr ials. Pract ice had no effects across tr ials, w i th a l l correlations be ing be low .08. Table 2.2. Correlations between the dependent measures in Experiment 1, of stop signal RT (SSRT), final RT (FINRT) and phase error (PE) for individual participants. Mean correlations across participants were calculated by averaging the Fisher's Z transformations of the correlations across participants and then re-transforming them. P N o . S S R T and P E S S R T and F I N R T F I N R T and P E 1 .09 .92 .04 2 .07 .89 .09 3 - . 0 1 .75 .06 4 - . 2 .73 - . 0 9 5 .06 .53 . .3 ' 6 .11 .7 .11 7 .03 .68 .1 8 - . 1 4 .61 - . 2 7 9 .08 .5 .04 10 .09 .73 .14 M Corre lat ion .02 .74 .05 Note. P no. = Part ic ipant number. 41 F i n a l l y , the stabi l i ty o f the m a i n measure o f S S R T was examined . Spec i f i ca l l y , we quest ioned whether the number o f trials used to measure mean S S R T w o u l d alter the result ing values. Therefore, mean S S R T was computed for the first 15, 2 0 , 30 , 4 0 , 50, 60, and 80 trials and compared to mean S S R T computed on a l l tr ials. W h e n entered into an A N O V A , no s igni f icant differences were observed between the measures (F<1). A l l mean S S R T measures w i t h 20 observations and above were w i t h i n 4 ms o f each other (302 ms) . M e a n S S R T o f the first 10 and 15 observations were 8 ms s lower . Correlat ions between S S R T measurements for the first 20 observations and above ranged f r o m .94 to .99 and were h igh ly s ignif icant . These analyses suggest that 20 observations are suff ic ient to gauge the mean S S R T value. In conc lus ion , the first study prov ided descript ive data ind icat ing that the present task cou ld indeed be used to gain measurement o f the in i t ia l stopping t imes. In accordance w i t h M o r e i n - Z a m i r and M e i r a n (2003), the pr imary measurements were S S R T and P E . S ince F I N R T correlated h igh ly w i th S S R T , but was also sensit ive to t rack ing performance (as indicated by the higher correlat ion w i t h P E ) , it was deemed a redundant dependent measure. In f o l l o w i n g experiments, w h e n poss ib le , F I N R T was calculated. H o w e v e r , it d i d not reveal any addit ional in format ion beyond S S R T and was often a more var iable measure leading to less power fu l analyses. A c c o r d i n g l y , F I N R T is not reported in the f o l l o w i n g study. Exper iment 2 The second experiment i l lustrated the types o f research questions that cou ld be investigated and further exp lored the task and the result ing data. In the previous 42 experiment, no manipulat ions were conducted. Thus, the present study set out to examine h o w performance on the task w o u l d appear when several condit ions were introduced and convent ional analyses (e.g., analyses o f var iance, A N O V A ) were required. L i k e w i s e , we examined the re l iab i l i ty o f the dependent measure a lgor i thm. The first manipulat ion invo l ved the type o f response required. In one task, as in the prev ious experiment, participants were instructed to stop as fast as possible w h e n observ ing the s ignal . In a second task, instead o f a stopping response, participants were asked to adjust their performance so as to increase their speed in response to the s ignal . Thus , participants increased the force as fast as poss ib le and accelerated the ve loc i ty o f the response marker for a few seconds. The second manipulat ion was the type o f s ignal used. The first s ignal repl icated the signal used in the previous experiment, that o f the target stopping. The second s ignal was the changing o f the target co lor w h i l e the target cont inued to rotate. It was hypothes ized that the target-stopping w o u l d be a more effect ive s ignal as previous w o r k has found that target locat ion change resulted in faster adjustments o f a po in t ing response as compared to a target co lor change (P isse l la , A r z i & Rossett i , 1998). B a s e d on compat ib i l i t y theory ( H o m m e l & P r i n z , 1997; Proctor & Reeve , 1990), it was also hypothesized that the target-stopping s ignal w o u l d be h i g h l y compat ib le w i t h the stopping response. Cor respond ing ly , an interaction was predicted where the target-stopping s ignal w o u l d be more effect ive than the co lor -change signal in the stop task compared to the accelerate task. Methods Participants. T w e l v e undergraduates f r o m the P s y c h o l o g y Department at the Un ive rs i t y o f B r i t i s h C o l u m b i a part ic ipated in exchange for course credit (mean age was 19.2, SD= 0.9). A l l were right handed, three were male , and a l l had no rmal or corrected-to -normal v i s i o n . A f t e r be ing in formed about the nature o f the exper iments, participants s igned an in fo rmed consent f o r m in accordance w i t h the ethical guidel ines o f the U n i v e r s i t y o f B r i t i s h C o l u m b i a . Procedure and Design. The present methods were s imi la r to Exper iment 1, w i th the f o l l o w i n g exceptions. T w o types o f signals were employed in each task. The first stop s ignal was the same as in Exper iment 1 and invo lved the target st imulus stopping. The second stop s ignal was a co lor change, where the target marker cont inued to rotate. T o ensure suff ic ient sal ience, target color changed to white . The stop signals were randomly intermixed w i t h i n each b l o c k o f tr ials. In addi t ion, there were n o w two separate tasks. The first task repl icated the previous study requi r ing that part icipants stop their t racking performance. The second task required that participants increased their appl ied force on the key in response to the s ignal , thus accelerating the response marker . Part ic ipants mainta ined the response marker at a faster rotational speed for several seconds, or for a suff ic ient t ime to c i rc le the screen twice. Response marker accelerat ion onset latency, A C C E L R T , was def ined as the t ime between st imulus onset and the onset o f appl ied force increase. Th is was conceptual ly s imi la r to S S R T , w h i c h was based on the onset o f force decrease (see analysis section above). The order o f the two tasks was counterbalanced across participants. A f t e r 10 practice tr ials, each part icipant completed 100 exper imental tr ials, 50 in each o f the tasks. Se l f - te rminat ing breaks were avai lable after the 17 t h and 34 t h trials in each b lock . 44 Results and Discussion In keep ing w i t h the conclus ions o f the prev ious exper iment, S S R T / A C C E L R T and P E were the m a i n dependent measures. Tr ia ls on w h i c h S S R T / A C C E L R T c o u l d not be computed re l iab ly , or resulted in values larger than 1010 ms or smal ler than 50 ms (1 .8%) , were subsequently dropped f r o m all analyses. Th is conf i rmed the re l iab i l i ty and v iab i l i t y o f the analysis a lgor i thm developed on the prev ious data set. A n A N O V A was per formed on the react ion t imes w i t h the factors o f response (stop versus accelerate) and s ignal (target stop versus co lor change). The A N O V A revealed a s igni f icant effect for response type, F ( l , l 1)=28.1, p<.01, w i t h mean S S R T be ing shorter than mean A C C E L R T (358 ms and 449 m s , respect ively) . L i k e w i s e , there was a m a i n effect for s ignal type, F ( l , l 1)=66, p<.01. M e a n S S R T / A C C E L R T was shorter when the st imulus was the target stopping as compared to the co lor change (364 ms and 442 m s , respect ively) . F i n a l l y , there was a marg ina l interaction, F ( l , l 1)=4.5, p<.06. P lanned compar isons indicated that in keeping w i th the predict ions, target stop led to faster mean S S R T than A C C E L R T (311ms versus 418 ms , respect ively) , F ( l , l 1)=26.9, p<.01. The c o l o r change also led to faster mean S S R T than A C C E L R T (405 ms versus 4 8 9 ms , respect ively) , albeit to a lesser degree, F ( l , 11 )=19.9, p<.01. A n A N O V A on P E indicated a s ignif icant inf luence for response type, F ( l , l 1)=6.7, p<.05, wi th t racking be ing worse for the stop task compared to the accelerate task (3.1° versus 2 .7°, respect ively) . The results o f Exper iment 2 p rov ided converg ing evidence for the results o f Exper iment 1. S S R T s were found to be faster than A C C E L R T s , and the target stopping proved to be a more effect ive st imulus than the co lor change albeit to a lesser degree in the accelerate task. The results demonstrate that the dependent measures can be used to 45 address questions about cogni t ive processes, and are consistent w i th prev ious response adjustment performance (P ise l la et a l . , 1998). Di f ferent signals m a y prove to be more salient overa l l , as the target stopping led to faster responses compared to the co lor change for both response types. Th i s not ion has yet to be exp lored i n the stopping literature. In addi t ion , these results demonstrate that the in i t iat ion o f different response changes require different latencies, w i t h the in i t iat ion o f a stop be ing faster than the in i t iat ion o f an accelerat ion. Conc lus ions T h e present study introduced and demonstrated the use o f a nove l t rack ing task designed to investigate response control measures in general and stopping i n part icular . The new task offers several advantages when compared to prev ious tasks ( Logan , 1994; M o r e i n - Z a m i r & M e i r a n , 2003) . U n l i k e the stop s ignal task, but s imi la r to the in i t ia l t racking task, the current task offers a measurement on every stop t r ia l , w i t h no need to rely on mathematical assumptions such as those used by the race m o d e l . The former in part icular can be s igni f icant when patients or groups such as ch i ldren or the elder ly are tested. Measurements on every trial also a l l o w for the ava i lab i l i t y o f addi t ional in format ion , such as the var iab i l i ty o f the stopping dist r ibut ion, w h i c h can be valuable when e x a m i n i n g groups suspected o f def ic ient inh ib i t ion (e.g., Ooster laan et a l . , 1998). The new task also offers several advantages when compared to the in i t ia l t rack ing task, such as more precise measurements, advanced temporal resolut ion and a more eff ic ient a lgor i thm for comput ing S S R T and A C C E L R T . Nevertheless, the task does require spec ia l i zed hardware to measure response force over t ime. L i k e w i s e , i f the t racking task 46 becomes too d i f f i cu l t , performance may become too var iable and the detection o f force change w o u l d require a more complex a lgor i thm. T w o important differences stand out between the stop s ignal task and the present t racking task. F i rst , wh i le stopping in the former task examines stopping an act ion before it is ini t iated, the present task leads participants to stop their act ion after it has been init iated. Ev idence f r o m a study compar ing performance in the two tasks supported the conc lus ion that they are sensit ive to the same stopping process ( M o r e i n - Z a m i r et a l . , 2004) . Second, almost a l l stop signal tasks have ut i l i zed a s imple button press or some discrete, a l l -o r -none response, wh i le the present task examines the un fo ld ing o f response force over t ime. The mon i to r ing o f response force, the core o f the present task, m a y prove benef ic ia l even in the use o f discrete responses, such as those observed in most o f the stopping literature (see also van den W i l d e n b e r g , van B o x t e l & van der M o l e n , 2003) . The two experiments descr ibed above demonstrated that numerous task parameters can be examined , inc lud ing st imulus and response adjustment characteristics. M a n y addit ional parameters cou ld be manipulated to alter the task and to address a variety o f research questions. F o r example , although the trigger for a s ignal was currently dependent on t rack ing performance, t racking durat ion can also determine s ignal onset. L i k e w i s e , t rack ing trials where no s ignal is presented can also be introduced, m a k i n g the task more s imi la r to the tradit ional stop s ignal task where stop signals are typ ica l l y present on on ly a minor i t y o f trials ( M o r e i n - Z a m i r et a l . , 2004) . Important ly , the nature o f the task can also be altered. F o r example , M o r e i n - Z a m i r and col leagues ( in press) m o d i f i e d the v isual contingencies so that the harder the part icipant pressed, the s lower the response marker m o v e d . In addi t ion, the task need not engage spatial t rack ing : 47 in order to measure response adjustment one can env is ion different v i sua l feedback (e.g., pressing on the sensor to inf late a ba l loon on the screen). In conc lus ion , the use o f several converg ing tasks p r o v i d i n g in depth informat ion on stopping performance, cou ld offer m u c h needed insight into the processes govern ing stopping in part icular , and response control i n general . 48 References B a n d , G . P. H . , van der M o l e n , M . W . , & L o g a n , G . D . (2003). Horse - race m o d e l s imulat ions o f the stop-s ignal procedure. Acta Psychologica, 112, 105 -142 . D e Jong , R., C o l e s , M . G . H . , L o g a n , G . D . , & Grat ton, G . (1990). In search o f the point o f no return: the control o f response processes. Journal of Experimental Psychology: Human Perception & Performance, 16, 1 6 4 - 1 8 2 . Henry , F. M . , & Har r i son , J . S . (1961) . Refractor iness o f a fast movement . Perceptual & Motor Skills, 13, 3 5 1 - 3 5 4 . H o m m e l , B . , & P r i n z , W . (1997). Theoretical issues in stimulus-response compatibility. N o r t h - H o l l a n d : E l sev ie r Sc ience B . V . Ladefoged , P. , S i l vers te in , R., & Papcun , G . (1973). Interrupti l i ty o f Speech. The Journal of the Acoustical Society of America, 54 , 1105 -1108 . L a p p i n , J . S . , & E r i k s e n , C . W . (1966). U s e o f a delayed s ignal to stop a v isua l react ion-t ime response. Journal of Experimental Psychology, 7 2 , 8 0 5 - 8 1 1 . L o g a n , G . D . (1982). O n the abi l i ty to inhibi t complex movements : a stop-s ignal study o f typewr i t ing . Journal of Experimental Psychology: Human Perception & Performance, 8, 7 7 8 - 7 9 2 . L o g a n , G . D . (1994). O n the abi l i ty to inhibi t thought and act ion: A users' guide to the stop s ignal parad igm. In D . Dagenbach & T. H . Car r (Eds. ) , Inhibitory processes in attention, memory, and language (pp. 2 1 4 - 2 4 9 ) . San D i e g o : A c a d e m i c Press. L o g a n , G . D . , & B u r k e l l , J . (1986). Dependence and independence in responding to double st imulat ion: a compar ison o f stop, change and dual - task paradigms. 49 Journal of Experimental Psychology: Human Perception & Performance, 12, 5 4 9 - 5 6 3 . L o g a n , G . D . , & C o w a n , W . B . (1984). O n the abi l i ty to inhib i t thought and act ion: A theory o f an act o f control . Psychological Review, 9 1 , 2 9 5 - 3 2 7 . M e g a w , E . D . (1972). D i rec t iona l errors and their correct ion in a discrete t racking task. Ergonomics, 15, 6 3 3 - 6 4 3 . M o r e i n - Z a m i r , S . , & M e i r a n , N . (2003). Ind iv idual stopping t imes and cogni t ive control : C o n v e r g i n g evidence f r o m a new stop s ignal paradigm. Quarterly Journal of Experimental Psychology, 56 , 4 6 9 - 4 9 0 . M o r e i n - Z a m i r , S . , N a g e l k e r k e , P. , C h u a , R., F ranks , I. M . , & K i n g s t o n e , A . F. (2004). Inh ib i t ing prepared and ongo ing responses: Is there more than one k i n d o f stopping? Psychonomic Bulletin & Review, 11, 103 -1040. M o r e i n - Z a m i r , S . , N a g e l k e r k e , P. , C h u a , R., F ranks , I. M . , & K i n g s t o n e , A . F. ( in press). C o m p a t i b i l i t y effects in stopping and response in i t iat ion in a cont inuous t racking task. Quarterly Journal of Experimental Psychology. N i g g , J . T. (2001). Is A D H D a d is inh ib i tory disorder? Psychological Bulletin, 127, 5 7 1 -598 . Ooster laan, J . , L o g a n , G . D . , & Sergeant, J . A . (1998). Response inh ib i t ion in A D / H D , C D , c o m o r b i d A D / H D + C D , anxious, and control ch i ld ren : a meta-analys is o f studies w i th the stop task. Journal of Child Psychology & Psychiatry, 3 9 , 4 1 1 -4 2 5 . O s m a n , A . , K o r n b l u m , S . , & M e y e r , D . E. (1986). The point o f no return in choice reaction t ime: contro l led and bal l is t ic stages o f response preparation. Journal of Experimental Psychology: Human Perception & Performance, 12, 2 4 3 - 2 5 8 . P i s e l l a , L. , A r z i , M . , & Y v e s , R. (1998). The t i m i n g o f co lour and locat ion processing in the motor context. Experimental Brain Research, 121, 2 7 0 - 2 7 6 . Proctor , R., & Reeve , T. G . (1990). Stimulus-response compatibility. A m s t e r d a m : E l sev ie r Sc ience Publ ishers , B . V . Rosentha l , R. (1991). Meta-Analytic Procedures for Social Research. N e w b u r y Park , C a l i f o r n i a : Sage Publ icat ions , Inc. S l a t e r - H a m m e l , A . T . (1960) . R e l i a b i l i t y , accuracy and refractoriness o f a transit react ion. Quarterly Research, 3 1 , 2 1 7 - 2 2 8 . Tannock , R. (2003). Neurophsycho logy o f attention disorders. In S . J . Sega lowi t z & I. R a p i n (Eds.) , Handbook of Neuropsychology (2nd E d i t i o n ed . , V o l . 8, Part II, pp. 7 3 5 - 7 8 4 ) . A m s t e r d a m : E lsev ier , van den W i l d e n b e r g , W . , van B o x t e l , G . J . M . , & van der M o l e n , M . W . (2003) . The durat ion o f response inh ib i t ion in the stop-s ignal parad igm varies w i t h response force. Acta Psychologica, 115 -129 . V i n c e , M . A . , & W e l f o r d , A . T. (1967). T i m e taken to change the speed o f a response. Nature, 2 1 3 , 5 3 2 - 5 3 3 . 51 C H A P T E R 3: Inhib i t ing prepared versus ongo ing responses 3 B y inh ib i t ing a response we can del iberately stop ourselves f r o m enact ing an undesired behaviour , such as v o i c i n g an inappropriate remark. There is m u c h appeal to studying direct measures o f response- inh ib i t ion , as this can lead to insights on how we control our actions, h o w this control changes w i t h age, and h o w it is affected by def ic iencies in response inh ib i t ion ( Logan , 1994; K o k , 1999; K r a m e r , H u m p h r e y , L a r i s h , L o g a n , & Strayer, 1994; Ooster laan, L o g a n & Sergeant, 1998). In the laboratory, stopping has been studied extensively us ing the countermanding procedure ( L a p p i n & E r i k s e n , 1966; L o g a n , 1994; L o g a n & C o w a n , 1984). Part ic ipants receive a response s ignal on every t r ia l , and occas ional l y , a stop s ignal f o l l o w s the go s ignal ind icat ing that the response should be wi thhe ld . B y vary ing the delay between the " g o " s ignal and the " s t o p " s ignal the probabi l i t y o f responding is manipulated : a short delay results in many successful stops, a l o n g delay in few stops (e.g., S l a t e r - H a m m e l , 1960). Per formance in this task is w e l l descr ibed by a horse-race m o d e l between go and stop processes ( L o g a n & C o w a n , 1984; L o g a n , C o w a n & D a v i s , 1984; O s m a n , K o r n b l u m & M e y e r , 1990, but also see M c G a r r y , Ingl is , & Franks , 2000) . I f the go process wins the race then there is a response, but i f the stop process wins no response occurs. B y assuming the two processes have independent f in i sh ing t imes, the race m o d e l a l lows an estimate o f stop s ignal reaction t ime ( S S R T ) to be der ived even though no overt act ion o f stopping was observed (Band , van der M o l e n , & L o g a n , 2 0 0 3 ; L o g a n , 1994). 3 A version of this chapter has been published: Morein-Zamir, S., Chua, R., Franks, I., Nagelkerke, P., & Kingstone, A . (2004). Inhibiting prepared and ongoing responses: Is there more than one kind of stopping? Psychonomic Bulletin and Review, 11, 1034-1040. 52 Interestingly, S S R T s t yp ica l l y converge around 2 0 0 - 3 0 0 ms (De Jong , C o l e s , L o g a n & Gratton, 1995; L o g a n , 1994; L o g a n & C o w a n , 1984). Th i s convergence has led to the conc lus ion that countermanding S S R T measures a generic stopping process ( B a n d & van B o x t e l , 1999; L o g a n , 1994; van B o x t e l , van der M o l e n , Jennings & B r u n i a , 2 0 0 1 ) 4 . Th i s conc lus ion is quite bo ld cons ider ing that to date the evidence has been der ived almost exc lus ive ly w i t h a countermanding task, w h i c h by def in i t ion can only address stopping a response before it is executed. Thus , a more conservat ive conc lus ion based on the extant data w o u l d be that a generic stopping process is engaged when one has to inhibit a response before it is carr ied out. Whether the stopping process that is be ing measured for an unexecuted response is generic to other forms o f stopping is an open question. T h i s issue is important for the development o f theories o f response inh ib i t ion that are grounded i n the countermanding procedure as w e l l as w i t h regard to the general izabi l i ty o f the countermanding procedure (Bark ley , 1997; L o g a n , 1994; L o g a n & C o w a n , 1984; N i g g , 2000) . Indeed recent theories o f inh ib i t ion dist inguish between several subsets o f inh ib i t ion , suggesting that ef fort fu l inh ib i t ion o f motor responses is not a unitary construct ( N i g g , 2000) . Hence , it is possib le that stopping at var ious stages o f act ion p lann ing and execut ion or stopping different actions altogether w o u l d engage different mechanisms and be subject to different constraints. T o measure whether stopping is a generic process, one must measure different forms o f response inh ib i t ion . The present study measured the negat ion o f a response before it is executed (i .e., countermanding) as w e l l as the negation of a response that is already being executed. Th is latter fo rm o f stopping, that by def in i t ion cannot be 4 De Jong and colleagues (1995) suggested a more complex inhibition mechanism involving central and peripheral stopping. However,, their results can also be explained by a single stop mechanism (Band & van Boxtel, 1999; van Boxtel et al., 2001). 53 measured us ing the c lass ic countermanding task, occurs rout inely in everyday l i fe . Whether it is stopping in mid -s t r ide at the edge o f a c rosswalk w h e n the l ight changes to red, or stopping in mid-sentence when the lecturer looks one's w a y , our da i l y l ives are r i fe w i th ongo ing responses that must be halted (see also L a d e f o g e d , S i lverste in & Papcun , 1973). M e a s u r i n g the terminat ion o f an action after it is already underway can be done w i t h a s imple pursuit task ( M o r e i n - Z a m i r & M e i r a n , 2003) . A f t e r a cont inuous t racking response has been init iated a stop s ignal is presented ind icat ing to part icipants that they must stop their t rack ing response as fast as possible . T r a c k i n g S S R T s are obtained by measur ing the t ime f r o m stop s ignal onset to the in i t ia l stopping or decelerat ion o f the ongo ing response (Henry & Har r i son , 1961; M o r e i n - Z a m i r & M e i r a n , 2003) . T o compare the countermanding and pursuit tasks, we matched procedures so that the two tasks di f fered on ly in the type o f stopping that was required. Thus , the perceptual characterist ics o f both tasks were h igh ly comparable , as were the attentional requirements. W e questioned whether the two forms o f stopping shared a c o m m o n response inh ib i t ion mechan ism by examin ing whether stopping per formance correlated across the two tasks. In addi t ion , we examined whether there w o u l d be any consistent differences between the stopping measures. In part icular , we exp lored whether S S R T s w o u l d be o f the same magnitude across tasks, and whether stop s ignal delay w o u l d inf luence S S R T s in the same manner. In sum, the countermanding task is l imi ted to measur ing a part icular type o f stopping - the inh ib i t ion o f an unexecuted response. The assumption however , heretofore untested, is that this countermanding measurement reflects a process o f inh ib i t ion that is 54 c o m m o n to a l l forms o f stopping, inc lud ing the inh ib i t ion o f an executed response. The present study put this c ruc ia l issue to the test. M e t h o d s Participants T e n undergraduates (8 female) part ic ipated; mean age was 19.7 years (5Z)=1.33). A l l had normal or cor rected- to -normal v i s i o n , and a l l but one were r ight -handed. Apparatus and Stimuli In both tasks the d isplay was a green response c i rc le (radius .3°) and.a blue target r ing (radius . 3 - . 4 0 5 0 ) m o v i n g around in an 8° radius c i rcular path. A st imulus f lash consisted o f a white square (.81x.81°) appearing direct ly over the target. S t i m u l i were presented on a 1 4 " V G A - m o n i t o r (640x480 p i x e l resolut ion, 60 H z refresh). Responses were executed us ing a telegraph key instrumented w i t h 2 strain gauges. A n a l o g signals f r o m the strain gauges were a m p l i f i e d ( N o r t h w o o d Instruments, m o d e l I A - 1 0 2 - 5 0 0 ) and sampled at 1000 H z by an A / D converter (Techmar Labmaster ) . Procedure Part ic ipants completed two sessions o f 480 trials on consecut ive days (task order counterbalanced). Stop and go trials were presented randomly , w i th se l f - terminat ing breaks every 48 trials. In the countermanding task each tr ial started w i th the st imul i stationary at the " 3 o ' c l o c k " pos i t ion o f an imaginary c i rc le (see F igure 3.1). A f t e r a random delay o f 2 0 0 - 6 0 0 ms the s t imul i began c i r c l i n g in un ison at 0.5 H z . A 100 ms f lash occurred 2 . 5 - 3 . 5 sec later (determined randomly ) . Part ic ipants were instructed to attend to the s t imul i and press the response key as fast as possib le when the f lash 55 appeared. The key was adjusted so that swi tch closure occurred w i t h 1 mi l l imete r o f key travel w i t h a 100 gram mass. The s t imul i cont inued to rotate unt i l 5 sec elapsed f r o m trial onset. These go trials composed 7 5 % o f the trials. O n the remain ing stop tr ials, the st imul i stopped at equal probabi l i t ies 30 , 110, 190 or 270 ms after f lash onset and remained stationary unti l the trial ended. In each b l o c k o f 48 tr ials, there were 3 stop trials in each delay. Part ic ipants were in formed that on some trials the c i rc les w o u l d stop (the stop s ignal ) and that they were to stop and not press the key . T h e y were also instructed at the beg inn ing and m i d w a y into the session not to delay responses to the f lash in order to improve their chances o f stopping (see L o g a n , 1994, p. 223 for instruct ion details). The sequence o f events in the t racking task was ef fect ive ly ident ica l , w i t h 2 5 % stop trials and the same delays between the f lash and stop s ignal as before. A t the beg inn ing o f the trial the two s t imul i appeared stationary and after the randomly determined per iod the target marker began to move. N o w however , participants contro l led the speed o f the response c i rc le (the target was st i l l cont ro l led by the program). The response c i rc le speed increased when participants increased their pressure on the key . A f i xed pressure o f 400 grams y ie lded the speed o f 0.5 H z , m i n i m u m pressure was 100 grams (0.125 H z ) and m a x i m u m pressure was 2 0 0 0 grams (2.5 H z ) . Part ic ipants were instructed to move the response st imulus to over lap the target, but also to stop as fast as poss ib le i f the target stopped. S topp ing was stressed as re leasing the pressure, without l i f t ing the f inger f r o m the key to ensure m a x i m a l s imi la r i t y between resident and remote effects ( H o m m e l , M i i s s l e r , Aschers leben & P r i n z , 2 0 0 1 , see d iscuss ion below) . The f lash occurred as in the countermanding task, but was irrelevant to the task. 56 Figure 3.1. The sequence of events on a stop trial during the stop signal task. The stimulus and response rings rotate along an imaginary circle, illustrated by the dashed line. Participants are instructed to press the key in response to the flash. The stimuli continue to rotate until the end of the trial. On stop trials only the events in the last frame occur where the stimuli stop abruptly, indicating that participants should withhold their keypress response. The tracking task is highly similar, however participants control the response stimulus speed by pressing on the key. On stop trials in this task, the stimulus ring stops, signaling that participants should stop pressing the key. Trial Onset Stimuli rotate for 2.5-3.5 sec 100 ms Flash Participants Press the Key Target Continues to Rotate If target stops (after a delay), participants should stop and not press the key . 57 The m a i n dif ference between the procedures was that i n the countermanding task, response and target s t imul i were a lways a l igned, whereas i n the pursuit task participants contro l led the response st imulus al ignment w i t h the target. T y p i c a l l y there was considerable over lap between the two s t imul i as participants t racked successful ly . A d d i t i o n a l l y , the f lash in the countermanding task served both as a go s ignal and a temporal w a r n i n g st imulus that the stop s ignal might occur soon. In the t rack ing task, the f lash served on ly as a warn ing st imulus for the stop s ignal . Results and D i s c u s s i o n Computing the dependent measures The nature o f the two tasks demanded that different data be co l lected . In the countermanding task, g o - R T s to the f lash were measured for each part ic ipant, as was probabi l i ty o f response in each delay (mean inh ib i t ion probabi l i ty was .53). S S R T was estimated by first rank order ing g o - R T s i n trials where no stop s ignal was presented, then determining the n t h react ion t ime, where n is the probabi l i ty o f responding i n a g iven delay mul t ip l ied by the number o f g o - R T s . Th is produced an estimate o f the t ime required to stop, relative to the onset o f the f lash. B y subtracting the delay, S S R T was obtained for that part icular delay (Logan , 1994). F o r example , i f the probabi l i t y to respond at the delay o f 30 ms o f participant 1 was .15 , then the S S R T for that delay was calculated by subtracting the delay value f r o m the 1 5 t h percenti le o f the g o - R T (e.g., subtracting 30 ms f r o m 300 ms , y i e l d i n g 270 ms) . Thus , S S R T s for ind i v idua l delays were not calculated by averaging. S S R T i was determined by averaging across the delays. A n alternative method to estimate S S R T subtracts the mean o f the inh ib i t ion funct ion 58 f r o m mean g o - R T to y i e l d S S R T 2 . The mean o f the inh ib i t ion funct ion was computed by m u l t i p l y i n g each stop s ignal delay w i th the probabi l i ty o f responding at the i t h delay minus the probabi l i t y o f responding at the i - l t h delay ( L o g a n , 1994). The two S S R T s were then averaged to attain the most rel iable measure o f countermanding S S R T (Logan & C o w a n , 1984) 5 . Figure 3.2. A response force profile of a representative stop trial in the tracking task. At first, as the target was stationary and the participant rested her finger, the response pressure was below the minimum necessary for motion. After the target began to move, the participant pressed yielding a steep elevation in pressure that was maintained as the target continued to rotate. Following the stop signal, the pressure decreased abruptly, falling to a resting level until the end of the trial. The vertical lines (from left to right) represent the flash onset, the stop signal onset, time detected by the algorithm as SSRT and time detected as Final RT. 600 500 H 400 300 200 100 0 Stop signal onset S S R T Flash onset Final RT 0 0.5 1 1.5 — i 1 1 — 2 2.5 3 3.5 4 4.5 5 5 To compute SSRTs using the race-model one must assume independence between stop and go processes. As the current go task is different from the more common choice tasks, we tested for independence between go-RTs and failures to respond (signal-respond) RTs (Logan, 1994; although see Band et al., 2003 for a critique of independence tests). The results revealed that stop-respond RTs were significantly faster than go-RTs, and that stop-respond RTs increased at longer delays in accordance with independence predictions. Furthermore, there were no significant differences between the observed stop-respond and stop-respond values predicted from go-RTs at the longest 3 delays (the low number of observations in the shortest delay precluded it from analysis). 59 In the t rack ing task, a prof i le o f the pressure on the response key was attained for each tr ial (see F igure 3.2). T r a c k i n g performance was measured by P E (phase error), i.e., the separation in degrees, a long the diameter o f the imaginary c i rc le , between the target and response s t imu l i . P E was sampled at 1000 H z for 1 sec before the f lash and averaged for that durat ion. Hence , the smal ler the P E , the better participants were at t rack ing the target. O n stop trials T r a c k i n g S S R T measured the t ime f r o m stop s ignal onset to the in i t iat ion o f pressure offset, and F i n a l R T measured the t ime f r o m stop s ignal onset to the terminat ion o f response pressure. The a lgor i thm comput ing T r a c k i n g S S R T used a threshold o f 250 grams to f i n d the m i d d l e o f the offset curve and then w o r k e d backwards to find a slope value o f less than 1 0 % peak slope. The same a lgor i thm detected F i n a l R T by m o v i n g fo rward f r o m the curve center to find a slope value o f less than 1 0 % o f peak s lope 6 . Is there a shared response-inhibUion mechanism: Do the SSRT measures converge? Corre lat ions were computed between a l l dependent measures (see Tab le 3.1). The cruc ia l result is the s igni f icant posit ive correlat ion o f .84 between countermanding S S R T and T r a c k i n g S S R T (for a l l s ignif icant correlations, p<.05)7. A person w h o is fast and eff ic ient at stopping an act ion before response in i t iat ion is fast and eff ic ient at stopping an act ion already be ing executed. The h igh correlat ion prov ides c o m p e l l i n g evidence that performance in the two tasks was mediated by the same inh ib i tory m e c h a n i s m . Th is result supports the conc lus ion that the response inh ib i t ion measured here applies across a 6 In agreement with Morein-Zamir and Meiran (2003), final RT (the time to complete a stop) was not a reliable measure of response inhibition in the tracking task , correlating significantly only with tracking PE (r=.83,e<.05). 7 Interestingly/tracking SSRT had higher correlations with both countermanding SSRT measures (r=.81 and r=.75) than countermanding SSRTs had between themselves (r=.72) indicating that Tracking SSRT is highly reliable. 60 variety o f situations and tasks ( L ogan , 1994; van B o x t e l et a l . , 2001) . G i v e n this corre lat ion, it is noteworthy that the go processes ( t racking P E and go R T ) were different w i th the correlat ion between them l o w and non-s ign i f icant . M o r e o v e r , the correlations Q between each go measure and a l l other measures were d iss imi la r . The differences between the go measures are important as they demonstrate that the tasks were distinct. Th is in turn emphasizes the f ind ing that the stop measures were h igh ly correlated despite the task differences. Table 3.1. Correlations between the dependent measures, of both the countermanding task and the continuous tracking task. 2 3 4 5 6 M SD 1. Countermand ing .53 .61 .61 .61 - . 1 2 432.8 38.3 G o - R T 2. Countermand ing .72* .93* . 8 1 * .49 237.6 24.1 S S R T , 3. Countermand ing . 9 2 * .75* .34 223.5 22.8 S S R T 2 4. Countermand ing .84* .45 230.5 21.7 M e a n S S R T 1 2 5. Cont inuous - . 3 1 " 261.4 27.5 T r a c k i n g S S R T 6. Cont inuous 7.24 3.07 T r a c k i n g Phase Er ro r N o t e - A l l measures are in ms other than continuous t rack ing phase error, w h i c h is in degrees. S S R T , stop s ignal reaction t ime. *p<.05. In spite o f the h igh correlat ion between the S S R T s , it cou ld be argued that the .84 correlat ion measured a general component not speci f ic to response inh ib i t ion , i.e., one that is also present in G o - R T . Indeed, although not s igni f icant , g o - R T d i d correlate pos i t i ve ly w i t h the stopping measures. T o test whether t racking S S R T inc luded a 8 We tentatively propose that the correlation between countermanding SSRT and tracking PE reflects a shared general source of executive control. Tracking necessitates a continuous goal-directed response that is continuously monitored and adjusted (Barkley, 1997). Nevertheless, tracking SSRT accounted for another 32% of countermanding variance over and above this component. 61 component speci f ic to response inh ib i t ion , a hierarchal regression analysis was per formed wi th g o - R T introduced first f o l l o w e d by t racking S S R T . G o - R T accounted for 3 7 % o f the var iance in countermanding S S R T , w i t h t racking S S R T account ing for an addit ional 3 5 % (p<.05). In contrast, w h e n t rack ing S S R T was introduced first to the regression, 7 1 % o f the var iance o f S S R T was accounted for. Th is demonstrates that the two S S R T measures are sensit ive to a m e c h a n i s m that appears to be speci f ic to response inh ib i t ion . Are there meaningful differences between the tasks? One might perhaps argue that the h igh correlat ion between countermanding and t rack ing S S R T s occurred mere ly because the tasks d i d not d i f fer i n a mean ingfu l way. F o r instance, one might suggest that the manipu lat ion o f task fa i led to engage different forms o f stopping, as once a t racking response is init iated it is " l o c k e d i n " and no longer constitutes an ongo ing response. T w o analyses address, and reject, this idea. F i rst , i f one response is mere ly prepared and the other actual ly ongo ing , it f o l l o w s that S S R T should be s ign i f icant ly longer for the latter because o f the addit ional t ime required to phys ica l l y implement the inh ib i t ion o f an ongo ing response (e.g., M i l l s , 1999). A t-test compar ing countermanding S S R T and t racking S S R T was s igni f icant , £(9)=6.5,/><.001, ref lect ing that t rack ing S S R T was s lower than countermanding S S R T by approx imate ly 30 ms (Table 1). The second analysis examined whether the delay between the go -s igna l and the stop-s ignal had s imi la r effects on countermanding and t rack ing inh ib i t ion . It is we l l establ ished that in the countermanding task, S S R T m a y decl ine s ign i f icant ly as the delay between the go and stop signals lengthens. Th is decl ine is general ly understood to reflect the race between go and stop responses. A s the delay increases, on ly faster stop latencies 62 w i n the race and s lower ones contribute less to S S R T (Band et a l . , 2 0 0 3 ; L o g a n & B u r k e l l , 1986). F o r a t racking response however , there is no race between go and stop processes because the go response is already occurr ing . A s a result, S S R T should not decl ine as the delay between go and stop signals increases. Th is is the pattern o f results that was found (Table 2). A n analyses o f var iance ( A N O V A s ) revealed that delay had a s igni f icant effect on S S R T , , F ( 3 , 2 7 ) = 6 8 . 4 , / K . 0 1 . A s the delay increased SSRTi decreased, and a l l N e w m a n - K e u l s compar isons were s igni f icant (p<.05). In contrast, longer stop s ignal delays d i d not lead to faster t racking S S R T . In fact, an A N O V A indicated that t rack ing S S R T was shortest at the 30 ms delay , F(3,27)=4.7,£><.01. Th is was conf i rmed by N e w m a n - K e u l s compar isons (no other compar i son was signif icant) . It is l i k e l y that the onset o f the flash (wh ich lasted for 100 ms and was therefore offset in the other delays) led to increased preparedness to stop and alert ing ( S l a t e r - H a m m e l , 1960). Table 3.2. Results of mean stop-signal reaction times (SSRTs, in milliseconds), as a function of delay in the Countermanding and Tracking tasks. D e l a y Task 30 110 190 270 Countermand ing 308 253 212 178 Cont inuous T r a c k i n g 243 266 262 274 Interestingly, the t rack ing task also sheds l ight on the inf luence o f delay on countermanding S S R T (e.g., L o g a n & B u r k e l l , 1986; L o g a n & C o w a n , 1984). A l ternat ive accounts to the race-horse account descr ibed above attribute the decl ine in 63 S S R T w i th delay to a refractory effect at short St imulus Onset A s y n c r h o n i e s ( S O A s ) or an increased preparedness to stop at longer delays ( B a n d et a l . , 2 0 0 3 ; L o g a n & B u r k e l l , 1986). These alternatives predict that t racking S S R T shou ld also decrease as delay increases, w h i c h it d i d not do (although one might argue that a refractory effect was absent because the f lash was not suff ic ient ly response-relevant in the t rack ing task). In sum, it seems reasonable to conclude that although the two S S R T measures appear sensit ive to the same inh ib i t ion process, they are susceptible to different processes afforded by the part icular context and demands o f the task. In addit ion to addressing different hypotheses set forth f r o m the countermanding procedure the t rack ing task has several useful advantages. F o r example , it does not require the assumption that stop and go processes are independent and S S R T data is col lected di rect ly on every tr ial ( M o r e i n - Z a m i r & M e i r a n , 2003) . Th is latter advantage m a y prove part icu lar ly useful when testing under t ime constraints, as fewer trials are needed. Furthermore, measures o f t racking S S R T var iab i l i ty are easi ly accessible. Recent l y , B a n d and col leagues (2003) indicated that the countermanding procedure cannot prov ide a reasonable estimate o f response inh ib i t ion var iab i l i t y , w h i c h is be l ieved to be indicat ive o f def ic ient response inh ib i t ion such as that found in A D H D (Oosterlaan e t a l . , 1998). Response inhibition in the tracking task It cou ld be argued that the release o f force measured by T r a c k i n g S S R T should be interpreted as an act ion rather than the terminat ion o f an ongo ing act ion. Th i s is especia l ly true for the current vers ion o f the t racking task, m o d i f i e d f r o m M o r e i n - Z a m i r and M e i r a n (2003). Consistent w i th this argument, N a i t o and M a t s u m u r a (1996) found no 64 signi f icant dif ference when countermanding constituted ref ra in ing f r o m a button press versus refra in ing f r o m the release o f a pressed key . In the present study, detracting pressure had important v i sua l consequences: the speed o f the response st imulus decreased, and w h e n pressure returned to basel ine (resting force) the st imulus stopped. There are several reasons w h y stopping here can be considered response inh ib i t ion . F i rst , the c o d i n g o f the response is d i rect ly and cont inuously contingent on the consequences o f the actions o f the part icipant. In part icular , releasing the pressure causes the response marker to stop m o v i n g . Furthermore, by coup l ing resident effects (the propr iocept ive and tacti le cues f r o m the f inger) w i t h the remote effects (the movement o f a v i sua l st imulus) participants w o u l d perceive increased pressure as an act ion and the release o f pressure as a terminat ion o f this act ion ( H o m m e l et a l . , 2001) . F o r the same reason, part icipants never l i f ted their f inger f r o m the key , ensur ing that they d id not perceive another act ion (stop-change) as stopping the marker . Ladefoged and col leagues (1973) made s imi lar assumptions when studying the terminat ion o f speech in response to an external signal (see also L o g a n , 1982 w h o examined the inh ib i t ion o f typing) . U l t i m a t e l y , the results support our interpretation that participants coded the release o f pressure as an act ion terminat ion. Conc lus ions The present study demonstrated that the countermanding and cont inuous t racking tasks measure a s imi la r stopping process, support ing the hypotheses that response w i t h h o l d i n g and response termination compr ise the same general response inh ib i t ion mechan ism. Th is a l lows one to general ize the f indings o f one task to the other, and lends 65 converg ing va l id i ty to both. Thus , this study supports the not ion that there exists a general m e c h a n i s m for motor response inh ib i t ion . It remains to be seen whether this response inh ib i t ion is part o f a unitary inhib i tory process that general izes to different effectors and act ion complex i t ies . One m a y also quest ion h o w this inh ib i tory process relates to other executive funct ions. O n the one hand, behaviora l inh ib i t ion has been hypothesized to be dist inct f r o m cognit ive inh ib i t ion ( inc lud ing negative p r i m i n g and directed forgetting) and interference control ( inc lud ing Stroop, f lanker and dual tasks) although a l l are considered execut ive, or effortful inh ib i t ion ( N i g g , 2000). A l te rnat i ve ly , there is evidence stressing a general executive funct ion ca l led ' i n h i b i t i o n ' that encompasses motor inh ib i t ion as w e l l as inh ib i t ion found in the Stroop and antisaccade tasks ( M i y a k e , F r i e d m a n , et a l . , 2000). Hence , the relat ionship between the present examples o f response inh ib i t ion and seemingly related tasks remains to be uncovered. 66 References B a n d , G . P. H . , & van B o x t e l , G . J . M . (1999). Inhibitory motor contro l in stop paradigms: rev iew and reinterpretation o f neural mechanisms. Acta Psychologica, 7 0 7 , 1 7 9 - 2 1 1 . B a n d , G . P. H . , van der M o l e n , W . , M . , & L o g a n , G . D . (2003). Horse - race mode l s imulat ions o f the stop-s ignal procedure. Acta Psychologica, 112, 105 -142 . B a r k l e y , R. A . (1997). B e h a v i o r a l inh ib i t ion , sustained attention, and execut ive funct ions: construct ing a u n i f y i n g theory o f A D H D . Psychological Bulletin, 121, 6 5 - 9 4 . C a b e l , D . W . J . , A r m s t o n g , I. T., R e i n g o l d , E. , & M u n o z , D . P. (2000) . Cont ro l o f saccade in i t iat ion in a coutnermading task us ing v isua l and auditory stop signals. Experimental Brain Research, 133, 4 3 1 - 4 4 1 . Cav ina -P ra tes i , C , B r i c o l o , E. , Pr io r , M . , & M a r z i , C . A . (2001). Redundancy gain in the stop-s ignal parad igm: Impl icat ions for the locus o f coact ivat ion in s imple reaction t ime. Journal of Experimental Psychology: Human Perception and Performance, 2 7 , 9 3 2 - 9 4 1 . C o l o n i u s , H . (1990). A not on the stop-s ignal parad igm, or h o w to observe the unobservable. Psychological Review, 97, 3 0 9 - 3 1 2 . D e Jong , R., C o l e s , M . G . H . , & L o g a n , G . D . (1995). Strategies and mechanisms in nonselect ive and selective inhib i tory motor control . Journal of Experimental Psychology: Human Perception and Performance, 21(3), 4 9 8 - 5 1 1 . D e Jong, R., C o l e s , M . G . H . , L o g a n , G . D . , & Gratton, G . (1990). In search o f the point o f no return: the contro l o f response processes. Journal of Experimental Psychology: Human Perception and Performance, 16, 1 6 4 - 1 8 2 . H e n r y , F. M . , & H a r r i o n , J . S . (1963) . Refractor iness o f a fast movement . Perceptual and Motor Skills, 7 3 , 3 5 1 - 3 5 4 . H o m m e l , B . , M u s s e l e r , J . , Aschers leben , G . , & P r i n z , W . (2002). The Theory o f Event C o d i n g ( T E C ) : a f ramework for percept ion and act ion p lann ing . Behavioral Brain Science, 24, 8 4 9 - 8 7 8 . James, W . (1890/1981). The principles of psychology: M a c m i l l a n / H a r v a r d Un ive rs i t y Press. Jennings, J . F., van der M o l e n , M . W . , B r o c k , K . , & Somsen , R. J . M . (1992). O n the synchrony o f stopping motor responses and de lay ing heartbeats. Journal of Experimental Psychology: Human Perception and Performance, 18, 4 2 2 - 4 3 5 . K o k , A . (1999). Var iet ies o f inh ib i t ion : manifestations in cogn i t ion , event-related potentials and aging. Acta Psychologica, 101, 129 -158. K r a m e r , A . F., H u m p h r e y , D . G . , L a r i s h , J . F., L o g a n , G . D . , & Strayer, D . L . (1994). A g i n g and inh ib i t ion : B e y o n d a unitary v i e w o f inh ib i tory process ing i n attention. Psychology and Aging, 9(4), 4 9 1 - 5 1 2 . Ladefoged , P. , S i lverste in , R., & Papcun , G . (1973). Interrupti l i ty o f Speech. The Journal of the Acoustical Society of America, 54, 1105 -1108 . L a p p i n , J . S . , & E r i k s e n , C . W . (1966) . U s e o f a delayed s ignal to stop a v i sua l react ion-t ime response. Journal of Experimental Psychology, 72(6), 8 0 5 - 8 1 1 . L o g a n , G . D . (1982). O n the abi l i ty to inhibi t complex movements : a stop-s ignal study o f typewr i t ing . Journal of Experimental Psychology: Human Perception and Performance, 8(6), 7 7 8 - 7 9 2 . L o g a n , G . D . (1994). O n the abi l i ty to inhibi t thought and act ion: A users' guide to the stop s ignal paradigm. In D . Dagenbach & T. H . Car r (Eds. ) , Inhibitory processes in attention, memory, and language (pp. 2 1 4 - 2 4 9 ) . San D i e g o : A c a d e m i c Press. L o g a n , G . D . , & B u r k e l l , J . (1986). Dependence and independence in responding to double st imulat ion: a compar ison o f stop, change and dual - task paradigms. Journal of Experimental Psychology: Human Perception and Performance, 12(A), 5 4 9 - 5 6 3 . L o g a n , G . D. , & C o w a n , W . B . (1984). O n the abi l i ty to inhib i t thought and act ion: A theory o f an act o f control . Psychological Review(9\), 2 9 5 - 3 2 7 . L o g a n , G . D. , C o w a n , W . B . , & D a v i s , K . A . (1984). O n the abi l i ty to inhibi t s imple and cho ice react ion t ime responses: a m o d e l and a method. Journal of Experimental Psychology: Human Perception and Performance, 10(2), 2 7 6 - 2 9 1 . M c G a r r y , T., & F ranks , I. M . (1997). A horse race between independent processes: E v i d e n c e for a phantom point o f no return in the preparation o f a speeded motor response. Journal of Experimental Psychology: Human Perception and Performance, 23(5), 1533-1542. M c G a r r y , T., Ingl is , T., & F ranks , I. M . (2000). Aga ins t a final ba l l i s t ic process in the control o f voluntary act ion : E v i d e n c e us ing the H o f f m a n n R e f l e x . Motor Control, 4 , 4 6 9 - 4 8 5 . 69 M i l l s , K . R. (1999). Magnetic stimulation of the human nervous system. N Y : O x f o r d U n i v e r s i t y Press. M i y a k e , A . , F r i e d m a n , N . P. , E m e r s o n , M . J . , W i t z k i , A . H . , Hower ter , A . , & Wager , T. D . (2000). The unity and divers i ty o f executive funct ions and their contr ibutions to c o m p l e x " f r o n t a l l o b e " tasks: A latent var iable analys is . Cognitive Psychology, 41, 4 9 - 1 0 0 . M o r e i n - Z a m i r , S . , C h u a , R. , & K ings tone , A . F. ( in Prep) . D i f fe rent tasks same stopping: a compar ison between the stop-s ignal and continuous t rack ing tasks. N a i t o , E. , & M a t s u m u r a , M . (1996). Movement - re la ted potentials associated w i th motor inh ib i t ion under different preparatory states dur ing performance o f two v isua l stop s ignal paradigms in humans. Neuropsychologia, 34, 5 6 5 - 5 7 3 . N i g g , J . T. (2000). O n Inhib i t ion/Dis inhib i t ion in Deve lopmenta l Psychopatho logy : V i e w s F r o m Cogn i t i ve and Personal i ty P s y c h o l o g y and a W o r k i n g Inhib i t ion T a x o n o m y . Psychological Bulletin, 126, 2 2 0 - 2 4 6 . Ooster laan, J . , L o g a n , G . D. , & Sergeant, J . A . (1998). Response inh ib i t ion i n A D / H D , C D , c o m o r b i d A D / H D + C D , anxious, and control ch i ld ren : a meta-analys is o f studies w i th the stop task. Journal of Child Psychology and Psychiatry, 39 , 4 1 1 -4 2 5 . O s m a n , A . , K o r n b l u m , S . , & M e y e r , D . E. (1986). The point o f no return in choice react ion t ime: contro l led and bal l is t ic stages o f response preparat ion. Journal of Experimental Psychology: Human Perception and Performance, 12(3), 2 4 3 - 2 5 8 . O s m a n , A . , K o r n b l u m , S . , & M e y e r , D . E. (1990). Does motor p r o g r a m m i n g necessitate response execut ion? Journal of Experimental Psychology: Human Perception and Performance, 1990(16), 183 -198 . Posner , M . I. (1978). Chornometric explorations of the mind. H i l l s d a l e , N J : L a w r e n c e E r l b a u m Assoc iates . R idder inkhof , K . R. , B a n d , G . P. H , & L o g a n , G . D . (1999) . A study o f adaptive behavior : effects o f age and irrelevant in format ion on the abi l i ty to inhibi t one's actions. Acta Psychologica, 101, 3 1 5 - 3 3 7 . R ieger , M . , & G a u g g e l , S. (1999). Inhibitory after-effects in the stop s ignal paradigm. British Journal of Psychology, 90, 5 0 9 - 5 1 8 . Schachar, R , M o t a , V . L. , G o r d o n , G . D . , Tannock , R., & K l i m , P. (2000) . C o n f i r m a t i o n o f an inh ib i tory control def ic i t in attention-deficit/hyperactivity disorder. Journal of Abnormal Child Psychology, 28, 2 2 7 - 2 3 5 . S l a t e r - H a m m e l , A . T. (1960). R e l i a b i l i t y , accuracy and refractoriness o f a transit react ion. Quarterly Research, 31, 2 1 7 - 2 2 8 . van B o x t e l , G . J . M . , van der M o l e n , M . W . , Jennings, J . R., & B r u n i a , C . H . M . (2001). A psychophys io log ica l analysis o f inhib i tory motor contro l in the stop-s ignal parad igm. Biological Psychology, 58, 2 2 9 - 2 6 2 . 71 C H A P T E R 4: C o m p a t i b i l i t y Ef fects in Stopping and Response Init iat ion in the Cont inuous T r a c k i n g T a s k 9 It has long been k n o w n that in human perceptual -motor performance the characteristics o f both the input in format ion and the motor response are important features (e.g., Fitts & Posner , 1967). In addit ion to the speci f ic characterist ics o f the s t imul i and responses, performance ef f ic iency is heav i l y dependent u p o n the relat ion between the two (Fitts & Posner , 1967). S t imulus response ( S - R ) compat ib i l i t y refers to part icular mappings between s t imul i and responses leading to enhanced performance (for rev iews see H o m m e l & P r i n z , 1997; Proctor & Reeve , 1990). In general , the more s imi la r i t y between s t imul i and response features - perceptual ly , conceptual ly or structural ly - the better the compat ib i l i t y between the two and hence the better the performance ( H o m m e l , 1997; K o r n b l u m , 1994; K o r m b l u m , H a s b r o u c q & O s m a n , 1990; K o r n b l u m & L e e , 1995). F o r the spec i f ic case o f spatial compat ib i l i t y , it has been demonstrated repeatedly that responses to s t imul i in a spatial array are faster and more accurate when the spatial relat ion between the st imulus and response alternatives is compat ib le (Fitts & De in inger , 1954; Fitts & Seeger, 1953). Investigations o f spatial S - R compat ib i l i t y have typ ica l l y invo l ved t w o - c h o i c e react ion t ime ( R T ) tasks in w h i c h v isua l s t imul i are presented to the left or right o f a central f i xat ion point , and discrete responses are made at a left or right locat ion. In compat ib le situations, left and right st imulus locat ions are mapped onto left 9 A version of this chapter is in press: Morein-Zamir, S., Chua, R., Franks, I., Nagelkerke, P., & Kingstone, A. (in press). Compatibility effects in stopping and response initiation in a continuous tracking task. Quarterly Journal of Experimental Psychology. 72 and right responses, respect ively , w h i l e in incompat ib le situations, the arrangement o f s t imul i to responses is reversed (e.g., Broadbent & Gregory , 1962; S i m o n , 1969). A l t h o u g h most studies on compat ib i l i t y have examined spatial S - R d imensions , many compat ib i l i t y effects have also been demonstrated w i t h non-spat ia l situations (e.g., Proctor & van Zandt , 1994; Sanders & M c C o r m i c k , 1993). F o r example , responses to a co lor patch are faster when the response entails press ing a same-co lo r key compared to a different co lor key (e.g., Hedge & M a r s h , 1975). Indeed, compat ib i l i t y effects have been demonstrated for a w ide variety o f S - R relations. F o r instance, compat ib i l i t y effects were found for v o c a l responses to digits and vowels (Sternberg, 1969), for postures such as grasping versus finger-spreading movements (Brass, B e k k e r i n g , W o h l s c h l a g e r & pr inz , 2 0 0 0 ; Sturmer, Aschers leben & Pr inz , 2000) , for actions such as rotat ing a steering wheel in response to a tone (Gu ia rd , 1983; M i c h a e l s & St ins, 1997), and even in s imple R T situations ( H o m m e l , 1996). S - R compat ib i l i t y effects are t yp ica l l y re lat ively large in magnitude and very stable, and consequently they have often been used to examine the relat ionships between percept ion and act ion ( H o m m e l 1997; A l l u i s i & W a r m , 1990). The present study investigated whether compat ib i l i t y effects can be found for the seemingly s imple responses o f act ion in i t iat ion and stopping. The mot ivat ion for our inquir ies or ig inated f r o m our interest in how ind iv iduals stop their actions. Stopping Research on act ion inh ib i t ion , that is , h o w ind iv iduals stop their actions, has usual ly been examined in the laboratory us ing the stop-s ignal pa rad igm ( L a p p i n & E r i k s e n , 1966; L o g a n , 1994). In this countermanding parad igm, part icipants per form a s imple or choice R T task. O n a subset o f trials a subsequent st imulus (the stop signal) is 73 presented, indicating that the planned response is to be stopped. That is, participants are to refrain from carrying out the planned action. Using this paradigm and a model that assumes a race between stop and go processes, researchers estimate the time necessary for stopping the planned action (Logan & Cowan, 1984). Research utilizing the stop-signal paradigm has suggested that stimuli signaling an interruption or the modification of some parameters of the response have privileged access to the response system. In contrast, stimuli that involve a new response require additional processing from the system (Logan & Cowan, 1984). In addition, different mechanisms appear to underlie action initiation and stopping. To cite just three examples, 1) go and stop measures exhibit different developmental trends (Bedard et al., 2002; Williams et al., 1999), 2) a deficient stopping process but not a deficient go process has been implicated in the performance of patients with disorders such as A D H D (attention deficit hyperactivity disorder, Oosterlaan, Logan & Sergeant, 1998) and, 3) unlike go measures, stopping has been found to be resistant to many cognitive processes such as increased practice, and expectations (Logan, 1994; Logan & Burkell, 1986). Nevertheless, it has also been acknowledged that the stop-signal paradigm and its accompanying race model have not adequately addressed the nature of the stopping process (cf. Logan, 1994; van den Wildenberg, van Boxtel & van der Molen, 2003). This is in part due to the significant fact that the stop-signal paradigm only provides estimates of a particular form of stopping response, that is, the stopping of a planned action. The countermanding paradigm does not avail itself easily to the examination of stopping an action after response execution is initiated. 74 Recently, we developed a continuous tracking task that allowed us to observe and measure stopping directly (Morein-Zamir, Nagelkerke, Chua, Franks & Kingstone, 2004; in press). This task measures the time required to stop an ongoing tracking response. On each trial, the participant first engages in a manual tracking task. Several seconds later, a stop-signal occurs and the participant is required to perform a speeded stop of this action (Morein-Zamir & Meiran, 2003). In the tracking task, participants control the speed of a response marker via a force-pressure sensor and are asked to track a moving target with the response marker. Previously, it had been proposed that interrupting an ongoing action would engage the same processes that are involved in preventing a planned action (e.g., Logan, 1982; Logan & Cowan, 1984). In strong agreement with this proposal, Morein-Zamir and colleagues found a high correlation of .84 between stopping times as measured by the tracking task and the countermanding task. Thus, it appears that stopping an ongoing response is mediated by similar mechanisms as stopping a planned response (Morein-Zamir et al., 2004). The present study set out to examine whether responses such as stopping an ongoing action are given to the same types of constraints such as action initiation, thus shedding light on the stopping process. This goes to establishing boundary conditions for the differentiation of the stopping process from the go process. The current study tested this issue by examining whether stopping is subject to S-R compatibility effects. Because S-R compatibility effects have been found for many different types of action initiations, this provided a particularly strong test of whether stopping and action initiations share 75 common constraints. Consequently, we examined whether action-based S-R compatibility effects would be evident in the initiation and termination of movements10. The stimuli were the starting or stopping of a visual marker, and responses were the initiation or stopping of a tracking response. We asked whether one type of stimulus signal is more effective for stopping a given action. In particular, does the degree of correspondence between the stop stimulus and stopping response play a role in executing a more efficient stopping response? Note that this question of compatibility has not been considered previously within the countermanding literature. S-R compatibility effects have generally been attributed to the cognitive codes used to represent stimulus and response sets, and thus implicate cognitive processes as the basis for compatibility effects (Proctor & Reeve, 1990). Given that S-R compatibility characterizes so many different responses and actions, we questioned whether stopping an ongoing action would also be susceptible to such an effect (Kornblum et al., 1990; Wallace, 1971). Although models describing action and goal codes do not explicitly refer to stopping, they can readily be generalized to any kind of perceivable action outcome. For example, Hommel and colleagues have proposed a theory of event coding (TEC, Hommel, Musseler, Aschersleben & Prinz, 2001) aimed at linking perception and action. Specifically, they have suggested that perceptions and actions are represented by feature codes referring to distal (external) event features that, among other things, can be the goal or consequence of an action (Hommel, 1997; Hommel, et al., 2001; Prinz, 1997). Thus, action codes can specify the intended outcome of an action. This framework can be expanded to include stopping. Given that S-R compatibility effects arise from the 1 0 In this chapter I use the terms 'stopping' and 'termination' interchangeably. 76 activation of a common code for the stimulus and response in a compatible situation, compatibility effects would be predicted to occur for both initiation responses and stopping responses (Hommel, 1997; Prinz, 1997). Thus, on the one hand, models such as TEC would suggest that actions are controlled by representations of action effects (Greenwald, 1970a,b; 1972; Hommel et al., 2001; Prinz, 1997). In this sense, stopping can be represented internally very much like an initiation response. On the other hand, the stopping literature considers the stopping process to be independent and unique from the go process (e.g., Logan, 1994). This lends support to the notion that stopping may not necessarily be governed by the same rules as go responses and their associated processes. For instance S-R translation or the activation of a response code may not be necessary for response terminations, as stopping does not necessarily entail carrying out a response but merely resorting to the interruption of one. Thus, insofar as stopping is unique it follows that unlike other types of responses it would not be sensitive to compatibility effects. In the following experiments, we examined whether compatibility effects would be evident for stopping a response. Experiment 1 Experiment 1 contrasted two types of S-R mapping conditions. In the first S-R mapping condition, a procedure similar to that used by Morein-Zamir and colleagues (2004, in press) was employed. This entailed a typical pursuit task (Sanders & McCormick, 1993). As illustrated in Figure 4.1 A , target and response marker stimuli were first stationary on a computer screen. Then, the target began to move. Participants responded by pressing a response key to track the target as fast as possible. Specifically, 77 they controlled the speed of the response marker by pressing on a response key so that the harder the force the faster the response marker moved. After a variable duration, the target stopped abruptly and participants had to stop by terminating their response. In the second S-R mapping condition (see Figure 4. IB), the response requirements were the same. However, the mapping between the stimulus and response was now reversed. Consequently, the visual events were basically opposite to those in the first condition. Initially, target and response markers were in motion. Then, the target marker stopped. This signaled to participants to respond by pressing the response key as fast as possible. On this occasion, the harder the participant pressed the slower the response marker moved. Here, in order to track the target, participants sought to press so as to keep the response marker stationary. After a variable duration, the target began moving and participants responded by stopping their response as fast as possible to allow the response marker to also begin moving. Thus, the mapping between the target and response was changed, but not the stimuli or responses themselves. In the first condition, participants responded to target movement by initiating a press and to a stop by terminating the press. In the second condition, participants responded to target movement by terminating a press and to a stop by initiating a press. In both types of S-R mapping, times to initiate the tracking response (in response to a moving or stopping target) and time to terminate it (in response to a stopping or moving target) were the primary measures. We expected to find an S-R mapping effect for the press initiation response, such that a target moving would be a more effective stimulus to initiate the tracking. What was unknown was whether a similar 78 type of effect would be found for the stopping response and whether the magnitude of this effect would be comparable to that found for response initiation. Methods Participants. Ten undergraduates (7 female) received course-credit for participating in the experiment; mean age was 17.9 years (SD=2A). A l l had normal or corrected-to-normal vision, and 8 were right-handed. Apparatus and Stimuli. The display was a green response circular marker (radius .3 cm) and a blue target ring (radius .3-.405 cm) moving around an 8.1 cm radius circular path. A computer controlled stimulus presentation and response collection. The target and response markers were presented on a 14" V G A monitor (640x480 pixel resolution, 60 Hz refresh). Performance was observed on a separate monitor. Responses were executed using a telegraph key instrumented with 2 strain gauges. Analog signals from the strain gauges were amplified (Northwood Instruments, model IA-102-500) and sampled at 1000 Hz by an A/D converter (Techmar Labmaster). For further details see Morein-Zamir et al. (in press). Procedure. There were two task conditions defined according to S-R mappings. Both task conditions were completed in a 1-hour session. The order of the two conditions was counterbalanced across participants. Participants sat approximately 60 cm from the screen and completed 30 trials in each task. There was a self-terminated break every 15 trials. Each condition began with up to 20 practice trials identical to the experimental trials. Trials using the first mapping condition began with both target and response markers at the "3 o'clock" position of an imaginary circle (see Figure 4.1). After a 79 random delay between 3000-5000 ms the target began to rotate around the imaginary circle at a constant speed. Participants controlled the speed of the response stimulus and were asked to press with their dominant hand on the response key so as to move the response marker to overlap the target. A fixed force corresponding to a 350-gram weight yielded the target speed of 0.5 Hz, minimum force was 100 grams (0.125 Hz) and maximum force was 2000 grams (2.5 Hz). Participants found the tracking easy to do and complied with the instructions. The target rotated between 7000-9000 ms and then stopped abruptly. Participants were instructed to respond as fast as possible when the target either moved or stopped. Stopping was stressed as terminating the force on the key, without lifting the finger from the key. Thus participants did not introduce another competing action, but simply stopped the press. This in turn would cause the response marker to stop on the screen. The target remained stationary for 3000-5000 ms after which the screen was blank and the trial ended. During this period, participants had to merely rest their finger on the response key so as to keep the response marker stationary. Trials with the second mapping condition began with both target and response markers rotating around the imaginary circle at 0.5 Hz . In this task, participants also controlled the speed of the response marker, but force on the response key caused the response marker to slow down. Resting their finger on the key resulted in the response marker rotating at 0.5 Hz . Following a random period of 3000-5000 ms the target stopped abruptly. Participants responded by aligning the response marker to overlap with the target stimulus. After aligning the two markers, constant force of 350 grams would lead to the response marker being stationary. Force above this threshold caused the response marker to reverse and move backwards. Hence, in order to track properly participants 80 again had to maintain the same constant force as in the previous mapping condition. After 7000-9000 ms the target began to move again and participants released the force from the key as fast as possible to allow the response stimulus to move at the same speed as the target. Again, participants rested their finger on the sensor until the end of the trial. Figure 4.1. A depiction of the procedure of Experiment 1. Events in Figure 4.1 A depict the procedure of compatible trials, and events in Figure 4.1 B depict the procedure in the incompatible trials. Figure 4.1A. 4±1 Sec -> 8±1 Sec -> 4±1 Sec Both target and response markers are stationary Target marker moves Participant tracks by controlling the speed of the response marker More pressure -> faster Target stops and participant stops pressing to make response marker stop. 81 Figure 4.1 B. 4±1 Sec 8±1 Sec 4±1 Sec -> Both target and response markers are moving Target marker stops Participant tracks by controlling the speed of the response marker More pressure -> slower (even reverse) Target moves and participant begins pressing to make response marker move. The two task conditions were hence identical in the nature of the response required from the participant, but different in terms of the visual display and the consequences of the participant's response. In both task conditions participants had to refrain from pressing in the first duration, press during the second duration and stop pressing in the final duration. Moreover, in both tasks participants were required to pursue the target with the response marker. Importantly, in the first mapping condition, 82 force on the response key led to movement of the response stimulus, while in the second mapping condition the same force led to a stationary response marker. Results Data Processing. For each trial in the tracking task, a profile of the force on the response key was obtained. Initiate RT was measured as the time interval from target change onset to the initiation of force onset, and Stop RT was measured as the time from target change offset to the initial termination of response force. The algorithm computing the RTs used a threshold of 250 grams to find the approximate middle of the onset and offset curves and then worked backwards to find a slope value of less than 10% peak slope. The performance during the tracking phase was also measured on each trial, and constituted the average force (in grams) applied on the force sensor for 1000 ms before the target signaled a stop. Analyses. A l l response latencies shorter than 100 ms and longer than 600 ms were omitted to remove anticipations and outliers (1.8% of the data). As performance complied with the instructions on every trial no trials were omitted due to incorrect responses. Thus, cell means based on approximately 30 measurements were computed for each subject for the following analyses. Our hypothesis pertained to the compatibility of the assignment of target events to the response events. Hence, the data were plotted as a function of target events and response events". As can be seen in Table 4.1, response initiation RTs were faster in response to the target moving as compared to the target stopping. In contrast, stopping RTs were faster to the target stopping as compared to the 1 1 Note that the first trial type included the compatible conditions for both dependent measures, while the second trial type included the incompatible conditions for both dependent measures. However, as our main hypothesis pertained to compatibility effects, the data were replotted as a function of target signal (move or stop). Different analyses were conducted with compatibility as a second factor instead of target signal and were found to demonstrate the same effects. 83 target m o v i n g . H e n c e , compat ib i l i t y effects were apparent for both response types but in the opposite d i rect ion. A repeated-measures analysis o f var iance ( A N O V A ) was conducted to test these conc lus ions , w i t h the factors o f target event (target m o v e d vs. target stopped) and response event (initiate vs. stop). The analysis indicated a signif icant m a i n effect for response w i t h initiate responses be ing s lower than stop responses, F ( l , 9 ) = 6 . 6 8 , / K . 0 0 1 , but no signi f icant effect for target event, F ( l , 9 ) = 1 . 0 5 , A/S£=262.15. Th is analysis also y ie lded a h igh l y s ignif icant interaction, F ( l , 9 ) = 1 4 9 . 1 5 , p<.00\. Planned compar isons indicated that for in i t iat ing a press, responses were s igni f icant ly faster to a target that began to m o v e as compared to a target that stopped, F ( l , 9 ) = l 13.22, p<.00\. In contrast, stopping responses were s igni f icant ly faster to a target that stopped than to a target that began to m o v e , F ( l ,9 )=30.47,/?< .001. A further compar i son indicated that the magnitude o f the S - R compat ib i l i t y effects for both types o f responses were not s ign i f icant ly different, F ( 1,9)= 1.05, M S £ = 5 2 4 . 3 0 . T r a c k i n g performance was compared between the two mapping, condi t ions, w i th the first and second m a p p i n g condit ions y i e l d i n g mean forces o f 353 grams and 356 grams, respect ively . A n analysis on the square root o f the average force conf i rmed that there was no s igni f icant dif ference, F ( l , 9 ) = 0 . 3 9 , MS is=0 .06 . The standard deviat ions o f the force over the same duration were 25.18 and 22.73 for the first and second m a p p i n g condit ions, respect ively . A compar ison o f the t ransformed values again conf i rmed that there was no signi f icant dif ference between the m a p p i n g condi t ions , F ( l , 9 ) = 0 . 9 9 , MSE=0A2. T o examine whether part icipants actual ly terminated their response, a final analysis compared the mean resting force before target movement to the mean force f o l l o w i n g the complet ion o f the stop response and d i d not find a signif icant di f ference, F ( l , 9 ) = 1 . 1 2 , MS!£=0.05. 84 Table 4.1. Results of Experiments 1 and 2: means and standard errors for initiate and stopping RTs (in ms) as a function of the target events (target moves versus target stops). Target Abso lu te M o v i n g Stopping D i f fe rence Response M SE M SE M SE Exper iment 1 Init iat ion 254 18 306 17 52 5 Terminat ion 275 12 233 10 42 8 Exper iment 2 Init iat ion 273 12 301 8 29 11 Terminat ion 239 8 206 7 33 10 Discussion The results o f Exper iment 1 demonstrated the existence o f S - R compat ib i l i t y effects for act ion in i t iat ion and stopping. Ini t iat ing a press was faster i n response to a target starting versus a target stopping. Converse ly , stopping a press was faster in response to the target stopping versus starting. Hence the results demonstrated compat ib i l i t y effects for both response types. The results o f the analys is also revealed that the magnitude o f the compat ib i l i t y effects for beg inn ing and stopping the press ing response were comparable . Th i s has several potential bearings on the stopping literature. F i rst and foremost, certain s t imul i are indeed more effect ive than others at in i t iat ing a stopping response. Spec i f i ca l l y , the target stopping is more effect ive than the target beg inn ing to move . P rev ious l y , it has been shown that stopping t imes are faster w h e n the st imulus is more salient (Cav ina -Praset i , B r i c o l o , P r io r & M a r z i , 2001) . H o w e v e r , the present result 85 is independent o f st imulus sal iency , as the target m o v i n g led to faster press in i t iat ion responses ind icat ing that it was not less salient overal l than the target stopping. The present f indings do not shed l ight on the under l y ing mechanisms responsible for S - R compat ib i l i t y effects, on ly that stopping is constrained by such mechanisms. Hence , before the stop response can be init iated, the st imulus t r igger ing it undergoes some fo rm o f translat ion ( E i m e r et a l . , 1995; Fitts & Seeger, 1953; Proctor et a l . , 1992; W a l l a c e , 1971). A l te rnat i ve ly , the T E C f ramework w o u l d suggest that instead o f a translation m e c h a n i s m , the S - R compat ib i l i t y w o u l d result f r o m the shared codes o f the st imulus and stopping representations ( H o m m e l et a l . , 2001) . In any case, the results indicate that it is l i k e l y that stopping an act ion is represented in terms o f its goals , or distal effects. In the present case, this w o u l d entail the p rox imal percept ion o f the release o f force by the f inger so that the f inger is perce ived as i m m o b i l e a long w i t h the distal percept ion o f the corresponding lack o f movement o f the response marker . The f i n d i n g o f equivalent rest ing force before and after the act ion supports this proposal . Ip s u m , the results impl icate an important s imi lar i ty between act ion in i t iat ion and stopping, as the effectiveness o f both response types depends at least i n part on the compat ib i l i t y between the s t imul i and the responses made to them. In the present experiment, compat ib le and incompat ib le tr ials were b l o c k e d . Th is p rov ided the opportunity for the format ion and maintenance o f long - te rm associations o f S - R mappings . H o w e v e r , it cou ld also promote special long - te rm strategies ( M a c L e o d , 1991). Such long - te rm strategies may mask potential differences between response init iat ions and stopping. The present results d i d not reveal a di f ference in the magnitude o f the compat ib i l i t y effect for the two types o f responses. It is poss ib le , however , that 86 there are differences between in i t iat ing a response and stopping it, but as the S - R codes c o u l d be developed and mainta ined over t ime, these potential di f ferences were masked. It fo l lows then that i f long - te rm S - R sets were not a l l owed to develop and be mainta ined, any differences between S - R compat ib i l i t y effects for starting and stopping should be uncovered. B y in te rmix ing compat ib le and incompat ib le tr ials, any dif ferences between the responses types should be revealed. Hence , w h e n the codes require more remapping and updat ing, select ive dif ferences m a y appear. Exper iment 2 Exper iment 2 was s imi la r to the previous experiment, w i t h the except ion that compat ib le and incompat ib le trials were n o w intermixed randomly . Part ic ipants d id not k n o w the tr ial type unt i l the tr ial began. Method Participants. T e n undergraduates (6 female) , who d id not part icipate in Exper iment 1, rece ived course-credit for part ic ipat ing in the exper iment ; mean age was 18.9 years (SD=\3). A l l had normal or corrected- to -normal v i s i o n , and 7 were right-handed. Procedure. The procedure was identical in a l l regards to the prev ious experiment, w i t h the f o l l o w i n g exceptions. The trials were randomly in termixed , and practice was p rov ided for both tr ial types at the onset on the experiment. Part ic ipants were instructed to observe the s t imul i on the screen at the onset o f each tr ial and respond accord ing ly . A s in the prev ious experiment, the response was to in i t ia l l y refrain f r o m any movement , then press to track the target and stop the t racking response. A s the in i t ia l response was always 87 to refrain f r o m any movement , participants were able to observe the markers , detect the trial type and per form accord ing ly . Results A l l responses shorter than 100 ms and longer than 600 ms were omitted (2 .3%) and means for the f o l l o w i n g analysis were computed for each subject. The dependent measures o f stop and init iate R T and t racking P E were computed as in Exper iment 1. A s seen in Table 4 . 1 , the results o f Exper iment 2 were s imi la r to those o f Exper iment 1. The target -mov ing st imulus was more effect ive for the init iate response and target-stopping st imulus was more effect ive for the stop response. A repeated-measures A N O V A was conducted to investigate these compat ib i l i t y effects. The first factor was target event (the target m o v e d vs. the target stopped), and the second factor was response (initiate vs. stop). The analysis indicated a s ignif icant m a i n effect for response w i t h init iate responses be ing s igni f icant ly s lower than stop responses, F ( l , 9 ) = 7 2 . 4 5 , j 9< . 0 0 1 . L i k e w i s e , there was a s igni f icant interaction between the two factors, F(\,9)=\ 10.72, p< .001 . P lanned compar isons indicated that for in i t iat ing responses, latencies were faster to a target that began to move as compared to a target that stopped, F ( l , 9 ) = 7 . 4 3 , p < . 0 1 . In contrast, for stopping, responses were faster to a target that stopped than to a target that began to m o v e , F ( l , 9 ) = l 1.78,/?<.01. A compar ison o f the magnitude o f the S - R compat ib i l i t y effects for both types o f responses revealed that they were not s ign i f icant ly different, F ( i , 9 ) = 0 . 0 2 , MSE- 683.16 . T r a c k i n g performance in the first and second m a p p i n g condit ions were 325 and 326 grams, respect ively , w i t h the corresponding standard deviat ions o f 16.46 and 11.29. A g a i n , no s igni f icant differences were found i n the transformed average t racking performance, F ( l , 9 ) = 0 . 3 6 , MSE= .02 , although the standard 88 deviat ion for the first m a p p i n g condi t ion was larger than the second cond i t ion , jp ( l ,9 )=7.92,p<.05. In addi t ion , the difference in rest ing force before and after the t racking was not s igni f icant , F ( l , 9 ) = 3 . 8 6 , M S £ = 8 4 . 2 3 . Discussion Exper iment 2 repl icated the key results o f Exper iment 1, in that compat ib i l i t y effects were observed w i th the target m o v i n g y i e l d i n g faster response ini t iat ions, wh i le the target stopping y ie lded faster response terminations. M o r e o v e r , as in the prev ious experiment, the magnitude o f the compat ib i l i t y effect was equivalent for both types o f responses. The in te rmix ing o f trials fa i led to reveal a s igni f icant d i f ference in terms o f compat ib i l i t y between the in i t iat ion and stopping o f responses. Thus , the associat ion between the st imulus and response codes proved to be quite f lex ib le and cou ld be easi ly updated between trials. C o m p a r i s o n between Exper iment 1 and Exper iment 2 A t first g lance, in te rmix ing the trials fa i led to inf luence the compat ib i l i t y effects for stopping and response in i t iat ion. W e compared the two experiments di rect ly to search for potential differences between them. A s the quest ion pertained to compat ib i l i t y effects per se, an A N O V A compared compat ib i l i t y between experiments. W h i l e mean response t ime d i d not d i f fer between experiments, F ( l , 9 ) = 0 . 7 1 , MSE= 4 4 0 8 . 4 1 , a s igni f icant interaction between Exper iment and compat ib i l i t y , F ( l , 1 8 ) = l 1.1 \,p<.0\, indicated that the compat ib i l i t y effect (across response types) was larger overal l for Exper iment 1 (46 ms) compared to Exper iment 2 (31 ms). Thus , the results o f the analysis lend some support to the hypothesis that the associations between the s t imul i and responses were 89 weakened by in te rmix ing the trials so that the m a p p i n g changed unpredictably . Hence , cont inuous updat ing o f the relat ionship between st imulus and response codes d i d appear to be in f luenced by in te rmix ing the tr ials. F i n a l l y , despite the increase in p o w e r the interaction between response and compat ib i l i t y was st i l l far f r o m s ign i f icance , F ( l , 9 ) = 0 . 7 5 , MSE= 6 0 2 . 6 1 , ind icat ing that the magnitude o f the compat ib i l i t y effect d id not s ign i f icant ly d i f fer for initiate responses and stopping responses. Genera l D i s c u s s i o n The present study has i l lustrated a nove l example o f S - R compat ib i l i t y . Stopping is faster in response to a target hal t ing , as compared to a target beg inn ing to move . L i k e w i s e , response in i t iat ion is faster to a target beg inn ing to m o v e , as compared to a target halt ing. The compat ib i l i t y effects were found consistently across two experiments, w i t h the magnitude o f the effects reduced somewhat w h e n compat ib le and incompat ib le trials were intermixed. A l t h o u g h stopping latencies were faster than response in i t iat ion latencies overa l l , the magnitude o f the compat ib i l i t y effects d i d not di f fer . Thus , it w o u l d appear that equivalent S - R compat ib i l i t y effects exist for stopping and for act ion in i t iat ion responses. Implications for stopping The present results add to a smal l but g r o w i n g body o f ev idence suggesting that many o f the constraints govern ing in i t iat ion o f responses also apply to stopping. F o r example , stopping latencies have been found to be faster to more salient s t i m u l i , demonstrat ing that they are inf luenced by perceptual processes (Cav ina -Praset i , et a l . , 2001) . Furthermore, selective stopping, that is , stopping in response to one signal but not another, resulted in 90 s ign i f icant ly s lower stopping t imes (Riegler et a l . , unpubl ished: in L o g a n , 1994); just as response latencies i n choice tasks and go/no-go tasks are s lower than s imple R T (e.g., van den W i l d e n b e r g et a l . , 2003) . It appears that response select ion processes constrain stop responses in the same manner as they do go responses. The present findings demonstrated that, s imi la r l y to response ini t iat ions, stopping is vulnerable to the speci f ic relat ionship between the st imulus and response codes. Thus, these results demonstrate a boundary condi t ion where attributes o f the stop are s imi la r to the go. A d d i t i o n a l boundary condit ions should be examined to further speci fy the characteristics o f the stop process, such as the extent o f its resistance to expectancies ( L o g a n & B u r k e l l , 1986). A s stated in the int roduct ion, theories o f percept ion and act ion , such as T E C , cou ld be extended to encompass the present results ( H o m m e l et a l . , 2001) . Spec i f i ca l l y , stopping an act ion can also be represented w i th a response code that refers to the perceived consequence o f act ion interruption. That is , since stopping also entails act ion effects such as not press ing a button or an act ion no longer occurr ing , act ion codes c o u l d apply to this f o r m o f inact ion. S u c h an extension to the theory w o u l d suggest that the internal representation o f a stop is susceptible to the same cognit ive contro l appl ied to representations o f other types o f actions. In other words , even act ion inh ib i t ion has action effects (see also Z iess ler & Nattkemper , 2001) . Future studies m a y explore further factors in f luenc ing the c o d i n g o f stopping by manipu lat ing its external consequences. It c o u l d be c l a i m e d that the release o f force is l ike an act ion in and o f itself, analogous to turn ing o f f a l ight by pressing a switch is an act ion. H o w e v e r , there is a c ruc ia l d i f ference between this example and the present case. The t rack ing task necessitates cont inuous ongo ing control o f one's actions. It is not suff ic ient to mere ly 91 initiate a press. Rather, one has to constantly monitor target and marker pos i t ion , and adjust one's force accord ing ly . Thus , at the t ime the target signals that a stop is required, an act ion is act ive ly be ing contro l led and it is this act ion that must be terminated (see also H e n r y & Har r i son , 1961; L a d e f o g e d et a l . , 1973; L o g a n , 1982; M o r e i n - Z a m i r & M e i r a n , 2 0 0 3 ; M o r e i n - Z a m i r et a l . , 2004, in press). The present task demands are also in keeping w i t h one o f the central assumptions behind many studies u t i l i z ing the stop s ignal task. That is , that stopping is a v a l i d measure o f a variety o f internal ly generated acts o f contro l , i n c l u d i n g the mod i f i ca t ion o f ongo ing act ions, w h i c h are abundant i n everyday behaviors (e.g., D e Jong et a l . , 1990; L o g a n , 1994; L o g a n & C o w a n , 1984). B y be ing able to direct ly access and measure the s topping process, the present study raises the poss ib i l i t y that other questions m a y also be accessible i n a w a y that was not immediate ly obv ious when app ly ing the tradit ional s top-s ignal parad igm. Studies us ing the standard methodology have typ ica l l y tr ied to gain insight into the stopping process indi rect ly , by invest igat ing h o w different forms o f inh ib i t ion inf luence estimated stopping R T s . In practice this has meant manipu lat ing characterist ics o f the go process and e x a m i n i n g h o w these inf luence stopping t imes (Logan , 1994). Th is can be done by us ing conf l ic t tasks such as the f lanker task where the target is f lanked by irrelevant s t imul i ind icat ing either the same (congruent) or the compet ing ( incongruent) go response (Er iksen & E r i k s e n , 1974). U s i n g the f lanker task, estimated stopping t imes to incongruent go s t imul i were found to have longer latencies than to congruent go s t imul i (e.g., K r a m e r et a l . , 1994; R i d d e r i n k h o f et a l . , 1999). The research us ing the f lanker task has conc luded that the inh ib i tory processes engaged by act ive ly s topping a response interact w i th the inhib i tory processes pass ive ly tr iggered w h e n suppressing the incorrect 92 go-response act ivat ion ( R i d d e r i n k h o f et a l . , 1999). In contrast, estimated stopping latencies do not appear to be inf luenced by inhib i tory processes engaged in spatial compat ib i l i t y . S topping performance is approx imately equal for spat ial ly compat ib le and spatial ly incompat ib le responses in tasks such as descr ibed in the Introduct ion section ( K o r n b l u m et a l . , 1990; L o g a n , 1981; L o g a n & I rwin , 2000) . These seeming ly inconsistent results cou ld be c lar i f ied i f the properties o f the stop process were examined by manipu lat ing the stop s ignal direct ly , as in the present study. N o t e that we are not advocat ing for the rejection o f the standard stop-s ignal parad igm. Rather we are propos ing that the tradit ional and present approaches can operate in a complementary fashion. F o r instance, the vast major i ty o f s top-s ignal studies have ut i l i zed a stop s ignal task w i th a v i sua l go signal and a s imple auditory stop s ignal ( N i g g , 2001) . The current results demonstrate that v isual stop-st imulus characteristics can faci l i tate or hinder stopping performance. Th is in turn raises the question o f whether manipu la t ing the stop s ignal w i l l also affect terminations o f prepared, but not executed, responses as estimated by the tradit ional paradigm. In accordance w i t h this proposal , a recent study indicated that estimated stopping t imes were faster when the stop s ignal was easy to d iscr iminate and was spatial ly compat ib le w i t h the response to be stopped (van den W i l d e n b e r g & van der M o l e n , 2004) . Implications for S-R compatibility The compat ib i l i t y effect observed in the present study can be regarded as a novel i l lustrat ion o f h o w to l i nk ideomotor compat ib i l i t y theories w i t h cont inuous tasks such as t racking. Ideomotor theory states that act ion performance and contro l are antic ipatory, and contro l led by representations o f intended act ion effects (G reenwald , 1970a, b; 93 H o m m e l et a l . , 2001) . Thus , S - R sets are ideomotor compat ib le i f the feedback f r o m the required response resembles the st imulus. In that sense, the in i t iat ion and stopping o f an act ion should also result in the in i t iat ion and stopping o f that act ion 's results. In the present case, this entai led the response marker beg inn ing and stopping its movement as w e l l as modulat ion o f the ongo ing force corresponding to an immediate change in t racking performance. Thus , compat ib le trials w o u l d be considered h igh l y ideomotor compat ib le . Interestingly, t racking performance i n the incompat ib le cond i t ion , but not the in i t iat ion and terminat ion o f the act ion , suggests that the actual t rack ing was also h igh ly ideomotor compat ib le . Th is is l i k e l y to result f r o m the immediate v isual feedback and the s imp l i c i t y o f the t rack ing task. A n addit ional point o f interest is in regards to the impl icat ions o f the current findings on S - R compat ib i l i t y as found in s imple R T tasks. The vast major i ty o f compat ib i l i t y effects have been examined wh i le us ing choice R T tasks. S - R compat ib i l i t y effects us ing tradit ional s imple R T tasks i n v o l v i n g spatial compat ib i l i t y are typ ica l l y not robust and are very smal l in magnitude ( M a r z i , B i s i a c c h i & N ico le t t i , . 1991). The present task cannot be considered a choice R T task in the tradit ional sense, yet robust compat ib i l i t y effects were noted. A l t h o u g h participants k n e w w h i c h response w o u l d be required and had ample t ime to prepare, they d i d not k n o w when w i t h i n a 2 0 0 0 ms range the response w o u l d be required. The findings converge w i t h H o m m e l ' s (1996) proposal that response readiness and not response uncertainty is a key determinant to compat ib i l i t y effects. H o m m e l also suggested that in s imple R T tasks, compat ib i l i t y effects m a y require at least two responses be held i n continuous readiness. In accordance w i t h this idea, in the present study both st imulus types (go/stop) and both response types (go/stop) were 94 present in a s ingle t r ia l . Thus participants had to code and translate the entire st imulus set w i th respect to the entire response set. It w i l l be interesting to examine the speci f ic roles temporal uncertainty and set uncertainty p lay in both spatial and non-spat ia l S - R compat ib i l i t y effects. F i n a l l y , the present results extend the range o f compat ib i l i t y effects observed w i t h m o v i n g s t imul i (e.g., B o s b a c h , P r i n z & K e r z e l , 2 0 0 4 , 2 0 0 5 ; Ehrenstein , 1994). Conclusion The present results indicate that stopping is susceptible to S - R compat ib i l i t y effects, s imi la r l y to response in i t iat ion. Th is impl ies that the constraints o f S - R compat ib i l i t y apply to response interruption in addit ion to response generation. The f ind ings also indicate that e x a m i n i n g stopping in an ongo ing task can prov ide converg ing evidence to ex ist ing tasks leading to a better understanding o f the stopping process. 95 References A l l u i s i , E . A . , & W a r m , J . S. (1990). Th ings that go together: a rev iew o f s t imulus -response compat ib i l i t y and related effects. In R. W . Proctor & T. G . Reeve (Eds.) , Stimulus-Response Compatibility (pp. 3 -30) . N o r t h - H o l l a n d : E l sev ie r Sc ience Publ ishers B . V . B e d a r d , A . , N i c h o l s , S . , Barbosa , J . A . , Schachar , R., L o g a n , G . D . , & Tannock , R. (2002). The development o f selective inhib i tory control across the l i fe span. , Developmental Neuropsychology, 2 7 , 9 3 - 1 1 1 . B o s b a c h , S . , P r i n z , W . , & K e r z e l , D . (2004). A S imon-e f fec t w i t h stationary m o v i n g s t imul i . Journal of Experimental Psychology: Human Perception and Performance, 30, 3 9 - 5 5 . B o s b a c h , S . , P r i n z , W . , & K e r z e l , D. (2005). Is d i rect ion posi t ion? P o s i t i o n - and di rect ion-based correspondence effects i n tasks w i t h m o v i n g s t imu l i . Quarterly Journal of Experimental Psychology, 58A, 4 6 7 - 5 0 6 . Brass , M . , B e k k e r i n g , H . , Woh lsch lager , A . , & P r i n z , W . (2000). C o m p a t i b i l i t y between observed and executed f inger movements : compar ing s y m b o l i c , spat ia l , and imitat ive cues. Brain and Cognition, 44, 124 -143 . Broadbent , D . E. , & Gregory , M . (1962). Donder 's B - and C - react ions and S - R compat ib i l i t y . Journal of Experimental Psychology, 63, 5 7 5 - 5 7 8 . Cav ina -P ra tes i , C , B r i c o l o , E. , Pr io r , M . , & M a r z i , C . A . (2001). Redundancy gain in the stop-s ignal parad igm: Impl icat ions for the locus o f coact ivat ion in s imple reaction 96 t ime. Journal of Experimental Psychology: Human Perception and Performance, 2 7 , 9 3 2 - 9 4 1 . D e Jong , R., C o l e s , M . G . H . , L o g a n , G . D . , & Gratton, G . (1990). In search o f the point o f no return: the control o f response processes. Journal of Experimental Psychology: Human Perception and Performance, 16, 1 6 4 - 1 8 2 . Ehrenstein , W . H . (1994). The S i m o n effect and v isual mot ion . Psychological Research, 56, 163-169. E i m e r , M . , H o m m e l , B . , & P r i n z , W . (1995). S - R compat ib i l i t y and response select ion. Acta Psychologica, 90, 3 0 1 - 3 1 3 . E r i k s e n , B . , & E r i k s e n , C . W . (1974). Ef fects o f noise letters in the ident i f icat ion o f target letters i n a non-search task. Perception and Psychophysics, 16, 143 -149 . F i t ts , P. M . , & De in inger , R. L . (1954). S - R compat ib i l i t y : correspondence among paired elements w i t h i n st imulus and response codes. Journal of Experimental Psychology, 48, 4 8 3 - 4 9 2 . F i t ts , P. M . , & Posner , M . I. (1967) . Human Performance. B e l m o n t , C A : B r o o k s - C o l e . F i t ts , P. M . , & Seeger, C . M . (1953). S - R compat ib i l i t y : spatial characterist ics o f st imulus and response codes. Journal of Experimental Psychology, 46, 199 -210 . G r e e n w a l d , A . G . (1970a). Sensory feedback mechanisms in performance contro l : W i t h specia l reference to the ideo -motor mechan ism. Psychological Review, 77, 7 3 - 9 9 . G r e e n w a l d , A . G . (1970b) . A double st imulat ion test o f ideomotor theory w i th impl icat ions for selective attention. Journal of Experimental Psychology, 84, 3 9 2 -398. 97 G r e e n w a l d , A . G . (1972) . Sensory feedback mechanisms in per formance cont ro l : w i t h specia l reference to the ideo -motor mechan ism. Psychological Review, 77, 7 3 - 9 9 . G u i a r d , Y . (1983). The lateral cod ing o f rotations: A study o f the S i m o n effect w i th wheel - rotat ion responses. Journal of Motor Behavior, 15, 3 3 1 - 3 4 2 . Hedge , A . , & M a r s h , N . W . A . (1975). The effect o f irrelevant spatial correspondences on two -cho ice response-t ime. Acta Psychologica, 39, 4 2 7 - 4 3 9 . Henry , F. M . , & Har r i son , J . S . (1961). Refractor iness o f a fast movement . Perceptual and Motor Skills, 7 3 , 3 5 1 - 3 5 4 . H o m m e l , B . (1996) . S - R compat ib i l i t y effects without response uncertainty. Quarterly Journal of Experimental Psychology, 49A, 5 4 6 - 5 7 1 . H o m m e l , B . (1997). T o w a r d an act ion-concept mode l o f st imulus- response compat ib i l i t y . In B . H o m m e l & W . P r i n z (Eds.) , Theoretical issues in stimulus-response compatibility (pp. 2 8 1 - 3 2 0 ) . N o r t h - H o l l a n d : E l sev ie r Sc ience B . V . H o m m e l , B . , M u s s e l e r , J . , Aschers leben , G . , & P r i n z , W . (2001). The Theory o f Event C o d i n g ( T E C ) : a f ramework for percept ion and act ion p lann ing . Behavioral Brain Science, 24, 8 4 9 - 8 7 8 . H o m m e l , B . , & P r i n z , W . (1997). Theoretical issues in stimulus-response compatibility. N o r t h - H o l l a n d : E l sev ie r Sc ience B . V . K o r n b l u m , S. (1994). The w a y irrelevant d imensions are processed depends on what they over lap w i th : The case o f Stroop- and S i m o n - l i k e s t imul i . Psychological Research, 56, 130 -135 . K o r n b l u m , S . , H a s b r o u c q , T.,' & O s m a n , A . (1990). D i m e n s i o n a l over lap: Cogn i t i ve basis for st imulus-response compat ib i l i t y - a mode l and taxonomy. Psychological Review, 97(2), 2 5 3 - 2 7 0 . K o r n b l u m , S . , & L e e , J . - W . (1995). St imulus - response compat ib i l i t y w i t h relevant and . irrelevant st imulus d imensions that do and do not over lap w i th the response, Journal of Experimental Psychology: Human Perception and Performance, 21, 8 5 5 - 8 7 5 . K r a m e r , A . F., H u m p h r e y , D . G . , L a r i s h , J . F., L o g a n , G . D . , & Strayer, D . L. (1994). A g i n g and inh ib i t ion : B e y o n d a unitary v i e w o f inhib i tory process ing in attention. Psychology and Aging, 9, 4 9 1 - 5 1 2 . L a d e f o g e d , P. , S i lverste in , R., & Papcun , G . (1973). Interrupti l i ty o f Speech. The Journal of the Acoustical Society of America, 54, 1105 -1108 . L a p p i n , J . S . , & E r i k s e n , C . W . (1966). U s e o f a delayed signal to stop a v isua l react ion-t ime response. Journal of Experimental Psychology, 72, 8 0 5 - 8 1 1 . L o g a n , G . D . (1981). At tent ion , automatic i ty , and the abi l i ty to stop a speeded choice response. In J . L o n g & A . D . Badde ley (Eds. ) , Attention and performance IX. H i l l s d a l e , N J : L a w r e n c e E r l b a u m . L o g a n , G . D . (1982). O n the abi l i ty to inhibit complex movements : a stop-s ignal study o f typewr i t ing . Journal of Experimental Psychology: Human Perception and Performance, 8, 7 7 8 - 7 9 2 . L o g a n , G . D . (1994). O n the abi l i ty to inhibi t thought and act ion: A users' guide to the stop s ignal paradigm. In D . Dagenbach & T. H . Car r (Eds. ) , Inhibitory processes in attention, memory, and language (pp. 2 1 4 - 2 4 9 ) . San D i e g o : A c a d e m i c Press. ( 99 L o g a n , G . D . , & B u r k e l l , J . (1986). Dependence and independence in responding to double st imulat ion: a compar ison o f stop, change and dual - task paradigms. Journal of Experimental Psychology: Human Perception and Performance, 12, 5 4 9 - 5 6 3 . L o g a n , G . D . , & C o w a n , W . B . (1984). O n the abi l i ty to inhib i t thought and act ion: A theory o f an act o f contro l . Psychological Review, 9 1 , 2 9 5 - 3 2 7 . L o g a n , G . D . , & I rw in , D . E. (2000). Don't look! Don' t touch! Inhibi tory contro l o f eye and hand movements . Psychonomic Bulletin & Review, 7, 107 -112 . M a c L e o d , C . M . (1991). H a l f a century o f research on the Stroop effect: an integrative rev iew. Psychological Bulletin, 109(2), 163 -203 . M a r z i , C . A . , B i s i a c c h i , P. , & N i c o l e t t i , R. (1991). Is interhemispher ic transfer o f v i suomotor in format ion asymmetr ic? E v i d e n c e f r o m a meta -analys is . Neuropsycholgia, 29, 1163 -1177. M i c h a e l s , C . F., & Stins, J . F. (1997). A n eco log ica l approach to st imulus- response compat ib i l i t y . In B . H o m m e l & W . P r i n z (Eds.) , Theoretical issues in Stimulus-response compatibility (pp. 333 -360) . A m s t e r d a m : E l sev ie r Sc ience B . V . M o r e i n - Z a m i r , S . , & M e i r a n , N . (2003). Ind iv idual stopping t imes and cogni t ive control : C o n v e r g i n g evidence f r o m a new stop s ignal paradigm. Quarterly Journal of Experimental Psychology, 56, 4 6 9 - 4 9 0 . M o r e i n - Z a m i r , S . , N a g e l k e r k e , P. , C h u a , R., F ranks , I. M . , & K i n g s t o n e , A . F. (2004). Inh ib i t ing prepared and ongo ing responses: Is there more than one k i n d o f stopping? Psychonomic Bulletin & Review, 11, 1034 -1040 . 100 M o r e i n - Z a m i r , S . , N a g e l k e r k e , P. , C h u a , R., F ranks , I. M . , & K i n g s t o n e , A . F. (in press). M e a s u r i n g onl ine vo l i t iona l response control w i t h a cont inuous t racking task. Behavioral Research Methods. N i g g , J . T. (2001) . Is A D H D a d is inh ib i tory disorder? Psychological Bulletin, 127, 5 7 1 -598. Ooster laan, J . , L o g a n , G . D . , & Sergeant, J . A . (1998). Response inh ib i t ion in A D / H D , C D , c o m o r b i d A D / H D + C D , anxious, and control ch i ld ren : a meta-analys is o f studies w i t h the stop task. Journal of Child Psychology and Psychiatry, 39, 4 1 1 -4 2 5 . P r i n z , W . (1997). W h y Donders has led us astray. In B . H o m m e l & W . P r i n z (Eds.) , Theoretical issues in stimulus-response compatibility (pp. 2 4 7 - 2 6 8 ) . N o r t h -H o l l a n d : E l sev ie r Science B . V . Proctor , R., & Reeve , T. G . (1990). Stimulus-response compatibility. A m s t e r d a m : E l sev ie r Sc ience Publ ishers , B . V . Proctor , R. , Reeve , T. G . , & van Zandt , T. (1992). Sal ient-features c o d i n g in response select ion. In G . E. Ste lmach & J . R e q u i n (Eds.) , Tutorials in motor behavior ( V o l . 2 , pp. 7 2 7 - 7 4 1 ) . A m s t e r d a m : N o r t h - H o l l a n d . Proctor , R., & van Zandt , T . (1994). Human factors in simple and complex systems. B o s t o n : A l l y n and B a c o n . \ R i d d e r i n k h o f , K . R , B a n d , G . P. H . , & L o g a n , G . D . (1999) . A study o f adaptive behavior : effects o f age and irrelevant in format ion on the abi l i ty to inhibit one's actions. Acta Psychologica, 101, 3 1 5 - 3 3 7 . 101 R ieg le r , G . L. (1986). O n . thee abi l i ty to inhibi t select ively : M e c h a n i s m s o f response interdict ion in choice reaction t ime. U n p u b l i s h e d M a n u s c r i p t c i ted i n : L o g a n , G . D . (1994). O n the abi l i ty to inhibi t thought and act ion: A users' guide to the stop s ignal parad igm. In D . Dagenbach & T. H . Car r (Eds.) , Inhibitory processes in attention, memory, and language (pp. 2 1 4 - 2 4 9 ) . San D i e g o : A c a d e m i c Press. Sanders, M . S . , & M c C o r m i c k , E. J . (1993). Human factors in engineering and design. N e w Y o r k : M c G r a w - H i l l Inc. S i m o n , R. J . (1969). React ions toward the source o f st imulat ion. Journal of Experimental Psychology, 81, 1 7 4 - 1 7 6 . Sternberg, S. (1969). The d iscovery o f process ing stages: extensions o f Donder 's method. In W . C . K o s t e r (Ed.) , Attention and Performance II ( V o l . 30 , pp. 2 7 6 - 3 1 5 ) . A m s t e r d a m : N o r t h - H o l l a n d P u b l i s h i n g C o m p a n y . Sturmer, B . , Aschers leben , G . , & Pr inz , W . (2000). Correspondence effects w i th manual gestures and postures: a study o f imi tat ion . Journal of Experimental Psychology: Human Perception and Performance, 26, 1746 -1759 . van den W i l d e n b e r g , W . , van B o x t e l , G . J . M . , & van der M o l e n , M . W . (2003). The durat ion o f response inh ib i t ion in the stop-s ignal parad igm varies w i th response force. Acta Psychologica(\\4), 115 -129. van den W i l d e n b e r g , W . , & van der M o l e n , M . W . (2004). A d d i t i v e factors analysis o f inhib i tory process ing i n the stop-s ignal parad igm. Brain and Cognition, 56,~253-266. W a l l a c e , R. J . (1971). S - R compat ib i l i t y and the idea o f a response code. Journal of Experimental Psychology, 88, 3 5 4 - 3 6 0 . 102 W i l l i a m s , B . R. , Ponesse, J . S . , Schachar , R. J . , L o g a n , G . D . , & Tannock , R. (1999). Deve lopment o f inhib i tory contro l across the l i fe span. Developmental Psychology, 35(1) , 2 0 5 - 2 1 3 . Z iess ler , M . , & Nat tkemper , D . (2001). E f fect ant ic ipat ion in act ion p lann ing . In W . P r i n z & B . H o m m e l (Eds. ) , Attention & Performance XIX(pp. 6 4 5 - 6 7 2 ) . O x f o r d : O x f o r d Un i ve rs i t y Press. 103 C H A P T E R 5: C o m p a r i n g stopping w i th another measure o f response contro l : the role o f p red ic tab i l i t y 1 2 O u r ab i l i t y to interrupt or rap id ly m o d i f y an act ion is one o f the ha l lmarks o f cogn i t i ve -motor control (De Jong, Co les , L o g a n & Gratton, 1990; L o g a n , 1994; L o g a n & C o w a n , 1984). Of ten this control a l lows us to direct our actions towards ach iev ing a part icular goal or alter ing the current actions when a prev ious goal is no longer appropriate. T o understand both the cognit ive and motor inf luences on such behaviors, researchers have u t i l i zed a variety o f tasks e x a m i n i n g response inh ib i t ion , one o f the most prominent be ing the stop s ignal (countermanding) task ( L o g a n , 1994; L o g a n & C o w a n , 1984; see also L a p p i n & E r i k s e n 1966). Th is task has been deve loped to address the process o f m o d i f y i n g responses and in part icular stopping them (Logan , 1994). In the stop s ignal task, one is asked to w i thho ld an act ion just before it is executed. A "go s t imulus" s igna l ing for a response to be executed, is occas iona l l y f o l l o w e d by a "stop s t imulus" that signals for the intended act ion to be w i thhe ld . B y v a r y i n g the delay between the go and stop s t imu l i , the stop s ignal task can prov ide an estimate o f the latency to inhibi t the prepared response ca l led stop signal react ion t ime ( S S R T ) . A s no overt movement w o u l d have been observed when participants managed to stop their actions, the est imation o f S S R T is based on a mathematical m o d e l ( B a n d , van der M o l e n & L o g a n , 2 0 0 3 ; L o g a n & C o w a n , 1984; O i l m a n , 1973). I 2 A version of this chapter is submitted: Morein-Zamir, S., Chua, R., Franks, I., Nagelkerke, P., & Kingstone, A . (submitted) Predictability influences stopping and response control. 104 T o date, the stop s ignal parad igm has been used to study m a n y different aspects o f stopping and response inh ib i t ion . F o r example , it has been u t i l i zed to examine developmental aspects o f stopping and its change across the l i fe span (e.g., B e d a r d , N i c h o l s , Barbosa , Schachar, L o g a n & Tannock , 2 0 0 2 ; K r a m e r , H u m p h r e y , L a r i s h , L o g a n & Strayer, 1994; W i l l i a m s , Ponesse, Schachar , L o g a n & Tannock , 1999). L i k e w i s e , it has been central in the study o f inh ib i t ion def ic iencies in special populat ions such as those suffer ing f r o m attention deficit/ hyperact iv i ty disorder ( A D / H D ; N i g g , 2 0 0 1 ; Ooster laan, L o g a n & Sergeant, 1998; Tannock , 2003) . The stop s ignal task has also been employed to ident i fy neural mechanisms that subserve stopping across a broad range o f measurement methodologies f r o m techniques as var ied as single ce l l record ing in non human primates to event related potentials and funct ional magnetic resonance i m a g i n g i n humans ( E R P and f M R I , respect ively ; B a n d & van B o x t e l , 1999; D e Jong et a l . , 1990, D e Jong , Co les & L o g a n , 1995; Hanes , Patterson & S c h a l l , 1998; R u b i a et a l , 2001) . M o r e o v e r , this m o d e l task has been appl ied to obtain a general index o f motor contro l (e.g., D e Jong , et a l . , 1990; M c G a r r y & Franks , 1997; O s m a n , K o r n b l u m & M e y e r , 1986). C lea r l y , the stop signal task has p layed an instrumental role i n the development o f an empi r i ca l and theoretical understanding o f the cognit ive underpinnings o f response adjustment. A t present, current theor iz ing has suggested that stopping possesses three key attributes. F i rst , a single mechanism underl ies a variety o f stopping behaviours ( B a n d & v o n B o x t e l , 1999; L o g a n & C o w a n , 1984, but see also D e Jong et a l . , 1995). Second, stopping is considered a generic instance o f response control in that not on ly does it general ize across different types o f stopping, but also to what are considered other acts o f response control ( Logan , 1994; L o g a n & C o w a n , 1984). E x a m p l e s o f other ^acts o f 105 response control in the present context include error correct ion, and the m o d i f i c a t i o n o f force or d i rect ion o f an ongo ing movement or other instances where an ex is t ing response is purposefu l l y adjusted (e.g., D a y & L y o n , 2 0 0 0 ; H e n r y & H a r r i s o n , 1961; M e g a w , 1972; V i n c e & W e l f o r d , 1967; for a rev iew see L o g a n & C o w a n , 1984). The advantage o f s tudy ing stopping over these other instances o f response adjustments has been that it is a clear and extreme behaviour that is easy to observe and possesses an establ ished protocol o f h o w it is to be measured (Logan , 1994). A th i rd characterist ic o f stopping assumed by current theor iz ing is that stopping is implemented by cogni t ive processes that are distinct f r o m the processes govern ing the go response ( L o g a n & C o w a n , 1984). U n l i k e new responses, w h i c h require s ignif icant process ing, the interruption or mod i f i ca t ion o f a response appears to have pr iv i leged access to the response system (Logan & C o w a n , 1984). Three examples o f empi r i ca l support for this not ion can be seen in that un l ike the go task, stopping appears to be insensit ive to s ignal predictabi l i ty ( L o g a n & B u r k e l l , 1986), as w e l l as to pract ice (Logan & B u r k e l l , 1986; W i l l i a m s et a l . , 1999) and demonstrates unique developmental trends (Bedard et a l . , 2 0 0 2 ; R idder inkhof , B a n d & L o g a n , 1999; W i l l i a m s et a l . , 1999). In sum, i m p l i c i t in m a n y studies is the assumption that stopping engages, in some sense, a general response control mechan ism that is governed by different constraints than those regulat ing the go response (Logan , 1994; L o g a n & C o w a n , 1984; N i g g , 2001) . The p ro found importance and demonstrated ut i l i ty that the stop s ignal task has had is beyond quest ion. Nevertheless, since the research on stopping is based p r imar i l y on one parad igm, theor iz ing lacks the abi l i ty to rely on converg ing evidence f r o m other tasks purport ing to measure the same processes. A s a result the k e y assumptions about 106 the nature o f stopping l isted above have not been direct ly or r igorous ly pursued. In part icular , when cons ider ing the generality o f stopping to other examples o f response adjustment, there is indeed some evidence support ing the not ion that stopping is a good representative as it demonstrates s imi lar reaction t imes (e.g., K u d o & O h t s u k i , 1998; P i s e l l a , A r z i & Rossett i , 1998). H o w e v e r , there is also evidence cha l leng ing this not ion, as differences can be found between var ious types o f response adjustments, l i m i t i n g the general izabi l i ty o f any one measure (e.g., D a y & L y o n , 2 0 0 0 ; V i n c e & W e l f o r d , 1967). In addi t ion , the evidence is unclear regarding the dist inct iveness o f s topping f r o m go response measures. O n the one hand , evidence f r o m domains such as the study o f A D / H D suggest that inh ib i t ing responses is at the core o f the def ic i t and therefore certainly distinct f r o m other types o f behaviours (Bark ley , 1997; Ooster laan et a l . , 1998). O n the other hand, the re lat ively id iosyncrat ic nature o f the stop s ignal task, the l im i ted success at invest igat ing its under l y ing mechanisms ( B a n d et a l . , 2 0 0 3 ; L o g a n , 1994; van den W i l d e n b e r g , van der M o l e n & L o g a n , 2002) and the f indings o f def ic ient go R T s in A D H D ( N i g g , 2 0 0 1 ; Tannock , 2003) , w o u l d suggest that the status o f stopping as distinct f r o m go ing is currently somewhat ambiguous. T o summar i ze , at present the possibi l i t ies del ineat ing the relat ionship between stopping and other measures o f response adjustment m a y not have been adequately explored. It is therefore uncertain to what extent the conclus ions acquired f r o m the stop s ignal literature general ize to other forms o f response adjustment. Fur thermore, the assumption that the processes govern ing stopping are distinct f r o m those govern ing go ing has also not been tested r igorously (Logan , 1994; Tannock , 2 0 0 3 ; but see M o r e i n - Z a m i r , N a g e l k e r k e , C h u a , Franks & K ings tone , in press-b) . It remains unclear whether f indings 107 supporting the assumption that stopping is distinct from going are a result of true characteristics of the stopping process or a result of specific task demands in the stop signal task. Evidence supporting or contrasting the status of stopping would have important ramifications for theories of response inhibition as well as for rehabilitation, diagnosis or treatment of conditions believed to stem from deficient control and inhibition (Barkley, 1997; Nigg, 2001; Sergeant, Geurts & Oosterlaan, 2002). Recently, a new task that measures stopping has been developed (Morein-Zamir, Nagelkerke, Chua, Franks & Kingstone 2004, in press-a, Morein-Zamir & Meiran, 2003). On each trial, participants track a moving target moving along a fixed circular trajectory by controlling the speed of a response marker with the use of a force sensor. The more force administered by the participant the faster the response marker moves. Stopping in this tracking task measures the time required to begin stopping after an action had already been underway for some time. Thus, on each trial the participant first engages in an action (manual tracking task), and may then be required to stop this ongoing action in response to an external stimulus as fast as possible. Using the new task, we were able to test the first assumption that the stop signal task has provided a measure that generalizes well to other forms of stopping (Morein-Zamir et al., 2004). We found a high correlation between measures of stopping on the stop signal and tracking tasks, supporting the view that interrupting a prepared response engages similar processes as stopping an ongoing action. Using the new tracking task, we have also found similarities between going and stopping in that both were influenced equally by manipulations of stimulus-response compatibility (Morein-Zamir et al., in press-b). In this sense, at least some factors 108 determining go ing performance inf luence stopping in m u c h the same way . Th is suggests that stopping m a y be more s imi la r to go ing than prev ious ly be l ieved . The a i m o f the present study was twofo ld . F irst , we u t i l i zed the t rack ing task to examine whether stopping does indeed general ize to another measure o f response adjustment. W e ut i l i zed the continuous t rack ing task to compare two response adjustment measures. The first was the stopping o f the t racking response as descr ibed above. Spec i f i ca l l y , we measured latency o f the in i t ia l force release leading to the stop o f the response marker i n response to an external s ignal . The second measure o f response adjustment required participants to increase their force to accelerate their response marker in response to an external s ignal . The second a i m o f the study was to test whether the two measures w o u l d be inf luenced s imi la r l y by the frequency o f the signals to stop or accelerate. One o f the most s t r ik ing pieces o f evidence in the stop s ignal literature support ing the distinctness o f stopping f r o m go ing has been that stopping is insensit ive to manipulat ions o f predictabi l i ty ( L o g a n , 1981; L o g a n & B u r k e l l , 1986; Ramautar , K o k & R idder inkhof , 2004) . Spec i f i ca l l y , L o g a n (1981) manipulated stop s ignal predictabi l i ty by va ry ing the frequency f r o m 1 0 % to 2 0 % and found that this d id not affect S S R T s . L i k e w i s e , L o g a n and B u r k e l l (1986) found no inf luence o f stop s ignal probabi l i ty on stopping speed when presenting the stop signals on 2 0 % , 5 0 % or 8 0 % o f the trials. Ramautar and col leagues (2004) repl icated this f i n d i n g w i th stop s ignal probabi l i t ies o f 2 0 % and 5 0 % . Th i s is in complete contrast to go R T s , for w h i c h extremely robust expectancy, effects (also named 'relat ive frequency effects' ) have been repeatedly reported (e.g., Ber te lson & T isseyre, 1966; D i l l o n , 1966; H a w k i n s & H o s k i n g , 1969; H y m a n , 1953). N a m e l y , the higher the 109 frequency o f a st imulus- response pair in a set o f tr ials, the faster and more accurate the response latencies. M a n i p u l a t i n g the predictabi l i ty o f the stop and accelerate signals a l l o w e d us to contrast the two types o f response adjustments, w h i l e testing whether they demonstrated a resistance to predictabi l i ty . Thus , we examined whether the previous finding o f stopping be ing resistant to the frequency o f the stop s ignal w o u l d general ize to other forms o f response contro l , namely response modi f i ca t ion result ing in acceleration o f the response marker . Th i s manipulat ion w o u l d also prov ide a test o f the assumption that stopping in part icular and response adjustment measures in general , are indeed distinct f r o m measures o f go R T s . The predict ion based on the prev ious f indings w o u l d be that both w o u l d remain unaffected by st imulus predictabi l i ty . O v e r the course o f the f o l l o w i n g exper iments, w e d iscovered that when a t radeoff was possible between stopping and accelerat ion, on ly the accelerat ion response was vulnerable to predictabi l i ty effects such that responses were faster w i th increasing frequency (Exper iment 1). H o w e v e r , w h e n a t radeoff between the two was prec luded by r e m o v i n g the compet i t ion between them, both stopping and accelerat ion appeared to be s imi la r l y inf luenced by predictabi l i ty (Exper iments 2 and 3). A d d i t i o n a l sequential and w i th in - t r ia l analyses examined expectancy on a br iefer t ime scale and revealed that predictabi l i ty effects s imi la r l y inf luenced stopping and accelerat ion. Hence , on ly under very speci f ic c i rcumstances does stopping appear to be resistant to predictabi l i ty . Exper iment 1 The present experiment set out to compare accelerat ion and stopping by e x a m i n i n g h o w they are in f luenced by predictabi l i ty . B a s e d on prev ious findings and the 110 p reva i l ing assumptions in the stopping literature, it was expected that both w o u l d be resistant to predictabi l i ty , as they are both response adjustments, considered instances o f response control ( L o g a n & C o w a n , 1984). Spec i f i ca l l y , on each t r ia l , part icipants tracked the target by cont ro l l ing the speed o f the response marker . A f t e r t rack ing for several seconds, on each tr ial a h igh or l o w auditory tone was presented ind icat ing that the part icipant was to stop or accelerate the response marker as fast as poss ib le . Stopping the marker was done by stopping the force on the key without overt movement , w h i l e accelerat ing the marker was done by increasing the force on the key . Thus , participants had to d iscr iminate between the two possib le signals and chose one o f two response modi f i cat ions that led to the appropriate change o f the ongo ing response. The frequency o f the accelerate s ignal and stop s ignal were var ied between b locks so that each o f the f o l l o w i n g values was presented for each: 2 5 % , 5 0 % , 7 5 % and 1 0 0 % . The remain ing trials in each b lock consisted o f the alternate s ignal . Thus far, we have discussed on ly predictabi l i ty as a factor that m a y be used to examine the inf luence o f expectancy on stopping and accelerat ion performance. S i m i l a r l y to prev ious studies (e.g., L o g a n , 1981) this was done across b l o c k s , as the frequency o f var ious signals was manipulated. H o w e v e r , expectancy cou ld be examined on a m u c h briefer t ime scale than across b locks . First , sequential effects, or after-effects, cou ld be examined , as expectancies have been a key concept in exp la in ing them (Berte lson, 1961; Soetens, B o e r & H e u t i n g , 1985). It has long been establ ished that response go performance is in f luenced by the part icular s ignal - response pair preceding it (Bertelson, 1965; Rabbit t & V y a s , 1973, 1974). The after-effects o f stopping on G o R T s have been observed in several studies (Cabe l et a l . , 2 0 0 0 ; R ieger & G a u g g e l , 1999). Genera l l y , a I l l robust effect is noted where G o R T s f o l l o w i n g a stop tr ial are m u c h s lower than f o l l o w i n g a go t r ia l . Unfor tunate ly , us ing the stop s ignal task it is not poss ib le to observe sequential effects on S S R T s as these are computed across b locks o f tr ials ( B a n d et a l . , 2 0 0 3 ; L o g a n & C o w a n , 1984). In the present study, the effects o f the prev ious tr ial cou ld be examined on stopping performance in the present tr ial as S T O P R T s c o u l d be computed on each tr ial ( M o r e i n - Z a m i r & M e i r a n , 2003) . The second w a y expectancy cou ld be examined was by observ ing the inf luence o f tr ial length. It was hypothesized that as tr ial length increased a fo reper iod - l ike effect w o u l d be observed, since expectancies that a s ignal was imminent w o u l d n o w increase ( N i e m i & Naatanen, 1981; W o o d r o w , 1914). In the present case, a l l s ignal trials were d i v i d e d post -hoc into shorter trials (wi th t rack ing durat ion ranging f r o m 3 - 4 sec) and longer trials (wi th t racking duration ranging f r o m 4 - 5 sec). It shou ld be noted that dur ing the exper iment tr ial length was presented on a cont inuum. Thus the d i v i s i o n into long and short trials is ar t i f i c ia l , conducted to highl ight another possib le m e c h a n i s m by w h i c h expectancies in f luenced performance. P rev ious l y , M o r e i n - Z a m i r and M e i r a n (2003) found that w h e n tr ial length was b l o c k e d , shorter tr ial length led to faster stopping, consistent w i t h the not ion that expectancies increase as stopping becomes more frequent. H o w e v e r , w h e n tr ial length was m i x e d (as in the present case), and the average frequency o f stop signals remained the same throughout a b lock , no effect for tr ial length was observed. In any case, g iven the many differences between the two studies (e.g., average tr ial length, t rack ing response and v isual feedback) , we tested for the possib le inf luence o f t r ia l length. 112 Methods Participants. Thi r teen undergraduates (10 female) received course-credit for part ic ipat ing in the experiment; mean age was 19.5 years (SD= 0.9). A l l had normal or cor rected - to -normal v i s ion , and a l l but one were r ight -handed. Apparatus and Stimuli. The v isua l d isp lay consisted o f a green response c i rc le (radius 0.3 c m ) and a blue target r i ng (radius o f 0 . 3 0 - 0 . 4 0 5 c m , w i t h a thickness o f 0.105 cm) m o v i n g around a f i xed 8.10 c m radius c i rcular path. One o f two tones was presented in each tr ia l . The first was a 1500 H z w i th a duration o f 100 m i l l i s e c o n d (ms), and the second was a 500 H z tone o f the same duration. A P e n t i u m II computer runn ing a dos-based program contro l led st imulus presentation and response co l lec t ion . S t i m u l i were presented on a 1 4 " V G A moni tor (640x480 p i xe l resolut ion, 60 H z refresh). A telegraph key instrumented wi th two strain gauges to measure depressive force measured the part ic ipant 's response. The strain gauge signal was a m p l i f i e d by a bridge transducer modu le ( N o r t h w o o d Instruments m o d e l l a - 1 0 2 ) p r o d u c i n g an ana log s ignal sampled by an A / D converter at 1000 H z (Techmar Labmaster P G H ) . Response rotation speed was proport ional to telegraph key force, the target rotation speed o f 0.50 H z corresponded to 3.43 N e w t o n o f force, equivalent to a 3 5 0 - g r a m mass resting on the telegraph key . Procedure and Design. Part ic ipants sat approx imately 60 c m f r o m the screen and completed a total o f 480 exper imental trials in two sessions o f approx imate ly 45 minutes. There was a sel f - terminated break every 20 trials. The exper iment consisted o f 5 condi t ions o f s ignal predictabi l i ty : 0 , 2 5 , 5 0 , 75 and 1 0 0 % stop trials cor responding to 100, 7 5 , 50, 25 and 0 % accelerate tr ials, respectively . In each cond i t ion the proport ion 113 was pre -determined but the order o f trials was random. A t the beg inn ing o f each b lock participants were in fo rmed o f whether there were more or less stop/accelerate trials compared to the prev ious b lock . E a c h session began w i t h up to 20 pract ice trials in the 5 0 % condi t ion . O n each t r ia l , both target and response s t imul i were presented at the " 3 o ' c l o c k " pos i t ion o f an imaginary c i rc le at the beg inn ing o f the tr ial (see F igure 5.1). A f t e r a random delay o f 1000 -3000 ms the target began to rotate around the imaginary c i rc le . Part ic ipants contro l led the speed o f the response st imulus and were asked to press on the response key w i t h their dominant index f inger so as to move the response st imulus to over lap the target. A f i xed pressure o f 350 grams y ie lded the speed o f 0.50 H z , m i n i m u m pressure was 100 grams (0.125 H z ) and m a x i m u m pressure was 2 0 0 0 grams (2.50 H z ) . Part ic ipants found the t racking moderately chal lenging , al though it was necessary to constantly mainta in attention to per form the task adequately. The target rotated for 3 0 0 0 -5000 ms after w h i c h a 100 ms h igh or l o w auditory tone was heard. T r a c k i n g durations were sampled randomly f r o m a rectangular distr ibut ion. F o l l o w i n g the tone, the target cont inued to rotate for an addit ional 1000 -2000 ms after w h i c h the tr ia l ended and a b lank screen was presented for 500 ms 114 Figure 5.1. A portrayal of the sequence of events in a trial in Experiment 1 . After a stationary period the target rotates along an imaginary circle, illustrated by the dashed line. Participants control the response stimulus speed by pressing on the key. On stop trials participants should respond to tone A by stopping the maker, while on accelerate trials, participants should respond to tone B by accelerating the marker. Both target and response markers are stationary Target marker moves Participant tracks by controlling the speed of the response marker More force -> faster Tone B -> participant presses harder to make response marker accelerate. Tone A participant stops pressing to make response marker stop. 115 H a l f the participants were instructed to stop as fast as possib le w h e n they heard the l o w tone and to accelerate as fast as possib le w h e n they heard the h igh tone. The instructions regarding tone meaning were reversed for the remain ing part ic ipants. Stopping was stressed as releasing pressure f r o m the key , without l i f t ing the f inger f rom the key (see M o r e i n - Z a m i r et a l . , in press-a for details). Th is in turn w o u l d cause the response st imulus to stop on the screen. D u r i n g this per iod , part icipants had to merely rest their f inger on the response key so as to keep the response st imulus stationary. A c c e l e r a t i o n was stressed as pressing on the key as fast as possib le to advance the response marker to overtake the target and maintain rapid movement o f the marker for at least two c i rc les . T o obtain a m i n i m u m o f 40 trials in each b lock , there were 4 0 , 160, 80 , 160 and 40 trials in the 0, 2 5 , 50 , 75 and 1 0 0 % stop trials condi t ions, respect ively . The data were co l lected over two sessions, w i th ha l f the trials in each condi t ion be ing co l lected on each day. The order o f the f ive condit ions was counterbalanced us ing a L a t i n square w i th the order for each part icipant be ing reversed in the second session. T h e order for each participant was randomly selected f r o m the 10 possible orders (tone assignment crossed, w i t h condi t ion order) , w i t h a m i n i m u m o f two subjects in each cond i t ion . Data Processing F o r each tr ial in the t racking task, a prof i le o f the force on the response key was obtained (see F igure 5.2). O n a l l t r ials, press R T was measured f r o m in i t iat ion o f target movement at the beg inn ing o f each tr ia l to the initiation o f pressure onset. In addi t ion , t racking performance was measured by sampl ing and averaging the phase error ( P E , in degrees) at 1000 H z for 1 sec before tone onset and then comput ing its absolute value. 116 The smal ler the phase error, the better the t racking performance. O n stop tr ials, S T O P R T was measured as the latency f r o m stop s ignal onset to the in i t iat ion o f pressure offset. O n accelerat ion tr ials , A C C E L R T was measured as the latency f r o m accelerat ion s ignal onset to the in i t iat ion o f pressure onset m a r k i n g the accelerat ion response. A t yp ica l response tr ial onset consisted o f a t ime interval w i t h zero force on the selector key , f o l l o w e d by a rap id onset peak ing at about 600 grams or more , f o l l o w e d by t rack ing at the target va lue o f 3 5 0 grams (see F igure 5.2). Stop tr ials contained a force release where a t rack ing force o f 350 grams decreased to zero grams. Acce le ra t ion trials contained a force increase after the stable t racking per iod f o l l o w e d by a decrease to zero grams at the end o f the accelerat ion per iod . Force onset and offset points were found by ana lyz ing the change i n force over t ime. Fo rce data was first d ig i ta l ly filtered w i t h a cutof f o f 10 H z . The a lgor i thm comput ing press R T , S T O P R T and A C C E L R T first calculated the force differential v i a a l o w pass 5 H z filter. Th is resulted in a posi t ive value when the force increased and a negative value when force decreased. Press R T and A C C E L R T were then determined as the point where the force di f ferent ial was above 100 grams/sec. S i m i l a r l y , S T O P R T was determined as the point where force crossed to be low a threshold o f 100 grams/second. A c c u r a c y descr ibed whether participants responded to the tr ial w i t h the correct response, and was computed based on the change in force between the onset o f the response and its offset. The offset was the point in t ime w h e n force stopped decreasing or increasing, for a stop or accelerat ion response, respect ively . The offset point was f o r m a l l y def ined as the point where the force dif ferential was above or b e l o w 100 grams/sec, ident i f y ing when the force first reached its m i n i m u m or m a x i m u m for the stop 117 or accelerate response, respect ively . Stop trials where the change i n force was posi t ive (above 200 grams) and accelerat ion trials where the change in force was negative (below - 2 0 0 grams) were def ined as erroneous, as participants responded w i t h the opposite response. N o t e that the change in force is on a different scale than the force di f ferent ia l used to compute S T O P and A C C E L R T and the offset points. C a r e f u l inspect ion o f the change in force revealed that very few trials lay above - 2 0 0 grams and b e l o w 200 grams (30 trials or less than 0 . 5 % o f the data set), ind icat ing that participants demonstrated a clear stop or accelerat ion on pract ica l ly a l l tr ials. Figure 5.2. Illustration of the pressure profile of two prototypical tr ials, denot ing the pressure exerted by the participant as a funct ion of t ime. At first, as the target was stationary and the participant rested her finger, the response pressure w a s below the min imum necessary for mot ion. After the target began to move, the participant pressed y ielding a steep elevation in pressure that was maintained as the target cont inued to rotate. Figure 5.2A depicts the scenar io where fo l lowing the s ignal , the pressure decreased abruptly, fal l ing to a resting level until the end of the tr ial . F igure 5.2B depicts the scenar io where fo l lowing the s ignal , the pressure increased abruptly for a few seconds indicating an accelerate response. The vertical markers (from left to right) represent the target movement onset, P ress RT, s ignal onset, and time detected by the algorithm as S T O P RT or A C C E L RT. Figure 5.2A Signal Force (grams) O Target Start • Response Onset • Signal A STOP RT 228 ms Time (seconds) 118 Figure 5.2A 2 1 0 0 T 1 7 5 0 1 4 0 0 S 1 0 5 0 4 7 0 0 3 5 0 3.0 4 . 0 5.0 Time (seconds) Results Tr ia ls where A C C E L or S T O P R T or accuracy cou ld not be computed (due to no isy t racking or participants cont inu ing to track) as w e l l as trials in w h i c h responses were faster than 100 ms or s lower than 900 ms were exc luded f r o m analys is . Th is resulted in the exc lus ion o f 2 . 2 % o f the data. Average Press R T and t racking P E were also computed in each condi t ion for each participant. Predictability Analyses. P re l iminary analyses o f the R T and accuracy data indicated that there were s igni f icant effects in both measures. Spec i f i ca l l y , there were R T effects in both tasks, but accuracy effects p r imar i l y in the accelerat ion task (see analyses be low) . A s performance was not matched across tasks in terms o f accuracy , this raised the concern that standard analysis o f R T may be inappropriate. N a m e l y , as can seen in Tab le 5 .1 , a speed-accuracy t rade-of f between R T and accuracy was l i k e l y to have occurred in some condit ions. F o r example , A C C E L R T s were shorter than S T O P R T s in 119 the 7 5 % , but longer i n the 1 0 0 % condi t ion . A t the same t ime, accelerat ion performance was less accurate in both condit ions. In order to enable a more direct interpretation o f the results as w e l l as to control for possible speed-accuracy tradeoffs and cr i ter ion-shi f ts , inverse e f f i c iency scores were used as our m a i n dependent measure. Inverse ef f ic iency scores are a standard w a y to combine R T and accuracy by d i v i d i n g R T b y the proport ion o f correct responses for a g iven condi t ion (Townsend & A s h b y , 1983). L o w e r values on this measure indicate more eff ic ient process ing and therefore better performance. A c c o r d i n g l y , accuracy and R T s o f adjustment responses in each c e l l were converted to inverse e f f i c iency scores. The inverse e f f i c iency scores were subjected to a repeated-measures analysis o f var iance ( A N O V A ) . The independent factors were task (stop or accelerate) and signal predictabi l i ty ( 2 5 % , 5 0 % , 7 5 % and 1 0 0 % ) . The analysis o f the inverse e f f i c iency scores indicated a m a i n effect for task, F ( l , 1 2 ) = 7 . 4 6 , p < . 0 5 , w i th stop be ing better than accelerat ion responses (403 versus 4 6 5 , respect ively) . Furthermore, there was a m a i n effect for predictabi l i ty , F(3,36)=8.25,/><.01, w i th decreasing scores as predictabi l i ty increased (493, 4 4 8 , 418 and 376 for 2 5 % , 5 0 % , 7 5 % and 1 0 0 % , respect ively) . F i n a l l y , there was an interaction between task and predictabi l i ty , i 7(3,36)=3.15,/j)<.05, as can be seen in F igure 5.3. S i m p l e m a i n effects indicated a s igni f icant effect for predictabi l i ty for both the stop response, F (3 ,36)= 10.82,p<.0\, and the accelerat ion response, F ( 3 , 3 6 ) = 5 . 2 2 , p < 0 1 . 120 Table 5.1. Resul ts of Experiments 1, 2 and 3: mean reaction time (RT in milliseconds) and accuracy (in percentages) for all conditions13. Stop Task Acce lerate Task 2 5 % 5 0 % 7 5 % 1 0 0 % 2 5 % 5 0 % 7 5 % 1 0 0 % R T (ms) 407 412 404 Exper iment 1 337 414 407 387 351 A c c u r a c y (%) ' 96 97 98 98 78 88 92 92 R T (ms) 447 450 427 Exper iment 2 338 461 453 4 0 9 343 A c c u r a c y (%) 98 97 99 99 96 97 98 92 R T (ms) 242 228 220 Exper iment 3 214 269 256 234 224 A c c u r a c y (%) 100 100 100 100 100 100 100 100 F o r stopping performance, compar isons between adjacent expectancy condit ions indicated that there were no signif icant differences between the 2 5 % , 5 0 % and 7 5 % condi t ions , w i t h the 1 0 0 % condi t ion be ing s igni f icant ly better i 7 ( l , 1 2 ) = 2 2 . 7 3 , p < . 0 1 . In contrast, for accelerat ion performance, the compar isons indicated that the effect stemmed f r o m the 2 5 % leading to worse performance than the 5 0 % condi t ion , F ( l , 1 2 ) = 8 . 2 2 , p<.02, w h i c h i n turn led to worse performance than the 7 5 % , F ( l , 1 2 ) = 6 . 5 2 , / ? < 0 2 5 . The 7 5 % expectancy condi t ion d id not s igni f icant ly di f fer f r o m the 1 0 0 % condi t ion (F<\). A n inspect ion o f the R T and accuracy scores suggested that the differences between stopping and accelerat ion were m a i n l y due to differences in accuracy (see Table 5.1). A n analysis o f accuracy us ing the same design as descr ibed above conf i rmed that performance was worse in the accelerat ion task, F ( l ,12)=33.76 ,/?< .01 . F o l l o w - u p analyses corroborated that w h i l e no signif icant effects for predictabi l i ty were observed for the stop task, F (3 ,36)<1 , a s ignif icant effect was noted for the accelerat ion task, The values presented in Table 1 were rounded from the group means for each condition. Note that as the values entered into the A N O V A for each individual were computed by the statistical program, they were accurate to several places after the decimal point. Thus, any inconsistencies between the reported values in the text and the Table 1 are due to repeated rounding errors. 121 F ( 3 , 3 6 ) = 5 . 7 3 , p < . 0 1 . D i f ferences in R T were also noted. Spec i f i ca l l y , a trend analysis indicated that w h i l e accelerat ion R T s had on ly a s ignif icant l inear component , 7 7 ( 1 , 1 2 ) = 1 2 . 8 8 , / J < . 0 1 , there were s ignif icant l inear, quadratic and cub ic components for stopping R T s , F ( l , 1 2 ) = 1 5 . 2 4 , / K . 0 1 , F ( l , 1 2 ) = 1 3 . 5 3 , / K . 0 1 , F ( 1 , 1 2 ) = 5 . 3 9 , / J < . 0 5 , respect i ve ly 1 4 . Sequential and trial length analyses. F o r the sequential analyses mean S T O P and A C C E L inverse e f f i c iency scores were computed for trials in the 5 0 % predictabi l i ty condi t ion where the preceding tr ial inc luded the same s ignal , versus a different s ignal . T h e analyses focused on the 5 0 % condi t ion as the number o f repetit ions and alternations was m a x i m a l and there were a suff ic ient number o f observations per ce l l . The A N O V A on S T O P and A C C E L performance inc luded the factor o f task (stop versus accelerate) and prev ious tr ial identity (different f r o m or same as current tr ial) . O f m a i n interest was whether the identity o f the prev ious trials w o u l d inf luence performance, and i f so w o u l d it do so equivalent ly for the two tasks. The results indicated a s igni f icant m a i n effect for prev ious t r ia l , F ( l , 1 2 ) = 2 0 . 5 8 , / ? < . 0 1 , w i t h more eff ic ient scores when the prev ious tr ial was the same as the current tr ial than when it was different ( 4 1 5 versus 4 7 9 , respect ively) . There was also a signif icant interaction between task and prev ious t r ia l , F ( l , 1 2 ) = 1 9 . 8 2 , p < . 0 1 , wh i le the main effect for task d i d not reach s igni f icance (p>.17) . The interaction resulted f r o m there be ing a smal l non s igni f icant effect o f prev ious tr ial 14 Although the trend analysis suggested differences in performance between the two tasks, it did not elucidate the nature of those differences. Specifically, due to the fact that trend analyses use a contrast it appeared misleading in the present case where stopping performance was equivalent in all conditions other than the 100% frequency condition. In addition, the present results suggest that the 100% condition may be qualitatively different from the 25%-75% conditions (see Discussion of Experiment 1). This is inconsistent with the assumption of trend analyses that the independent variable is continuous. Consequently, trend analysis was not used as the primary method of analysis in this study. 122 identity for stopping, F<\, and a large and s ignif icant effect for accelerat ion, F ( l ,12)=45.32,/?<.01 (13 versus 117, respectively) . F o r the tr ial length analyses, tr ial length was categor ized as short ( ranging f r o m 3 -4 sec) or long ( ranging f r o m 4 - 5 sec). A l t h o u g h per formed post -hoc , i n this and in a l l f o l l o w i n g experiments there was no s igni f icant di f ference between the number o f trials i n each category (al l F ' s < l ) . A n A N O V A was conducted on S T O P and A C C E L inverse e f f i c iency scores w i t h task, predictabi l i ty (25, 50 , 7 5 , and 1 0 0 % ) and tr ial length (short versus long) as wi th in -subject factors. In addit ion to the prev ious ly reported effects o f task and predictabi l i ty , there was a m a i n effect for tr ial length, F ( l , 1 2 ) = l 1.68,p<.001, w i t h better performance when tr ia l length was l o n g compared to w h e n it was short (416 versus 4 4 9 , respect ively) . There was also an interaction between predictabi l i ty and tr ial length, F (3 ,36)=5.75, p<.0l, result ing f r o m the effect o f tr ial length gradual ly d i m i n i s h i n g f r o m the 2 5 % condi t ion to no effect at the 1 0 0 % condi t ion (70, 2 9 , 30 and -2 for condit ions 2 5 % - 1 0 0 % , respect ively) , F ( l ,12)=6 .41 ,/5< .05 . N o other effects were s ignif icant . Additional performance measures. A n analysis us ing task and predictabi l i ty as factors on Press R T d i d not demonstrate any signif icant effects (p>.25 for a l l effects), w i t h mean Press R T o f 246 ms for both tasks. L i k e w i s e , a s imi la r analysis on mean t racking P E d i d not show any s ignif icant effects (/?>.35 for both m a i n effects andp>A6 for the interaction) and mean t racking P E was 5.45° and 5.22° for the stop and accelerate tasks respect ively . N o consistent patterns emerged for either measure. Sequential analyses conducted w i t h Press R T and t rack ing P E as dependent measures also d i d not show any s igni f icant effects for either measure. A n a l y s e s o f tr ia l length conducted on 123 tracking P E also indicated no significant effects. Trial length was not expected to influence Press R T , as at the initiation of the trial, its length is unknown. A confirmatory analysis indicated that this indeed was the case. Figure 5.3. The results of Experiment 1. Mean inverse efficiency (IE) scores are plotted as a function of signal predictability for stop and acceleration performance. 700 n 1 0) 600 H 200 -i 1 1 1 25% 50% 75% 100% Frequency - • - S t o p - " -Acce le ra te Discussion The results indicated that stopping and acceleration were not influenced in a similar manner by signal predictability. Stopping performance did not appear to be influenced when signal predictability was varied between 25%, 50% and 75%, although performance improved significantly when a signal appeared on 100% of the trials. In contrast, acceleration performance improved as signal predictability increased from 25% to 75%, leveling off with no significant change at 100% compared to 75%. The sequential analysis dovetailed with the main finding with no effect for the identity of the previous trial on stopping but a large effect on acceleration performance. 124 S topp ing performance repl icated the prev ious f ind ings reported in the stop s ignal l iterature, where s ignal predictabi l i ty ranged f r o m 1 0 % to 8 0 % and no change was found (Logan , 1981; L o g a n & B u r k e l l , 1986; Ramautar et a l . , 2004) . Interestingly however , a large inf luence o f s ignal predictabi l i ty was observed w h e n a s ignal appeared on every tr ial ( 1 0 0 % ) . In other words , when the task changed f r o m a choice task to a s imple response type task, a considerable improvement in stopping performance was observed. Th is is consistent w i th the w e l l established f ind ing that the introduct ion o f two or more options leads to s l o w i n g and decreased accuracy in performance, often attributed to the engagement o f a central dec is ion process (e.g., Donders 1868/1969). Studies u t i l i z ing the stop s ignal task w o u l d not be able to detect this effect s ince the task w o u l d not be feasible wi th stop signals presented on a l l tr ials. Acce le ra t ion performance appeared s imi lar to go performance prev ious ly reported in the literature i n that it was in f luenced by s ignal predictabi l i ty (Berte lson & T isseyre, 1966; H a w k i n s & H o s k i n g , 1969). A l t h o u g h the effect is l inear w i t h l o w to moderate s ignal predictabi l i ty , it levels o f f w i t h s ignal predictabi l i ty above 7 5 % . W e suspect this negative accelerat ion pattern is due to performance be ing at ce i l i ng w i t h the h igh s ignal predictabi l i ty . T r i a l length demonstrated an effect for both stopping and accelerat ion, w i th longer trials lead ing to better performance. Interestingly, tr ial length interacted w i t h predictabi l i ty equal ly for s topping and accelerat ion. S p e c i f i c a l l y , the inf luence o f tr ial length was m u c h larger when the l i k e l i h o o d o f a relevant s ignal was the lowest (at 2 5 % ) . Hence , the more probable the s ignal , the less benefit f rom the increased probabi l i ty over the length o f the tr ial (see also d iscuss ion o f condi t ional probabi l i ty effects in R e q i u n , Brener & R i n g , 1991). Th i s indicates that short - term expectations occur r ing w i th in a tr ial 125 are in f luenced by long - te rm expectations occur r ing over the course o f mul t ip le trials (e.g., T re i sman & W i l l i a m s , 1984). In a previous study ( M o r e i n - Z a m i r & M e i r a n , 2003) tr ial length d i d not inf luence stop performance when the tr ial lengths were m i x e d . H o w e v e r , the t rack ing durations used were m u c h longer, and the short and long durations d i d not over lap. O v e r a l l , this result demonstrates h o w expectancies m a y inf luence performance in many ways w i t h i n the same task, i l lustrat ing the intr icacies o f task demands. In the stop s ignal literature a mechan ism has been suggested that a l lows for the resistance o f stopping performance to stop s ignal predictabi l i ty . The studies va ry ing predictabi l i ty found that w i t h higher frequencies o f stop s ignals , there was an increase in the probabi l i ty o f stopping ( L o g a n , 1981; L o g a n & B u r k e l l , 1986; Ramautar et a l . , 2004) . H o w e v e r , these changes cou ld be accounted for complete ly by the s l o w i n g o f the go R T s , w h i l e S S R T s remained unchanged. It is poss ib le that the p rev ious ly found resistance o f stopping to predictabi l i ty is due to id iosyncrasies o f the stop s ignal task. L o g a n (1994) conc luded that participants appeared to have increasingly p laced a greater importance on the stopping task as the stop s ignal appeared more frequently. H e n c e , participants delayed their go R T s for an increased probabi l i ty o f stopping. S topp ing performance w o u l d remain unaltered w h i l e go performance w o u l d be adjusted accord ing ly . L o g a n (1994) conc luded that a h igh frequency o f stop signals w o u l d lead to a t radeoff between the go and stop processes that in turn cou ld promote unwanted strategies. Thus , the stop signal m a y invo l ve a strategic component where participants adjust their go responses depending on stop s ignal characterist ics, such as its probabi l i ty . 126 In the present continuous t racking task, demands are quite different f r o m those o f the stop s ignal task. N a m e l y , there is l ittle conf l ic t as to whether to continue t rack ing (i.e., going) or to stop. In fact, t rack ing performance was not in f luenced b y the predictabi l i ty o f the stop s ignal suggesting that participants d i d not s low or alter their t rack ing performance in order to improve their stopping performance (see also M o r e i n - Z a m i r & M e i r a n , 2003) . Thus , the expected tradeoff where participants sacr i f ice performance on the go task to enhance their performance on the stop task was not found. In other words, participants d id not sacr i f ice their t racking performance as their expectancies that they w i l l have to alter it changed. O n the other hand, a different type o f t radeoff was possible between the stopping and acceleration responses. A t rack ing response was required on each tr ia l , however either a stop or an accelerate s ignal required one o f the two response adjustments. M o r e o v e r , there was a negative correlat ion between stopping and accelerat ion, a l l o w i n g for a tradeoff between them. If a stop was expected then an accelerat ion was less expected and participants cou ld have prepared a stop at the expense o f accelerat ion performance. Th is poss ib i l i t y is tested i n the f o l l o w i n g experiment. Exper iment 2 Exper iment 2 set out to test whether a t radeoff between stopping and accelerat ion was l i k e l y to account for the results o f Exper iment 1. T o this end, the task was m o d i f i e d so that in either o f two sessions, on ly one response w o u l d be required. In one session participants were required to stop exc lus ive ly to one tone. L i k e w i s e , they were asked to ignore the alternate tone presented on the remainder o f the trials and to cont inue to track w h e n it was presented. In another session, participants accelerated to the second tone, 127 w h i l e ignor ing the first tone. Hence , on ly the task requirements changed wh i le st imulus presentation remained s imi la r to that o f Exper iment 1 w i t h tone predictabi l i ty vary ing between b locks . In the present case, m i n i m a l tradeoff between stopping and accelerat ion responses was predicted as participants per formed on ly one type o f response change in any g iven session. U n d e r these condit ions it was o f interest to see whether stopping w o u l d be resistant to the manipu lat ion o f predictabi l i ty as prev ious ly observed. L i k e w i s e , it was unclear whether stopping and accelerat ion performance w o u l d n o w be s imi lar . The tradeoff account w o u l d predict that without a tradeoff, both stop and accelerat ion w o u l d n o w be s imi la r l y in f luenced by predictabi l i ty . A l te rnat ive ly , s topping m a y remain resistant to the predictabi l i ty manipu lat ion , w h i l e accelerat ion per formance w o u l d improve as it became more predictable. Methods Participants. S ixteen undergraduates (9 female) received course-credit for part ic ipat ing in the experiment; mean age was 20 years (SD= 1.9). A l l had normal or corrected- to -normal v i s i o n , and al l were r ight -handed. Apparatus and Stimuli. The apparatus and s t imul i were the same as those used in Exper iment 1 w i t h the f o l l o w i n g changes. A 3 2 0 C D T T o s h i b a laptop n o w contro l led the experiment. Part ic ipants ' response force was measured on the selector button o f an older model M a c i n t o s h computer mouse. The computer mouse was m o d i f i e d to accept a pressure sensor (F lex i fo rce A 2 0 1 - 1 ) y i e l d i n g an analog voltage that was sampled by an A / D converter (P ico A D C - 2 1 2 ) at 1000 H z . 128 Procedure and Design. The d isp lay was s imi la r to that o f the prev ious experiment, however n o w participants completed the stopping task i n one session and the accelerat ion task i n the other. F o r each task, they were asked to respond to one tone, and to ignore the other. Th is meant that when the irrelevant tone was heard, they were to continue t rack ing . The pa i r ing o f the h igh and l o w tones remained constant w i th in part icipant, but was counterbalanced across participants as was the order o f relevant task (stop or accelerate task first). In total one participant was ran in each o f the 16 possible orders (4 predictabi l i ty condit ions x 2 tones x 2 task orders). D u r i n g each one-hour session 4 b locks were administered in w h i c h the percentage o f relevant s ignal trials was 2 5 % , 5 0 % , 7 5 % or 1 0 0 % . The predictabi l i ty condi t ion b locks were counterbalanced in a L a t i n square design. C o n d i t i o n 2 5 % consisted o f 120 tr ials, 5 0 % o f 60 tr ials, 7 5 % o f 40 trials and 1 0 0 % o f 30 trials, result ing in 30 stop trials in each condi t ion . There was a self - terminated break every 30 trials ( in the 7 5 % condi t ion , the break occurred every 2 0 trials). The events i n each tr ial were ident ica l to those descr ibed in the prev ious experiment. Results Tr ia ls where R T and accuracy measures cou ld not be computed (due to noisy tracking) as w e l l as trials in w h i c h responses were faster than 100 ms or s lower than 900 ms were exc luded f r o m analys is . Th is resulted in the exc lus ion o f 1 . 4 % o f the data. The average press R T , t rack ing P E , as w e l l as inverse e f f i c iency scores for S T O P and A C C E L R T s in each predictabi l i ty condi t ion were computed for each part ic ipant, us ing the same methods detai led above. L i k e w i s e , means for the sequential and tr ial length analyses were computed as in Exper iment 1. 129 Predictability Analyses. The inverse ef f ic iency scores were subjected to an A N O V A , w i t h the factors o f response (stop or accelerate) and predictabi l i ty ( 2 5 % , 5 0 % , 7 5 % and 1 0 0 % ) . The results o f the A N O V A on inverse e f f i c iency scores indicated there was on ly a s igni f icant m a i n effect o f s ignal predictabi l i ty , F ( 3 , 4 5 ) = 1 9 . 5 9 , / K . 0 1 . Responses became faster as s ignal predictabi l i ty increased (471, 4 7 5 , 4 2 8 , 360 for 2 5 % , 5 0 % , 7 5 % and 1 0 0 % , respect ively) . N o other effects were s igni f icant ( F ' s < l ) . P lanned compar isons indicated that performance on 2 5 % and 5 0 % d i d not di f fer s ign i f icant ly (F<\), but that performance on 7 5 % was better, F ( l ,15)=7.05,/?<.02, w i t h performance at 1 0 0 % be ing s igni f icant ly better, F ( l ,15)=23.09,/?<.01. A f o l l o w - u p analysis examined whether the order o f the tasks inf luenced performance. The order d i d not have a signif icant effect; neither d i d it interact w i t h any other factor (al l F ' s < l ) . F igure 5.4 il lustrates the effects o f both stopping and accelerat ion performance as a funct ion o f s ignal predictabi l i ty . A n a l y s i s o f the corresponding R T data supported the results o f the inverse e f f i c iency score analys is , w i t h a main effect o f predictabi l i ty , F (3 ,45)=37.77, p<.0\. The remain ing effects were not s ignif icant (F<1). A n a l y s i s o f the accuracy data showed on ly a marg ina l effect for task, F ( l , 1 5 ) = 3 . 9 3 , p < . 0 7 . A n addi t ional A N O V A examined the accuracy on trials where part icipants were required to continue t rack ing , w i t h the factors o f relevant tone proport ion (25, 50 and 7 5 % ) and the response that was to be prevented. There was a s igni f icant effect o f proport ion , w i t h accuracy decreasing as proport ion o f relevant tone increased, F(2,30)=7.2,/?<.001. M e a n accuracy was 94, 89 and 7 5 % for proport ions 2 5 , 50 , and 7 5 % , respect ively . Compar i sons revealed that at 7 5 % , accuracy was s ign i f icant ly worse 130 than at 5 0 % , F ( l ,15)=23.4 ,/?< .01 , w h i c h in turn was s igni f icant ly worse than at 2 5 % , F ( l , 1 5 ) = 1 8 . 1 , / K . 0 1 . Figure 5.4. The results of Experiment 2. Mean inverse efficiency (IE) scores are plotted as a function of signal predictability for stop and acceleration performance. 700 n • • O 600 -o o (fi 500 H 200 -I 1 1 1 • — 25% 50% 75% 100% Frequency - • - S t o p - " -Acce le ra te Sequential and trial length Analyses. The mean S T O P and A C C E L inverse e f f i c iency scores for the sequential analysis were computed as in the previous experiment. The A N O V A on S T O P and A C C E L performance inc luded the factor o f task (stop versus accelerate) and prev ious tr ial identity (different f r o m or same as current trial) . The results indicated on ly a s ignif icant m a i n effect for prev ious t r ia l , i 7 ( l , 1 2 ) = 9 . 6 0 , p<.0\, w i t h more eff ic ient scores when the previous tr ial was the same as the current tr ial than when it was different (454 versus 4 9 1 , respect ively) . T o examine the effect o f tr ial length, an A N O V A was conducted on S T O P and A C C E L inverse e f f i c iency scores w i t h task, predictabi l i ty (25, 5 0 , 7 5 , and 1 0 0 % ) and tr ial length (short versus long) as with in -subject factors. In addi t ion to the prev ious ly 131 reported effect o f predictabi l i ty , there was a m a i n effect for tr ial length, F ( l , 1 2 ) = 1 8 . 7 3 , j t K . 0 0 1 , w i t h better performance when tr ia l length was l o n g compared to w h e n it was short (411 versus 4 5 6 , respectively) . N o other effects were s igni f icant i n this analysis . Additional performance measures. A s in Exper iment 1, the A N O V A on t racking P E w i t h task and signal predictabi l i ty as factors y ie lded no s igni f icant results and mean P E was 4.85° and 5.59° for the stop and accelerate tasks, respect ively . H o w e v e r , Press R T was found to be s igni f icant ly faster dur ing the stopping task as compared to the accelerat ion task, F ( l ,15)=6.86,/><05. M e a n Press R T was 224 ms and 238 ms for the stopping and accelerat ion tasks, respectively . The sequential and tr ial length analyses conducted w i t h Press R T and t racking P E as dependent measures indicated no signif icant effects for either measure. Discussion The results o f Exper iment 2 indicated that accelerat ion and stopping performance were s i m i l a r l y inf luenced b y predictabi l i ty , w i th performance b e c o m i n g increasingly better as predictabi l i ty increased. A l t h o u g h performance d i d not s ign i f icant ly change w i th the s ignal v a r y i n g f r o m 2 5 % to 5 0 % , it d i d improve w i t h the s ignal va ry ing over the range o f 5 0 % to 1 0 0 % . L i k e w i s e , the abi l i ty to suppress the irrelevant response was in f luenced by predictabi l i ty in a s imi la r manner for both accelerat ion and stopping. Thus, under the condit ions avai lable in Exper iment 2 , stopping was in f luenced by predictabi l i ty when s ignal f requency var ied between 5 0 % and 1 0 0 % , un l ike Exper iment 1 and previous findings (e.g., L o g a n & B u r k e l l , 1986). A s in Exper iment 1, accelerat ion was inf luenced by predictabi l i ty , albeit in a s l ight ly different manner. The results o f the sequential and trial length effects were consistent w i t h the m a i n analyses, demonstrat ing s imi la r effects 132 for stopping and accelerat ion. C o m p a r i n g stopping and accelerat ion resulted in on ly a sl ight divergence in the Press R T measure, w i th responses in the stopping session being somewhat faster. Th is m a y indicate possible differences between the two response adjustments w i th after-effects for accelerat ion but not stopping found under l imi ted condit ions. H o w e v e r , at this stage we chose caut ion due to the large number o f compar isons in the study. Together w i t h the previous results, the present results p rov ide support for the tradeoff hypothesis . In Exper iment 1, the two responses competed w i t h each other as either c o u l d be tr iggered on a g iven tr ial . W i t h a t radeoff between the two compet ing responses, stopping appeared imperv ious w h i l e acceleration was vulnerable to expectations (as manipulated by predictabi l i ty ) . In Exper iment 2 , the l i k e l i h o o d o f compet i t ion was reduced and therefore the l i ke l ihood o f a t radeoff between the two responses was decreased. Cor responding ly , stopping and accelerat ion were both in f luenced by predictabi l i ty and in a s imi lar manner. Th is suggests that the differences observed in Exper iment 1 between the two response adjustments were due to task demands and not due to true funct ional differences. The m a i n di f ference between Exper iment 1 and Exper iment 2 was that in the latter although one o f two tones cou ld be presented dur ing a t r ia l , part icipants k n e w in advance that on ly one o f the tones was relevant to their performance. Nonetheless , participants st i l l had to d ist inguish between two tones and ignore the irrelevant one when it appeared. A c c u r a c y performance on trials where the tone was irrelevant suggests that participants st i l l maintained their expectancies on these trials. O n a s imi la r note, Bedard and col leagues (2002) examined selective control us ing the stop s ignal task. They 133 required part icipants to stop when one tone was presented, but to ignore another. Bedard and col leagues suggested that us ing a selective stop-s ignal w o u l d increase the complex i t y o f the stopping process, a l though they d i d not compare select ive to non-select ive stopping performance. H e n c e , in the present case, even though accelerat ion and stopping performances were s imi lar , it m a y be that the condit ions are not representative o f those used in the vast major i ty o f stop s ignal studies, where a non-se lect ive s ignal is presented. Exper iment 3 Exper iment 3 a imed to examine stopping and accelerat ion performance when no selective response adjustment was required. Thus, the task was changed so h a l f the participants stopped in response to one possible tone, w h i l e the other h a l f accelerated in response to the tone. Tone predictabi l i ty var ied between b locks , and on the remain ing tr ials, no tone was presented and participants were required to track throughout the entire tr ial durat ion. The l i k e l i h o o d o f a tradeoff between stopping and accelerat ing was n o w reduced to a m i n i m u m as participants were not even aware o f the alternate opt ion. It was hypothesized that both stopping and accelerat ion w o u l d be inf luenced by predictabi l i ty in an ident ical manner, as a t radeoff w o u l d not be possible . M o r e o v e r , both stopping and accelerat ion performance w o u l d improve overal l as they were n o w less cogni t i ve ly demanding (Bedard et a l . , 2002) . Methods Participants. Twenty - fou r undergraduates (10 female) received course-credit for part ic ipat ing in the experiment. M e a n age was 19.5 (SD=1.3). A l l had normal or 134 cor rected - to -normal v i s i o n , and 23 were r ight -handed. Part ic ipants in neither group were in fo rmed o f the alternative role o f the tone. Apparatus and Stimuli. The apparatus and st imul i were the same as in previous experiments w i t h the f o l l o w i n g exceptions. N a m e l y , on ly a s ingle tone o f 1000 H z was presented. T o one group o f part icipants, this served exc lus ive ly as a stop s ignal , wh i le to the second group the tone served exc lus ive ly as an accelerat ion s igna l . Procedure. The design and procedure were s imi la r to those o f Exper iment 2. In each condi t ion the proport ion was pre-determined but the order o f trials was random. O n signal tr ials, as in previous experiments, the target rotated for a durat ion between 3 0 0 0 -5000 ms after w h i c h a tone was presented. The target cont inued to rotate for an addit ional 1000 -2000 ms after w h i c h the tr ial ended. O n no -s igna l tr ials, on ly t rack ing was required and participants t racked for a randomly vary ing duration o f 4 0 0 0 - 6 0 0 0 ms. Results Tr ia ls i n w h i c h S T O P and A C C E L R T s were faster than 50 ms or s lower than 500 ms were exc luded f r o m the analys is , result ing in the exc lus ion o f 1 . 7 7 % o f the data. The average press R T , t rack ing P E , and S T O P or A C C E L R T were computed i n each predictabi l i ty condi t ion for each participant. Inverse e f f i c iency scores are presented as in the prev ious exper iments, a l though participants commit ted no errors and so this data is ident ical to the R T data. Predictability Analyses. Inverse ef f ic iency scores were subjected to an A N O V A wi th the factors o f task (stop or accelerate) as a between group factor and predictabi l i ty o f the signal ( 2 5 % , 5 0 % , 7 5 % and 1 0 0 % ) as a repeated-measures factor. The predictabi l i ty o f s ignal trials had a s igni f icant inf luence on responses, F (3,66)=16.2,/?<.001. F igure 5.5 135 indicates that as the proportion of signal trials increases, inverse efficiency scores improved (256, 242, 227 and 219 for 25%, 50%, 75%, 100%, respectively). Planned comparisons indicated that 50% was better than 25%, F(l,22)=8.5,/J<.01, and 75% was better than 50%, F(l,22)=l 1.2,/?<.01. The 75% condition was marginally significantly better than the 100% condition, F(l,22)=3,/j»<.09. There was no main effect for response, nor was there a significant interaction (all p values>.25). Figure 5.5. The results of Experiment 3. Mean inverse efficiency (IE) scores for stop and acceleration performance are plotted as a function of signal predictability. 300 -, 100 i 1 1 1 25% 50% 75% 100% Frequency - • - S t o p - " -Acce lera te Sequential and Trial Length Analyses. Sequential effects were computed as in the previous experiments, however now the design included task as a between-subjects factor and previous trial identity as a within-subjects factor. In addition, the previous trial could either be the same (a signal trial) or different (a no-signal trial). The results of the A N O V A again revealed a main effect for the previous trial identity, F(l,22)=14.73, p<.00\, where A C C E L and STOP performance was more efficient when the previous 136 trial was the same as the current trial than when it did not involve the same response (230 versus 252, respectively). Trial length was examined in an A N O V A which included task as a between-subjects factor and predictability and trial length as within-subject factors. In addition to the effects reported above, there was a main effect for trial length, F(l,22)=24.17, p<.001, with responses being more efficient when trial length was long compared to when it short (230 versus 241, respectively). An interaction was observed between task and trial length, F(l,22)=5.39, p<.05, resulting from the effect of trial length being larger for the stopping task than for the acceleration task (16 versus 6, respectively). Additional Performance Measures. The analysis on Press RT with predictability and task as factors indicated no significant effects (all p values>.35), with mean Press RT being 227 ms. The analysis on tracking PE revealed a main effect for response, with better tracking performance in the stop task (4.56°) than the acceleration task (7.01°), F(l,22)=5.11, p<.Q5. Likewise, there was a significant effect of signal predictability, F(3,66)=5.06,p<.05. Mean tracking PE was 5.06°, 6.27°, 4.79° and 6.82° for the 25%, 50%, 75% and 100% conditions, respectively. However, follow-up comparisons indicated that the main effect stemmed from an unsystematic difference between 25% and 75% on the one hand and the 50% and 100% on the other (p-values<.05). As before, trial length did not influence tracking PE or Press RT, and neither did previous trial influence tracking PE. The analysis on Press RT revealed a main effect for the previous trial, with Press RT being faster when there was no signal in the previous trial (222 ms) than when there was a signal to modify the response in the previous trial 137 (231 ms), ^(1,22)^1 l.l ,p<.01. This effect was equivalent for stopping and acceleration, as there was no significant interaction. Discussion The results of Experiment 3 indicated that both stopping and acceleration were influenced by predictability, and in a similar manner with both becoming faster as predictability increased. Stopping was influenced by predictability, this time within the range 25-75% where it was previously found to be resistant to such manipulations (Logan & Burkell, 1986). Again the sequential effects supported the signal predictability findings, with similar effects for the two response adjustments. These results provide further support for the tradeoff account, as again when no tradeoff is afforded, stopping and acceleration performance are similar. Stopping and accelerating did differ slightly in Experiment 3 with a somewhat larger trial length effect for stopping. However this seemed more to be a difference in the magnitude of an existing effect, rather than a qualitative difference. A final point of interest is that stopping and acceleration performances in Experiment 3 were much better than in the previous two experiments. This provides supporting evidence for the notion that selective stopping is harder, being much more cognitively demanding, than a stop-all situation (Bedard et al., 2002). General Discussion This study aimed to test whether stopping was unique in that it was resistant to predictability (e.g., Logan, 1981, 1994; Logan & Burkell, 1986). Furthermore, the study examined whether stopping performance generalized to another measure of response adjustment. To this end, signal frequency was manipulated and stopping and acceleration 138 performance were compared in a continuous t racking task ( M o r e i n - Z a m i r et a l . , 2004, in press-a) . In Exper iment 1 a stop or acceleration s ignal was presented on each tr ial . Predictabi l i ty was var ied so that each s ignal c o u l d appear on 2 5 % , 5 0 % , 7 5 % and 1 0 0 % o f the tr ials. The results indicated that wh i le stopping performance was not in f luenced by predictabi l i ty w i t h i n the range o f 2 5 % - 7 5 % , accelerat ion performance improved . In Exper iment 2 , a-signal was st i l l presented on every t r ia l , however i n each session part icipants were required to respond exc lus i ve l y to one s ignal either b y stopping or accelerat ing, and to ignore the irrelevant s ignal . Th is t ime, the results for both stopping and accelerat ion performance were s imi lar , as both improved as predictabi l i ty increased. In Exper iment 3 , on ly one s ignal appeared ind icat ing a stop for one group o f participants and an acceleration for another group. A g a i n , an equivalent effect for predictabi l i ty was noted for both stopping and accelerat ion. A d d i t i o n a l analyses indicated that expectancies also operated on a smal ler t ime-scale as both stopping and accelerat ion were inf luenced by task demands on the previous tr ial as w e l l as by tr ial length. S topp ing and acceleration performances were better when tr ial length was shorter and, w i t h the except ion o f stopping in Exper iment 1, w h e n the prev ious tr ial had s imi la r task demands. Stopping can be influenced by predictability The present results are compat ib le w i th the t rade-of f account p rev ious l y used to exp la in stopping performance i n the stop s ignal task ( Logan , 1994). The t rade-of f account proposed that participants are g iven to strategic effects generated by task demands. A s expectancies for a stop s ignal rise and stopping becomes more predictable, participants become more l i k e l y to stop. H o w e v e r , in order to increase their chances o f successful stopping, participants sacr i f ice performance o f the compet ing response wh i le stopping 139 latency remains constant (see D i s c u s s i o n o f Exper iment 1 for further details) . In Exper iment 1, where the stop response cou ld be traded o f f against the accelerat ion response, s topping performance d i d not alter w i t h 2 5 - 7 5 % s ignal probabi l i t y . Th is appeared consistent w i t h evidence that stopping is imperv ious to manipulat ions that inf luence go performance ( R i d d e r i n k h o f et a l . , 1999; W i l l i a m s et a l . , 1999). H o w e v e r , subsequent experiments went on to show that stopping, l ike many other response types, can indeed be vulnerable to predictabi l i ty and strategic effects. W h e n a t rade-of f was less l ike ly or not poss ib le , stopping was inf luenced by predictabi l i ty (Exper iments 2 and 3). Furthermore, i n Exper iments 1 and 2 w h e n a stop s ignal was expected on all t r ials, a dramatic improvement was noted. F i n a l l y , we found that stopping performance cou ld be in f luenced both by expectancies that were carr ied between successive trials and by expectancies operating w i t h the duration o f a single t r ia l . The present results have clear impl icat ions regarding stopping behavior and the stop s ignal l iterature. The f indings oppose one o f the key pieces o f ev idence regarding the dist inct iveness o f s topping f r o m go ing , and spec i fy that the observed resistance o f stopping to predictabi l i ty is l i k e l y to stem f r o m the tradeoff and not f r o m a true funct ional dif ference. Th is interpretation w o u l d suggest that previous studies e m p l o y i n g the stop s ignal task d i d not f i n d predictabi l i ty effects because the task a lways a l l o w e d for the t rade-of f between the go and stop responses (e.g., L o g a n & B u r k e l l , 1986). It remains an open quest ion whether the t rade-of f can be prevented in the stop s ignal task and whether in this case stopping w o u l d be vulnerable to expectancies. The or ig ina l t rade -of f account d id not speci fy what w o u l d happen in the absence o f a trade-off , but the present results suggest that stopping w o u l d be inf luenced by expectancies in the same manner as the go 140 response. One possib le w a y to test this hypothesis is to ut i l i ze a d y n a m i c staircase a lgor i thm to determine the delay between the go and stop signals instead o f f i xed delays (e.g., O s m a n et a l . , 1986). A d j u s t i n g the delay on a t r ia l -by - t r ia l basis that is contingent on performance m a y reduce the l i k e l i h o o d o f a strategic t radeoff between stopping and go ing . In this case one w o u l d expect to f ind rel iable effects o f s top -s ignal probabi l i ty on stopping latencies. The present results have further impl icat ions regarding prev ious f indings us ing the stop s ignal task, as they emphasize the not ion that task characterist ics p l a y a cruc ia l role in determining stopping performance. L o g a n (1994) recommended the use o f a stop signal f requency o f 2 5 % to a v o i d expectancy and strategic affects (see also L a p p i n & E r i k s e n , 1966). Ostens ib ly , this f requency a l lows for a suff ic ient number o f stop trials w h i l e at the same t ime participants are less l i k e l y to sacr i f ice performance on the compet ing response. Despi te the recommendat ion , many studies operat ing under a variety o f constraints have employed stop s ignal frequencies ranging f r o m 3 3 % to 5 0 % . F o r example , D e Jong and col leagues (1990, 1995) w h o used event-related potentials and E M G measures to examine stopping employed 3 9 % and 5 0 % stop signals in their studies. R u b i a and col leagues ( R u b i a et a l . , 1999) w h o tested stopping per formance us ing f M R I employed 5 0 % stop signals exc lus ive ly . Thus, studies subject to pract ical constraints have e m p l o y e d more frequent stop signals than recommended. In none o f these studies was the predictabi l i ty o f the stop signals addressed, and the impact o f stop s ignal f requency on go performance remained moot. O n a related note, the c o m m o n use o f low predictabi l i ty i n the stop s ignal task may not be representative o f situations where a stop 141 is h igh l y probable. The present f indings also suggest that i f a t radeoff were less l i k e l y to occur , the use o f a higher frequency o f stop s ignal w o u l d lead to better stopping. Stopping is similar to other response adjustment measures The current results also shed l ight on whether stopping general izes w e l l to other forms o f cont ro l , namely those requir ing an adjustment o f an ongo ing act ion (Logan , 1994; L o g a n & C o w a n , 1984; M o r e i n - Z a m i r et a l . , 2 0 0 4 , i n press-a) . S p e c i f i c a l l y , stopping was compared w i t h accelerat ing the response marker i n the t rack ing task. B o t h required participants to alter an act ion that was constantly be ing contro l led and moni tored . O v e r a l l , s topping and accelerat ion performances were s imi lar , lend ing support to the c l a i m that they are comparable examples o f response contro l . Th is was evident when no tradeoff between them existed, as in Exper iments 2 and 3. W h e n differences appeared such as the tr ial length effects in Exper iment 3 , they appeared to be a matter o f degree rather than po int ing to true funct ional differences. O n l y i n Exper iment 1 was there a di f ference as stopping p roved to be quite resistant to the inf luence o f predictabi l i ty and sequential effects, w h i l e accelerat ion d i d not. The t rade-of f account suggests that i n this case a stop was more l i k e l y to be prepared at the expense o f an accelerat ion. Th is result prov ides some support to the suggestion that the relat ionship between stopping and other, subtler forms o f contro l might not be s imple and straightforward ( Logan , 1994). W h a t m a y be the source o f the difference in Exper iment 1 between stopping and acceleration? Stopp ing is absolute in the sense that the end result no longer requires further contro l : once a stop is completed no further act ion is required and the task is completed. Other types o f response adjustment, such as accelerat ion, require addit ional act ion f o l l o w i n g the execut ion o f the control process. Th is d ist inct ion is also consistent 142 with models of control that assume some executive system managing subordinate systems to realize intended actions in a hierarchal manner (e.g., Norman & Shallice, 1986). Stopping may simply require less cognitive effort than acceleration and thus is initiated more effectively. Such a view would purport that stopping does not functionally differ from acceleration. That stopping was more efficient than acceleration in Experiment 1 would only be due to a difference in the degree of control required. In the case of stopping, motor planning is minimized as it merely entails the termination of force while acceleration requires an increase of that force. A difference in the extent of the adjustment necessary is also consistent with a model proposed by Logan and Cowan (1984) where stopping and other forms of response control simply differ in the degree of control they engage. According to the model, stop signals and signals to modify response parameters all have privileged access to the response system as opposed to signals requiring a new or sufficiently different response. According to this view, the amount of adjustment of the response would determine the degree of privileged access to the system. Future studies may utilize additional manipulations to further compare types of response adjustment. Likewise, to test whether the differences between stopping and acceleration stem from differences in the degree of difficulty, stopping may be altered to become more demanding. On stopping in stopping tasks Use of the continuous tracking task enabled the current study to focus on issues not easily addressed using the stop signal task and the accompanying race model. First, we were able to directly compare stopping and a different response adjustment measure under similar task demands. Second, the examination of predictability was addressed 143 across a larger range, w h i c h inc luded the use o f a stop s ignal on every t r ia l . Th i s revealed that w i t h 1 0 0 % predictabi l i ty , stopping is even more eff ic ient than w h e n it is required on on ly a port ion o f the trials. Thus , the typ ica l stop s ignal task m a y underestimate S S R T s as it c o m m o n l y presents a stop s ignal on on ly 2 5 % o f the trials. T h i r d , a stopping measure on each tr ial enabled us to d iscover that, s imi la r to the go process, stopping appears to be inf luenced by the type o f previous t r ia l . It is l i k e l y that this is also the case in the stop s ignal task. Th is i n turn further i l lustrates the vu lnerabi l i ty o f response m o d i f i c a t i o n to strategic inf luences. In fact, the stop s ignal task w o u l d appear to be quite vu lnerable to strategic inf luences (as observed by L a p p i n & E r i k s e n , 1966). The importance o f strategic inf luences in the stop signal task are i l lustrated by the s igni f icance o f the w o r d i n g o f the instructions as w e l l as by the dramatic s l o w i n g o f go R T s w i t h the introduct ion o f stop signals (e.g., Cav ina -P ra tes i , B r i c o l o , Pe l legr in i & M a r z i , 2 0 0 4 ; M c G a r r y & F ranks , 1997; R ieger & G a u g g e l , 1999; van den W i l d e n b e r g , van B o x t e l & van der M o l e n , 2003) . The latter may be counteracted by the introduct ion o f r igorous feedback or extensive practice but is not often implemented (e.g., C o l o n i u s , Ozy r t & A r n d t , 2 0 0 1 ; K r a m e r et a l , 1994; Ramautar et a l . , 2004) . The t rade-of f account suggests that the race between the stop and go processes can be inf luenced by the relat ive we ight ing o f each. A l t h o u g h it has been demonstrated that go performance is the one that is adjusted, it does not preclude the poss ib i l i t y that the go performance cou ld be preserved. It remains unclear whether under such condi t ions , stopping w o u l d st i l l be imperv ious to expectancies in the stop s ignal task, although the present results suggest that it w o u l d not. 144 The use o f converg ing tasks w o u l d a l l o w for a r igorous analysis o f strategic components that inf luence performance. It is evident that the cont inuous t racking task and the stop s ignal task both examine stopping. In part icular , in both cases an external s ignal determines that an act ion should no longer be carr ied out and that the c i rcumstances warrant that one either prevent or terminate the act ion. Nonetheless, there are also obv ious differences between the tasks. The most important dif ference is that w h i l e in the t racking task an ongo ing response is terminated, in the stop s ignal task it is w i thhe ld and prevented altogether ( M o r e i n - Z a m i r et a l . , 2004) . Recent evidence has also suggested that stopping a p lanned act ion is less vulnerable to P R P effects than stopping an executed act ion , al though such differences may also be attributable to the differences in the actions be ing stopped (Horstmann, 2003) . Ano ther notable divergence is the relat ionship between the stop and go responses a l l o w e d w i t h i n each task. In the t racking task, participants can first track and later mod i f y their response, and conf l ic t appears m i n i m a l . In contrast, i n the stop s ignal task where typ ica l l y on ly a button closure is recorded, performance on a g i v e n tr ial is registered in a d ichot ic fashion: either the button is pressed and a swi tch is c losed or there is no recorded act ion. Y e t , as act ion in i t iat ion is a smooth and continuous process, it is quite possible that a movement starts yet is terminated before the swi tch is c losed. In fact, precisely such situations have been observed when r igorous measures, such as e lect romyograms ( E M G s ) measur ing musc le act iv i ty are u t i l i zed ( M c G a r r y & F ranks , 1997; M c G a r r y , Ingl is & F ranks , 2000) . The tradit ional race m o d e l has l im i ted success in exp la in ing such results, unless a cont inuous response is d i v ided post -hoc so that to a point it is considered a successful stop and after that point it is considered a successful go ( L o g a n & C o w a n , 145 1984). F o r example , D e Jong and col leagues (1990, 1995) e m p l o y e d a cutof f on a cont inuous gr ip response where pressure under 2 5 % o f m a x i m a l gr ip force was considered a stop w h i l e pressure above this value was considered a go response. Thus, responses that were init iated but d id not reach the cutof f were c lass i f ied as stop responses (see also M c G a r r y & F ranks , 2003 for a related discussion) . Consequent ly , it may be benef ic ia l to v i e w the two stopping tasks as ex ist ing a long a cont inuum. O n the one extreme both the stop and go process are sequential w i th l itt le conf l i c t between them, and one process m a y exist without the presence o f the other. A t the other end o f the cont inuum the stop and go processes exist in paral le l and compete. The situation m a y be further compl i ca ted by the. introduct ion o f addit ional response modi f i cat ions such as the change o f one response into another, as was the case o f stop-change in prev ious studies and accelerat ion in the present study (e.g., L o g a n & B u r k e l l , 1986). In addi t ion to the tradit ional benefits found when us ing converg ing tasks, such as increasing va l id i t y and a better understanding o f the id iosyncrasies o f each, further insight into the construct o f stopping can be found. B o t h stopping tasks a i m to gain understanding o f how people control their actions. A c t i o n contro l in the present sense refers to intent ional ly m o d i f y i n g a p lanned or ongo ing act ion to suit a change i n the environment. T h e present study suggests that adjusting an ongo ing act ion, be l ieved to be a c lassic example o f act ion control ( Logan & C o w a n , 1984), can at t imes be v i e w e d as a response in and o f itself. F o r example , in the 1 0 0 % expectancy cond i t ion , part icipants responded to the s ignal by stopping every response and thus stopping in and o f i tsel f became a p reva i l i ng response as in a s imple R T task. L i k e w i s e , in Exper iment 1 the s ignal entai led a choice between stopping and accelerat ing, w h i l e i n Exper iment 2 146 participants had to respond to one signal and ignore another. The m a i n di f ference between the present stop and accelerate responses and a s imple act ion in i t iat ion is that here, w h e n the signal appeared, participants were in the midst o f execut ing a continuous task that required constant contro l and mon i to r ing o f their act ion and the resul t ing propr iocept ive and v i sua l feedback ( M o r e i n - Z a m i r et a l . , 2004, i n press-b) . In the case o f s imple act ion in i t iat ion, participants go f r o m inact ion to do ing something. S i m i l a r l y to the present case, s topping in the stop signal task is also a response type, one where the outcome is ref ra in ing f r o m c los ing a swi tch (see also H o m m e l , M u s s e l e r , Aschers leben & P r i n z , 2 0 0 2 ; M o r e i n - Z a m i r et a l . , in press-b) . In fact, task requirements as presented in Exper iments 1 and 2 can be presented w i t h i n the c lassic stop s ignal parad igm w i t h stop and stop-change signals presented w i th in the same b l o c k or select ive stopping, respect ively (e.g., Bedard et a l . , 2002) . Thus , response adjustment, and stopping in part icular , can be v i e w e d as a more f lu id construct since the same act ion m a y be v i e w e d as an example o f a s imple response at t imes, but at other t imes as an example o f act ion mod i f i ca t ion or contro l ( L o g a n & C o w a n , 1984). It w o u l d appear that task demands and task meaning a long w i t h act ion consequences determine the pos i t ion o f each speci f ic case a long this cont inuum. In conc lus ion , the present study has i l lustrated the benefits o f us ing converg ing tasks as w e l l as demonstrat ing the inf luence o f cogni t ive factors on response adjustment measures such as stopping and accelerat ion. Th is study has also i l lustrated the importance o f strategic inf luences on such measures, as w h e n tradeoffs are prevented both stopping and accelerat ion were s imi la r l y in f luenced by expectancies. References 147 B a n d , G . P. H . , & van B o x t e l , G . J . M . (1999). Inhibitory motor contro l in stop paradigms: rev iew and reinterpretation o f neural mechanisms. Acta Psychologica, 7 0 7 , 1 7 9 - 2 1 1 . B a n d , G . P. H . , van der M o l e n , M . W . , & L o g a n , G . D . (2003). Horse - race mode l s imulat ions o f the stop-s ignal procedure. Acta Psychologica, 112, 105 -142 . B a r k l e y , R. A . (1997). Behav io ra l inh ib i t ion , sustained attention, and execut ive funct ions: construct ing a u n i f y i n g theory o f A D H D . Psychological Bulletin, 121, 6 5 - 9 4 . B e d a r d , A . , N i c h o l s , S . , Barbosa , J . A . , Schachar, R., L o g a n , G . D . , & Tannock , R. (2002). The development o f selective inhib i tory control across the l i fe span. Developmental Neuropsychology, 21, 9 3 - 1 1 1 . Berte lson , P. (1961). Sequential redundancy and speed in a serial two -cho ice responding task. Quarterly Journal of Experimental Psychology, 13, 9 0 - 1 0 2 . Ber te lson , P. (1965). Ser ica l choice react ion-t ime as a funct ion o f response versus s igna l -and-response repetit ion. Nature, 206, 2 1 7 - 2 1 8 . Ber te lson , P . , & T isseyre , F. (1966). C h o i c e reaction t ime as a funct ion o f st imulus versus response relative frequency o f occurrence. Nature, 212, 1 0 6 9 - 1 0 7 0 . C a b e l , D . W . J . , A r m s t o n g , I. T., R e i n g o l d , E. , & M u n o z , D . P. (2000) . Cont ro l o f saccade in i t iat ion in a countermanding task us ing v isual and auditory stop signals. Experimental Brain Research, 133, 4 3 1 - 4 4 1 . Cav ina -P ra tes i , C , B r i c o l o , E. , Pe l leg r in i , B . , & M a r z i , C . A . (2004). A t what stage o f manual v isua l react ion t ime does interhemispheric t ransmiss ion occur : contro l led or bal l is t ic? Experimental Brain Research, 155, 2 2 0 - 2 3 0 . C o l o n i u s , FL , Ozyr t , J . , & Arndt , P. A . (2001). Countermanding saccades w i t h auditory stop signals: testing the race m o d e l . Vision Research, 41, 1 9 5 1 - 1 9 6 8 . D a y , B . L. , & L y o n , I. N . (2000). Vo luntary mod i f i ca t ion o f automatic a rm movements evoked by mot ion o f a v isual target. Experimental Brain Research, 130, 159 -168. D e Jong , R. , C o l e s , M . G . FL , & L o g a n , G . D . (1995) . Strategies and mechanisms in nonselect ive and selective inhib i tory motor control . Journal of Experimental Psychology: Human Perception and Performance, 21, 4 9 8 - 5 1 1 . D e Jong , R., C o l e s , M . G . FL , L o g a n , G . D . , & Gratton, G . (1990) . In search o f the point o f no return: the control o f response processes. Journal of Experimental Psychology: Human Perception and Performance, 16, 164 -182 . D i l l o n , P. J . (1966). St imulus versus response decis ions as determinants o f the relative frequency effect in dis junct ive react ion- t ime performance. Journal of Experimental Psychology, 7 7 , 3 2 1 - 3 3 0 . Donders , F. C . (1868/1969). O n the speed o f mental processes. Acta Psychologica, 30, 4 1 2 - 4 3 1 . Hanes , D. P. , Patterson, W . F., & S c h a l l , J . D . (1998). R o l e o f frontal eye fields in countermanding saccades: v i sua l , m o v e m e n t , and f i xat ion act iv i ty . Journal of Neurophysiology, 79, 8 1 7 - 8 3 4 . H a w k i n s , H . L. , & H o s k i n g , K . (1969). St imulus probabi l i ty as a determinant o f discrete choice reaction t ime. Journal of Experimental Psychology, 83, 4 3 - 4 4 0 . H e n r y , F. M . , & Har r i son , J. S . (1961). Refractor iness o f a fast movement . Perceptual and Motor Skills, 73, 351-354. H o m m e l , B . , Musse le r , J., Aschers leben , G., & P r i n z , W. (2002). The Theory o f Event . C o d i n g ( T E C ) : a f ramework for percept ion and act ion p lann ing . Behavioral Brain Science, 24, 849-878. Hors tmann, G. (2003). The psycho log ica l refractory per iod o f stopping. Journal of Experimental Psychology: Human Perception and Performance, 29, 965-981. H y m a n , R. (1953). S t imulus informat ion as a determinant o f react ion t ime. Journal of Experimental Psychology, 45, 189-196. K r a m e r , A . F., H u m p h r e y , D . G., L a r i s h , J. F., L o g a n , G. D . , & Strayer, D . L. (1994). A g i n g and inh ib i t ion : B e y o n d a unitary v i e w o f inh ib i tory process ing in attention. Psychology and Aging, P, 491-512. K u d o , K . , & O h t s u k i , T. (1998). Funct iona l m o d i f i c a t i o n o f agonist -antagonist e lect romyographic act iv i ty for rapid movement inh ib i t ion . Experimental Brain Research, 122, 23-30. L a p p i n , J. S . , & E r i k s e n , C . W. (1966). U s e o f a delayed signal to stop a v isua l react ion-t ime response. Journal of Experimental Psychology, 72, 805-811. L o g a n , G. D . (1981). A t tent ion , automatic i ty , and the abi l i ty to stop a speeded choice response. In J. L o n g & A . D . Badde ley (Eds.) , Attention and performance IX. Hi l l sdate , NJ: L a w r e n c e E r l b a u m . L o g a n , G. D . (1994). O n the abi l i ty to inhibit thought and act ion: A users' guide to the stop s ignal parad igm. In D . Dagenbach & T. H . Car r (Eds. ) , Inhibitory processes in attention, memory, and language (pp. 214-249). San D i e g o : A c a d e m i c Press. 150 L o g a n , G . D . , & B u r k e l l , J . (1986). Dependence and independence i n responding to double st imulat ion: a compar ison o f stop, change and dual - task paradigms. Journal of Experimental Psychology: Human Perception and Performance, 12, 5 4 9 - 5 6 3 . L o g a n , G . D . , & C o w a n , W . B . (1984). O n the abi l i ty to inhib i t thought and act ion: A theory o f an act o f control . Psychological Review, 91, 2 9 5 - 3 2 7 . M c G a r r y , T., & Franks , I. M . (1997). A horse race between independent processes: E v i d e n c e for a phantom point o f no return in the preparat ion o f a speeded motor response. Journal of Experimental Psychology: Human Perception and Performance, 23, 1 5 3 3 - 1 5 4 2 . M c G a r r y , T., & Franks , I. M . (2003). O n the nature o f stopping an earl ier intended voluntary act ion. Motor Control, 7, 155 -198. M c G a r r y , T., Ingl is , T . J . , & Franks , I. M . (2000). Aga ins t a final ba l l i s t ic process in the control o f voluntary act ion: Ev idence us ing the H o f f m a n n R e f l e x . Motor Control, 4, 4 6 9 - 4 8 5 . M e g a w , E. D . (1972). D i rec t iona l errors and their correct ion in a discrete t rack ing task. Ergonomics, 15(6), 6 3 3 - 6 4 3 . M o r e i n - Z a m i r , S . , & M e i r a n , N . (2003). Ind iv idual s topping t imes and cogni t ive control : C o n v e r g i n g evidence f r o m a new stop s ignal parad igm. Quarterly Journal of Experimental Psychology, 56, 4 6 9 - 4 9 0 . M o r e i n - Z a m i r , S . , N a g e l k e r k e , P. , C h u a , R., F ranks , I. M . , & K i n g s t o n e , A . F. (2004). Inh ib i t ing prepared and ongo ing responses: Is there more than one k i n d o f stopping? Psychonomic Bulletin & Review, 11, 103 -1040 . M o r e i n - Z a m i r , S . , N a g e l k e r k e , P. , C h u a , R., F ranks , I. M . , & K i n g s t o n e , A . F. ( in press-a) . M e a s u r i n g onl ine vo l i t iona l response control w i th a cont inuous t rack ing task. Behavior Research Methods. M o r e i n - Z a m i r , S . , N a g e l k e r k e , P. , C h u a , R., F ranks , I. M . , & K i n g s t o n e , A . F. ( in press-b) . C o m p a t i b i l i t y effects in stopping and response in i t iat ion in a cont inuous t rack ing task. Quarterly Journal of Experimental Psychology. N i e m i , P. , & Naatanen, R. (1981). Foreper iod and s imple react ion t ime. Psychological Bulletin, 89, 133 -162 . N i g g , J . T. (2001). Is A D H D a d is inhib i tory disorder? Psychological Bulletin, 127, 5 7 1 -598. N o r m a n , D . A . , & Shal l i ce , T. (1986). At tent ion to act ion: w i l l e d and automatic control o f behaviour . In R. D a v i s o n , G . Shwartz & D . Shapi ro (Eds. ) , Consciousness and self regulation: advances in research and theory ( V o l . 4) . N e w Y o r k : P l e n u m . O i l m a n , R. T . (1973). S i m p l e reactions w i th random countermanding o f the "go" s ignal . In S . K o r n b l u m (Ed.) , Attention and Performance IV (pp. 5 7 1 - 5 8 1 ) . N e w Y o r k : A c a d e m i c Press. Ooster laan, J . , L o g a n , G . D. , & Sergeant, J . A . (1998). Response inh ib i t ion in A D / H D , C D , c o m o r b i d A D / H D + C D , anxious, and control ch i ld ren : a meta-analys is o f studies w i t h the stop task. Journal of Child Psychology and Psychiatry, 39 , 4 1 1 -4 2 5 . O s m a n , A . , K o r n b l u m , S . , & M e y e r , D . E. (1986). The point o f no return in choice react ion t ime: contro l led and bal l is t ic stages o f response preparat ion. Journal of Experimental Psychology: Human Perception and Performance, 12, 2 4 3 - 2 5 8 . P i s e l l a , L. , A r z i , M . , & Rossett i , Y . R . . (1998). The t i m i n g o f co lour and locat ion process ing in the motor context. Experimental Brain Research, 121(3), 2 7 0 - 2 7 6 . Rabbi t t , P. M . A . , & V y a s , S. M . (1973). W h a t is repeated in the "repetit ion effect"? In S. K o r n b l u m (Ed.) , Attention and Performance IV (pp. 3 2 7 - 3 4 2 ) . N e w Y o r k : A c a d e m i c Press. Rabbi t t , P. M . A . , & V y a s , S . M . (1974). Interference between b inary c lass i f icat ion judgments and some repetit ion effects i n a serial choice react ion t ime task. Journal of Experimental Psychology, 103, 1181 -1190 . Ramautar , J . R., K o k , A . , & R idder inkhof , K . R. (2004). Ef fects o f stop-s ignal probabi l i ty in the stop-s ignal parad igm: the N2/P3 complex further val idated. Brain and Cognition, 56, 2 3 4 - 2 5 2 . R e q u i n , J . , Brener , J . , & R i n g , C . (1991). Preparat ion for act ion. In J . R. Jennings & M . G . H . C o l e s (Eds.) , Handbook of cognitive psychophysiology: Central and autonomic nervous system approaches (pp. 357 -448) . O x f o r d , E n g l a n d : John W i l e y & Sons. R idder inkhof , K . R. , B a n d , G . P. H . , & L o g a n , G . D . (1999). A study o f adaptive behavior : effects o f age and irrelevant in format ion on the ab i l i t y to inhibi t one's actions. Acta Psychologica, 101, 3 1 5 - 3 3 7 . R ieger , M . , & G a u g g e l , S. (1999). Inhibitory after-effects in the stop s ignal parad igm. British Joiwnal of Psychology, 90, 5 0 9 - 5 1 8 . R u b i a , K . , Overmeyer , S . , Tay lo r , E. , B r a m m e r , M . J . , W i l l i a m s , S . C . R., S i m m o n s , A . , et a l . (1999). Hypof ronta l i t y in attention def ic i t hyperact iv i ty disorder dur ing 153 higher order motor contro l : a study w i t h f M R J . American Journal of Psychiatry, 156, 8 9 1 - 8 9 6 . R u b i a , K . , R u s s e l l , T., Overmeyer , S . , B r a m m e r , M . J . , B u l l m o r e , E. T., Sharma, T., et a l . (2001). M a p p i n g motor inh ib i t ion : conjunct ive brain act ivat ions across different versions o f go/no-go and stop tasks. Neurolmage, 13, 2 5 0 - 2 6 1 . Sergeant, J . A . , Geurts , H . , & Ooster laan, J . (2002). H o w spec i f ic is a def ic i t o f executive funct ion ing for At tent ion -Def ic i t/Hyperact iv i ty D isorder? Behavioral Brain Research, 130, 3 - 2 8 . Soetens, E. , B o e r , L . C , & Huet in g , J . E . (1985). Expectancy or automatic faci l i tat ion? Separat ing sequential effects in two -cho ice reaction t ime. Journal of Experimental Psychology: Human Perception and Performance, 11, 5 9 8 - 6 1 6 . Tannock , R. (2003). Neurophsycho logy o f attention disorders. In S . J . Sega lowi t z & I. R a p i n (Eds.) , Handbook of Neuropsychology (2nd E d i t i o n ed. , V o l . 8, Part II, pp. 7 5 3 - 7 8 4 ) . T o w n s e n d , J . T., & A s h b y , F. G . (1983). Stochastic modeling of elementary psychological processes. London/ N e w Y o r k : C a m b r i d g e Un i ve rs i t y Press. T re i sman , M . , & W i l l i a m s , T. C . (1984). A theory o f cr i ter ion setting w i t h an appl icat ion to sequential dependencies. Psychological Review, 91, 6 8 - 1 1 1 . van den W i l d e n b e r g , W . , van B o x t e l , G . J . M . , & van der M o l e n , M . W . (2003) . The durat ion o f response inh ib i t ion in the stop-s ignal pa rad igm varies w i t h response force. Acta Psychologica, 114, 115 -129. van den W i l d e n b e r g , W . , van der M o l e n , M . W . , & L o g a n , G . D . (2002) . R e d u c e d response readiness delays stop s ignal inh ib i t ion . Acta Psychologica, 11, 155 -169. 154 V i n c e , M . A . , & W e l f o r d , A . T. (1967). T i m e taken to change the speed o f a response. Nature, 213, 5 3 2 - 5 3 3 . W i l l i a m s , B . R., Ponesse, J . S . , Schachar, R. J . , L o g a n , G . D . , & Tannock , R. (1999). Deve lopment o f inhib i tory contro l across the l i fe span. Developmental Psychology, 35, 2 0 5 - 2 1 3 . W o o d r o w , H . (1914). The measurement o f attention. Psychological Monographs, 17, 1-158. 155 C H A P T E R 6: Genera l D i s c u s s i o n Th i s dissertation a imed to address some o f the pr inc ip les i n v o l v e d in how we intent ional ly m o d i f y our actions, i f by adjusting them s l ight ly or by terminat ing them complete ly to better fit our goals in a cont inuously chang ing envi ronment . The studies compr i s ing this dissertation focused on some o f the unresolved issues in the stopping literature. In these studies I developed and ut i l i zed an alternative task that required participants to stop an ongo ing act ion after it was init iated. Th i s task a l l o w e d for a direct observat ion o f stopping a contro l led act ion on each tr ial where a stop s ignal was presented. The studies, set out w i th three m a i n aims. F i rst , to prov ide converg ing evidence to earl ier results found in the stop s ignal literature. Second, to c r i t i ca l l y examine the key assumptions in ways prev ious ly not v iable . A n d f ina l l y , to suggest new avenues o f research that c o u l d , in turn, ut i l i ze the strengths and weaknesses o f both tasks to gain a deeper understanding o f stopping in part icular and response control in general . S u m m a r y Chapter 1 discussed w h y prev ious studies have invest igated stopping. In addi t ion, it introduced the stop s ignal parad igm w h i c h has been a central method for invest igat ing response control behaviors such as stopping and response adjustments. F i n a l l y , Chapter 1 ident i f ied three m a i n assumptions o f the stopping literature and establ ished the need for converg ing evidence. Chapter 2 presented a new t rack ing task, designed to measure elements o f response contro l , and in part icular response adjustments such as stopping. In 156 this task, participants t racked a v isual target by manual l y pressing on a force sensor to y i e l d a trace o f force over t ime. In addi t ion to detai l ing the hardware speci f icat ions, an a lgor i thm for determining the latencies o f response adjustments was descr ibed. Chapter 2 also i l lustrated the use o f the new t racking task in two experiments. Exper iment 1 demonstrated the re l iab i l i t y o f the data produced by the task. Exper iment 2 examined some o f the issues that c o u l d be addressed us ing the n e w task, such as manipu lat ing the type o f response adjustment required and the type o f st imulus s igna l ing the response adjustment. The results established the usefulness and potential o f the task for gauging response control w i t h i n the context o f stopping. Chapters 3 through 5 ut i l i zed the new t rack ing task to c r i t i ca l l y test the three main assumptions o f stopping research. Chapter 3 tested the assumpt ion that stopping as measured by the stop s ignal parad igm generalizes to other forms o f stopping. Stopping performance in the stop s ignal task was compared to stopping performance in the new cont inuous t rack ing task. W h i l e st imulus presentations were c lose ly matched across paradigms, the two tasks d i f fered in the type o f stopping required. In the stop s ignal parad igm, response inh ib i t ion latency was measured prior to response execut ion, i.e., it was inferred f r o m the successful w i t h h o l d i n g o f a " g o " response. In the t rack ing task, stopping was earned out after response execut ion, i.e., it was measured as the t ime to begin stopping the continuous t racking response. The results indicated that stopping latencies between the two paradigms were h igh ly correlated, o f fer ing strong evidence that stopping an unexecuted response engages s imi lar mechanisms as stopping an ongo ing response. Th is p rov ided direct support for the assumption that there is a generic stopping m e c h a n i s m that is accessed by both the stop s ignal parad igm and the t rack ing task. 157 Hence , Chapter 3 substantiated a c o m m o n stopping construct encompass ing a range o f stopping behaviors. Chapter 4 examined the field's assumption that stopping is distinct f r o m go ing and governed by different mechanisms. T w o experiments exp lored stopping performance us ing a new st imulus-response ( S - R ) compat ib i l i t y effect spanning act ion in i t iat ion and stopping. In the compat ib le condi t ion , part icipants pressed the sensor in response to the target m o v i n g and stopped pressing in response to the target stopping. In the incompat ib le cond i t ion , part icipants stopped press ing in response to the target m o v i n g and init iated pressing in response to the target stopping. Stopping and response in i t iat ion demonstrated st imulus- response compat ib i l i t y effects o f the same magnitude w i t h faster responses under compat ib le compared to incompat ib le condi t ions , regardless o f whether compat ib le and incompat ib le trials were b l o c k e d or m i x e d . These findings indicated that stopping is in f luenced by st imulus-response properties such as compat ib i l i t y , in the same manner as response in i t iat ion. Th is in turn suggested that stopping m a y be governed by constraints s imi la r to those o f other responses. Thus, Chapter 4 p rov ided evidence against the not ion that stopping is governed by different constraints than go ing . Chapter 5 investigated the assumption that stopping is dist inct f r o m go ing us ing a different method o f invest igat ion. Spec i f i ca l l y , it was questioned whether stopping is imperv ious to predictabi l i ty , as prev iously reported i n the stop s ignal literature (e.g., L o g a n & B u r k e l l , 1986). M o r e o v e r , Chapter 5 examined the assumpt ion that stopping is a representative measure o f response control . Part ic ipants stopped their ongo ing t racking in response to auditory signals on 2 5 , 50 , 7 5 , or 1 0 0 % o f the tr ials. S topp ing was contrasted w i th accelerat ing, a different f o r m o f response mod i f i ca t ion , where part icipants accelerated the marker in response to auditory signals. In Exper iment 1, on each tr ial 158 participants either stopped or accelerated, a l l o w i n g a t rade-of f between the two responses. In Exper iments 2 and 3 , participants on ly stopped or on ly accelerated thereby decreasing the l i k e l i h o o d o f a t radeoff between stopping and accelerat ing. W h e n a trade-o f f was poss ib le , stopping was resistant to predictabi l i ty . H o w e v e r , when there was little or no tradeoff, expectancies inf luenced stopping and accelerat ing s imi la r l y . These f indings conf l ic t w i t h the established v i e w that stopping is insensit ive to expectancies, p r o v i d i n g further support to the m a i n conc lus ion o f Chapter 4 , that u l t imately stopping may not be distinct f rom go ing . In addi t ion, when response tradeoffs were prevented, the results were in agreement w i t h the third assumption that stopping is representative o f other response adjustment measures, such as accelerat ion. Conc lus ions and Impl icat ions o f the Present Resul ts The present studies developed and demonstrated the ut i l i t y o f the continuous t rack ing task in invest igat ing stopping and response mod i f i ca t ion behavior . Th is body o f w o r k provides expl ic i t tests o f previous key assumptions in the stopping literature. In testing the key assumptions, the current body o f w o r k presented clear support for the assumpt ion that stopping is generic ( A s s u m p t i o n #1). L i k e w i s e , it p r o v i d e d support for the assumption that stopping is representative o f other instances o f response mod i f i ca t ion (Assumpt ion #3). H o w e v e r , it d i d not support the second assumpt ion that stopping is distinct f r o m response in i t iat ion measures such as go ing . M o r e o v e r , the present f indings suggest that some o f the prev ious evidence in the stop s ignal literature support ing this second assumption was due to the id iosyncrasies o f that task and not to true funct ional differences between stopping and go ing . In other words , characterist ics o f the stop s ignal 159 parad igm and its part icular task demands are responsible for the observed performance, rather than the stop and go mental processes. Consequent ly , future invest igations into stopping should c lear ly delineate w h y stopping is be ing examined instead o f go ing , and in what ways is stopping performance expected to be different f r o m standard go ing performance. The current results have several methodolog ica l and pract ica l impl icat ions for future research investigations o f stopping and response contro l . F i rst these present studies emphasize the need to explore the boundary condit ions o f the findings f r o m the standard stop s ignal task. The current v i e w o f stopping has been fo rmed i n part by the fact that stopping, as measured b y the stop s ignal task, was found to be resistant to many exper imental manipulat ions such as practice and expectancy effects. The present results conf i rmed some o f these f indings , w i t h stopping again be ing found to be resistant to pract ice effects (Chapter 2). H o w e v e r other f indings were not con f i rmed , w i th stopping be ing found to be sensit ive to expectancy effects (Chapter 5). Thus , the results suggest that speci f ic task demands o f the stop s ignal task and the fact that S S R T is estimated mathemat ica l l y are l i k e l y to be responsible for some o f the prev ious f ind ings regarding the resistance o f S S R T s to different manipulat ions. The present results also emphasize the importance o f the strategies adopted by the participants in the stop s ignal task (Chapter 5). Th i s is important as differences in S S R T that arise between groups (one o f the most robust types o f effects observed in this literature, M o r e i n - Z a m i r & K ings tone , in preparation) m a y result direct ly f r o m different strategies be ing adopted by different groups and not f r o m actual differences in stopping abi l i ty (e.g., Casada & R o a c h e , 2005) . M o r e o v e r , evidence pertaining uniquely to S S R T measurements in the stop s ignal task 160 w i th no converg ing evidence f r o m addit ional tasks e x a m i n i n g response control should be treated w i t h caut ion , as they m a y not result f r o m factors re lat ing to response control . B y be ing able to d i rect ly access and measure the stopping process, the present studies raise the poss ib i l i t y that other questions may also be accessible in ways that were not immediate ly obv ious when app ly ing the tradit ional stop-s ignal parad igm. F o r instance, direct manipulat ions o f the stop s ignal i tsel f (as in Chapter 4) have been re lat ively scarce ( M o r e i n - Z a m i r & K ings tone , in preparation). The vast major i ty o f stop-s ignal studies have ut i l i zed a stop s ignal task w i th a v i sua l go s ignal and a s imple auditory stop s ignal ( N i g g , 2001) . Results f r o m Chapter 4 indicated that carefu l ly chosen characteristics o f the stop-st imulus can facil itate or h inder s topping performance. Thus , exp lor ing the effects o f alter ing the stop signal itself, rather than manipu la t ing the go responses, m a y prov ide valuable in format ion on effect ive response m o d i f i c a t i o n . F i n a l l y , Chapter 4 ascertained that certain s t imul i in the environment are l i k e l y to faci l i tate response control depending on their characteristics in relat ion to per formance, (e.g., a st imulus stopping in the envi ronment w i l l faci l i tate stopping behavior ) . Th i s entails that human-computer interfaces where the human part icipant must exhib i t a h igh level o f response control (e.g., remote surgery, or f l ight controls) , m a y be designed to m a x i m i z e the compat ib i l i t y between the speci f ic response modi f i cat ions that are essential for performance and the st imulus elements convey ing the need for m o d i f i c a t i o n . The present results also have several theoretical impl icat ions . The data are compat ib le w i t h the h ierarchica l v i e w o f act ion generation and contro l . A s noted in Chapter 1, acts o f control are def ined as the interactions in a hierarchal scheme between an executive or supervisory system that forms intentions and c o m m a n d s , and subordinate 161 systems that interpret the commands and execute them. A c c o r d i n g to this f ramework , signals to stop or m o d i f y on ly some parameters o f an act ion have p r i v i leged access to the subordinate systems. The present data prov ide on ly l imi ted support for the not ion that stopping has p r i v i leged access to this hierarchy. Spec i f i ca l l y , the present results add to a smal l but g r o w i n g body o f evidence suggesting that many o f the constraints govern ing in i t iat ion o f responses also apply to stopping. In this sense, it w o u l d appear that stopping and other response modi f i cat ions are carr ied out by the superv isory system in m u c h the same w a y as response init iat ions. M o r e o v e r , response modi f i cat ions in general can be construed as responses in and o f themselves, governed by constraints s imi la r to those that come into p lay in response in i t iat ion. Thus , the present studies do not support the general not ion that response inh ib i t ion has pr i v i leged access to the subordinate systems in the hierarchy. A d d i t i o n a l evidence support ing this conc lus ion can be found in the double step parad igm u s i n g manua l response adjustments (e.g., Georgopou los , K a l a s k a & M a s s e y , 1981; Pau l ignan , M a c K e n z i e , M a r t e n u i k & Jeannerod, 1991; P e l i s s o n , Preb lac , Gooda le & Jeannerod, 1986). In a typ ica l double step reaching task, the spatial locat ion o f a f reaching movement is m o d i f i e d in response to a locat ion shift o f an external target. Th is literature has demonstrated that m o d i f y i n g the spatial components o f a reaching or po in t ing movement can be implemented very rapid ly , often w i t h i n 1 0 0 - 1 2 0 m s (e.g., Preblanc & M a r t i n , 1992). Th is is inconsistent w i t h the m u c h s lower estimates required to stop a movement , found in both the c lassic stop s ignal and t rack ing tasks, where the latencies typ ica l l y vary f r o m 200 to 300 ms. It appears that rapid adjustment latencies in the double step parad igm are in part due to the relat ively smal l and constrained 162 adjustments. B u t more important ly , the rapid latencies result f r o m the very h igh degree o f S - R compat ib i l i t y between the new response and the target, w h i c h faci l i tates an automatic and swif t adjustment (Day & L y o n , 2000) . In contrast, more contro l led (top-down) adjustments o f an ongo ing reaching movement , w i t h a lesser degree o f S - R compat ib i l i t y , result in much s lower adjustments (Day & L y o n , 2000) , that are s imi la r to stopping latencies. Note that the importance o f S - R compat ib i l i t y i n response mod i f i ca t ion , is in l ine w i t h the m a i n f indings o f Chapter 4. H e n c e , p r i v i l eged access to the h ierarchical response control system is not granted to smal l modi f i cat ions o f ex ist ing responses. Rather , p r i v i leged access appears to depend on the automatic i ty o f the response mod i f i ca t ion , as determined b y the degree o f S - R compat ib i l i t y (see also C a b e l , A r m s t r o n g , R e i n g o l d & M u n o z , 2000) . Th is evidence then prov ides further support to the present w o r k ' s conc lus ion that stopping does not have p r i v i leged access to the response system. M o r e o v e r , this present body o f w o r k suggests that the current " theory o f s topp ing" w o u l d be better construed as a "general theory o f response c o n t r o l " i n c l u d i n g a l l instances o f response mod i f i ca t ion and response in i t iat ion, rather than a f ramework for response inh ib i t ion and stopping. Cor responding ly , it is not meaningfu l to target stopping behav ior in isolat ion but rather the broader array o f response modi f i ca t ions that the theory addresses should be investigated. The present results also have impl icat ions regarding several recent studies that employ the stop s ignal task and assume it engages a unique control m e c h a n i s m o f inh ib i t ion (e.g., A r o n & P o l l d r a c k , 2006) . B y focus ing on a single task and award ing the behavior it measures pr i v i leged status, this approach denies other tests o f response control and l imi ts the scope o f response control research. 163 It is beyond any. doubt that the stop signal task has been an extremely useful method to measure response control . M o r e o v e r , it has helped shape current theor iz ing in m a n y areas, ranging from saccadic response control to the conceptual izat ion o f def ic ient response inh ib i t ion in A D H D (e.g., Hanes & Carpenter, 1999; N i g g , 2001) . Undoubted ly , it w i l l continue to prov ide valuable informat ion in the future. A t the same t ime, the data f r o m this dissertation suggests that researchers should be m i n d f u l o f the l imitat ions o f the task and the related horserace m o d e l , as w e l l as the l imitat ions o f the conceptual f ramework in w h i c h it has operated. W i t h m a n y interesting questions remain ing to be answered, it w o u l d be exceedingly valuable to adopt an expansive approach us ing converg ing tasks and addit ional instances o f response contro l . Outstanding Questions and Future D i rect ions Th is f ina l sect ion discusses some o f the questions raised by the current series o f experiments, and points to several avenues for future inqui ry . F i rst , appl icat ions o f the current results to the stop s ignal parad igm are discussed. Second , m o d i f i e d versions o f the t rack ing task are proposed in order to further investigate response control i n normal ind iv iduals . F i n a l l y , the relevance o f the t racking task to invest igat ing inh ib i t ion dysfunct ion is h igh l ighted, and an instance where the task has been successful ly appl ied is descr ibed. Extending the Present Conclusions From Within the Stop Signal Paradigm. In order to further establish the present conclus ions and prov ide converg ing va l id i ty , these conclus ions may be used to generate investigations us ing the stop s ignal task. Chapters 4 and 5 g ive rise to clear predict ions regarding stopping behavior in the 164 stop s ignal task. Chapter 4 suggests that characteristics o f the s ignal , such as s t imulus -response ( S - R ) compat ib i l i t y can inf luence stopping. It is predicted that s t imul i that are compat ib le w i t h stopping, such as the w o r d 's top ' in red i n k w i l l i l l i c i t faster S S R T s than incompat ib le s t i m u l i , such as the w o r d ' g o ' in green ink. Th is study cou ld reinforce the conc lus ion that stopping is indeed inf luenced by constraints k n o w n to operate on go responses. L i k e w i s e , such a study cou ld demonstrate an effect on S S R T resul t ing f r o m a direct exper imental manipu lat ion on the stop s ignal , further support ing the v iab i l i t y o f the estimated S S R T measure in such c i rcumstances. Chapter 5 suggests that i f the tradeoff between stopping and go ing is prevented, it w o u l d be possib le to observe the effects o f s ignal predictabi l i ty on S S R T s . One method for d o i n g this i n the stop s ignal task is to prevent participants f r o m s l o w i n g their go R T i n order to improve their chances o f stopping. Th is can be implemented by manipu lat ing instructions, t ra in ing regimes, catch trials and feedback together w i th a p a y o f f scheme (Osman , K o r n b l u m & M e y e r , 1986, 1990; M c G a r r y & F ranks , 1997; R idder inkhof , B a n d & L o g a n , 1999). In addi t ion, hav ing ident i f ied strategy as p l a y i n g an important role in behavior dur ing the stop s ignal task, one can n o w explore h o w it inf luences performance. F o r example , the balance between go and stop processes can be manipulated as w e l l as the perce ived degree o f conf l ic t between them in order to c lar i fy h o w strategy comes into p lay . In fact, contro l versions o f the task e m p l o y i n g two conf l i c t ing responses other than go versus stop (e.g., a left response versus a right response) m a y help i l lustrate w h i c h i conc lus ions are speci f ic to response inh ib i t ion and w h i c h are spec i f ic to response conf l ic t . W h e n keep ing in m i n d the strengths and l imitat ions o f the stop s ignal task, wh i le 165 not be ing m i s l e d by task id iosyncrasies, a deeper understanding o f the stopping process in a l l its varieties and levels o f complex i t y may be afforded. Investigating Response Control with Variations of the Tracking Task. The present series o f studies has demonstrated the usefulness o f the t rack ing task. Th is task m a y n o w be employed as a v iable tool to further investigate addi t ional characteristics o f response control . F o r example , the t rack ing task m a y be used to further explore the s imi lar i t ies and differences between various instances o f response adjustment. Such an invest igat ion w o u l d further address the issues discussed in Chapter 5 regarding the general izabi l i ty o f stopping to addit ional response adjustment measures, such as accelerat ion. T o this end, manipu lat ing the degree o f d i f f i cu l t y o f response adjustments in the t rack ing task m a y reveal systematic differences between stopping and accelerat ion, or re inforce the s imi la r i t y between them. Response control m a y be further investigated us ing the t rack ing task by e x a m i n i n g whether the degree o f conf l ic t between t rack ing and stopping components alters performance. In the present t racking task, there is very l i tt le conf l ic t between t racking and stopping (see Chapter 5 Genera l D iscuss ion) . The degree o f conf l i c t may be systematical ly manipulated by alter ing t racking prepotency. Th i s can be implemented by manipu lat ing instructions (prov id ing more emphasis on the importance o f t rack ing and adding a sense o f urgency) , by p rov id ing feedback and by pos i t ing a concrete goal (such as a v i s ib le target locat ion to w h i c h the participant must track). S u c h a systematic manipulat ion o f the degree o f inh ib i t ion required to overcome the prepotency o f the ongo ing response, c o u l d also prov ide a clearer understanding o f stopping as a measure o f 166 se l f regulat ion. Th i s in turn cou ld prov ide a convenient method in w h i c h to examine this construct in patients w i t h impai red sel f - regulat ion. A th i rd avenue o f invest igat ion proposes us ing the t rack ing task to probe the general i ty o f response mod i f i ca t ion and inh ib i t ion pr inc ip les across different product ion systems and motor responses. One o f the m a i n reasons that stopping behav ior is considered an index o f motor control is because it appears to be exhib i ted in different product ion systems in a s imi la r manner (stopping is easi ly observed in speech as we l l as in hand, foot, or eye movements) . D e v e l o p i n g t racking tasks adapted to addit ional product ion systems, such as voca l motor responses w i l l permit this k e y assumption to be investigated. T o date on ly a single study exp l ic i t l y and systematical ly exp lored vo l i t iona l termination o f speech (Ladefoged, S i lverste in & Papcun , 1973). A cont inuous voca l task w o u l d prov ide further means to investigate stopping an ongo ing act ion and compar ing it to response mod i f i ca t ion . Such a task m a y employ cont inuous speech or related voca l responses such as h u m m i n g or s ing ing interrupted by signals to stop or m o d i f y performance. M o r e o v e r , the use o f verbal s t imul i in a v o c a l t rack ing task c o u l d also faci l i tate the invest igat ion o f the relat ionship between semantic mean ing and action control . In sum, a compar ison o f response control i n different product ion systems w o u l d be valuable to the understanding o f motor control in general and in determin ing the necessity for over lap between different f ie lds . Applying the Tracking Task to Inhibition Dysfunction. Another l ine o f invest igation ar is ing f rom the present studies is to apply the t racking task to speci f ic instances where behavior is inf luenced by inhib i tory dysfunct ion. One p r ime example o f this is to examine stopping behavior w i t h the t rack ing task in 167 ch i ld ren w i t h A D H D . It is c o m m o n l y be l ieved that the disorder d i rect ly results f r o m or heav i l y invo lves d i f f icu l t ies i n suppressing inappropriate actions (e.g., B a r k l e y , 1997). A c c o r d i n g l y ch i ldren and adults w i th A D H D re l iab ly demonstrate s lowed S S R T s (e.g., N i g g , 2 0 0 1 ; Ooster laan, L o g a n & Sergeant, 1998). A p p l y i n g the t rack ing task in this instance has the added benefit o f addressing several unresolved concerns in this research domain . F o r example , several theories predict that it is not on ly stopping latency but also stopping var iab i l i ty that is impai red in A D H D (e.g., Tannock , 2003) . H o w e v e r , stopping var iabi l i ty cannot be examined re l iab ly us ing the stop s ignal task ( B a n d , van der M o l e n & L o g a n , 2003) . The t racking task is ideal ly suited to test whether A D H D patients demonstrate general s lowed performance in stopping behavior , as w e l l as testing whether they demonstrate increased var iabi l i ty in stopping. In a study designed to test these issues, a group o f A D H D ch i ld ren and age and gender matched controls per formed the two stopping tasks ( M o r e i n - Z a m i r , H o m m e r s o n , Johnston & K i n g s t o n e , 2 0 0 6 , in preparation). One was the c lass ic stop s ignal task, wh i le the other was a m o d i f i e d vers ion o f the continuous t racking task where the ch i ldren were asked to track a spaceship on the screen unt i l a sudden a larm indicated they should stop t racking. The results indicated that in both tasks go performance and var iab i l i t y d id not s ign i f icant ly d i f fer between A D H D and control ch i ldren , w h i l e stopping latencies were s ign i f icant ly s lowed in the A D H D chi ldren. Importantly , stopping latencies in the spaceship t racking task were also more var iable for the A D H D ch i ld ren . These results offer c o m p e l l i n g support for the heretofore untested predic t ion that stopping is both delayed and more var iable in ch i ldren w i t h A D H D . M o r e o v e r , a l though the two tasks i n v o l v e d very different procedures, the results indicated a s igni f icant correlat ion o f .55 168 between SSRT measures. This latter finding further confirms the conclusion presented in Chapter 3 of this dissertation, that the two tasks engage the same generic stopping mechanism. This logic can be applied to additional cases where response inhibition as measured by stopping is believed to be impaired, as in the very young and very old, following frontal lobe damage, or in disorders such as schizophrenia, and obsessive compulsive disorder. Furthermore, response control is also believed to be impaired following damage to additional neural substrates such as the basal ganglia (e.g., Gauggel, Rieger & Feghoff, 2004; van den Wildenberg, van Boxtel, van der Molen, Bosch, Speelman & Brunia, in press). The tracking task could be used to test the exact impairments in each case by providing multiple measures of performance in a relatively short amount of time. Closing Remarks The series of studies contained in this thesis have conveyed the development and application of a novel task to investigate stopping behavior as part of a general effort to understand how individuals volitionally control their actions. By critically applying this novel task to test the prevailing assumptions in the field, this body of work sheds new light on past studies as well as illuminating new avenues for future investigations of stopping performance. 169 References A r o n , A . R., & Po ld rack , R. A . (2006). Cor t i ca l and subcort ical contr ibut ions to stop s ignal response inh ib i t ion : role o f the subthalamic nucleus. Journal of Neuroscience, 26, 2 4 2 4 - 2 4 3 3 . B a r k l e y , R. A . (1997) . B e h a v i o r a l inh ib i t ion , sustained attention, and execut ive funct ions: construct ing a u n i f y i n g theory o f A D H D . Psychological Bulletin, 121, 6 5 - 9 4 . C a b e l , D . W . J . , A r m s t o n g , I. T., R e i n g o l d , E. , & M u n o z , D . P. (2000) . Cont ro l o f saccade in i t iat ion in a countermanding task us ing v isua l and auditory stop signals. Experimental Brain Research, 133, 431 - 4 4 1 . Casada , J . H . , & R o a c h e , J . D . (2005). B e h a v i o r a l inh ib i t ion and act ivat ion i n posttraumatic stress disorder. The Journal of Nervous and Mental Disease, 193, 102-109 . D a y , B . L. , & L y o n , L . N . (2000). Vo lun ta ry modi f i ca t ion o f automatic a rm movements evoked by mot ion o f a v i sua l target. Experimental Brain Research, 130, 159 -168. G a u g g e l , S . , R ieger , M . , & Feghoff , T. A . (2004). Inhib i t ion o f ongo ing responses in patients w i t h Parkinson's disease. Journal of Neurology, Neurosurgery and Ps, 75, 5 3 9 - 5 4 5 . Georgopou los , A . P. , K a l a s k a , J . F., & M a s s e y , J . T. (1981). Spat ial trajectories and react ion t imes o f a imed movements : effects o f pract ice, uncertainty, and change in target locat ion. Journal of Neurophysiology, 46, 7 2 5 - 7 4 3 . Hanes , D . P. , & Caipenter , R. H . S. (1999). Countermanding saccades in humans. Vision Research, 3 9 , 2 7 7 7 - 2 7 9 1 . Ladefoged , P. , S i lverste in , R. , & Papcun , G . (1973). Interrupti l i ty o f Speech. The Journal of the Acoustical Society of America, 54, 1105 -1108. L o g a n , G . D . , & B u r k e l l , J . (1986). Dependence and independence in responding to double s t imulat ion : a compar ison o f stop, change and dual - task paradigms. Journal of Experimental Psychology: Human Perception and Performance, 12(A), 5 4 9 - 5 6 3 . M c G a r r y , T., & Franks , I. M . (1997). A horse race between independent processes: E v i d e n c e for a phantom point o f no return in the preparation o f a speeded motor response. Journal of Experimental Psychology: Human Perception and Performance, 23(5), 1533 -1542. M o r e i n - Z a m i r , S . , H o m m e r s e n , P. , C h e n , M . , Johnston, C , & K i n g s t o n e , A . F. (2006). A novel motor control task reveals slowed and more variable stopping in children with ADHD. Poster presented at the Cogn i t i ve Neurosc ience , San Franc isco . M o r e i n - Z a m i r , S . , H o m m e r s e n , P. , Johnston, C , & K i n g s t o n e , A . F. ( in preparation). A nove l motor control task reveals s lowed and more var iable stopping in chi ldren w i t h A D H D . M o r e i n - Z a m i r , S . , & K i n g s t o n e , A . F. ( in preparation). E x a m i n i n g the stop s ignal : a cr i t ica l rev iew. N i g g , J . T. (2001). Is A D H D a d is inh ib i tory disorder? Psychological Bulletin, 127, 5 7 1 -598. Ooster laan, J . , L o g a n , G . D . , & Sergeant, J . A . (1998). Response inh ib i t ion in A D / H D , C D , c o m o r b i d A D / H D + C D , anxious, and control ch i ld ren : a meta-analys is o f 171 studies w i th the stop task. Journal of Child Psychology and Psychiatry, 39 , 4 1 1 -4 2 5 . O s m a n , A . , K o r n b l u m , S . , & M e y e r , D. E. (1986). The point o f no return in choice react ion t ime: contro l led and bal l is t ic stages o f response preparation. Journal of Experimental Psychology: Human Perception and Performance, 12(3), 2 4 3 - 2 5 8 . O s m a n , A . , K o r n b l u m , S . , & M e y e r , D . E. (1990). D o e s motor p r o g r a m m i n g necessitate response execut ion? Journal of Experimental Psychology: Human Perception and Performance, 1990(16), 183 -198 . Pau l ignan , Y . , M a c K e n z i e , C , M a r t e n i u k , R., & Jeannerod, M . (1991) . Select ive perturbation o f v i sua l input dur ing prehension movements 1. the effects o f chang ing object pos i t ion . Experimental Brain Research, 83, 5 0 2 - 5 1 2 . Pe l i sson , D. , Preb lanc , C , G o o d a l e , M . A . , & Jeannerod, M . (1986). V i s u a l control o f reach ing movements without v i s ion o f the l i m b . Experimental Brain Research, 62, 3 0 3 - 3 1 1 . Preb lanc , C . , & M a r t i n , O . (1992) . A u t o m a t i c contro l dur ing hand reach ing at undetected t w o - d i m e n s i o n a l target displacements. Journal of Neurophysiology, 67, 4 5 5 - 4 6 9 . R idder inkhof , K . R , B a n d , G . P. H . , & L o g a n , G . D . (1999). A study o f adaptive behavior : effects o f age and irrelevant in format ion on the ab i l i t y to inhib i t one's actions. Acta Psychologica, 101, 3 1 5 - 3 3 7 . Tannock , R. (2003). Neurophsycho logy o f attention disorders. In S . J . Sega lowi t z & I. R a p i n (Eds. ) , Handbook of Neuropsychology (2nd E d i t i o n ed . , V o l . 8, Part II, pp. 7 5 3 - 7 8 4 ) . A m s t e r d a m : E lsev ier . 172 van den W i l d e n b e r g , W . , van B o x t e l , G . J . M . , van der M o l e n , M . W . , B o s c h , D . A . , Spee lman , J . D . , & B r u n i a , C . H . M . ( in press). S t imula t ion o f the subthalamic reg ion facil itates the select ion and inh ib i t ion o f motor responses in Park inson 's disease. Journal of Cognitive Neuroscience. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0092957/manifest

Comment

Related Items