Contingency learning is not affected by con
ict experience: Evidencefrom a task con
ict-free, item-speci
c Stroop paradigm
Yulia Levin
a,
, Joseph Tzelgov
a
Department of Psychology and Zlotowski Center for Neuroscience, Ben-Gurion University of the Negev, Israel
b
Department of Psychology and Zlotowski Center for Neuroscience, Department of Brain and Cognitive Sciences, Ben-Gurion University of the Negev, and Achva Academic College, Israel
a b s t r a c ta r t i c l e i n f o
Article history:
Received 12 February 2015Received in revised form 14 December 2015Accepted 15 December 2015Available online xxxx
A contingency learning account of the item-speci
c proportion congruent effect has been described as an asso-ciative stimulus
–
response learning process that has nothing to do with controlling the Stroop con
ict. As sup-portive evidence, contingency learning has been demonstrated with
response con
ict
-free stimuli, such asneutralwords.However,whatgivesrisetoresponsecon
ictandtoStroopinterferenceingeneralis
task con
ict
.The present study investigated whether task con
ict can constitute a trigger or, alternatively, a booster to thecontingency learning process. This was done by employing a
“
task con
ict-free
”
condition (i.e., geometricshapes) and comparing it with a
“
task con
ict
”
condition (i.e., neutral words). The results showed a signi
cantcontingency learning effect in both conditions, refuting the possibility that contingency learning is triggered bythepresenceof ataskcon
ict. Contingencylearningwasalsonotenhancedbythetaskcon
ict experience,indi-catingitscompleteinsensitivitytoStroopcon
ict(s).Thus,theresultsshowednoevidencethatperformanceop-timization asa resultof contingency learning isgreater undercon
ict,implying that contingency learning isnotrecruited to assist the control system to overcome con
ict.© 2015 Elsevier B.V. All rights reserved.
Keywords:
Automaticity of readingItem-speci
c proportion congruent effectInformational con
ictTask con
ictContingency learningCognitive control
1. Introduction
Readingis anacquired human ability to decode and interpret visuallexicalsymbols.Inadults,thisabilityisknowntobeautomatic,thatis,itoccurs whenever a lexical stimulus is encountered. The most dramaticdemonstration of the automaticity of the reading process is an
interfer-enceeffect
obtainedintheStrooptask(Stroop,1935).Inthistask,partic-ipantshavetonamethecolorofvisuallypresentedwords(e.g.,blueforthe stimulus RED presented in blue ink) while ignoring their meaning(e.g., the word RED). There is no need to read the words to accomplishthetask and yet readingoccurs, asevidenced by slower response timesfor incongruent stimuli (e.g., RED in blue ink) than for neutral letterstrings (e.g., XXXX in blue ink). The fact that reading takes place inspite of the fact that it is not required, and even interferes with perfor-mance, demonstrates its automaticity (Perlman & Tzelgov, 2006).The interference, or con
ict, produced by the automatic perfor-mance of the irrelevant reading task has been shown to be a target of cognitivecontrol.Thatis,whenacon
ictbecomestoostrong,cognitivecontrol is able to reduce it. Much evidence for that ability of the cogni-tive system came from a bulk of studies that manipulated the propor-tion of congruent vs. incongruent stimuli to control the Stroop effect(e.g. Logan, 1985; Logan & Zbrodoff, 1979). The main
nding of thesestudies, or what is known as the
“
list-wide proportion-congruent ef-fect
”
, was that the magnitude of the observed interference effect wassmallerwhentheexperiencedcon
ictwastoostrong(i.e.,largepropor-tion of incongruent trials in the list).Several models have been proposed to explain the mechanism bywhich con
ict is reduced in the Stroop task (a con
ict-monitoringframework; Blais, Robidoux, Risko, & Besner, 2007; Botvinick, Braver,Barch, Carter, & Cohen, 2001; Botvinick, Cohen, & Carter, 2004; DePisapia & Braver, 2006). According to Botvinick et al.'s (2001);
Botvinick et al.'s (2004) con
ict-monitoring architecture, increasingthe proportion of incongruent trials raises the amount of (response)con
ict (i.e., stronger competition between the response activated bythecolor-namingprocessandtheirrelevantresponseactivatedbyread-ing). The elevation in con
ict is detected by the con
ict-monitoringunit, which in turn signals units responsible for control exertion. Thecontrol is achieved through focusing attention on the relevant task.This way the irrelevant reading task does not get much attention andthe con
ict it produces is considerably reduced.It has also been proposed that the control system is not just able toreduce the con
ict accumulated at the list level, but is also
exibleenough to reduce the con
ict produced by speci
c items in the list(Bugg, Jacoby, & Chanani, 2011; Bugg, Jacoby, & Toth, 2008; see alsoBlais et al., 2007). The
“
item-speci
c proportion-congruent
”
effect( Jacoby, Lindsay, & Hessels, 2003; Jacoby, McElree, & Trainham, 1999)demonstrates that when the proportion of incongruent stimuli is
⁎
Corresponding author at: Department of Psychology, Ben-Gurion University of theNegev, P.O.B. 653, Beer Sheva 84105, Israel.
E-mail address:
levinyu@post.bgu.ac.il (Y. Levin).http://dx.doi.org/10.1016/j.actpsy.2015.12.0090001-6918/© 2015 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
Acta Psychologica
journal homepage: www.elsevier.com/locate/actpsy
manipulated at the level of speci
c words, the words mostly presentedas incongruent stimuli tend to produce less interference than thosemostly presented as congruent stimuli.
1.1. Cognitive control, learning and what is in between
Theprocessofcognitivecontrolisassumedtoproceedinawaythatcan be described as
“
automatic
”
, that is, without assuming any hiddenagency deciding when, where and how to intervene. As such, for thelast couple of decades there has been some tension in this research
eldtodifferentiatebetweenthe
“
automaticcontrol
”
andsimplelearn-ing mechanisms, or to de
ne how much the former may be relying onthe latter.According to the proposal of Verguts and Notebaert (2008, 2009), asimple learning process may be in fact
“
in service of control
”
. Speci
-cally, their proposal holds that the goal of control (i.e., con
ict reduc-tion) can be achieved through associative (Hebbian) learning thatbinds together all currently active (i.e., task-relevant) representations.Thatis,accordingtothisaccountthecontrolisactuallybasedonalearn-ingprocess.Itisimportantforthepresentdiscussiontonotethatexceptfor extending the general con
“
where
”
tointervene,itsharesmostofitsotherfeatures.Thus,whenthecon
ictissuf
cientlyreduced, less learning occurs, which means learning in this situation isdependentonandguidedbythemagnitudeoftheexperiencedcon
ict.In contrast, there are suggestions that learning does not represent amechanism
“
in-service-of control
”
, but separate cognitive phenomenathat sometimes might
“
mimic
”
the effects of control. That is, learningisassumedtoproduceanindependent(confounding)effectonreactiontime(RT)thathappenstolookliketheeffectofcon
ictreductionattrib-uted to cognitive control. In the context of the Stroop task, Schmidt,Crump, Cheesman, and Besner (2007) proposed that the item-speci
cproportion-congruent effect might be driven by such learning that hasnothing to do with (controlling the) con
ict. This
contingency learning
account of the item-speci
c proportion-congruent effect is based onthe fact that in the original study of Jacoby et al.'s (2003), as well as inthe recent replications (Hazeltine & Mordkoff, 2014), faster responseswere also observed for congruent items in a mostly congruent condi-tion, as compared to the condition where the probability of an itemappearing in a congruent or incongruent color was equal.
1
This resultcannotbeaccounted forbyassumingtheinterventionofcognitivecon-trol,sincecongruentitemsdonotproduceresponsecon
contingency learning
account (Schmidt, 2013a, 2013b;Schmidt & Besner, 2008; Schmidt et al., 2007; for contrasting viewssee Bugg & Hutchison, 2013; Bugg et al., 2011; Hutchison, 2011; seealso Abrahamse, Duthoo, Notebaert, & Risko, 2013; Atalay & Misirlisoy,2012;Bugg,2014),theitem-speci
cproportion-congruenteffectisbet-ter described as a speeding-up observed for the words frequentlyappearing in a speci
c (be it congruent or incongruent) color, and isduetothefactthatmanipulationofproportionsattheitemlevelcreatescontingencies between speci
c words andresponses. Thesecontingen-cies are learned and subsequently used to predictresponses. For exam-ple,ifREDfrequentlyappearsinblueink,thelearnedassociationwouldbe
“
if the word is RED then push the
‘
blue
’
button
”
. Note, in contrast tothe learning-based control view (Verguts & Notebaert, 2008, 2009),contingency learning is not assumed to be aided by response con
ict,but rather to represent a general ability to bind stimuli and responseson the basis of their existing correlations. The mechanism of contin-gency learning as implemented in the parallel-episodic processingmodel (Schmidt, 2013a) has no feature that is able to measure the re-sponsecon
ict,norhasitapropertyallowingforallocationofattentionin an adaptive manner, and yet it successfully simulates the pattern of the item-speci
c proportion-congruent effect.However, according to recently reported data, which will bediscussed shortly, there might be a third type of control
–
learning rela-tionshipthatcomesrightinbetweenthetwoaforementionedproposalsand which is at the focus of the present study. Recent studies showedthat implicit learning processes might not be completely independentof con
ict as suggested by Schmidt et al. (2007) for contingency learn-ing. However, the way they depend on con
ict does not
ict-monitoring framework) either. Deroost,Vandenbossche, Zeischka, Coomans, and Soetens (2012) presented aprobabilistic sequence of the colors in the Stroop task, which was im-plicitly learned by the participants. They found that sequence learningdidnothelptoreducethecon
ict(i.e.,theStroopeffect).Stroopcon
icthowever, was shown to enhance the expression
2
of learning: the ac-quired sequence knowledge was used more under con
ict(i.e., incongruent) conditions than under con
ict-free (i.e., congruentand neutral) conditions. Boostingeffects of con
ict on implicit learninghave also been reported in other studies. Deroost and Soetens (2006)observed a larger sequence learning effect for participants who weretrained with incompatible (i.e., con
icting) than compatible stimulus
–
response mappings. Similarly, Zhao, Ngo, McKendrick, and Turk-Browne (2011) found that engaging in a secondary (i.e., interfering)taskduringthetrainingphase,asopposedtopassiveviewing,improvedstatistical learning. Finally, Vandenbossche, Coomans, Homble, andDeroost (2014) reported a larger sequence learning effect for agedadults under high-interference (i.e., a dual task performed in the samemodality) than under low-interference (i.e., a cross-modal dual task)training condition.To summarize, the implicit sequence learning was not found
“
toserve control
”
by
reducing the con
ict
(Deroost et al., 2012) as assumedfor learning-based control (Verguts & Notebaert, 2008, 2009). Yet, theobserved enhancement and stronger reliance on implicit learning incon
ict environments (Vandenbossche et al., 2014; Zhao et al., 2011;seealsoDeroost&Soetens,2006;Koch,2007,Exp.1)speaksforthepos-sibilitythatimplicitlearningprocessesmightneverthelesscontributetocognitive control. However, this may happen not by reducing the con-
optimization of thetaskperformance
:
“
Optimizationoftaskperformancewasaccomplishedby an increased reliance on implicit sequence knowledge under highcon
ict.Thisindicatesthatimplicitlearningprocessescanbe
exiblyre-cruitedtosupportcognitivecontrol
”
(p.15).Thisideamightseemnovelin the domain of cognitive control, since the latter is traditionally de-scribed as the process that is sensitive to the amount of con
ict andthat aims to minimize this con
ict when it gets too strong. However,considering the control process more broadly makes it is perfectlyclear that the reduction of the con
ict is only needed to ensure a goodlevel of performance in the ongoing task. Stated otherwise, the
nalgoal of the control process is to protect the performance from thecon
ict-relateddecline.Bugg's(2014)studyprovidesempiricalsupportfor such a view of cognitive control. It was shown in a series of experi-ments that in a high-con
ict context, cognitive control was only en-gaged as a
“
last resource
”
, when stimulus
–
response associations didnot allow maintaining a suf
cient level of performance. Thus, the mag-nitudeofthecon
ictseemsonlytomatterwhenithasadetrimentalef-fect on performance. This emphasizes the importance of theperformance rather than con
ict per se in the context of controlengagement.Onewaytopreservetherequiredperformancewhencon
ictarises,as suggested by the con
ict-monitoring theory, is by reducing the
1
2
As opposed to
acquisition
of learning that was not affected by the amount of con
ict(manipulated by the proportion of congruent trials) at the training phase.40
Y. Levin, J. Tzelgov / Acta Psychologica 164 (2016) 39
–
45
con
ict.Thisisobviouslythemostef
cientwaythatsolvestheproblemfrom the root
—
if you want to protect the performance from interfer-ence, eliminate the cause of the interference. However, as the resultsof the aforementioned studies suggest, there might be another way to
“
assist
”
maintaining good performance in the face of con
ict, namely,recruitment of the implicit learning processes and stronger reliance onwhat was learned by these processes. For example, learning what but-ton should be pushed when the word that appears is
“
red
”
(i.e., contingency learning) allows shortening the time that is neededto respond, thus promoting better performance. Such performance-boosting learning seems to be especially useful when performance hasbeen damaged by con
ict. That is, it is conceivable that in con
ict situ-ations,thecontrolsystem,alongwithitsattemptstoreducethecon
ict,might also engage/enhance such an implicit learning process that mayhelp to maintain performance in a faster and resource-saving way.
1.2. Possible contribution of the contingency learning to cognitive control
RegardingStroopcontingencylearning,ithasbeenclaimedthatthislearning process does not represent learning-based control (as pro-posed by Verguts & Notebaert, 2008, 2009). This claim was supportedby empirical data demonstrating a contingency learning effect withneutral color-unrelated words (Schmidt et al., 2007). By showing thatthe effect of contingency learning was observable even when the con-gruency variable was not manipulated, the authors claimed that it hadto be regarded as an independent process that had nothing to do withcontrol.Oneof theaims of thecurrentstudyistocompletetheattemptofSchmidtetal.todisregardcontingencylearningasamechanismcon-tributing to cognitive control. Whereas Schmidt et al. only consideredone way in which learning might contribute to control
—
the way sug-gested by Verguts and Notebaert
—
we suggest that there is another op-tion that should be considered as well. Speci
cally, we investigatedwhether contingency learning in the Stroop task might represent the
“
assisting
”
type of the control
–
learning relationship, where the contin-gency learning is recruited not to reduce the con
ict, but to help thecontrol system optimize performance (Deroost et al., 2012) in some of the trials.What gives rise to the latter possibility is a notion of the multiple-con
ict nature of Stroop interference. Using neutral words in Schmidtet al.'s (2007) and Schmidt and Besner's (2008) experiments did not
eliminate the Stroop con
ict completely. This is because response con-
ict is not the only con
ict known to contribute to Stroop interference(Goldfarb & Henik, 2007; MacLeod & MacDonald, 2000; for contrastingviews see Melara & Algom, 2003). In fact, response con
ict is a directoutcomeoftheparallelactivationoftwocolorconceptsinthesemanticnetwork, which is frequently referred to as the
informational con
ict
3
Theactivationofonecolorconceptrepresentstheretrievaloftheword'smeaningasaresultofthereadingprocess,whereastheactivationofthesecondcolorconceptrepresentsprocessingofthecoloroftheinkthewordiswrittenin.Note,however,thatevenwhenthereadingdoesnotresultinactiva-tion of the irrelevant color concept, for example, when color-unrelatedstimuli such as neutral words (e.g., DOG) or letter strings (e.g., LGFD/XXXX) are used, the interference effect, though of smaller magnitude,is still observed (Brown, 2011; Klein, 1964; Sharma & McKenna,1998). This fact indicates that the
main
origin of Stroop interference isnot an informational incompatibility that leads eventually to theresponse con
ict, but a competition between two possible tasks thatone can perform when the stimulus has lexical properties
—
reading,which is automatically activated while not being a part of the task re-quirement (Tzelgov, 1997), and the color-naming task.
4
The competi-tion between the two tasks has been conceptualized by MacLeod andMacDonald (2000) as well as by Goldfarb and Henik (2007) to be the
task con
ict
, the existence of which has been well-documented by be-havioral (Entel, Tzelgov, Bereby-Meyer, & Shahar, 2014; Kalanthroff,Goldfarb, & Henik, 2013) as well as neuroimaging data (Bench et al.,1993; see also Aarts, Roelofs, & van Turennout, 2009; Steinhauser &Hübner, 2009).Asevidentfromthisdiscussion,interferenceduetotaskcon
ictcanbeobtainedaslongasthestimuluscanberead,regardlessofwhetheritis related to color or not (for an analysis further delineating the contri-bution of each type of con
ict to the Stroop effect see Levin & Tzelgov,2015). Hence, using color-unrelated words in Schmidt and Besner's(2008)andSchmidtetal.'s(2007)studieseliminatedtheresponsecon-
ict but not the main component of the Stroop interference
—
the taskcon
ict that was experienced by the participants in every trial. There-fore, in this experiment the contingency learning effect was only ob-served under a con
ict condition. This raises a question whetherconstant presence of the task con
ict in Schmidt and Besner's andSchmidtetal.'sexperimentscouldenhance(oreventrigger)thecontin-gencylearningprocessin order to optimizeperformance by facilitatingthe RTs on some of the trials. Answering this question and refuting thepossibilitythatcontingencylearningmaybeinvolvedin
“
assisting
”
con-trolbyoptimizingperformancerequiresevidencethatitcanalsobeob-served in a Stroop situation where no task con
ict is produced by thestimuli.Tothatend,inoneconditionweemployedanitem-speci
cpar-adigm with stimuli that could not be read (i.e., geometric shapes), andthe second condition was identical to that used by Schmidt et al.(2007) and used neutral words as stimuli. Comparing the performancein these conditions would allow
rst, to test whether contingencylearning is triggered
5
by con
ict experience. If that is the case, weshould not expect to observe the contingency learning effect in theshapescondition,butonlyinthewordscondition.Second,ifthecontin-gency learning effect is observed with shapes stimuli, it is possible totest whether its effect, as has been shown for other implicit learningprocesses, is boosted by the con
ict experience. The latter would ex-press itself in a larger contingency learning effect under a task con
ict(i.e., neutral words) condition. Results supportive of at least one of thesehypotheses would indicatethatcontingency learningis recruitedas a performance-optimizing tool assisting the control system to over-come con
ict.
2. Method
2.1. Participants
Forty-two undergraduate students at Ben-Gurion University of theNegev (25 females and 17 males, mean age = 25 years old,
SD
= 2.3),who were native speakers of Hebrew, participated in the experimentand were paid 25 NIS. All participants reported having normal orcorrected-to-normal vision acuity, as well as normal color vision. No
3
NotethatinatypicalStroopexperiment,eachcolorconceptrequiresadifferentvocalresponse, so in such experiments, response con
ict and informational con
ict are con-founded (but for a manual response design allowing decoupling of this confound see DeHouwer,2003).Thus,inthepresentarticlewedonotdistinguishbetweenthesetwotypesof con
ict and when mentioning
“
response con
ict
”
we refer to both the informationaland response con
icts.
4
Note,by
“
mainorigin
”
wedonotmeanthetaskcon
ictcontributesmostoftheStroopeffectinaclassiccolor-wordtask,butthatithasakeyroleininitiationoftheinterference.Simply put, withoutengagingintheirrelevantreading task, therewould benoStroopin-terference no matter what stimulus type (incongruent/neutral word) is used.
5
ict onlyaffects theex-pressionbutnottheacquisitionoflearning,webelievethatthe
“
triggering
”
isalesslikelyscenario than the
“
enhancement
”
.However,sincewe use anotherparadigmthan theoneusedoriginallybyDeroostetal.,itwouldnotbeexperimentallycorrectnottoconsiderthispossibility as well.41
Y. Levin, J. Tzelgov / Acta Psychologica 164 (2016) 39
–
45
Recompense su curiosidad
Todo lo que desea leer.
En cualquier momento. En cualquier lugar. Cualquier dispositivo.
Sin compromisos. Cancele cuando quiera.