This article was downloaded by: [Northwestern University]On: 28 July 2012, At: 08:25Publisher: Psychology PressInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Cognitive NeurosciencePublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/pcns20

More than a feeling: Pervasive influences of memorywithout awareness of retrievalJoel L. Voss a , Heather D. Lucas b & Ken A. Paller ba Department of Medical Social Sciences, Northwestern University Feinberg School ofMedicine, Chicago, IL, USAb Department of Psychology, Northwestern University, Evanston, IL, USA

Accepted author version posted online: 15 Mar 2012. Version of record first published: 30Mar 2012

To cite this article: Joel L. Voss, Heather D. Lucas & Ken A. Paller (2012): More than a feeling: Pervasive influences ofmemory without awareness of retrieval, Cognitive Neuroscience, 3:3-4, 193-207

To link to this article: http://dx.doi.org/10.1080/17588928.2012.674935

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form toanyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses shouldbe independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims,proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly inconnection with or arising out of the use of this material.

Discussion Paper

More than a feeling: Pervasive influences of memorywithout awareness of retrieval

Joel L. Voss1, Heather D. Lucas2, and Ken A. Paller2

1Department of Medical Social Sciences, Northwestern University Feinberg School of Medicine,Chicago, IL, USA2Department of Psychology, Northwestern University, Evanston, IL, USA

The subjective experiences of recollection and familiarity have featured prominently in the search forneurocognitive mechanisms of memory. However, these two explicit expressions of memory, which involveconscious awareness of memory retrieval, are distinct from an entire category of implicit expressions of memorythat do not entail such awareness. This review summarizes recent evidence showing that neurocognitive processingrelated to implicit memory can powerfully influence the behavioral and neural measures typically associated withexplicit memory. Although there are striking distinctions between the neurocognitive processing responsible forimplicit versus explicit memory, tests designed to measure only explicit memory nonetheless often capture implicitmemory processing as well. In particular, the evidence described here suggests that investigations of familiaritymemory are prone to the accidental capture of implicit memory processing. These findings have considerableimplications for neurocognitive accounts of memory, as they suggest that many neural and behavioral measuresoften accepted as signals of explicit memory instead reflect the distinct operation of implicit memory mechanismsthat are only sometimes related to explicit memory expressions. Proper identification of the explicit and implicitmechanisms for memory is vital to understanding the normal operation of memory, in addition to the disruptedmemory capabilities associated with many neurological disorders and mental illnesses. We suggest that futureprogress requires utilizing neural, behavioral, and subjective evidence to dissociate implicit and explicit memoryprocessing so as to better understand their distinct mechanisms as well as their potential relationships. Whensearching for the neurocognitive mechanisms of memory, it is important to keep in mind that memory involves morethan a feeling.

Keywords: Implicit memory; Explicit memory; Recollection; Familiarity; Awareness.

So many people have come and goneTheir faces fade as the years go byYet I still recall as I wander onAs clear as the sun in the summer sky . . .

—Donald Scholz and Boston, More thana feeling (1976)

Life is enriched by the ability to conjure long-pastepisodes back to mind in vivid detail. Yet not allexpressions of memory involve this experience ofconscious recollection. Indeed, differences betweenrecollection and another memory experience knownas familiarity are often emphasized in contemporary

COGNITIVE NEUROSCIENCE, 2012, 3 (3–4), 193–226 (includes Discussion Paper, Commentaries, and Reply)

Correspondence should be addressed to: Joel L. Voss, Department of Medical Social Sciences, Northwestern University Feinberg School ofMedicine, Abbott Hall, Suite 729, 710 N. Lake Shore Drive, Chicago, IL 60611, USA. E-mail: joel-voss@northwestern.edu

© 2012 Psychology Press, an imprint of the Taylor & Francis Group, an Informa businesswww.psypress.com/cognitiveneuroscience http://dx.doi.org/10.1080/17588928.2012.674935

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

memory research. The experience of familiarity isstrikingly illustrated by the butcher-on-the-bus phe-nomenon, which occurs when an individual (such asyour butcher) is encountered out of context (such as ona bus instead of in the butcher shop) and can be recog-nized but not identified (Mandler, 1980). Seeing thebutcher in this circumstance triggers a strong feeling ofknowing devoid of the contextual recollection thatwould be associated with successful identification.This acontextual familiarity arises with little or noeffort, although further retrieval attempts may lead torecollection of past encounters in the butcher shop thatwould support identification. A substantial portion ofmemory research over the past 40 years has focused onthe cognitive and neural characteristics of recollectionand familiarity (Yonelinas, 2002). In many ways, thisresearch has been revolutionary, as memory researchprior to the cognitive revolution largely avoided con-sideration of subjective experience (Mandler, 2008).Efforts to identify the cognitive and neural bases ofthese experiences have produced many fundamentalinsights into the mechanisms of memory.

We contend that a recent overemphasis on recollec-tion and familiarity, however, has caused much confu-sion and is impeding further progress. The chiefdownside of this recollection/familiarity focus is thatforms of memory without awareness of retrieval havebeen relatively ignored. These nonconscious or impli-cit expressions of memory occur with behavioral, cog-nitive, and neural signatures that are reliably producedin many circumstances, despite the fact that individualsare generally unaware of their occurrence (Dew &Cabeza, 2011; Eichenbaum & Cohen, 2001; Schacter,1987; Squire, 2004). As such, a near-exclusive focuson conscious, or explicit, expressions of memory, suchas recollection and familiarity, has led to a gross over-simplification of the complex relationships betweenneurocognitive processing and subjective experiences.The immediate ramifications of this situation aredistortions in our understanding of relationshipsbetween memory processing and memory experiences.Negative impacts are broadening as findings and meth-ods from memory research are increasingly applied toother psychological variables. Accordingly, we arguethat implicit memory must be given adequate consid-eration in cognitive neuroscience, in order to elucidatethe complexity whereby neurocognitive processinggives rise to subjective memory experiences. Some ofthis processing also falls into the category of implicitmemory processing and does not produce subjectiveexperiences. Although subjective states such as recol-lection and familiarity are important, there is more thana feeling to consider when searching for the neurocog-nitive basis of memory.

In this review, we will first describe some of thenegative impacts that have been spawned by an over-emphasis on subjective memory experiences. We willthen outline research that illustrates the broad reach ofimplicit memory processing, with a focus on resultsfrom experiments conducted in our laboratories. Thesefindings and the methods used to achieve them haveimportant ramifications for our understanding of theneurocognitive basis of memory and for the ways inwhich different memory processes can be characterizedexperimentally. Implicit memory processing can bedifficult to characterize, but doing so is nonethelessimportant as this type of processing can influencebehavior in ways that are powerful and under-appreciated. It is important to keep in mind what is atstake in the accurate identification of explicit andimplicit mechanisms of memory. Although our reviewfocuses primarily on the methodological issuesinvolved in distinguishing memory mechanisms fromthe conscious awareness of memory, these generalconsiderations are highly applicable to the attempt toidentify causes for disrupted memory in many neuro-logical disorders and mental illnesses. Indeed, our the-oretical and methodological suggestions are highlyrelevant to ERP investigations of memory disruptionsin conditions such as schizophrenia (e.g., Guillaume,Guillem, Tiberghien, & Stip, 2012) and Alzheimer’sdisease (e.g., Ally, McKeever, Waring, & Budson,2009), and we therefore suggest a strong focus on thedistinction between explicit and implicit memory pro-cessing in future studies on mechanisms for disruptedmemory.

BUTCHERING THE BUTCHER-ON-THE-BUS EXPERIENCE: THEOVERSIMPLIFICATION OFFAMILIARITY MEMORY

The subjective experiences of recollection and famil-iarity are relatively easy to identify; an individual rel-ives past events when recollection is experienced and“knows”when only familiarity is felt. Methods such asthe remember/know procedure are therefore sufficientto identify these experiences during memory testing(Yonelinas, 2002). Because these are such well-definedsubjective states, it is tempting to think that uncoveringthe mechanisms (i.e., neural processing events) respon-sible for each state should be equally straightforward.However, this endeavor is far from straightforwardif, as we contend, there is not a simple one-to-onemapping between the requisite neural events andexperiences of either recollection or familiarity. Thecomplexity of relationships between neural processing

194 VOSS, LUCAS, PALLER

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

and subjective experience must not be overlooked. Wedo not dispute that subjective states can be linked toneural processing. Rather, we oppose the notion thatrecollection and familiarity will each be associatedwith, for instance, a particular neural signal, becausethis proposition oversimplifies the complex originsof subjective qualities. Indeed, despite tremendousefforts, the neural mechanisms responsible for evenrelatively simple experiences, such as conscious visualperception, are still poorly understood (Lau &Rosenthal, 2011; Leisman & Koch, 2009), and recol-lection and familiarity are far more complex. Asdescribed below, the neural processing that accompa-nies subjective experience in memory experimentsdepends on factors such as the context of the retrievalevent, the nature of the retrieval cue, and other vari-ables that are relatively unrelated to the experienceitself. The operating principles of the neural systemsthat accomplish memory appear to be dictated by thenature of the representations that they support and thekind of computations performed on these representa-tions rather than by the resulting subjective experiences(for other specific examples, see Cowell, Bussey, &Saksida, 2010; Eichenbaum & Cohen, 2001; Henke,2010; Ranganath, 2010).

The realization that many forms of neurocognitiveprocessing can be relevant to the subjective “end state”of retrieval is relatively well appreciated for recollec-tion compared to familiarity. That is, there is a betterappreciation in the field that recollection is likely multi-faceted, varying in the type of neurocognitive proces-sing responsible and the relevant brain regions andbased on the content of the recollection experience.For instance, various findings have emphasized dis-tinctions between retrieval processing generally rele-vant for episodic memory and neural activationsspecifically associated with recollection. To brieflysummarize a substantial literature, various regions inprefrontal, parietal, and medial temporal cortex areimplicated in various encoding and retrieval functionsthat support explicit memory, including recollection(Buckner & Wheeler, 2001; Eichenbaum & Cohen,2001; Gabrieli, 1998; Simons & Spiers, 2003). In con-trast, processing particularly related to recollectionhas been reported in the form of activity in somemedial temporal regions, such as the hippocampus(Eichenbaum, Yonelinas, & Ranganath, 2007), aswell as in primary sensory cortex (Danker &Anderson, 2010). It is especially intriguing that activityis observed in sensory-specific cortex in a mannerconsistent with the contents of the recollection experi-ence. That is, olfactory cortex is active when recollec-tion involves previously smelled odorants (Gottfried,Smith, Rugg, & Dolan, 2004), auditory cortex is active

when recollection involves previously heard sounds(Goldberg, Perfetti, & Schneider, 2006), and so on,even when these modalities are not subject to externalinput during retrieval (i.e., no odorants or sounds pre-sented). These sensory-specific effects have beenreported for many stimulus categories (reviewed inDanker & Anderson, 2010), and are consistent withthe notion that the specific regions of cortex responsi-ble for originally experiencing an event are reactivatedwhen that experience is recalled. Of course, there ismuch still to learn about the mechanisms for recollec-tion. Nevertheless, attempts to account for recollectionmechanisms generally appreciate the complexity of therelevant retrieval events. Recollection is not associatedwith a specific all-purpose neural correlate, but insteadinvolves different kinds of neurocognitive processingdepending on the exact nature of the recollectionexperience and the situation in which it is produced.A tacit appreciation of this complex relationshipbetween the recollection experience and the neuralprocessing responsible for it is demonstrated by thefact that it is not common for memory researchers toclaim that recollection has occurred whenever “brainactivation X” is observed in an experiment. That is,it is extremely unlikely that the reverse inference(Poldrack, 2006) of the recollection experience isbased on some observed pattern of brain activity (infact, we are aware of only a few such inferences in theliterature). This is a good thing, given that even somefairly well-established “signatures” of recollection,such as activation of the hippocampus, have recentlybeen shown to occur even in the absence of subjectiveawareness of memory (i.e., during implicit expressionsof memory that do not involve the experience of recol-lection; Hannula & Ranganath, 2009; Rose, Haider,Salari, & Buchel, 2011). Indeed, some current theoriz-ing suggests that recollection may be supported byautomatic and implicit processing in the hippocampus,followed by the emergence of more widespread corti-cal interactions that support the subjective state ofrecollection (Moscovitch, 2008), although some evi-dence indicates that the second stage of this retrievalprocess need not follow the first (e.g., Hannula &Ranganath, 2009). In sum, recollection is too complexto be reduced to a one-to-one relationship with a parti-cular neural signal or brain structure.

In contrast, overly simplistic accounts of familiarityare common. At some level, this is not surprising giventhe definition of the familiarity experience: Familiarityis a feeling of knowing that is devoid of the contextualdetail that, if present, would be experienced instead asrecollection. Therefore, familiarity is defined primarilyin opposition to recollection. Indeed, the vast majorityof relevant neuroimaging investigations identify neural

MORE THAN A FEELING 195

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

signals of familiarity via an “exclusion” method, bywhich any brain activity that is related to memory thatis not associated with recollection is thereby attributedto familiarity. In fact, this means of definingfamiliarity has been the standard approach in almostall cognitive neuroscience experiments on recollectionand familiarity (as outlined by Paller, Voss, & Boehm,2007). However, little evidence exists to support thevalidity of this exclusion approach. We propose thatthis practice is extremely problematic, because manykinds of neural processing can be related to memoryyet are not necessarily associated with familiarity,including implicit processing. The exclusion approachwill lump all of this processing into the familiaritycategory without adequately testing whether it is func-tionally related to familiarity.

A prime example of the fruit of the exclusionapproach is the erroneous link between familiarityand a particular event-related brain potential (ERP)known as the FN400 (also known as the mid-frontalold/new effect; Rugg & Curran, 2007). The FN400 is anegative deflection at approximately 400 ms that issometimes maximal at frontal electrode locations.Several recognition memory experiments have foundthat item repetition affects FN400 amplitude indepen-dently of experimental manipulations that influence theexperience of recollection. By contrast, recollection istypically found to vary in conjunction with other ERPs,particularly the late positive complex (LPC) (seebelow1). Thus, in keeping with the aforementionedexclusion approach, effects on FN400 have been attrib-uted to familiarity. For example, among the most citedpieces of evidence for this attribution is a study thatutilized a plurality switch from study to test (e.g., studythe word “apple” and be tested on the word “apples,”thus requiring memory for the plurality “source”;Curran, 2000), to attempt to dissociate the neural cor-relates of recollection and familiarity. In that study,both FN400 amplitude and familiarity varied for stu-died compared to unstudied words, but neitherdepended on plurality. In contrast, the plurality switchreduced the experience of recollection as well as theLPC. The author therefore concluded that FN400

constitutes a neural correlate of familiarity. However,this is a faulty conclusion, because FN400 could havereflected any memory processing unaffected by theplurality switch and unrelated to recollection—suchas implicit memory processing—instead of or in addi-tion to familiarity. The same case can be made forvirtually every other experiment putatively linkingFN400 to familiarity (as reviewed in Paller et al.,2007). Although the possibility of implicit processingduring explicit memory tests (and vice versa) has longbeen appreciated (Richardson-Klavehn & Bjork,1988), this possibility is overwhelmingly ignoredwith regard to the interpretation of relevant neuraldata. It is especially problematic that implicit memoryprocessing is not measured in studies claiming to sepa-rate FN400 correlates of familiarity from implicitmemory (e.g., Curran & Doyle, 2011; Jager,Mecklinger, & Kipp, 2006; Woodruff, Hayama, &Rugg, 2006), and therefore these studies do not meettheir stated goals of separating neural correlates offamiliarity from implicit memory processing. Indeed,even findings of correlations between subjective famil-iarity strength and FN400 amplitude (e.g., Woodruffet al., 2006) constitute evidence from the exclusionapproach, given that implicit and explicit memory,although neurocognitively distinct, are often correlatedin strength (Paller et al., 2007), and no measures ofimplicit memory have been included to validate theputative exclusive link between FN400 and familiarity.Indeed, as reviewed below, testing in circumstances inwhich implicit memory is decoupled from familiarityabolishes the correlation between FN400 and subjec-tive familiarity strength (Voss & Paller, 2009c), thusshowing the relevance of implicit memory for FN400.We therefore believe that the link between FN400 andfamiliarity has arisen as a direct result of the oversightof implicit memory.

Based on this kind of evidence, FN400 has becomewidely accepted as a generic neural correlate of famil-iarity (Rugg & Curran, 2007). This assumption has sothoroughly permeated the field that it has become thenorm to infer that familiarity occurs whenever anFN400 effect is observed; that is, to infer the experi-ence of familiarity based only on FN400 without anyother direct evidence for familiarity (such as subjectivereport or behavioral performance). This inferencepervades the memory literature as well as experimentson other psychological variables (e.g., Czernochowski,Mecklinger, & Johansson, 2009; Ecker, Arend,Bergstrom, & Zimmer, 2009; Klonek, Tamm,Hofmann, & Jacobs, 2009; Mecklinger, Brunnemann,& Kipp, 2011; Nyhus & Curran, 2009; Opitz &Cornell, 2006; Rosburg, Mecklinger, & Frings, 2011;Speer & Curran, 2007). For instance, Rosburg and

1 Note that many studies have made overly simplistic interpreta-tions of the LPC as a unique and general correlate of recollection(e.g., Curran & Doyle, 2011); yet, there are other, more nuancedinterpretations as well (e.g., Finnigan, Humphreys, Dennis, &Geffen, 2002; Marzi & Figgiano, 2010). Nevertheless, although themethods commonly used to link recollection to LPC potentials(e.g., source memory vs. item memory comparisons) are far superiorto the exclusion approach used to link familiarity to FN400, and farless likely to be confounded by implicit memory processing, it is farless common to assume a mutually exclusive relationship betweenrecollection and LPC.

196 VOSS, LUCAS, PALLER

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

colleagues (2011) identified FN400-like effects whensubjects demonstrated a simple decision heuristicknown as the “recognition heuristic,” and on thissole basis they made the erroneous inference that famil-iarity is part of the mechanism for this heuristic.Furthermore, familiarity has long been thought toinvolve particular cognitive qualities, such as relativeautomaticity and rapid onset relative to recollection(Yonelinas, 2002); these qualities are also ofteninferred based on FN400 without any independentevidence (e.g., manipulations to show that effects areautomatic and not influenced by strategy, intentional-ity, or other factors). The common assumption is thusthat of a one-to-one mapping between FN400 andfamiliarity. However, as described next, there is directand incontrovertible evidence against this assumptionthat renders conclusions regarding familiarity in theaforementioned studies invalid.2

The one-to-one mapping between familiarity andFN400 necessary to predicate the inference of famil-iarity from FN400 requires a unique relationship,whereby (1) variations in FN400 effects are alwaysassociated with similar variations in the familiarityexperience (and not other experiences), and (2) varia-tions in the familiarity experience are always asso-ciated with similar variations in FN400 effects (andnot other ERP effects). Neither of these conditionsholds, based on very straightforward counterevidence.The experience of familiarity can occur for stimuli ofmany varieties, including words and nameable picturesas well as novel stimuli, such as geometrical patterns,that have never been seen before the experiment. Iffamiliarity is universally associated with FN400, theneffects on FN400 should generalize to all of thesestimulus categories. However—as described in moredetail in the next section—robust FN400 effects areobserved due to repetition of words and nameablepictures (Paller et al., 2007; Rugg & Curran, 2007),but FN400 effects are not observed for many types ofnovel stimuli (e.g., Voss & Paller, 2009c), even whenfamiliarity is strong (e.g., Danker et al., 2008).Furthermore, changing arbitrary features of stimuli,such as coloration, has no influence on familiarity, yetit influences FN400 (Groh-Bordin, Zimmer, & Ecker,2006). Therefore, variations in FN400 effects are notalways associated with similar variations in familiarity(condition 1). Finally, self-reported feelings of increas-ing familiarity strength had no relationship to effects on

FN400 for geometric patterns (Voss & Paller, 2009c),thus demonstrating that variations in the familiarityexperience are not always associated with similarvariations in FN400 effects (condition 2). In summary,very straightforward evidence weighs against both ofthe conditions that would need to be met in order toestablish FN400 as a generic indicator of familiarityand to permit inferences of familiarity based on FN400.

Given these considerations, it is reasonable to won-der how progress can be made in accurately identifyingneural mechanisms of familiarity, and, more specifi-cally, what relationship FN400 potentials have with theexperience of familiarity, if any. We have alreadyreviewed the simple evidence against the notion thatFN400 is a direct measure of the familiarity experiencein all circumstances. However, the fact remains thatFN400 and familiarity often co-occur. Therefore, it isworth considering other possible ways that FN400could relate to familiarity, such as (1) FN400 is a directmeasure of a neurophysiological process that serves asa precursor to familiarity in certain circumstances, or(2) it is a direct measure of a process that often occurs atroughly the same time as familiarity but is not a pre-cursor to familiarity. In the next section, we will reviewevidence that was obtained in an effort to address thisissue, emphasizing results from our laboratories. Whenreviewing this evidence, it is important to keep in mindthe ramifications these different outcomes have for ourunderstanding of familiarity memory as well as forrelationships between implicit memory processingand subjective awareness. Outcome (1) would suggestthat although FN400 may signal one specific precursorto familiarity, the extent to which this or other precur-sors contributed to familiarity varies with contextual/situational factors. Outcome (2) would suggest that theFN400 reflects other processes—such as implicitmemory processes—that are common during recogni-tion testing, and have been misattributed to familiaritysimply because they sometimes occur contempora-neously. Determining how FN400 and familiarity arerelated is therefore central to better understandingfamiliarity memory as well as the nature of relation-ships between memory processing and subjectivememory experiences more generally.

CONCEPTUAL IMPLICIT MEMORYPROCESSING DURING EXPLICIT

MEMORY TESTING

To explore the functional significance of FN400, it isfirst necessary to consider the multiple ways in whichmemory can be expressed, and how the relevant neu-rocognitive processing can relate—or not relate—to

2 There is also a strong argument against over-reliance on reverseinference of cognitive processing from neural signals in general, andfor an established set of criteria for making relatively valid reverseinferences (Poldrack, 2006) that have not been met for attempts toinfer familiarity from FN400.

MORE THAN A FEELING 197

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

subjective memory experience. We take as a startingpoint the very intuitive notion that various kinds ofneurocognitive processes can transpire during a mem-ory test, including those that support expressions ofmemory for which the test was not designed to capture.In other words, although tests of explicit memory areintended to provide measurements of processes rele-vant to the experiences of recollection and familiarity,neurocognitive processing unrelated to these experi-ences can nonetheless occur and can influence neuraland/or behavioral outcomes. As previously mentioned,a particularly relevant category of processing is thatwhich supports implicit expressions of memory.Implicit memory does not involve subjective experi-ences of memory retrieval, and is characterized asmemory processing that occurs when participants donot realize that their behavior has been influenced bypast experience. For instance, in priming tests—whichare commonly used to measure item-specific implicitmemory—procedures typically include initial presen-tations of specific items followed by tests in whichparticipants perform ostensibly non-mnemonic taskson repeated items intermixed with new items. In thesetasks, participants generally respond faster or moreaccurately to repeat items even when they evince noexplicit memory for the prior encounters. It has longbeen acknowledged that the implicit memory

processing that supports performance in these primingtests can be operative during tests designed to measureexplicit memory, and vice versa, such that tests do notgenerally provide “process pure” measures of eithermemory type (e.g., Richardson-Klavehn & Bjork,1988). Again, this is because repeating items duringmemory tests can have a multitude of effects on pro-cessing (Figure 1), including effects unrelated to theparticular behavioral outcome that is measured.Therefore, neural measures obtained during explicittesting need not correspond only to forms of explicitmemory such as familiarity or recollection, even ifthese are the only behavioral measures of memoryprocessing collected during the test. Instead, neuralcorrelates of repetition can reflect forms of processingrelated to implicit memory. This implicit memory pro-cessing may contribute to behavioral performance, butit also may not. Thus, considerable effort is needed todisentangle explicit and implicit memory processingand their neural correlates (see also Voss & Paller,2008a). Indeed, we have argued—using evidencedescribed below—that familiarity has been erro-neously associated with FN400 precisely because ithas not been disentangled from implicit memory pro-cessing, and, in fact, FN400 actually reflects a perva-sive type of implicit memory.

Our work in this area was originally motivated bythe suggestion by Olichney and colleagues (2000) thatN400 repetition effects in explicit memory tests mayreflect conceptual implicit memory. This suggestionarose because these N400 repetition effects were (1)relatively intact in individuals with impaired explicitmemory due to amnesia and (2) similar in several waysto the widely studied N400 correlates of semantic/con-ceptual processing (Kutas & Federmeier, 2011).Repeating a word, as well as other manipulations thatcause facilitated processing of word meaning, causes apositive shift in the amplitude of N400 effects (thesepositive shifts in amplitude are often called “N400reductions” in the literature on N400 priming effects,given that the N400 is a negative-going ERP peak thatbecomes more positive; i.e., it has a reduced negativepeak). Word repetition in memory experimentsinvolves a similar positive amplitude shift for FN400,which we argue results because both ERP effectsreflect conceptual implicit memory. In the primingliterature, conceptual implicit memory involves facili-tated processing of conceptual stimulus attributes(e.g., the meaning of a word or object), and is oftencontrasted with perceptual implicit memory, whichinvolves facilitated processing of physical form.Conceptual implicit memory often has been separatedfrom perceptual implicit memory by changing the phy-sical form of a stimulus across repetitions, but not the

Figure 1. Repetition influences many types of neurocognitive pro-cessing. Viewing a simple stimulus engages a multitude of neuro-cognitive processing steps. Likewise, repetition influences many ofthese same steps and elicits other types of processing as well. Some ofthese are shown for words and word repetition, highlighting the factthat the processing affected by repetition may or may not be related tothe relatively few outcomemeasures used in a particular memory test.Figure reproduced from Paller et al. (2007) with permission fromElsevier.

198 VOSS, LUCAS, PALLER

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

meaning (e.g., by presenting the same word in differentfonts; Schacter, Dobbins, & Schnyer, 2004). Thehypothesis that FN400 potentials are related to N400and reflect concomitant conceptual implicit memoryprocessing during recognition testing, rather thanfamiliarity per se, is of significance for identifyingvalid relationships between memory expressions andneural processing. Indeed, a demonstrated linkagebetween FN400 and conceptual implicit memorywould imply that implicit memory processing is ubi-quitous during explicit memory testing, such that it isoperative in virtually all situations in which familiarityfor meaningful stimuli has been studied. Conceptualimplicit memory thus may be an important but rela-tively uninvestigated part of the neurocognitive basisof memory.

Our first direct investigations in this area soughtto use behavioral measures of both conceptualimplicit memory and of the experience of familiar-ity to disentangle relevant neural processing duringa single memory test (as opposed, for instance, tomaking comparisons based on results from differenttest formats). Famous faces studied with relevantbiographical information, a source of relevant con-ceptual information that could influence later con-ceptual priming, were later identified faster thanfamous faces studied without this information(Voss & Paller, 2006). This conceptual primingprovided a behavioral measure of conceptual impli-cit memory processing during ERP recording. Later,subjects categorized the same famous faces usingratings of the experience of familiarity. Neural cor-relates of repetition recorded in response to famousfaces during the test for conceptual priming weresorted according to their association with concep-tual priming (studied with vs. without information)and their association with familiarity (high vs. lowfamiliarity ratings). A “double dissociation” ofneural correlates of conceptual priming and famil-iarity was thus identified: the magnitude of FN400potentials was associated with conceptual primingbut not familiarity, whereas the magnitude of ERPeffects occurring after FN400 and with a posteriordistribution––the LPC (see Voss & Paller, 2008b)––was associated with familiarity but not with con-ceptual priming. Furthermore, individual differencesin conceptual priming magnitude were strongly cor-related with individual differences in FN400,whereby subjects with stronger conceptual primingeffects also displayed larger FN400 effects. Theseindividual differences in conceptual priming werenot related to LPC amplitude. In contrast, indivi-dual differences in familiarity were associated withLPC amplitude, but not with FN400. Notably, these

selective associations between FN400 and concep-tual priming were identified for a subset of stimuliwith familiarity held constant (i.e., all given onefamiliarity rating level), whereas selective associa-tions between LPC and familiarity were identifiedwith conceptual priming held constant (i.e., just forfaces primed with biographical information). Theseresults provide compelling evidence that conceptualimplicit memory measured during a priming test isassociated with FN400, whereas familiarity in thesame circumstances is associated with a distinctERP correlate (LPC). A follow-up study using simi-lar methods in conjunction with fMRI also found adissociation between conceptual priming and famil-iarity (Voss, Reber, Mesulam, Parrish, & Paller,2008b). Conceptual priming was associated withactivity reductions in left inferior frontal cortex,whereas familiarity was associated with activityenhancements in right parietal cortex. These fMRIfindings support the link between FN400 and con-ceptual implicit memory, given that left inferiorfrontal cortical activity reductions are associatedwith conceptual priming for a variety of stimuluscategories and testing circumstances (Donaldson,Petersen, & Buckner, 2001; Schacter, Wig, &Stevens, 2007). The same comparisons made duringthe priming test isolated both FN400 and theseconceptual-priming-related response reductions,therefore suggesting their linkage and dissociatedthem from neural correlates of familiarity.

These experiments thus established that conceptualpriming and familiarity can co-occur during a primingtest, yet produce distinct neural correlates. We nextsought to determine whether this pattern extends torecognition memory tests similar to those normallyused to associated familiarity with FN400. The primarygoal was to identify neural correlates of conceptualimplicit memory during recognition memory testingand to compare them to neural correlates of familiarityobtained during the same test. This was a novel com-parison, because familiarity and conceptual implicitmemory are likely to be correlated under typical recog-nition testing circumstances (Paller et al., 2007). Thatis, when words or nameable pictures are used (as iscommon in recognition studies) familiarity can occur,but so can implicit memory for the conceptual aspectsof these meaningful stimuli. To avoid conflating neuralcorrelates of familiarity and conceptual implicit mem-ory, we used stimuli that differed from item to item intheir ability to support conceptual implicit memory—yet all stimuli could be recognized with familiarity.Specifically, we determined that novel visual shapes(termed “squiggles” due to their inclusion of curvedline segments) evoke meaningful associations in a very

MORE THAN A FEELING 199

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

idiosyncratic manner. That is, any individual will findmeaning only in a subset of squiggles, and there is ahigh level of variability across individuals with regardto which squiggles are perceived as meaningful.Moreover, this subset is highly consistent for any indi-vidual across delays of up to approximately 1 year(Voss & Paller, 2007). Critically, only those squigglesthat cue meaningful associations have the capacity tosupport conceptual implicit memory with repetition.Indeed, meaningful squiggles were found to supportconceptual priming in tests involving repeated ratingsof meaningfulness, when measured in priming tests,whereas meaningless squiggles were not (Voss,Federmeier, & Paller, 2011; Voss & Paller, 2007;Voss, Schendan, & Paller, 2010). In contrast, bothmeaningful and meaningless squiggles supported per-ceptual priming to the same extent in a task involvingperceptual judgments, thus indicating a selective asso-ciation between meaningfulness and conceptual prim-ing. Furthermore, both meaningful and meaninglesssquiggles can support recognition based on familiarity,and roughly the same proportion of squiggles of thetwo types yield familiarity responses (Voss et al., 2011;Voss & Paller, 2007). When subjects indicate theexperience of familiarity in a recognition memorytest, we reasoned that meaningful squiggles wouldengage neural signals of familiarity plus neural signalsof conceptual implicit memory, to the extent that con-ceptual implicit memory was operative during recogni-tion testing. In contrast, relatively meaninglesssquiggles would engage neural signals of familiarity,but not of conceptual implicit memory. Therefore, bymaking comparisons across meaningful and meaning-less squiggles that were matched for familiarity, wecould isolate neural signals of conceptual implicitmemory during recognition testing.

During recognition memory testing, familiarity wasapproximately matched for meaningful and meaning-less squiggles. ERP correlates of familiarity includedFN400 and LPC for meaningful squiggles, but onlyLPC for meaningless squiggles (Voss & Paller, 2007).This pattern of results indicates that FN400 potentialswere neural correlates of conceptual implicit memoryoperative during recognition testing selectively for themeaningful squiggles. Moreover, we also measuredconceptual implicit memory during a priming testusing the same stimuli, and found that the magnitudeof conceptual priming for meaningful squiggles (mea-sured as the repetition-related reduction in responsetime during the conceptual priming test) correlatedwith FN400 magnitude (Voss, Schendan, et al., 2010),thereby further supporting the link between FN400 andconceptual implicit memory irrespective of test format.Results from a follow-up study using fMRI also

support this conclusion (Voss et al., 2011). Using thesame general paradigm, neural correlates of concep-tual priming were identified selectively for meaning-ful squiggles. These correlates included activityreductions in regions of cortex strongly associatedwith the representation of meaningful objects, suchas anterior temporal cortex and anterior fusiform/parahippocampal cortex (Martin, 2007), as well asthe same left inferior prefrontal cortex regions gener-ally associated with conceptual priming and identifiedin our previous experiment that examined priming forfamous faces (Voss et al., 2008b). During a recogni-tion memory test, familiarity was associated withactivity reductions in the same regions only for mean-ingful squiggles. In contrast, both meaningful andmeaningless squiggles were associated with activityenhancements in prefrontal and parietal corticalregions that are commonly associated with explicitmemory and that have been dissociated from concep-tual priming (e.g., Donaldson et al., 2001).

To summarize these experiments, neural correlatesof conceptual implicit memory for meaningful squig-gles included FN400 potentials as well as canonicalfMRI correlates of conceptual priming. During recog-nition memory tests for these stimuli, FN400 and fMRIactivity reductions associated with conceptual primingoccurred when subjects made familiarity responses, butonly for the meaningful squiggles which are also cap-able of supporting conceptual priming. Althoughmeaningless squiggles were endorsed with familiaritywith similar prevalence, the neural correlates of famil-iarity for these items did not include FN400 or fMRIactivity reductions in the same regions. In contrast,familiarity for both meaningful and meaningless squig-gles was associated with LPC potentials and fMRIactivity enhancements in the explicit retrieval network.We can therefore conclude that meaningfulness can beused to dissociate behavioral and neural signals ofconceptual implicit memory and familiarity(Figure 2). Conceptual implicit memory for squigglesis associated with FN400 and fMRI activity reductionsin conceptual processing regions, and can occur forrelatively meaningful stimuli both during conceptualpriming tests and during recognition memory tests.

We also sought to show that these findings are notspecific to squiggles and generalize across stimuluscategories. We therefore used a very similar approachwith uncommon words. Definitions of these wordswere generally unknown to subjects. In fact, weexcluded any words with known definitions for eachindividual we tested (an average of about 10% of thewords). As with squiggles, the remaining words variedidiosyncratically in meaningfulness across participants(Voss, Lucas, & Paller, 2010). Conceptual implicit

200 VOSS, LUCAS, PALLER

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

memory occurred only for meaningful words, as indi-cated by significant effects in tests of conceptual prim-ing for these words but not for relatively meaninglesswords. During a recognition memory test, familiarityfor meaningful words was associated with FN400 aswell as LPC. Familiarity for meaningless words wasalso associated with LPC, but FN400 was absent(Figure 2B). Therefore, we replicated the findingsobtained with squiggles in that conceptual implicitmemory was associated with FN400 potentialsrecorded during a recognition memory test.

Comparisons have also been made between ERPcorrelates of recognition memory for stimuli of differ-ent categories altogether that vary greatly in meaning-fulness. For instance, recognition memory paradigmshave included items that evoke relatively little in theway of meaningful associations, such as novel faces(Yovel & Paller, 2004; MacKenzie & Donaldson,2007; but see Donaldson & Curran, 2007), complexand/or novel geometric patterns (De Chastelaine,Friedman, Cycowicz, & Horton, 2009; Voss & Paller,2009c), and Gabor patches (Danker et al., 2008), and

Figure 2. Stimulus meaningfulness can be used to dissociate ERP signals of conceptual implicit memory and familiarity. (A) The magnitudes offamiliarity, conceptual fluency, and FN400 are shown along a continuum of strong (green) to weak (red) according to variations in stimulusmeaningfulness. In repetition paradigms, stimuli that are inherently high in meaning (left) produce familiarity, conceptual fluency, and FN400(Paller et al., 2007; Rugg&Curran, 2007), whereas stimuli that are minimally meaningful (right) produce familiarity, but not conceptual fluency orFN400 (e.g., Danker et al., 2008; Voss & Paller, 2009c; Yovel & Paller, 2004). Stimuli that vary idiosyncratically across individuals inmeaningfulness (middle) support familiarity irrespective of rated meaningfulness, but produce conceptual fluency and FN400 only when ratedas relatively meaningful (Voss, Lucas, et al., 2010; Voss & Paller, 2007; Voss, Schendan, et al., 2010). FN400 therefore tracks conceptual fluencyrather than familiarity. (B) Example brain potentials are shown for visual words that were matched for familiarity but varied in the degree to whichthey were thought by the viewer to be meaningful (Voss, Lucas, et al., 2010). FN400 effects (relative to a new-word baseline) were observed onlyfor meaningful words, and FN400 was thus associated with conceptual implicit memory instead of with familiarity. LPC brain potentials weregreater than baseline for words endorsed as familiar from both meaningfulness categories and were therefore associated with familiarity.

MORE THAN A FEELING 201

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

ERP correlates of familiarity in these circumstanceshave not included FN400. These images should notbe expected to support conceptual implicit memory,and the absence of FN400 is therefore consistent withthe link between FN400 and conceptual implicit mem-ory. Nonetheless, this conclusion should be interpretedwith caution given that it can be problematic to com-pare neural correlates across stimulus categories (i.e.,common stimuli vs. novel stimuli) and across experi-ments (i.e., those that examine common stimuli vs.those that examine novel stimuli). However, the selec-tive association of FN400 with meaningfulness wasidentified in at least one study in which comparisonsbetween common meaningful stimuli and novel, rela-tively meaningless stimuli were made within the samesubjects (Danker et al., 2008). In general, these resultstherefore reinforce the differences in FN400 based onsubjective variations in meaningfulness of squigglesand uncommon words.

Finally, we also recently sought to determinewhether FN400 potentials signal conceptual implicitmemory for the stimuli most often used in recognitionmemory experiments: common words (Voss &Federmeier, 2011). Meaningful/meaningless compari-sons used in our experiments with squiggles anduncommon words would not adequately capture varia-tions in conceptual implicit memory for commonwords, given that all common words are relativelyhigh in meaningfulness and would be expected to read-ily support conceptual priming. Instead, we took adifferent approach, and used a manipulation of short-term conceptual priming often referred to as “semanticpriming.” This method of priming is a standard way tomanipulate the N400 correlate of conceptual proces-sing (Kutas & Federmeier, 2011). For instance, N400amplitude and conceptual processing of the word “doc-tor” varies according to whether it immediately followsthe related word “nurse” or the unrelated word “shoe.”We therefore focused on ERP correlates of familiarity-based recognition for two categories of words: (1)those that were primed by an immediately precedingrelated word, and (2) those that were not primed(immediately preceded instead by an unrelated word).We reasoned that neural correlates of conceptual prim-ing would be enhanced selectively for the primedwords during this recognition test. To the extent thatfamiliarity was similar for these two categories, neuralcorrelates of conceptual priming could therefore beseparated from those of familiarity. Subjects madevalence judgments (positive/neutral) to each word, fol-lowed by a recognition memory judgment usingremember, know, guess, and new responses.Consistent with our hypotheses, familiarity-basedrecognition was nearly identical for primed words and

words that were not primed. Familiarity for both wordcategories was associated with nearly identical FN400as well as LPC effects (the LPC effects were left-lateralized, as is often the case for recognition memoryexperiments using words). The conceptual primingmanipulation significantly increased the magnitude ofFN400 for the primed words, but did not influencefamiliarity. Therefore, we concluded that conceptualimplicit memory was indeed operative during recogni-tion memory testing for words, and was indicated byFN400. Familiarity was also operative—though it wasunaffected by semantic priming—and was indexed byLPC. Even when using common words as stimuli, forwhich familiarity and conceptual implicit memory areoften correlated (Figure 2A), these two memoryexpressions can be disentangled.

Based on these findings from many experiments,we conclude that conceptual implicit memory pro-cessing can be indexed by FN400 potentials (atleast in the circumstances we have investigated).These potentials appear to selectively associatewith conceptual implicit memory during primingtests specifically designed to measure conceptualimplicit memory, and also during recognition testsintended to measure explicit memory. Therefore,conceptual implicit memory processing is pervasiveduring tests intended to measure familiarity andrecollection; it is so pervasive, in fact, that itsneural correlates have been erroneously assignedto those of familiarity. In many circumstances,especially involving common words and nameableimages, familiarity and conceptual implicit memoryare correlated. Nonetheless, we have shown thatthey can be disentangled and linked to distinctneural correlates.

The findings we have summarized thus far suggestthat conceptual implicit memory is not a necessaryprecursor of familiarity. Familiarity and its neural cor-relates are nearly identical for relatively meaningfuland meaningless images, yet conceptual implicit mem-ory and its neural correlates are only present for mean-ingful images (Figure 2). Furthermore, a primingmanipulation that enhanced conceptual implicit mem-ory and its neural correlates did not enhance familiarityor its neural correlates (Voss & Federmeier, 2011). Itremains to be seen whether conceptual implicit mem-ory can support or influence familiarity in some cir-cumstances. Indeed, some behavioral evidencesuggests that manipulations of conceptual implicitmemory can sometimes influence familiarity (seebelow and Dew & Cabeza, 2011). However, the con-sistent patterns of dissociation between conceptualimplicit memory and familiarity described here suggestthat implicit memory is by no means necessary to

202 VOSS, LUCAS, PALLER

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

produce familiarity. Moreover, in the following sec-tion, we review evidence that implicit memory proces-sing can drive behavioral responses on explicitmemory tests in the absence of recollection or famil-iarity. These findings thus raise the possibility thatwhat appear to be influences of implicit memory onfamiliarity can sometimes reflect a direct impact ofimplicit memory processing on behavioral perfor-mance during explicit memory tests, without anyneed for familiarity to mediate the linkage betweenimplicit memory processing and behavior (just asimplicit memory influences behavior during primingtests, without any necessary explicit feelings offamiliarity).

EXPLICIT MEMORY IN NAME, IMPLICITMEMORY IN NATURE

The preceding section presented arguments supportingthe notion that neural signals during explicit memorytests can sometimes reflect implicit memory proces-sing. This notion calls into question the assumptionthat the neural measures one observes are necessarilylinked to behavioral and/or subjective qualities ofmemory that are expressed at the same time. We willnow review another set of findings that demonstratesan even stronger way in which memory tests are not“process pure.” In these experiments, behavioralexpressions of memory during tests intended to mea-sure explicit memory do not reflect explicit memory atall, but instead are determined by implicit memoryprocessing. These results demonstrate the ability ofimplicit memory processes to guide behavioral choicesduring memory testing—not because they interact withexplicit memory, but because they can direct mnemo-nic behaviors without involving the sense of awarenessthat characterizes explicit memory.

It is commonly assumed that performance in recog-nition memory tests reflects explicit memory proces-sing, partly because confident recognition responsescan be dissociated from implicit memory processing(e.g., Conroy, Hopkins, & Squire, 2005; Stark &Squire, 2000; Wagner, Gabrieli, & Verfaellie, 1997).However, there is also evidence that responses duringrecognition testing can be influenced by implicit mem-ory, such as when increases in perceptual and concep-tual fluency cause an increased tendency to endorsefluent items as studied (e.g., Keane, Orlando, & Ver-faellie, 2006; Verfaellie & Cermak, 1999;Whittlesea &Williams, 2000; Wolk et al., 2005). Some findingssuggest that this influence is particularly strong forfamiliarity responses (e.g., Rajaram & Geraci, 2000),consistent with the notion that the experience of

familiarity can arise when the sensation of fluency isattributed to prior experience (Whittlesea & Williams,2000). However, an alternative possibility is that influ-ences of implicit memory on recognition responses arenot accompanied by the phenomenology of familiarityor of any conscious memory experience. In otherwords, it could be that a false link has been madebetween implicit memory and familiarity merelybecause individuals in most of these memory experi-ments do not have the choice to respond “guess,” butcan only respond with “recollection,” “familiarity,” or“new” (or just “old” or “new”). In this way, the linkbetween these behavioral responses and the experienceof familiarity is inferred based on the widespreadassumption that performance in these tests reflectsonly explicit memory phenomena. Indeed, in oneexperiment individuals were given the option torespond “guess”—signaling no subjective experienceof memory retrieval—in addition to the standard recol-lection and familiarity options (Tunney & Fernie,2007), and the relationships between test cue fluencyand memory responses were measured. Fluency effectsin this experiment were restricted to guess responses,and did not influence familiarity. This demonstrationsuggests that implicit memory processing may notnecessarily lead to an increased sense of familiarity,and that experimenters have potentially identified falseassociations between fluency and familiarity because“guess” or “no awareness” response options wereabsent. Instead, implicit memory processing maydirectly drive behavior during recognition testing,influencing accuracy without simultaneously engen-dering any subjective memory experience.

To test these ideas, we assessed recognition memoryfor complex geometrical patterns (“kaleidoscopeimages”) in order to limit the subject’s ability to invokesemantic/conceptual encoding or elaborative retrievalstrategies that promote explicit memory (Figure 3A).Indeed, we found that subjects rarely find these stimulito be meaningful, endorsing less than 8% of images asbeing meaningful whatsoever, with an average ratingvalue corresponding to “no meaning whatsoever,”when the sole task was to attempt to find meaning inthe images (Voss & Paller, 2009c). We also developedrecognition test parameters intended to either increaseor decrease the relevance of implicit memory proces-sing for accurate responding. These parametersallowed us to determine whether influences of implicitmemory on recognition engender subjective experi-ences of familiarity, or, conversely, occur withoutawareness of these influences. Divided attention duringencoding has been found to reduce subsequent explicitmemory without affecting subsequent perceptualimplicit memory (Mulligan, 1998). We therefore used

MORE THAN A FEELING 203

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

both full-attention and divided-attention encoding con-ditions in order to manipulate the ability of subjects toengage in study operations that support explicit

memory. In addition, we alternately used either a yes/no format for testing recognition or a forced-choiceformat, wherein each target that was repeated fromthe study session was presented alongside an unstudiedfoil with a very similar appearance (Figure 3A). Thelatter format was intended to enhance the ability touse differences in visual fluency between the targetand the foil as a signal for old/new discrimination.Even though some targets would be more fluent thanothers, each target would tend to be more fluent than itscorresponding foil. Indeed, it has long been appre-ciated that perceptual discrimination can be performedwithout awareness only for highly similar stimuli dur-ing forced-choice format tests (Adams, 1957), becausegross perceptual stimulus differences enhance the roleof awareness in discrimination.

In several experiments, we identified striking influ-ences of implicit memory on recognition performancewhen the aforementioned testing parameters encour-aged a reliance on fluency (Vargas, Voss, & Paller,2012; Voss, Baym, & Paller, 2008a; Voss & Paller,2009a, 2010b). Furthermore, these influences occurredwithout awareness of retrieval on the part of subjects.We instructed subjects that “guess” responses were tobe made during tests only when recognition responseswere unaccompanied by awareness of retrieval––thatis, when the subjects were blindly guessing.Nonetheless, on the forced-choice tests, these guesseswere highly accurate. In fact, they were significantlymore accurate than were familiarity responses accom-panied by awareness of retrieval (Voss & Paller, 2009a,2010b). In contrast, guess response accuracy duringyes-no format tests was no better than chance (Vosset al., 2008a). Because forced-choice guess responseswere devoid of any experience of explicit memory,including recollection and familiarity, we reasonedthat highly accurate guesses were based on implicitmemory for the perceptual attributes that allowedfluency-based discrimination of targets from foilsselectively during forced-choice testing.

Several findings support this implicit-memoryaccount of accurate guess responding:

1. Performance in all experiments that used forced-choice testing increased when study was per-formed with divided attention as opposed to fullattention. This pattern is opposite to the deleter-ious effects that dividing attention during encod-ing commonly has on explicit expressions ofmemory (Mulligan, 1998). In contrast, yes-noperformance showed the standard effect ofreduced accuracy with divided attention.Furthermore, guesses were highly accurate forboth full- and divided-attention encoding.

Figure 3. Distinct neural signals of recognition based on familiarityversus perceptual implicit memory. (A) An example target/foil pair ofkaleidoscope images is shown. One image from the pair was studied,and later presented alongside its matched foil during forced-choicerecognition testing. (B) ERP signals of three distinct expressions ofmemory for kaleidoscope images, including (1) recognition based onfamiliarity, (2) recognition based on highly accurate guess responseswithout any awareness of memory retrieval, and (3) enhanced identi-fication speed and accuracy indicative of perceptual priming (Voss &Paller, 2009a, 2010a). Our interpretation that highly accurate guessresponses during forced-choice recognition testing were based onperceptual implicit memory is supported by the striking similaritiesbetween ERP correlates of accurate guesses and ERP correlates ofperceptual priming. Notably, ERP correlates of familiarity were dis-tinct from both ERP correlates of highly accurate guessing and ERPcorrelates of perceptual priming. Panel B adopted from Voss & Paller(2010b) with permission from Cold Spring Harbor Laboratory Press.

204 VOSS, LUCAS, PALLER

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

Subjects were instructed to memorize kaleido-scope images for the upcoming tests, and atten-tion was divided by having subjects perform aconcomitant 1-back task using auditory numericdigits, involving an odd/even judgment duringtrial n for the digit from trial n – 1. Divided-attention encoding reduced the prevalence ofexplicit memory responding, and yet increasedoverall accuracy by making highly accurateguesses more prevalent. Therefore, when explicitmemory was reduced by divided attention duringencoding, guess responses based on intact per-ceptual implicit memory exerted greater influ-ences on overall performance.3

2. This divided-attention advantage on forced-choiceformat tests was eliminated when gross perceptualdifferences were introduced between targets andfoils, therefore presumably limiting the relevanceof implicit fluency signals (Voss et al., 2008a; seealso Migo, Montaldi, Norman, Quamme, &Mayes, 2009).

3. The divided-attention advantage on forced-choicetests was also eliminated when subjects weregiven an extended period of time to respond dur-ing the test (Voss et al., 2008a), as this presumablyencouraged deliberation and explicit-memory-based responding.

4. Likewise, the high accuracy of guess responseswas eliminated when subjects were encouragedto adopt an explicit retrieval strategy, whereasguess responses remained highly accurate whensubjects were encouraged to respond withoutattempting explicit retrieval (Voss & Paller,2010b; see also Jeneson, Kirwan, & Squire,2010, for a similar trade-off between prevalentexplicit memory responding and guessaccuracy).

Based on these findings, we concluded that ourtesting parameters were suitable for identifying influ-ences of implicit memory processing on a recognition

test of the variety commonly assumed to measure onlyexplicit memory. Furthermore, when implicit memoryinfluenced performance, this influence was limited toguess responses that conveyed a lack of retrievalawareness. Experiments on implicit recognition collec-tively indicate that highly accurate guesses duringrecognition testing are most prevalent when (1) sub-jects follow instructions to minimize explicit memorystrategies and make many guess responses, (2) manip-ulations at study such as rTMS or divided attention areused to reduce strategies that aid explicit memory,(3) the test format emphasizes perceptual informationby utilizing a forced-choice format, and (4) responsesare made without much deliberation during test (basedon either response deadlines or instructions to guessfreely).

The neural signals related to highly accurate guesseslend additional support to the interpretation that theseresponses are based on perceptual implicit memory.Highly accurate guesses were associated with greaterERP negativity for targets compared to foils at occipitaland left frontal recording sites from approximately200–400 ms after stimulus onset (target and foil ERPswere separated by an alternating-presentation forced-choice design that produced the same behavioraleffects as the original behavioral experiments). In con-trast, both recollection and familiarity were associatedwith greater ERP positivity at distinct locations andlatency intervals (Voss & Paller, 2009a). Furthermore,a negative fronto-occipital effect similar to that identi-fied for accurate guesses was also related to behavioralmeasures of perceptual implicit memory for the samekaleidoscope images in another experiment. Perceptualimplicit memory was identified as faster and moreaccurate responding for repeated compared to newkaleidoscope images during a priming test involvingperceptual judgments of color composition, and themagnitude of these negative fronto-occipital ERPsscaled with response speed such that greater magnituderelated to faster responding (Voss & Paller, 2010a).Thus, ERP effects associated with perceptual implicitmemory during a priming test were similar to thoseidentified when subjects made highly accurate guessresponses during forced-choice recognition testing(Figure 3B).

Fronto-occipital negative ERP effects similar tothose we have found in association with highly accu-rate guess responses have also been linked to implicitmemory and have been dissociated from ERP corre-lates of explicit memory in other experimental circum-stances (e.g., Paller, Hutson, Miller, & Boehm, 2003).Notably, these negative repetition effects could becaused by reduced neural responses in frontal andoccipital cortex, which have been associated with

3 Note that we do not propose any necessary role for dividedattention encoding in producing these effects. This manipulation ismerely one of many that can be used to reduce explicit memory.Indeed, recent evidence shows that similar effects can be obtainedusing repetitive transcranial magnetic stimulation (rTMS) to disruptleft prefrontal cortical regions important for effortful encoding opera-tions that promote explicit memory. Temporary prefrontal disruptionusing rTMS just before encoding was associated with significantlyenhanced accuracy of guess responses during forced-choice formattests (Lee, Blumenfeld, & D’Esposito, 2011). Without rTMS, guessresponses were no better than chance, whereas guess responses afterrTMS were significantly more accurate than chance and approxi-mately as accurate as in our experiments.

MORE THAN A FEELING 205

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

repetition-related perceptual processing fluency(Schacter et al., 2007). Indeed, recent behavioral find-ings from our laboratory support the notion that theERP correlates of implicit recognition partly reflectfluent processing in visual cortical regions. Using alateralized presentation paradigm, we found that theaccuracy of guess responses was significantly influ-enced by consistency of visual hemifield from studyto test (Vargas et al., in press). Whereas guessresponses during forced-choice testing for kaleido-scopes presented in the same visual hemifield at studyand at test were highly accurate, the accuracy of guessresponses for kaleidoscopes presented in differentvisual hemifields at study and at test was no betterthan chance. One interpretation of these findings isthat highly accurate guesses under these circumstancesdepend on processing in contralaterally organizedvisual cortical regions, although additional evidence(e.g., complementary lateralized neural repetitioneffects) is needed to rule out alternative possibilities.Nonetheless, these findings show that study-test per-ceptual overlap is an important factor in highly accu-rate guess responses, thus further implicating the roleof perceptual fluency.

Based on these multiple behavioral and neuralfindings, we conclude that implicit memory proces-sing of perceptual stimulus attributes can have power-ful influences on performance in what are ostensiblytests of explicit memory. Furthermore, this implicitmemory processing can have a direct influence onperformance. In other words, implicit memory pro-cessing need not influence performance via an indir-ect influence on familiarity or recollection. Indeed,only guess responses indicating no awareness ofmemory retrieval demonstrate effects consistent withimplicit memory processing. Furthermore, theseguess responses can be dissociated from explicitmemory responses, including familiarity, on beha-vioral and neural grounds. Whereas familiarityresponses are less accurate than recollectionresponses, guess responses can be more accuratethan familiarity responses, and of similar accuracy torecollection responses (Voss & Paller, 2009a, 2010b).This U-shaped function across recognition confidencelevels is not consistent with the alternative interpreta-tion that guess responses merely reflect weak famil-iarity and/or weak recollection. Furthermore, neuralcorrelates of highly accurate guess responses werestrikingly dissociable from neural correlates of bothrecollection and familiarity (Voss & Paller, 2009a),and instead resembled neural correlates of implicitmemory in a priming test (Voss & Paller, 2010a). We

thus conclude that behavioral responses during arecognition test can be directly sensitive to perceptualimplicit memory processing, without any necessaryrole for awareness of memory retrieval. The couplingof behavior to neural signals of memory thereforedoes not appear to depend on any attributional processinvolving awareness, just as is the case for mostpriming tests of implicit memory in which subjectsdemonstrate faster or more accurate responses with-out acknowledging any influence (including attribu-tions) stemming from prior experiences.

Note that highly specialized testing circumstanceswere needed in order to identify these strong influencesof implicit memory on recognition performance. It istherefore important to consider whether responses aresimilarly influenced by implicit memory processing incircumstances more characteristic of explicit memorytesting. Although we do not currently know the answerto this question, some considerations suggest a possiblerole for more general influences of implicit memoryprocessing on recognition performance. Although con-fidence and accuracy are usually correlated duringrecognition testing, with high confidence associatedwith high accuracy and low confidence with low accu-racy (Heathcote, 2003; Yonelinas, 2001), this correla-tion is not universal. It is therefore possible thatperformance is influenced by implicit memory proces-sing on at least a subset of trials, perhaps especiallythose involving minimal explicit memory. Furthe-rmore, the association between confidence and accu-racy is sometimes inverted, with higher confidenceassociated with lower accuracy and lower confidenceassociated with higher accuracy (Dobbins, Kroll, &Liu, 1998; Heathcote, Bora, & Freeman, 2010;Heathcote, Freeman, Etherington, Tonkin, & Bora,2009; Tulving, 1981). These confidence–accuracyinversions are consistent with the effects we note ofhighly accurate guesses, as there is a fundamentaldecoupling of performance and awareness in eithercase. Notably, all reported confidence–accuracy inver-sions have been identified by forced-choice formattests with stimuli of relatively high target/foil percep-tual similarity. These testing parameters are perhapsgenerally ideal for enhancing influences of implicitmemory processing on recognition performance, sug-gesting that performance could be driven to someextent by implicit memory pervasively in tests intendedto measure explicit memory whenever some of theseparameters are included (e.g., in forced-choice recog-nition tests, including delayed match-to-sample ornonmatch-to-sample tests, and perhaps wheneverexplicit memory is relatively weak).

206 VOSS, LUCAS, PALLER

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

CONCLUSIONS

We hope that our descriptions of the pervasive opera-tion of implicit memory processing will contribute tobetter mechanistic understanding of both implicit andexplicit memory. In our view, aspects of memory pro-cessing associated with subjective awareness havebeen overemphasized, at the expense of accuratedescriptions of these mechanisms. The findings wereview here underscore why it is important not to pre-maturely assume relationships between neural mea-sures and self-reported subjective memory states.Experiences such as familiarity often occur at thesame time as implicit memory processing, and onlyby teasing apart these separate processes can relevantmechanisms be identified. Furthermore, it is just asimportant to refrain from assuming that tests intendedto measure explicit memory necessarily do so, as over-all performance can reflect combinations of both expli-cit and implicit memory processing. Prematureassumptions based on an overemphasis on self-reported memory experiences have already generatedconsiderable confusion in the memory literature(e.g., with respect to familiarity memory and implicitmemory, as reviewed here, and with respect to thedevelopment of animal models of explicit memory, assuggested by Voss & Paller, 2009b). The reach of thesemissteps is being amplified as memory paradigmsbecome increasingly important for mechanisticdescriptions of more general psychological functions(e.g., Rosburg et al., 2011, as described above).

It is therefore important that future investigations ofmemory processing refrain from drawing over-generalconclusions solely on the basis of neural correlates,

subjective reports, or assumptions about what kind ofmemory processing contributes to performance in aspecific test. Instead, all of these phenomena—neuralmeasures, subjective report, and behavioral perfor-mance—should be better integrated in memory experi-ments (see also Paller, Voss, & Westerberg, 2009, forrelated suggestions). In the experiments reviewed here,for example, careful assessments of results from self-reports of memory experiences, performance in recog-nition and priming tests, and neural measures obtainedin relation to all of these factors were necessary todisentangle familiarity and conceptual implicit mem-ory and to demonstrate separate influences of explicitand implicit memory processing during recognitiontests. In this way, we identified clear distinctionsbetween implicit and explicit memory. Conceptualimplicit memory occurred contemporaneously withfamiliarity, but was not a necessary precursor to famil-iarity experiences. Furthermore, perceptual implicitmemory sometimes determined responding duringrecognition testing, but without any awareness ofmemory retrieval or contribution from explicit mem-ory. Implicit memory and explicit memory are co-active in many circumstances, yet can be distinguishedby their distinct neural mechanisms considered in con-junction with their unique behavioral ramifications andsubjective qualities. Performance and subjectiveexperiences during any given test of memory likelyinvolve a complex interplay of both implicit and expli-cit memory processing. Overall, our results are consis-tent with a fundamental dissociation of implicit andexplicit memory processing, and show that implicitneurocognitive processing plays vital roles in memory,despite the fact that it occurs without a feeling.

MORE THAN A FEELING 207

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

Commentaries

Intermixing forms of memoryprocessing within the functionalorganization of the medialtemporal lobe memory system

Howard EichenbaumCenter for Memory and Brain, Boston University,Boston, USAE-mail: hbe@bu.edu

http://dx.doi.org/10.1080/17588928.2012.689958

Abstract:Voss et al. discuss evidence indicating an intermixingof implicit and explicit memory processing, and of familiarityand recollection, in tests of memory. Here I support this view,and add that the anatomy of cortical-medial temporal lobepathways indicates a hierarchical and bidirectional functionalorganization of memory in which implicit memory processingcontributes to familiarity, and implicit memory and familiarityprocessing inherently contribute to recollection. Rather thanlook for new ways to separate these processes, it may be asimportant to understand how they are integrated.

Voss and colleagues are to be applauded for raisingserious discussion about the relationships betweenmultiple dichotomies in memory, explicit vs. implicitmemory, recollection vs. familiarity, perceptual vs.conceptual priming, and questioning whether formalbehavioral and electrophysiological assays supportthese distinctions or instead indicate overlap amongthem. Their review, focused on ERPs as the metric ofmemory processing, favors the latter and the authorsconclude that measures of explicit memory also cap-ture implicit processing and other aspects of overlap.

Many studies have shown the combined contribu-tions of implicit and explicit processing, as well asfamiliarity and recollection, of the same material. Forexample, recent work has revealed implicit memorythrough eye movement patterns as subjects scannovel, familiar, and altered pictures, independent of,but intermixed with, explicit recall (Ryan, Althoff,Whitlow, & Cohen, 2000; Hannula & Ranganath,2009). Similarly, recent studies in animals haveshown that recollection-like and familiarity-like pro-cesses contribute in concert to object recognition(Eichenbaum, Fortin, Sauvage, Robitsek, & Favorik,

2010). The question Voss et al. raise is whether thecombined contributions of implicit/explicit or familiar-ity/recollection occur in parallel or interact.

The answer, I suggest, lies in a consideration of theanatomical basis for contributions of various forms ofmemory processing. The medial temporal lobe memorysystem involves three stages of interconnected structures:(1) Multiple higher-order single modality and multimo-dal cortical areas sending outputs that converge on (2)The parahippocampal cortical areas, including perirhinal,parahippocampal and entorhinal cortex, and outputs ofthese areas project to (3) Subdivisions of the hippocam-pus; note that major outputs of the hippocampus are sentback to the parahippocampal region and thence back tocortical areas that were the origins of medial temporalinput (Eichenbaum et al., 2007; see Figure 1).

There is substantial evidence that perceptual primingis supported by reactivation of modality-specific sen-sory representations in single modality cortical areas,whereas conceptual priming is supported by reactivationof material-specific multimodal areas (e.g., Keane,Gabriele, Fennema, Growdon, & Corkin, 1991). Thereis also substantial evidence that perirhinal cortex sup-ports familiarity whereas the hippocampus is critical torecollection (Eichenbaum, Yonelinas, & Ranganath,2007). Surely, these areas do not support these functionsin isolation, but bidirectional interactions between suc-cessive stages in this system are required.

Given this organization, it should be clear that dis-tinct types of processing interact in a hierarchical andinteractive fashion. The regeneration of modality-specific sensory representations reflected in perceptualpriming contributes to the retrieval of more abstractand multimodal representations reflected in conceptualpriming. Both of these implicit forms of memory arethe main inputs to the medial temporal lobe structures.Thus, implicit memories are the substance of informa-tion processing by parahippocampal areas, includingperirhinal cortex. Also, familiarity likely arises frombidirectional interactions between perirhinal cortex andappropriate higher-order cortical areas, and therebyintermixes implicit and explicit processing of mem-ories (e.g., a familiar face). Then the outcome of mem-ory processing in perirhinal cortex, as well as otherparahippocampal areas, are the substance of informa-tion processing within the hippocampus, which relates

208 COMMENTARIES

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

memories with each other and with the context inwhich they occur (Eichenbaum, 2004). Thus, recollec-tion arises from bidirectional interactions between thehippocampus and parahippocampal areas, and perhapsthroughout the entire cortical-hippocampal system.

From this perspective, the hierarchical organizationand two-way interactions within the system are fullyexpected to intermix the contributions of several formsof implicit and explicit processing within everydaymemory as well as formal tests of memory. So, inaddition to studying the distinctions between theseforms of memory processing, it may be as useful ormore useful to examine further how they are integrated.

* * *

You can feel it all over: Manysignals potentially contribute tofeelings of familiarity

Jason R. Taylor and Richard N. HensonMRC Cognition and Brain Sciences Unit,Cambridge, UKE-mail: jason.taylor@mrc-cbu.cam.ac.uk

http://dx.doi.org/10.1080/17588928.2012.689966

Abstract: Voss, Lucas, and Paller provide a thought-provoking summary of their recent research showing thatneural effects which are often attributed to (explicit) feelingsof familiarity can instead be attributed to the (implicit)effects of conceptual priming. Here, we discuss research

that shows effects of priming on (putative) behavioral andneural measures of familiarity, and consider a slightlydifferent interpretation: That multiple neurocognitiveprocesses can serve as signals to prior experience with atest item (i.e., can influence judgments of familiarity), andthe set of signals that will be interpreted as familiaritydepends on the experimental context.

Voss et al. review recent research showing thatbehavioral and neural effects that are typicallyattributed to “familiarity”, an explicit memory judg-ment, can instead be attributed to conceptual prim-ing, an example of implicit memory. We aresympathetic to the view that the influence of impli-cit memory on direct tests of memory is oftenunderestimated, particularly in relation to concur-rent neuroimaging data. To underscore this point,we discuss some research that uses masked primesto influence the processing fluency of test cues in arecognition memory paradigm. Our interpretation ofthese effects differs in detail, if not in spirit, fromthat proposed by Voss et al.

As Voss et al. note, previous exposure to an itemincreases fluency of processing on subsequent encoun-ters with the same item—a classic implicit memoryeffect. Although this increase in fluency due to priorexposure can influence participants’ performance with-out their awareness, such fluency could also contributeto feelings of familiarity and hence influence explicitmemory judgments. Jacoby and Whitehouse (1989)found evidence for just this sort of effect: Repetitionprimes presented briefly immediately beforerecognition-memory test items increased the likelihoodthat participants would judge those items as “old”. Theincreased tendency to judge primed items as “old”occurred even for items that had not been previously

modality-specificcortical areas

multimodalcortical areas

Parahippocampalregion (includingperirhinal cortex)

Hippocampus

Perceptual priming

Conceptual priming familiarity recollection

Figure 1. An anatomically based model of cortical-medial temporal functional organization.

This work was supported by the UK Medical Research Council(MC_A060_5PR10)

© 2012 Medical Research Council

COMMENTARIES 209

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

studied, suggesting that processing fluency was being(mis)attributed to memory. Subsequent studies havefound that this fluency manipulation selectivelyincreases “Know” and not “Remember” responses,suggesting that fluency is interpreted as familiarity(Rajaram, 1993; Woollams, Taylor, Karayanidis, &Henson, 2008).

The finding that processing fluency can influencefamiliarity underscores Voss et al.’s warning that thecontribution of implicit memory processes must beconsidered before conclusions about putative measuresof explicit memory are drawn. However, not all suchmeasures of explicit memory can be entirely explainedby priming: In an ERP version of the Jacoby-Whitehouse paradigm, we found that effects of primingand of familiarity occurred in the same time-window(300–500ms), but had different topographical distribu-tions over sensors, indicating that their neural sourceswere not identical (Woollams et al., 2008). This dis-sociation between ERP effects of repetition primingand of familiarity is perhaps unsurprising since, asVoss et al. point out, familiarity is a catch-all category,operationally defined as recognition without retrievalof context. Prior exposure to an item might increasefluency at any level of processing—perceptual, lexical,conceptual, etc.—each subserved by different neuralsources (which may be difficult to distinguish withEEG), and each with the potential to serve as a validsignal of familiarity (e.g., conceptual priming has alsobeen claimed to increase familiarity, Rajaram &Geraci, 2000; though see Taylor, Buratto, & Henson,submitted). The short stimulus onset asynchrony(SOA) masked repetition priming used by Woollamset al. likely emphasized perceptual fluency, which mayhave only been one of multiple neural signals thatcontributed to familiarity.

A second likely source of differences between ERPeffects of priming and familiarity is the fluency-attribution heuristic itself, or in Voss et al.’s terms, themechanism by which the memory process comes to beinterpreted as a memory experience. This attributionmechanism appears to be under conscious control:Participants are able to discount fluency arising fromobvious non-mnemonic sources, such as when repeti-tion primes are clearly visible, resulting in a reversal ofthe effect of priming on memory (Jacoby &Whitehouse, 1989). Indeed, whether and how any onetype of fluency is used as a memory signal may dependon the broader experimental context, such as the type ofinformation emphasized by the explicit memory

instructions (retrieval orientation), or the presence ofother sources of fluency. For example, masked concep-tual primes increase correct “remember” responses, butonly when repetition primes are also present in theexperiment (Taylor & Henson, in press).

In summary, we agree with Voss et al.’s generalposition that a closer look at the memory experienceof familiarity can reveal the action of underlyingimplicit memory processing. Evidence from a recog-nition memory paradigm in which test-cue proces-sing fluency is manipulated by priming suggeststhat fluency at multiple levels of processing cansignal that an item has been encountered previously.Future work is needed to identify the circumstancesthat determine which set of fluency signals will beattributed to memory in any given experimentalcontext.

* * *

On the contribution ofunconscious processes torecognition memory

Anne M. ClearyColorado State University, Department ofPsychology, Fort Collins, CO, USAE-mail: Anne.Cleary@colostate.edu

http://dx.doi.org/10.1080/17588928.2012.689965

Abstract: Voss et al. review work showing unconsciouscontributions to recognition memory. An electrophysiologicaleffect, the N300, appears to signify an unconscious recognitionprocess. Whether such unconscious recognition requires highlyspecific experimental circumstances or can occur in typicaltypes of recognition testing situations has remained a question.The fact that the N300 has also been shown to be the soleelectrophysiological correlate of the recognition-without-identification effect that occurs with visual word fragmentssuggests that unconscious processes may contribute to a widerrange of recognition testing situations than those originallyinvestigated by Voss and colleagues. Some implications of thispossibility are discussed.

Voss, Lucas and Paller review work showingunconscious contributions to recognition memory(e.g., Voss & Paller, 2009a). As they note, the

210 COMMENTARIES

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

circumstances under which such unconscious influ-ences were previously shown were highly specific;these include minimizing reliance on explicit memory,emphasizing relative perceptual information withforced-choice testing, and utilizing guess responses.The authors suggest that whether such unconsciousinfluences play a role in more standard recognitionmemory tasks remains to be determined.

Recent research suggests that such unconsciousprocesses may, in fact, contribute on tasks that moreclosely approximate standard recognition testing situa-tions. Ryals, Yadon, Nomi and Cleary (2011) examinedthe event-related potential (ERP) correlates of therecognition-without-identification effect, which is thefinding that when subjects cannot identify the itemsabout which they are being questioned at test, they canstill discriminate between those that were studied andthose that were not (e.g., Cleary, 2006; Cleary&Greene,2000; Cleary, Langley & Seiler, 2004; Cleary, Winfield& Kostic, 2007). Ryals et al. (2011) examined the ERPcorrelates of the recognition-without-identificationeffect found with visual word fragments (e.g., Cleary& Greene, 2000). Subjects studied words and weretested with visual word fragments, half from studiedand half from unstudied words. For each test fragment,subjects attempted to identify the word and gave a yes-no judgment indicating whether the word was studied(even if it was unidentified). The only ERP correlate tothe recognition-without-identification effect was thetype of N300 effect that Voss and Paller (2009) arguedsignified unconscious recognition. Regarding what thisimplies about the recognition-without-identificationeffect, one should be hesitant to use the very reverseinference that Voss et al. caution against. However,Ryals et al. (2011) also found that the N300 old-newdifference was only obtained among those subjects whoactually showed the behavioral recognition-without-identification effect; subjects who failed to show thebehavioral effect showed no ERP old-new differencesamong unidentified items. This suggests that the N300effect is indeed related to the behavioral recognition-without-identification effect, and raises the interestingpossibility that the recognition-without-identificationeffect shown with visual word fragments reflects thesame type of unconscious processing shown in Vossand Paller’s (2009) study. If so, then this would suggestthat unconscious contributions can occur in a widerrange of recognition memory testing situations thaninitially shown by Voss and Paller. In Ryals et al.’s

(2011) study, yes-no judgments were given on an item-by-item basis at test, the studied stimuli were wordsencoded under full attention conditions, and the testinginstructions did not at all emphasize trying not to rely onexplicit memory.

The idea that unconscious processes may contri-bute to a wider range of recognition memory deci-sions than those investigated by Voss and Paller(2009) has far-reaching implications. For example, itis common in the recognition literature to assume thatsource memory indicates the experience of consciousrecollection. However, Starns, Hicks, Brown andMartin (2008) demonstrated above-chance sourcememory in cases where subjects failed to even recog-nize a test item as old. Moreover, Kurilla andWesterman (2010) showed above-chance sourcememory for unidentified visual word fragmentswhen the usual recognition-without-identificationeffect was not even shown; in this case, above-chancesource memory occurred both when there was a fail-ure of recognition (i.e., old-new discrimination) and afailure of identification of the test stimuli. Takentogether, the findings of Voss and Paller (2009) andof Ryals et al. (2011) suggest the possibility that theabove-chance source recognition shown in these stu-dies resulted from unconscious processing. Thatsource memory might sometimes be unconscious isa notion that undermines the assumption, held bymany, that above-chance source memory reflects theexperience of conscious recollection.

* * *

Reconsidering the use of“explicit” and “implicit” as termsto describe task requirements

Jennifer D. RyanRotman Research Institute, Baycrest, Toronto,CanadaE-mail: jryan@rotman-baycrest.on.ca

http://dx.doi.org/10.1080/17588928.2012.689963

Abstract: Conscious and unconscious expressions ofmemory—explicit and implicit memory, respectively—maybe used to support performance in a given task, even when the

COMMENTARIES 211

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

task demands do not ostensibly require one or the other.Workfrom Voss, Lucas and Paller reveal that just as indirect taskscan capture the influence of explicit memory, direct tasksof memory can capture implicit memory mechanisms.Consequently, tasks cannot be truly labeled as explicit orimplicit; as such labels presuppose, perhaps erroneously, thenature of memory that supports performance.

Philosophers and psychologists have long remarkedthat there exist unconscious, as well as consciousexpressions of memory (c.f. de Biran, 1804/1929),and much research in the past fifty years has beendedicated to comparing and contrasting the two. Bynecessity, these concepts of memory were initiallyexamined using different techniques; conscious mem-ory was examined using direct enquiry, whereasunconscious memory was inferred through a changein behavior as a result of prior experience. Researchsuggested that while amnesic patients who had damageto the medial temporal lobe could not explicitly com-ment on prior learning episodes, they could neverthe-less benefit from prior exposure in some situations, andtherefore show memory “implicitly” (Schacter, 1987).Consequently, conscious and unconscious expressionsof memory were attributed to different systems—explicit and implicit memory systems, respectively(Schacter & Tulving, 1994).

It has become common practice for researchers toequate the particular tasks used to investigate explicitand implicit memory systems (e.g., free recall, prim-ing) with the systems themselves. Thus, the notion ofexplicit and implicit memory, and their distinct under-lying neural systems, has given rise to the notion ofexplicit and implicit memory tasks. Such an approach isnot only an oversimplification, as any given task is notlikely to be process- (or rather, memory-) pure (Ryan &Cohen, 2003), but it can lead to erroneous conclusionsregarding the nature of memory and the underlyingcontributions to cognition and behavior. For instance,performance on associative priming, often described asan implicit memory task, has been shown to be con-taminated by explicit memory under certain conditions(Hannula and Greene, 2012), complicating interpreta-tions regarding the organization of normal memory andthe nature of the impairment in amnesia.

Work that collected behavioral (i.e., eye movementmeasures) or neural responses in addition to explicitreports showed that on any given trial within the same

task, memory may be indexed via behavioral or neuralresponses with or without concomitant consciousawareness for the stored representations (Ryan et al.,2000; Hannula & Ranganath, 2009). Thus, it cannotbe guaranteed that explicit or implicit memory isbeing solely investigated across all trials, even if theoverall task is labeled as such. Even if participants arenot directly asked to examine the contents of theirmemories, they may nevertheless use whatever infor-mation is at their disposal in order to successfullycomplete a given task, including memories for priorlearning sessions for which they had consciousaccess. As a result, the onus is on any researcherwho wishes to comment upon the nature of implicitmemory to provide evidence that precautionswere taken to rule out any vestiges of consciouscontribution.

Now, Voss, Lucas, and Paller provide a compellingcase for the converse situation: Those who wish tocomment solely upon the nature of explicit memoryought to ensure that there is no confounding influenceof implicit memory. Their work demonstrates thatunder certain circumstances, accurate recognitionresponses in a direct enquiry task, typically ascribedas an explicit memory task, may be driven bymemoriesthat are not available for conscious access. Moreover,the underlying neural mechanisms typically attributedto responses of familiarity, and therefore of explicitmemory in general, may additionally capture mechan-isms underlying implicit memory.

The cautionary tale provided by Voss, Lucas andPaller will hopefully force researchers to reconsiderthe use of the terms explicit and implicit to describetasks when uncertainty exists regarding whetherany or all memory representations that are usedto support performance are available to consciousaccess. Instead, research should describe whenperformance on any given task fundamentally relieson memory representations that are available to con-scious awareness, compared to when performancedoes not require such conscious access. Such a dis-cussion can then truly reveal the nature of memory,its interaction with conscious awareness, and how itmay be used to support ongoing cognition andbehavior.

* * *

212 COMMENTARIES

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

Familiarity and conceptualimplicit memory: Individualdifferences and neural correlates

Wei-chun Wang and Andrew P. YonelinasDepartment of Psychology and Center for Mind andBrain, University of California, Davis, USAE-mail: wwa@ucdavis.edu

http://dx.doi.org/10.1080/17588928.2012.689968

Abstract: Voss, Lucas, and Paller point out that explicitrecognition tests can be supported by implicit processes,and that conceptual implicit memory may be reflected inERP correlates of familiarity-based recognition. Here, weargue that an examination of individual differencesindicates that familiarity is coupled with conceptual implicitmemory across participants, and that fMRI and patient dataindicate that the perirhinal cortex is critical for both forms ofmemory. We suggest that the same process that leads an itemto come to mind readily in conceptual implicit tests may alsolead the item to seem familiar in explicit recognition tests.

Voss and colleagues propose that memory tasks are notprocess pure and that explicit recognition judgmentscan be influenced by implicit memory. We are in agree-ment on this point and its implications for interpretingstudies of implicit and explicit memory. Models claim-ing that implicit and explicit memory reflect dissoci-able systems (e.g., Squire, 2004) have beenincreasingly challenged by evidence that these formsof memory are not entirely distinct, and that both can besupported by the same brain regions (e.g., Dew &Cabeza, 2011). One recent challenge involves parsingthe relationship between conceptual implicit memoryand familiarity-based recognition. Tests of conceptualimplicit memory and familiarity are found to be sensi-tive to the same behavioral manipulations (for a review,see Yonelinas, 2002), and as Voss and colleagues pointout they can be related to similar ERP effects. Oneinterpretation of these results is that explicit tests arecontaminated by implicit processes, or conversely thatimplicit tests are contaminated by explicit processes.However, we suggest another possibility: The sameprocess that leads an item to seem familiar in an explicitrecognition test can also lead an item to come to mindreadily in a conceptual implicit test.

We have taken two approaches to examine thishypothesis. First, if the two are related, then they shouldbe correlated across individuals. That is, participantswho exhibit high conceptual priming scores shouldalso have high familiarity estimates. To test this, wefirst had participants incidentally encode words by mak-ing abstract/concrete judgments. Afterwards, they maderecognition confidence judgments in a test containing amixture of studied and new words. Receiver-operatingcharacteristics were examined to provide estimates ofrecollection and familiarity (Yonelinas, 1994). In the lastphase, participants completed an implicit-free associa-tion task in which they were presented with non-studiedcue words and asked to produce the first related wordthat came to mind. Cues were selected to be associatedwith studied words from the previous tasks and unstu-died baselines. Priming was measured as the proportionof studied words generated relative to baseline. Acrossparticipants, overall recognition performance was corre-lated with conceptual implicit memory performance,and this correlation was driven by the fact that familiar-ity was correlated with conceptual priming, whereasrecollection was not (Wang & Yonelinas, in press).

Secondly, if familiarity and conceptual implicitmemory rely on the same process, then they may bedependent on the same neural regions. The perirhinalcortex has been implicated in numerous fMRI studiesof both familiarity-based recognition (for a review, seeDiana, Yonelinas, & Ranganath, 2007) and conceptualimplicit memory (e.g., Voss, Hauner, & Paller, 2009).Nonetheless, these correlational studies leave open thequestion of whether this region plays a necessary rolein these forms of memory. However, left perirhinaldamage can lead to impairments in familiarity-basedrecognition (Bowles et al., 2007; Yonelinas et al.,2002), as well as conceptual implicit memory (Wang,Lazzara, Ranganath, Knight, & Yonelinas, 2010).Moreover, the region of maximal lesion overlap inthat latter study included the left perirhinal cortex, aregion that was found to be related to increased activa-tion during encoding for subsequently primed items inhealthy participants (Wang, et al., 2010).

The simplest conclusion to be drawn from theseresults is that the same process that supportsfamiliarity-based recognition also supports conceptualimplicit memory. This does not mean that familiarityand conceptual implicit memory are identical, as pre-vious results have indicated that familiarity can besupported by both perceptual and conceptual factors(for a review, see Yonelinas, 2002). Nor does it mean

Supported by NIMH MH59325 and NRSA training grantMH096346.

COMMENTARIES 213

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

that there are no differences in the brain regions neces-sary for these two forms of memory (e.g., the responseand decision demands of recognition tasks are quitedifferent from those of implicit tasks, see Donaldson,Petersen, & Buckner, 2001). Finally, future work maydemonstrate that different subregions within the peri-rhinal cortex may be differentially involved in famil-iarity and conceptual implicit memory.

* * *

“Implicit contamination” extendsacross multiple methodologies:Implications for fMRI

Ilana T. Z. Dew and Roberto CabezaCenter for Cognitive Neuroscience, DukeUniversity, Durham, USAE-mail: ilana.dew@duke.edu

http://dx.doi.org/10.1080/17588928.2012.689972

Abstract: The article “More than a feeling: Pervasiveinfluences of memory without awareness of retrieval”reviews evidence from ERP studies of recognition memorythat the FN400 effect typically ascribed to familiarity mayindex implicit memory that occurs during recognition testing.We find their argument compelling, and contend that thispotential “implicit contamination” is not unique to ERPstudies. We suggest an analogous problem affecting fMRIstudies, focusing particularly on the perirhinal cortex.Resolving this issue is critical for understanding therelationship between memory and the medial temporal lobes.

The memory literature has dedicated considerableattention to the idea of explicit contamination, inwhich spontaneous memory awareness may compli-cate both behavioral and neural indices of implicitmemory (e.g., Roediger & McDermott, 1993).However, relatively little attention has been paid tothe reverse phenomenon, what we might term implicitcontamination. The article “More than a feeling:Pervasive influences of memory without awareness ofretrieval” (Voss, Lucas & Paller) identifies this as acommon confound in ERP studies, and reviews evi-dence that the FN400 effect typically ascribed to famil-iarity may index implicit memory that occurs duringrecognition testing. The evidence put forth makes a

compelling case that, if we are to understand the elec-trophysiological basis of memory, ERP studies ofrecognition memory must measure and control forconcurrent implicit memory.

Importantly, we contend that implicit contaminationis not unique to ERP studies. Evidence is mounting thatan analogous problem affects studies using functionalmagnetic resonance imaging (fMRI), with particularimplications for the perirhinal cortex (PrC). A vastbody of research agrees that the PrC makes an impor-tant contribution to recognition memory, particularlyfamiliarity memory (Brown & Aggleton, 2001;Eichenbaum, Yonelinas, & Ranganath, 2007; Henson,Cansino, Herron, Robb, & Rugg, 2003). During retrie-val, the PrC typically shows repetition suppression,with greater neural activity during novel relative tostudied stimuli, or parametric decreases in neural activ-ity with increasing familiarity (e.g., Daselaar, Fleck, &Cabeza, 2006; Gonsalves, Kahn, Curran, Norman, &Wagner, 2005; Montaldi, Spencer, Roberts, & Mayes,2006). A predominant explanation is that the PrCresponds to novel items (Brown & Aggleton, 2001), inturn decreasing activations for repeated items and produ-cing an efficient neural system for learning (Fernandez&Tendolkar, 2006). The link to familiarity memory couldthus be supported by this sensitivity to item representa-tions, independently from the contextual detail thatunderlies recollection (Eichenbaum et al., 2007).

Separately, however, the PrC has been linked withconceptual implicit memory. PrC reductions have beenobserved during repeated relative to novel semantic deci-sions (O’Kane, Insler, & Wagner, 2005) and the magni-tudes of conceptual priming and PrC reductions have beenlinked directly (Voss, Hauner, & Paller, 2009).Additionally, encoding-related PrC activity has beenshown to predict later conceptual priming (Wang,Lazzara, Ranganath, Knight, & Yonelinas, 2010). Theselinks between PrC and implicit memory raise the questionofwhether increased fluency associatedwith re-processingrepeated informationmay be integral to the function of thePrC during familiarity memory. That is, a neural regionoverwhelmingly interpreted to index item-specific proces-sing could alternatively reflect fluency-based processingthat occurs incidentally with item repetition. Fluencycould consequently account for cases in which famil-iarity and conceptual priming respond similarly to thesame experimental manipulations (Yonelinas, 2002).

This alternative interpretation of perirhinal functionis consistent with the logic put forth by Voss et al., andwe argue that fMRI investigations of recognition mem-ory must similarly isolate and identify the potential

214 COMMENTARIES

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

contribution of fluency. Resolving this issue is criticalfor clarifying the relationship between memory and themedial temporal lobes.

* * *

It is time to fill in the gaps left bysimple dissociations

Craig E. L. StarkCenter for the Neurobiology of Learning andMemory, University of California, Irvine, USAE-mail: cestark@uci.edu

http://dx.doi.org/10.1080/17588928.2012.689967

Abstract: Our field has been dominated by the quest fordissociations. A number of the dissociations have gotten usfar and the implicit/explicit one is an excellent example. Itholds in the vast majority of cases and has furthered ourunderstanding of memory considerably. There are nowseveral very interesting exceptions to this basic rule thathighlight how systems may interact and how independenceis an option or a default, but not a necessity. We should notthrow away the fundamental dissociation, but nor should wefail to learn from the interesting exceptions.

Voss, Lucas, and Paller raise a number of excellentpoints in their paper. While I might take issue with theclaim that there has been a “near-exclusive focus onconscious, or explicit expressions of memory” (there isa vast amount of literature that studies implicit forms ofmemory like categorization, priming, etc.), this is nottheir central tenet. Their central tenet is that the bordersbetween implicit and explicit memory, interactionsbetween them, and the fact that any task will reflectsome mixture of processes has been neglected to ourdetriment. I agree wholeheartedly with this view.

Studies of severely amnesic patients like H.M. andE.P. have taught us many things as have the studies ofrodent and primate animal models of amnesia. Onething they have reliably shown is that implicit andexplicit forms of memory can be independent and wecan design tasks in which the behavioral outcome isdictated largely, if not entirely, by one form of memory.In my own work, despite trying to push patient E.P. totap into his intact implicit memory for words,

recognition memory remained steadfastly at chance(49.9% correct) across numerous experiments, evenwhen a stem-completion probe immediately precededthe recognition memory probe for each item (Stark &Squire, 2000). That these forms of memory can beindependent is a clear contribution to our understand-ing of memory. However, this does not mean that weshould adopt a Fodorian view (Fodor, 1983) of com-plete modularity as these demonstrations do not showthat they must be independent.

Amnesia and its animal models have taught us otherthings as well. For example, we can see that thesevarious memory processes typically occur in parallel.Even if the behavior is driven largely or entirely by onememory process or system, others will continue toshow learning and memory (Packard & McGaugh,1996). Learning is typically ubiquitous and automaticand comes in many forms. Further, these studies com-bine with decades of cognitive psychology to showunequivocally that the instructions (and/or taskdemands) matter a great deal and help push the beha-vior to be driven by different memory processes orsystems. The instructions push—and they bias—butthis needn’t be an all-or-none process.

Let me use as an example their observation ofimplicit memory effects on a recognition memorytask (e.g., Voss & Paller, 2010). The authors havereplicated this effect several times (see the manuscriptfor review), but it also has a notable replication failure(Jeneson, Kirwan, & Squire, 2010). As noted, a clearlegacy of amnesia (and of cognitive psychology) is thatwe use instructions and task demands as ways of bias-ing participants to have behavior driven by one form ofmemory largely over others. Why should it be surpris-ing that there is something about the experiment(instructions, stimuli, context, mood, etc.) that canpush subjects to have their behavior driven by implicitmemory more than explicit memory despite the moreovert task demands?While it certainly appears to be thecase that the vast majority of the time when participantsperform standard laboratory recognition memory tests,“explicit” memory and the medial temporal lobe are inthe driver’s seat, why should it be surprising that underspecific circumstances this can be over-ridden? Isn’tthis exactly what the initial reports (and now decades ofresearch) into implicit vs. explicit memory have taughtus? Despite external similarities, simply changing theinstructions and/or demands can have a huge effect onwhat memory process or system drives behavior. Whatexactly is driving their effect (Voss & Paller, 2010)?

COMMENTARIES 215

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

Finding out the key factors and boundary conditionswill push us forward.

Their data do not overturn or call into question thedissociations in amnesia or the notion that these sys-tems can operate separately if pushed to do so. Theydon’t indicate even that under typical circumstances,implicit memory drives recognition memory perfor-mance to a significant degree. These basic ideas arewell-established and their findings show us more detail.They show us that while the dissociation captures much,it is not perfect and that there can be exceptions to therule. Other studies have shown exceptions to the dis-sociation rule as well. Chun and Phelps’ (1999) demon-stration of an implicit visual search task relying onmedial temporal lobe (MTL) structures (but seeminglynot the hippocampus per se; Manns & Squire, 2001)demonstrates this as does the often-used visual pairedcomparison task (which is implicit in nature but reliesupon the MTL).

Our dissociations have taken us far. It is time for usto learn from the interesting cases of how and when thedissociations break down and how and when the manymemory processes and systems interact.

* * *

The impact of fluency on explicitmemory tasks in amnesia

Scott M. Hayes1,2 and Mieke Verfaellie11Memory Disorders Research Center, VA BostonHealthcare System and Boston University School ofMedicine, Boston, USA2Neuroimaging Research Center, VA BostonHealthcare System, Boston, USAE-mail: smhayes@bu.edu

http://dx.doi.org/10.1080/17588928.2012.689974

Abstract: Distinguishing implicit and explicit memory anddelineating their relationship has haunted memory researchers

for decades, and Voss et al. provide an impressive overview oftheir work examining these issues. We briefly comment on thefollowing: (1) There is evidence indicating that implicitmemory impacts cued recall, in addition to recognition; (2)Fluency can manifest as priming in implicit memory or it canbe experienced as familiarity (in association with attributionprocesses) in recognition tasks; and (3) The impact of fluencyon accuracy of “guess” responses during recognition memoryin normal subjects is reminiscent of similar effects onrecognition in amnesia.

Voss, Lucas, and Paller (VLP) provide a detaileddescription of a primary challenge facing memoryresearchers: Integrating neural data, behavioral perfor-mance, and subjective report. Their review motivatesconsideration of the broader distinction between impli-cit and explicit memory processes and their relation-ship. They argue that: (1) Conceptual priming has adistinct neural correlate from familiarity-based recog-nition; and (2) There is no one-to-one mappingbetween these two measures. Despite these differences,they emphasize the “accidental capture” of implicitmemory in recognition.

Influences of implicit memory on explicit memoryperformance are not unique to recognition but are alsoobserved in cued recall. In amnesic patients, similarword stem completion performance has been observedunder explicit cued recall and “opposition” instructions(to isolate implicit memory, subjects are asked toexclude studied items). This has been taken as evidencethat their explicit memory performance is largely basedon implicit memory (Cermak, Mather, & Hill, 1997).Similar implicit effects on cued recall have beendescribed in normal cognition (McCabe, Roediger, &Karpicke, 2011).

Unlike VLP, we do not consider this “capture”of implicit memory in explicit memory tasks to beaccidental. Rather, we postulate that enhanced flu-ency (facilitated processing) that manifests as prim-ing in implicit memory can be experienced asfamiliarity during recognition. Thus, we postulatea more direct link between fluency and feelings offamiliarity. VLP might argue against such a viewbased on the failure to obtain a one-to-one mappingbetween priming and familiarity, but it is importantto consider that whereas priming is a direct conse-quence of fluent processing, familiarity requires anadditional process whereby fluency is attributed toa memorial source.

This work was supported by the Department of Veterans AffairsRehabilitation Research and Development Service and ClinicalScience Research and Development Service.

Thisworkwas authored as part of theContributor's official duties asan Employee of the United States Government and is therefore a workof theUnited StatesGovernment. In accordancewith 17U.S.C. 105, nocopyright protection is available for such works under U.S. Law.

216 COMMENTARIES

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

Whether fluency is attributed to previous experiencewill depend on both task and individual factors. Whenan alternate source of fluency is readily apparent, afeeling of familiarity may not result because there isno discrepancy between the experienced and expectedfluency (Whittlesea & Leboe, 2003). Informative inthis regard is a comparison of results obtained byVoss and Federmeier (2011) and Rajaram and Geraci(2000), as both examined the effect of semantic prim-ing on recognition. In the former study, which used along stimulus onset asynchrony (SOA) betweensemantic prime and target (5.5–7s), conceptual fluencydid not affect recognition. By contrast, in the latterstudy, which used a short SOA (250ms), conceptualfluency did affect recognition. With a long SOA, therewas ample time to note the relationship between primeand target, and thus the fluency was expected. With ashort SOA, the experienced fluency was higher thanexpected, and this discrepancy was attributed to famil-iarity with the target. Familiarity responses are alsoaffected by subjects’ willingness to utilize a fluencysignal and the criterion set for making familiarityjudgments. Manipulating information about thealleged proportion of study items in a recognitiontest (70% vs. 30%; although in fact the proportionwas held constant) improved memory discriminationin amnesic patients, with the enhancement reflectinga selective enhancement of familiarity-based recog-nition (Verfaellie, Giovanello, & Keane, 2001).Reliance on fluency signals in recognition may alsobe influenced by the processing approach encour-aged by the task (Voss & Paller, 2009a).

An intriguing finding presented by VLP concernsinstances in which fluency enhances accuracy of“guess” responses in recognition paradigms. The influ-ence of fluency on guess responses in healthy subjectsis reminiscent of similar effects in recognition tasks inamnesic patients. Indeed, the conditions outlined asbeing conducive to fluency effects are characteristicof amnesia. In the context of normal cognition, whenfluency effects translate into the experience of famil-iarity or simply into guess responses will be an impor-tant question for future study. Here as well, attention tofactors that influence willingness to engage an attribu-tional process may be useful in understanding the linkbetween fluency and its varied expressions in recogni-tion memory.

* * *

Electrophysiological correlatesof memory processes

Edward L. Wilding and Lisa H. Evans

Cardiff University Brain Research Imaging Centre(CUBRIC), School of Psychology, Cardiff, Wales,UKE-mail: wildinge@cardiff.ac.uk

http://dx.doi.org/10.1080/17588928.2012.689971

Abstract: In this commentary we highlight what are to ourminds conflicting findings that have been employed to arguefor different functional accounts of the mid-frontal event-related potential (ERP) old/new effect. We also offer ourviews on the difficulties associated with measuring concep-tual priming as well as familiarity, and reemphasise that theseissues are only a sub-set of those to consider when assessingthe ERP literature that is germane to the question of theproceses that the mid-frontal ERP old/new effect indexes.

It is acknowledged widely that the terms used to describememory tasks should not overlap with those used todescribe the processes engaged during completion ofthe tasks. This distinction emphasizes that implicit aswell as explicit memory processes might be active irre-spective of whether task instructions in memory studiesorient participants to information encoded previously.Voss et al.’s target article is a reminder of this point,with one articulation of their wider argument for theimportance of implicit memory research cast in terms ofwhether the functional significance of an event-relatedpotential (ERP) modulation—the FN400 or mid-frontalERP old/new effect—is aligned more closely with anexplicit memory process (familiarity) or an implicit pro-cess (conceptual priming).Among the data they cover in their target article,

Voss and colleagues highlight three of their ERP stu-dies in which measures of conceptual priming as wellas measures of explicit memory were available (Voss &Paller, 2006, 2007; Voss, Lucas, & Paller, 2010). Ineach case, a link between the mid-frontal old/neweffect and conceptual priming would stem from varia-tions in the old/new effect alongside variations in prim-ing when familiarity is equated.

In one of these three reports, Voss and Paller acknowl-edged that their measure of explicit memory cannotisolate familiarity (Voss & Paller, 2006). In the

COMMENTARIES 217

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

remaining two, visual inspection of the averaged datasuggests that familiarity estimates are equivalent whileconceptual priming varies only if accompanying differ-ences between estimates of recollection are not consid-ered (Voss & Paller, 2007; Voss, Lucas, & Paller, 2010).Voss et al.’s arguments would be stronger were they toshow that familiarity estimates in these studies remainequivalent after concomitant differences between esti-mates of recollection are controlled for. The data pointsin their 2010 paper appear to be those for which the claimof statistically equivalent contributions from familiaritywhile priming varies is most likely to be maintained.

Were this speculation to be correct, then the datawould be at odds with those reported by Stenberg andcolleagues (Stenberg et al., 2009). In a modified recog-nition memory task for names, they showed that themid-frontal old/new effect varied with name frequencybut not with whether the names referred to famous orunknown individuals. Estimates of familiarity derivedfrom receiver operating characteristic (ROC) data werecoupled with frequency rather than fame, while in aseparate assessment of conceptual priming, fame butnot frequency promoted priming. Lucas, Voss, & Paller

(2010) have already commented on these results, iden-tifying their reservations about the findings as well asadditional outcomes that would bolster Stenberg et al.’sclaims. In a subsequent response, Stenberg and collea-gues reported outcomes that appear to meet at leastsome of the criteria Lucas et al. set (Stenberg et al.,2010). These appear to us to be data points that meritfurther discussion.

Convergence on a consensual account for thefunctional significance of the mid-frontal ERP old/new effect is important. The focus in this briefcommentary is on only a subset of the issues thatwe think merit further consideration, another of notebeing the marked heterogeneity in analysis strategiesacross studies where ostensibly the same effect isbeing analyzed (for a similar comment see Voss &Federmeier, 2011). It is to be hoped that the debatein this Special Issue will result in progress towardsidentification of data points and perhaps futureexperiments that have the potential to adjudicatebetween alternative interpretations.

* * *

218 COMMENTARIES

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

Reply to Commentaries

On the pervasive influences of implicit memory

Heather D. Lucas1, Ken A. Paller1, and Joel L. Voss2

1Department of Psychology, Northwestern University, Evanston, IL, USA2Department of Medical Social Sciences, Northwestern University Feinberg School of Medicine,Chicago, IL, USA

In Voss, Lucas, and Paller (2012) we reviewed evidence that behavioral, neural and subjective phenomena oftenconsidered to be hallmarks of explicit memory can be powerfully influenced by processing related to implicitmemory. We argued that, as a result, behavioral and neuroimaging experiments designed to measure explicitmemory are sometimes prone to capture aspects of implicit memory instead. The nine commentaries published inresponse to our target article largely resonated with these conclusions. Here we highlight the numerous additionalinsights offered by these commentaries regarding the circumstances and precise mechanisms that may characterizerelationships between implicit and explicit memory, and we describe similarities and differences with ourinterpretations.

Keywords: Implicit memory; Explicit memory; Recollection; Familiarity; Awareness.

We are fortunate that so many excellent commentarieswere written in response to our target article (Voss,Lucas, & Paller, 2012). They largely resonated withour viewpoints and showed how our focus—influencesof implicit memory processing on behavioral, neural,and subjective qualities associated with explicit mem-ory—is being pursued by many laboratories and withmany approaches. For example, Cleary described howthe recognition-without-identification effect studied inher laboratory is similar to the implicit recognitioneffects we have characterized, and suggested thatthese phenomena might stem from similar neuralmechanisms. Taylor and Henson highlighted fluencyeffects on recognition behavior and ERPs producedwhen recognition test cue fluency is manipulated viamasked priming. Commentaries by Dew and Cabezaand by Wang and Yonelinas described research on therole of fluency in generating neural responses in peri-rhinal cortex during recognition testing, whileEichenbaum put forth a well-reasoned neuroanatomi-cal account of possible interactions between implicitand explicit memory processes. These contributions

extend the issues raised in our article beyond the spe-cific tasks and measures we described.

Although we emphasized effects on familiarity,commentaries by Hayes and Verfaellie and by Clearyemphasized that implicit memory processing maymore broadly influence all manner of explicit memoryexpression. For instance, implicit memory processinghas been shown to influence performance in tasksdesigned to isolate recollection (Cermak, Mather, &Hill, 1997; McCabe, Roediger, & Karpicke, 2011;Starns, Hicks, Brown, & Martin, 2008; Kurilla andWesterman, 2008). The assumption that subjectiveexperiences of recollection mediate performance onrecall and source memory tasks is arguably evenmore widespread than the assumption that familiarityexperiences dictate all recollection-free recognitiondecisions. That neither assumption is safe stronglybolsters the argument put forth by Ryan that it isproblematic to use the terms “implicit” and “explicit”to describe memory tasks—even recall tasks—becausethese terms presuppose the degree to which consciousawareness mediates task performance.

http://dx.doi.org/10.1080/17588928.2012.697454 REPLY TO COMMENTARIES 219

Correspondence should be addressed to: Heather D. Lucas, Department of Psychology, Northwestern University, 2029 Sheridan Road,Evanston, IL 60201, USA. E-mail: hdlucas@gmail.com

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

Other commentaries provided alternate ways ofthinking about relationships between implicit memoryprocessing and explicit memory performance. In parti-cular, some commentaries took issue with the idea thatsuch influences should be described as “accidental” orthe product of “contamination.” Commentaries byTaylor and Henson, by Wang and Yonelinas, and byHayes and Verfaellie suggested that these influencesinstead reflect a common source for implicit andexplicit memory expressions. Indeed, the notionthat the same fluency signals that produce implicitpriming effects can also produce experiences of famil-iarity on recognition tests is a central tenet of long-standing fluency-attribution accounts of familiarity(e.g., Jacoby & Dallas, 1981; Whittlesea, Jacoby, &Girard, 1990). We entertained this possibility in ourtarget article, and we agree that priming and familiaritymay derive from the same fluency signals under somecircumstances. As a point of clarification, we used theterm “accidental” not to suggest that fluency effects onfamiliarity are somehow an accident. Rather, we usedthis term to suggest that experimenters seeking to iso-late neural correlates of explicit memory can acciden-tally measure implicit memory instead, if they fail totake into account that some combination of the twotypes of processing are engaged in many circum-stances. It is no accident that individuals sometimesinterpret fluency signals as familiarity; rather, suchattributions constitute one way that memory decisionscan be made.

Although we agree that similar mechanisms canunderlie implicit and explicit memory expressions insome circumstances, we adhere to the key position wetook in our article—the assumption that familiarityexperiences mediate the relationship between fluencyand recognition decisions in any given situation mustnot be accepted without sufficient evidence. For exam-ple, the Jacoby and Whitehouse (1989) findings usingmasked priming of recognition test cues are frequentlyinvoked as evidence for fluency attribution accounts,particularly when taken together with subsequent find-ings that masked priming affects “Know” but not“Remember” responses. However, the option to“guess” or to rate decision confidence is seldom offered.As described in our target article, “Know” decisionsunder such circumstances may function as a catch-allcategory that includes not only responses associatedwith familiarity, but also guess responses that stemfrom implicit memory processing. Interestingly, onestudy that did include an option to guess (Tunney &Fernie, 2007) showed that masked priming increased thepercentage of old responses that were guesses, and notthose associated with familiarity or recollection.Without the “guess” response option, the influence of

implicit memory likely would have been erroneouslyattributed to familiarity. In a similar vein, Hayes andVerfaellie pointed out interesting ways in which theinfluence of fluency on “guess” responses is reminiscentof effects of fluency on recognition performance inamnesic patients. For example, Verfaellie, Giovanello,and Keane (2001) found that encouraging a liberalresponse criterion improved memory discrimination inamnesic patients.Whereas these and similar findings areoften interpreted as influences on familiarity, it is plau-sible that implicit recognition is at work instead. Indeed,we have found that liberal encouragement increasesimplicit recognition indicated with the “guess” optionbut has no influence on familiarity experiences separatefrom guessing (Voss & Paller, 2010).

It is important to emphasize that showing that amanipulation selectively affects “Know” or some gen-eral measure of non-recollective “old” responses is notthe same as showing that its effect is selective tofamiliarity. The research we described in our targetarticle illustrated that other mechanisms can producesuch effects without involving experiences of familiar-ity. Of course, these findings are in no way at odds withthe notion that fluency can contribute to familiarityexperiences under certain circumstances, and weagree with Hayes and Verfaellie that identifying thefactors that modulate whether fluency leads to implicitrecognition versus familiarity is an exciting topic forfuture research.

We suggest that our ability to answer these and otherquestions about relationships between implicit andexplicit memory hinges on our ability to separate theirrespective neural signals. In our target article, wedescribe dissociations indicating that neural signals offamiliarity can be reliably separated from those relatedto perceptual and conceptual fluency. The commentarybyWilding critiqued some of the specific data points weused to support the distinction between familiarity andconceptual fluency. In several experiments, we linkedFN400 ERPs with conceptual fluency, contradicting awidespread (and, we argue, incorrect) notion that FN400potentials index a generic familiarity signal.Specifically, we reviewed evidence that: (1)Associations between FN400 potentials and familiarityduring recognition were limited to situations in whichconcurrent conceptual fluency effects were likely, and(2) FN400 potentials correlated with behavioral mea-sures of conceptual fluency on priming tests, whereasother ERPs (the late-positive complex) correlated withbehavioral measures of familiarity. Thus, our position isthat FN400 potentials reflect conceptual fluency and donot directly or universally reflect familiarity.

Wilding expressed two concerns about this conclu-sion. The first was whether our studies were successful

220 LUCAS, PALLER, VOSS

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

in matching familiarity between situations that did anddid not permit conceptual fluency. Second, Wildingcited findings that seem to contradict our interpretations(Stenberg, Hellman, Johansson, & Rosén, 2009).Because behavioral measures of familiarity are imper-fect, there cannot be complete certainty that our neuralcontrasts of conditions in which familiarity occurredwith versus without conceptual fluency were 100%matched. Nevertheless, much of the evidence wedescribed contradicts the interpretation that our reporteddissociations between conceptual fluency and familiar-ity were due to subtle differences in familiarity. Somestimuli fail to evokemeaningful associations but can stillbe recognized. We found that such meaningless stimuliunquestionably supported familiarity-based recognition,yet did not produce FN400 ERPs. For stimuli that werenot conceptually meaningful in the study by Voss,Lucas, and Paller (2010), for example, high-confidence“Know” hits outnumbered false alarms by a factor ofnine, and yet no FN400 effects were present when thesetrials were compared with correct rejections (CRs).Thus, FN400 potentials cannot provide a reasonableindex of familiarity for these words. In contrast,familiarity-based recognition for conceptually meaning-ful items in our study occurred with indistinguishablydifferent accuracy, yet was associated with robustFN400 effects. Our target article also describes conver-ging evidence from other labs and paradigms that sti-muli that are devoid of conceptual fluency yetrecognized with familiarity do not produced FN400effects. For these reasons, we do not share the concernexpressed by Wilding about our findings, though weagree that there is value in continuing to find newways to dissociate familiarity from conceptual fluency.

RegardingWilding’s second concern, we appreciatethat Stenberg and colleagues (2009) attempted to dis-sociate ERP correlates of familiarity and conceptualfluency, but we remain skeptical of their interpretationsdue to methodological shortcomings (Lucas, Voss, &Paller, 2010). Their conclusions rest on the problematicassumption that conceptual fluency was not facilitatedby repetition of non-famous names. If this assumptionwere correct, then a dissociation between FN400 andconceptual fluency would have been achieved throughthe finding that FN400 differences between hits andCRs for non-famous names did not differ significantlyfrom the corresponding contrast for famous names. Onthe contrary, participants taking memory tests fre-quently manufacture meaning for initially unfamiliarstimuli as an intuitive mnemonic strategy, such thatunfamiliar stimuli perceived as meaningful can reliablyand robustly support conceptual priming (e.g., Voss &

Paller, 2007; Voss et al, 2010; Voss, Schendan, &Paller, 2010) as well as neural repetition effects inbrain regions that closely match effects of conceptualpriming for real objects and words (Voss, Federmeier,& Paller, in press). Moreover, it is precisely thosestimuli for which meaning is successfully invokedthat are likely to be remembered later. Thus, the useof Hit/CR contrasts to examine familiarity in non-famous names surely led to the selective inclusion oftrials corresponding to the names that held the mostmeaning for each participant—in other words, thenames with the capacity to support conceptual fluencyand cue successful recognition.

Although Stenberg and colleagues did attempt to ver-ify their claim that conceptual fluencywas enhanced onlyfor repeated famous names usingpriming tests of speededfame and frequency, only one of these tests—speededfame judgments—evinced priming for either category ofnames. Whereas this priming was indeed selective tofamous names, it is logical to question the utility of thismeasure for indexing analogous priming for both types ofstimuli given that famous and non-famous names requireopposite responses in order to be correct on a famejudgment task. Should fluency with non-famous namesbe expected to facilitate their identification as non-famous? In fact, fluency in this context might lead to atleast some degree of a false sense of fame for non-famousnames, thus impairing performance. Furthermore, a sec-ond pillar of the arguments of Stenberg and colleagues—the finding that FN400 potentials, but not priming effects,were greater for low-frequency relative to high-frequencynames—should be regarded with caution given theknown effects of frequency on conceptual priming forwords. Repetition of low-frequency words producesgreater facilitation on priming tests when compared torepetition of high-frequency words (e.g., Duchek &Neely, 1989; Forster & Davis, 1984; Scarborough,Cortese, & Scarborough, 1977) and there is little reasonto suspect that this would not also be the case for names.Accordingly, we remain unconvinced by Stenberg andcolleagues’ findings, but sympathize with Wilding’s callfor further investigation of this issue.

In summary, these discussions illustrate that there ismerit in putting further effort toward dissociating explicitrecognition memory from fluency in order to study thesephenomena in isolation (see also the comment by Starkfor a reminder of the ways that such dissociations havecontributed to our understanding of memory disorders).These efforts will remain relevant even as we follow theadvice put forth by Eichenbaum, Stark, and other com-mentators tobegin to“fill in thegaps” leftbydissociationsby focusing on interactions among memory systems.

ON THE PERVASIVE INFLUENCES OF IMPLICIT MEMORY 221

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

References from the Discussion Paper,the Commentaries, and the Reply

Adams, J. K. (1957). Laboratory studies of behavior withoutawareness. Psychological Bulletin, 54, 383–405.

Ally, B. A., McKeever, J. D., Waring, J. D., & Budson, A. E.(2009). Preserved frontal memorial processing for pic-tures in patients with mild cognitive impairment.Neuropsychologia, 47, 2044–2055.

Bowles, B., Crupi, C., Mirsattari, S. M., Pigott, S., Parrent,A. G., Pruessner, J. C., Yonelinas, A. P., & Kohler, S.(2007). Impaired familiarity with preserved recollectionafter anterior temporal-lobe resection that spares hippo-campus. Proceedings of the National Academy ofSciences, 104, 16382–16387.

Brown, M. W., & Aggleton, J. P. (2001). Recognition mem-ory: What are the roles of the perirhinal cortex and hippo-campus? Nature Reviews: Neuroscience, 2(1), 51–61.

Buckner, R. L., & Wheeler, M. E. (2001). The cognitiveneuroscience of remembering. Nature ReviewsNeuroscience, 2, 624–634.

Cermak, L. S., Mather, M., & Hill, R. (1997). Unconsciousinfluences on amnesics’ word-stem completion.Neuropsychologia, 35(5), 605–610.

Chun, M. M., & Phelps, E. A. (1999). Memory deficits forimplicit contextual information in amnesic subjects withhippocampal damage. Nature Neuroscience, 2(9), 844–847.

Cleary, A. M. (2006). Relating familiarity-based recognitionand the tip-of-the-tongue phenomenon: Detecting aword’s recency in the absence of access to the word.Memory & Cognition, 34, 804–816.

Cleary, A. M., & Greene, R. L. (2000). Recognition withoutidentification. Journal of Experimental Psychology:Learning, Memory, and Cognition, 26, 1063–1069.

Cleary, A. M., Langley, M. M., & Seiler, K. R. (2004).Recognition without picture identification: Geons as com-ponents of the pictorial memory trace. PsychonomicBulletin & Review, 11, 903–908.

Cleary, A. M., Winfield, M. M., & Kostic, B. (2007).Auditory recognition without identification. Memory &Cognition, 35, 1869–1877.

Conroy, M. A., Hopkins, R. O., & Squire, L. R. (2005). Onthe contribution of perceptual fluency and priming torecognition memory. Cognitive, Affective, & BehavioralNeuroscience, 5, 14–20.

Cowell, R. A., Bussey, T. J., & Saksida, L. M. (2010).Components of recognition memory: Dissociable cogni-tive processes or just differences in representational com-plexity? Hippocampus, 20, 1245–1262.

Curran, T. (2000). Brain potentials of recollection and famil-iarity. Memory & Cognition, 28, 923–938.

Curran, T., & Doyle, J. (2011). Picture superiority doublydissociates the ERP correlates of recollection andfamiliarity. Journal of Cognitive Neuroscience, 23,1247–1262.

Czernochowski, D., Mecklinger, A., & Johansson, M.(2009). Age-related changes in the control of episodicretrieval: An ERP study of recognition memory inchildren and adults. Developmental Science, 12,1026–1040.

Danker, J. F., & Anderson, J. R. (2010). The ghosts of brainstates past: Remembering reactivates the brain regions

engaged during encoding. Psychological Bulletin, 136,87–102.

Danker, J. F., Hwang, G.M., Gauthier, L., Geller, A., Kahana,M. J., & Sekuler, R. (2008). Characterizing the ERP old-new effect in a short-term memory task.Psychophysiology, 45, 784–793.

Daselaar, S. M., Fleck, M. S., & Cabeza, R. (2006). Tripledissociation in the medial temporal lobes: Recollection,familiarity, and novelty. Journal of Neurophysiology, 96(4), 1902–1911.

de Biran, M. (1929). The influence of habit on the faculty ofthinking. Baltimore, MD: Williams & Wilkins. (Originalwork published 1804)

De Chastelaine, M., Friedman, D., Cycowicz, Y. M., &Horton, C. (2009). Effects of multiple study-test repetitionon the neural correlates of recognition memory: ERPsdissociate remembering and knowing. Psychophysiology,46, 86–99.

Dew, I. T., & Cabeza, R. (2011). The porous boundariesbetween explicit and implicit memory: Behavioral andneural evidence. Annals of the New York Academy ofSciences, 1224, 174–190.

Diana, R. A., Yonelinas, A. P., & Ranganath, C. (2007).Imaging recollection and familiarity in the medial tem-poral lobe: A three-component model. Trends inCognitive Sciences, 11, 379–386.

Dobbins, I. G., Kroll, N. E. A., & Liu, Q. (1998).Confidence—accuracy inversions in scene recognition:A remember-know analysis. Journal of ExperimentalPsychology: Learning, Memory, & Cognition, 24,1306–1315.

Donaldson, D. I., & Curran, T. (2007). Potential (ERP) stu-dies of recognition memory for faces. NeuroImage, 36,488–489.

Donaldson, D. I., Petersen, S. E., & Buckner, R. L. (2001).Dissociating memory retrieval processes using fMRI:Evidence that priming does not support recognition mem-ory. Neuron, 31, 1047–1059.

Duchek, J. M., & Neely, J. H. (1989). A dissociative wordfrequency x levels-of-processing interaction in episodicrecognition and lexical decision tasks. Memory &Cognition, 17, 148–162.

Ecker, U. K., Arend, A. M., Bergstrom, K., & Zimmer,H. D. (2009). Verbal predicates foster conscious recol-lection but not familiarity of a task-irrelevant perceptualfeature—an ERP study. Consciousness & Cognition,18, 679–689.

Eichenbaum, H. (2004). Hippocampus: Cognitive processesand neural representations that underlie declarative mem-ory. Neuron, 44, 109–120.

Eichenbaum, H., & Cohen, N. J. (2001). From conditioningto conscious recollection: Memory systems of the brain.New York, NY: Oxford University Press.

Eichenbaum, H., Fortin, N., Sauvage, M., Robitsek, R. J.,and Farovik, A. (2010). An animal model of amnesiathat uses Receiver Operating Characteristics (ROC)analysis to distinguish recollection from familiarity def-icits in recognition memory. Neuropsychologia, 48,2281–2289.

222 REFERENCES

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

Eichenbaum, H., Yonelinas, A. P., & Ranganath, C. (2007).The medial temporal lobe and recognition memory.Annual Review of Neuroscience, 30, 123–152.

Fernandez, G., & Tendolkar, I. (2006). The rhinal cortex:‘Gatekeeper’ of the declarative memory system. Trendsin Cognitive Science, 10(8), 358–362.

Finnigan, S., Humphreys, M. S., Dennis, S., & Geffen, G.(2002). ERP ‘old/new’ effects: Memory strength anddecisional factor(s). Neuropsychologia, 40, 2288–2304.

Fodor, J. A. (1983). The modularity of mind: An essay onfaculty psychology. Cambridge, MA: MIT Press.Retrieved from http://cognet.mit.edu/library/books/view?isbn=0262560259

Forster, K. I., & David, C. (1984). Repetition priming andfrequency attenuation in lexical access. Journal ofExperimental Psychology: Learning, Memory, andCognition, 10, 680–698.

Gabrieli, J. D. (1998). Cognitive neuroscience of humanmemory. Annual Review of Psychology, 49, 87–115.

Goldberg, R. F., Perfetti, C. A., & Schneider, W. (2006).Perceptual knowledge retrieval activates sensory brainregions. Journal of Neuroscience, 26, 4917–4921.

Gonsalves, B. D., Kahn, I., Curran, T., Norman, K. A., &Wagner, A. D. (2005). Memory strength and repetitionsuppression: Multimodal imaging of medial temporalcortical contributions to recognition. Neuron, 47(5),751–761.

Gottfried, J. A., Smith, A. P., Rugg, M. D., & Dolan, R. J.(2004). Remembrance of odors past: Human olfactorycortex in cross-modal recognition memory. Neuron, 42,687–695.

Groh-Bordin, C., Zimmer, H. D., & Ecker, U. K. (2006). Hasthe butcher on the bus dyed his hair? When color changesmodulate ERP correlates of familiarity and recollection.NeuroImage, 32, 1879–1890.

Guillaume, F., Guillem, F., Tiberghien, G., & Stip, E. (2012).ERP investigation of study-test background mismatchduring face recognition in schizophrenia. SchizophreniaResearch, 134, 101–109.

Hannula, D. E., & Greene, A. J. (2012). The hippocampusreevaluated in unconscious learning and memory: At atipping point? Frontiers in Human Neuroscience, 6:80.doi: 10.3389/fnhum.2012.00080.

Hannula, D. E., & Ranganath, C. (2009). The eyes have it:Hippocampal activity predicts expression of memory ineye movements. Neuron, 63, 592–599.

Heathcote, A. (2003). Item recognition memory and theROC. Journal of Experimental Psychology: Learning,Memory, & Cognition, 29, 1210–1230.

Heathcote, A., Bora, B., & Freeman, E. (2010). Modelingchoice-similarity effects in episodic recognition. Journalof Memory and Language, 62, 183–203.

Heathcote, A., Freeman, E., Etherington, J., Tonkin, J., &Bora, B. (2009). A dissociation between similarity effectsin episodic face recognition. Psychonomic Bulletin andReview, 16, 824–831.

Henke, K. (2010). A model for memory systems based onprocessing modes rather than consciousness. NatureReviews Neuroscience, 11, 523–532.

Henson, R. N., Cansino, S., Herron, J. E., Robb, W. G., &Rugg, M. D. (2003). A familiarity signal in humananterior medial temporal cortex? Hippocampus, 13(2),301–304.

Jacoby, L. L., & Dallas, M. (1981). On the relationship betweenautobiographical memory and perceptual learning. Journalof Experimental Psychology: General, 110, 306–340.

Jacoby, L. L., & Whitehouse, K. (1989). An illusion ofmemory: False recognition influenced by unconsciousperception. Journal of Experimental Psychology:General, 118, 126–135.

Jager, T., Mecklinger, A., & Kipp, K. H. (2006). Intra- andinter-item associations doubly dissociated the electrophy-siological correlates of familiarity and recollection.Neuron, 52, 535–545.

Jeneson, A., Kirwan, C. B., & Squire, L. R. (2010).Recognition without awareness: An elusive phenomenon.Learning & Memory, 17, 454–459.

Keane,M.M., Gabrieli, J. D., Fennema,A. C., Growdon, J. H.,Corkin, S. (1991). Evidence for a dissociation betweenperceptual and conceptual priming in Alzheimer’s disease.Behavioral Neuroscience, 105, 326–342.

Keane, M. M., Orlando, F., & Verfaellie, M. (2006). Increasingthe salience of fluency cues reduces the recognition memoryimpairment in amnesia. Neuropsychologia, 44, 834–839.

Klonek, F., Tamm, S., Hofmann,M. J.,& Jacobs, A.M. (2009).Does familiarity or conflict account for performance in theword-stem completion task? Evidence from behaviouraland event-related-potential data. Psychological Research,73, 871–882.

Kurilla, B. P., &Westerman, D. L. (2008). Processing fluencyaffects subjective claims of recollection. Memory &Cognition, 36, 82–92.

Kurilla, B. P., &Westerman, D. L. (2010). Source memory forunidentified stimuli. Journal of Experimental Psychology:Learning, Memory, and Cognition, 36, 398–410.

Kutas, M., & Federmeier, K. D. (2011). Thirty years andcounting: Finding meaning in the N400 component ofthe event-related brain potential (ERP). Annual Reviewof Psychology, 62, 621–647.

Lau, H., & Rosenthal, D. (2011). Empirical support forhigher-order theories of conscious awareness. Trends inCognitive Sciences, 15, 365–373.

Lee, T., Blumenfeld, R. S., & D’Esposito, M. (2011).Disruption of prefrontal cortical function improves mem-ory performance without awareness. Paper presented atthe Annual Meeting of the Society for Neuroscience,Washington, DC.

Leisman, G., & Koch, P. (2009). Networks of consciousexperience: Computational neuroscience in understand-ing life, death, and consciousness. Reviews in theNeurosciences, 20, 151–176.

Lucas, H. D., Voss, J. L., & Paller K. A. (2010). Familiarity orconceptual priming? Good question! Comment onStenberg, Hellman, Johansson, and Rosén (2009).Journal of Cognitive Neuroscience, 22, 615–617.

MacKenzie, G., & Donaldson, D. I. (2007). Dissociatingrecollection from familiarity: Electrophysiological evi-dence that familiarity for faces is associated with a poster-ior old/new effect. NeuroImage, 36, 454–463.

Mandler, G. (1980). Recognizing: The judgment of previousoccurrence. Psychological Review, 87, 252–271.

Mandler, G. (2008). Familiarity breeds attempts.Perspectives on Psychological Science, 3, 390–399.

Manns, J. R., & Squire, L. R. (2001). Perceptual learning,awareness, and the hippocampus. Hippocampus, 11,776–782.

REFERENCES 223

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

Martin, A. (2007). The representation of object concepts inthe brain. Annual Review of Psychology, 58, 25–45.

Marzi, T. & Viggiano, M. P. (2010). When memory meetsbeauty: Insights from event-related potentials. BiologicalPsychology, 84, 192–205.

McCabe, D. P., Roediger, H. L., & Karpicke, J. D. (2011).Automatic processing influences free recall: Convergingevidence from the process dissociation procedure andremember-know judgments. Memory & Cognition, 39,389–402.

Mecklinger, A., Brunnemann, N., & Kipp, K. (2011). Twoprocesses for recognition memory in children of earlyschool age: An event-related potential study. Journal ofCognitive Neuroscience, 23, 435–446.

Migo, E., Montaldi, D., Norman, K. A., Quamme, J., &Mayes, A. (2009). The contribution of familiarity torecognition memory is a function of test format whenusing similar foils. Quarterly Journal of ExperimentalPsychology, 62, 1198–1215.

Montaldi, D., Spencer, T. J., Roberts, N., & Mayes, A. R.(2006). The neural system that mediates familiarity mem-ory. Hippocampus, 16(5), 504–520.

Moscovitch, M. (2008). The hippocampus as a “stupid,”domain-specific module: Implications for theories of recentand remote memory, and of imagination. CanadianJournal of Experimental Psychology, 62, 62–79.

Mulligan, N. W. (1998). The role of attention during encod-ing in implicit and explicit memory. Journal ofExperimental Psychology: Learning, Memory, andCognition, 24, 27–47.

Nyhus, E., & Curran, T. (2009). Semantic and perceptualeffects on recognition memory: Evidence from ERP.Brain Research, 1283, 102–114.

O’Kane, G., Insler, R. Z., & Wagner, A. D. (2005).Conceptual and perceptual novelty effects in human med-ial temporal cortex. Hippocampus, 15(3), 326–332.

Olichney, J. M., Van Petten, C., Paller, K. A., Salmon, D. P.,Iragui, V. J., & Kutas, M. (2000). Word repetition inamnesia: Electrophysiological measures of impaired andspared memory. Brain, 123, 1948–1963.

Opitz, B., & Cornell, S. (2006). Contribution of familiarityand recollection to associative recognition memory:Insights from event-related potentials. Journal ofCognitive Neuroscience, 18, 1595–1605.

Packard, M. G., & McGaugh, J. L. (1996). Inactivation ofhippocampus or caudate nucleus with lidocaine differen-tially affects expression of place and response learning.Neurobiology of Learning & Memory, 65(1), 65–72.doi:10.1006/nlme.1996.0007

Paller, K. A., Hutson, C. A., Miller, B. B., & Boehm, S. G.(2003). Neural manifestations of memory with and with-out awareness. Neuron, 38, 507–516.

Paller, K. A., Voss, J. L., & Boehm, S. G. (2007). Validatingneural correlates of familiarity. Trends in CognitiveSciences, 11, 243–250.

Paller, K. A., Voss, J. L., & Westerberg, C. E. (2009).Investigating the awareness of remembering. Perspectiveson Psychological Science, 4, 185–199.

Poldrack, R. A. (2006). Can cognitive processes be inferredfrom neuroimaging data? Trends in Cognitive Sciences,10, 59–63.

Rajaram, S. (1993). Remembering and knowing: Two meansof access to the personal past. Memory & Cognition, 21,89–102.

Rajaram, S., & Geraci, L. (2000). Conceptual fluency selec-tively influences knowing. Journal of ExperimentalPsychology: Learning, Memory and Cognition, 26(4),1070–1074.

Ranganath, C. (2010). A unified framework for the functionalorganization of the medial temporal lobes and the phe-nomenology of episodic memory. Hippocampus, 20,1263–1290.

Richardson-Klavehn, A., & Bjork, R. A. (1988). Measures ofmemory. Annual Review of Psychology, 39, 475–543.

Roediger, H. L., & McDermott, K. B. (1993). Implicit mem-ory in normal human subjects. In F. Boller & J. Grafman(Eds.), Handbook of neuropsychology (Vol. 8, pp. 63–131).Amsterdam: Elsevier.

Rosburg, T., Mecklinger, A., & Frings, C. (2011). When thebrain decides: A familiarity-based approach to the recog-nition heuristic as evidenced by event-related brain poten-tials. Psychological Science, 22, 1527–1534.

Rose, M., Haider, H., Salari, N., & Buchel, C. (2011).Functional dissociation of hippocampal mechanism dur-ing implicit learning based on the domain of associations.Journal of Neuroscience, 31, 13739–13745.

Rugg, M. D., & Curran, T. (2007). Event-related potentialsand recognition memory. Trends in Cognitive Sciences,11, 251–257.

Ryals, A. J., Yadon, C. A., Nomi, J. S., & Cleary, A. M.(2011). When word identification fails: ERP correlates ofrecognition without identification and of word identifica-tion failure. Neuropsychologia, 12, 3224–3237.

Ryan, J. D., Althoff, R. R., Whitlow, S., & Cohen, N. J.(2000). Amnesia is a deficit in declarative (relational)memory. Psychological Science, 11, 454–461.

Ryan, J. D., & Cohen, N. J. (2003). Evaluating the neurop-sychological dissociation evidence for multiple memorysystems. Cognitive, Affective, & Behavioral Neuroscience,3, 168–185.

Scarborough, D. L., Cortese, C., & Scarborough, H. S.(1977). Frequency and repetition effects in lexical mem-ory. Journal of Experimental Psychology: HumanPerception and Performance, 3, 1–17.

Schacter, D. L. (1987). Implicit memory: History and currentstatus. Journal of Experimental Psychology: Learning,Memory and Cognition, 13, 501–518.

Schacter, D. L., Dobbins, I. G., & Schnyer, D. M. (2004).Specificity of priming: A cognitive neuroscience perspec-tive. Nature Reviews Neuroscience, 5, 853–862.

Schacter, D. L., & Tulving, E. (1994). What are the memorysystems of 1994? In D. L. Schacter & E. Tulving (Eds.),Memory systems 1994. (pp. 1–38). Cambridge, MA: MITPress.

Schacter, D. L., Wig, G. S., & Stevens, W. D. (2007).Reductions in cortical activity during priming. CurrentOpinion in Neurobiology, 17, 171–176.

Simons, J. S., & Spiers, H. J. (2003). Prefrontal and medialtemporal lobe interactions in long-term memory. NatureReviews Neuroscience, 4, 637–648.

Speer, N. K., & Curran, T. (2007). ERP correlates of famil-iarity and recollection processes in visual associativerecognition. Brain Research, 1174, 97–109.

Squire, L. R. (2004). Memory systems of the brain: A briefhistory and current perspective.Neurobiology of Learningand Memory, 82, 171–177.

Squire, L. R. (2004). The medial temporal lobe. AnnualReview of Neuroscience, 27, 279–306.

224 REFERENCES

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

Stark, C. E., & Squire, L. R. (2000). Recognition memoryand familiarity judgments in severe amnesia: No evidencefor a contribution of repetition priming. BehavioralNeuroscience, 114, 459–467.

Starns, J. J., Hicks, J. L., Brown, N. L., & Martin, B. A.(2008). Source memory for unrecognized items:Predictions from multivariate signal detection theory.Memory & Cognition, 36, 1–8.

Stenberg, G., Hellman, J., Johansson, M., & Rosén, I. (2009).Familiarity or conceptual priming: Event-related poten-tials in name recognition. Journal of CognitiveNeuroscience, 21, 447–460.

Stenberg, G., Johansson, M., Hellman, J., Rosén, I. (2010).“Do you see yonder cloud?” –On priming concepts, a newtest, and a familiar outcome. Reply to Lucas et al.:“Familiarity or conceptual priming? Good question!Comment on Stenberg, Hellman, Johansson, and Rosén(2009)”. Journal of Cognitive Neuroscience, 22, 618–620.

Taylor, J. R., Buratto, L. G., & Henson, R. N. (submitted).Behavioural and neural evidence for masked conceptualpriming of recollection.

Taylor, J. R., & Henson, R. N. (in press). Could maskedconceptual primes increase recollection? The subtletiesof measuring recollection and familiarity in recognitionmemory. Neuropsychologia.

Tulving, E. (1981). Similarity relations in recognition. Journalof Verbal Learning and Verbal Behavior, 20, 479–496.

Tunney, R. J., & Fernie, G. (2007). Repetition priming affectsguessing not familiarity. Behavioral and Brain Functions,3, 40–46.

Vargas, I. M., Voss, J. L., & Paller, K. A. (2012). Implicitrecognition based on lateralized perceptual fluency. BrainSciences, 2, 22–32.

Verfaellie, M., & Cermak, L. S. (1999). Perceptual fluency as acue for recognition judgments in amnesia.Neuropsychology,13, 198–205.

Verfaellie, M., Giovanello, K. S., & Keane, M. M. (2001).Recognition memory in amnesia: Effects of relaxingresponse criteria. Cognitive, Affective & BehavioralNeuroscience, 1, 3–9.

Voss, J. L., Baym, C. L., & Paller, K. A. (2008a). Accurateforced-choice recognition without awareness of memoryretrieval. Learning & Memory, 15, 454–459.

Voss, J. L., & Federmeier, K. D. (2011). FN400 potentials arefunctionally identical to N400 potentials and reflectsemantic processing during recognition testing.Psychophysiology, 48(4), 532–546.

Voss, J. L., Federmeier, K. D., & Paller, K. A. (2011). Thepotato chip really does look like Elvis! Neural hallmarksof conceptual processing associated with finding novelshapes subjectively meaningful. Cerebral Cortex.doi:10.1093/cercor/bhr315

Voss, J. L., Hauner, K. K. Y., & Paller, K. A. (2009).Establishing a relationship between activity reduction inhuman perirhinal cortex and priming. Hippocampus, 19,773–778.

Voss, J. L., Lucas, H. D., & Paller, K. A. (2010). Conceptualpriming and familiarity: Different expressions of memoryduring recognition testing with distinct neurophysiologi-cal correlates. Journal of Cognitive Neuroscience, 22,2638–2651.

Voss, J. L., & Paller, K. A. (2006). Fluent conceptual proces-sing and explicit memory for faces are electrophysiologi-cally distinct. Journal of Neuroscience, 26, 926–933.

Voss, J. L., & Paller, K. A. (2007). Neural correlates ofconceptual implicit memory and their contamination ofputative neural correlates of explicit memory. Learning &Memory, 14, 259–267.

Voss, J. L., & Paller, K. A. (2008a). Brain substrates ofimplicit and explicit memory: The importance of concur-rently acquired neural signals of both memory types.Neuropsychologia, 46, 3021–3029.

Voss, J. L., & Paller, K. A. (2008b). Neural substrates ofremembering: Electroencephalographic studies. In H.Eichenbaum (Ed.), Learning and memory: A comprehen-sive reference (vol. 3). Oxford, UK: Elsevier.

Voss, J. L., & Paller, K. A. (2009a). An electrophysiologicalsignature of unconscious recognition memory. NatureNeuroscience, 12, 349–355.

Voss, J. L., & Paller, K. A. (2009b). Recognition withoutawareness in humans and its implications for animalmodels of episodic memory. Communicative andIntegrative Biology, 2, 203–204.

Voss, J. L., & Paller, K. A. (2009c). Remembering andknowing: Electrophysiological distinctions at encodingbut not retrieval. NeuroImage, 46, 280–289.

Voss, J. L., & Paller, K. A. (2010). What makes recognitionwithout awareness appear to be elusive? Strategic factorsthat influence the accuracy of guesses. Learning &Memory, 17(9), 460–468. doi:10.1101/lm.1896010

Voss, J. L., & Paller, K. A. (2010a). Real-time neural signalsof perceptual priming with unfamiliar geometric shapes.Journal of Neuroscience, 30, 9181–9188.

Voss, J. L., & Paller, K. A. (2010b). What makes recognitionwithout awareness appear to be elusive? Strategic factorsthat influence the accuracy of guesses. Learning &Memory, 17, 460–468.

Voss, J. L., Reber, P. J., Mesulam, M. M., Parrish, T. B., &Paller, K. A. (2008b). Familiarity and conceptual primingengage distinct cortical networks. Cerebral Cortex, 18,1712–1719.

Voss, J. L., Schendan, H. E., & Paller, K. A. (2010). Findingmeaning in novel geometric shapes influences electrophy-siological correlates of repetition and dissociatesperceptual and conceptual priming. NeuroImage, 49,2879–2889.

Wagner, A. D., Gabrieli, J. D., & Verfaellie, M. (1997).Dissociations between familiarity processes in explicitrecognition and implicit perceptual memory. Journal ofExperimental Psychology: Learning, Memory, andCognition, 23, 305–323.

Wang, W. C., Lazzara, M. M., Ranganath, C., Knight, R. T.,& Yonelinas, A. P. (2010). The medial temporal lobesupports conceptual implicit memory. Neuron, 68,835–842.

Wang, W. C., & Yonelinas, A. P. (in press). Familiarity isrelated to conceptual implicit memory: An examina-tion of individual differences. Psychonomic Bulletin &Review.

Whittlesea, B. W. A., Jacoby, L. L., & Girard, K. (1990).Illusions of immediate memory: Evidence of an attribu-tional basis for feelings of familiarity and perceptualquality. Journal of Memory and Language, 29, 716–732.

Whittlesea, B. W. A., & Leboe, J. P. (2003). Two fluencyheuristics (and how to tell them apart). Journal of Memoryand Language, 49(1), 62–79.

Whittlesea, B. W., & Williams, L. D. (2000). The sourceof feelings of familiarity: The discrepancy-attribution

REFERENCES 225

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012

hypothesis. Journal of Experimental Psychology:Learning, Memory, & Cognition, 26, 547–565.

Wolk, D. A., Schacter, D. L., Berman, A. R., Holcomb, P. J.,Daffner, K. R., & Budson, A. E. (2005). Patients withmild Alzheimer’s disease attribute conceptual fluency toprior experience. Neuropsychologia, 43, 1662–1672.

Woodruff, C. C., Hayama, H. R., & Rugg, M. D. (2006).Electrophysiological dissociation of the neural correlatesof recollection and familiarity. Brain Research, 1100,125–135.

Woollams, A. M., Taylor, J. R., Karayanidis, F., & Henson,R. N. (2008). Event-related potentials associated withmasked priming of test cues reveal multiple potentialcontributions to recognition memory. Journal ofCognitive Neuroscience, 20, 1114–1129.

Yonelinas, A. P. (1994). Receiver operating characteristics inrecognition memory: Evidence for a dual process model.

Journal of Experimental Psychology: Learning, Memory,and Cognition, 20, 1341–1354.

Yonelinas, A. P. (2001). Consciousness, control, and confi-dence: The 3 C’s of recognition memory. Journal ofExperimental Psychology: General, 130, 361–379.

Yonelinas, A. P. (2002). The nature of recollection and famil-iarity: A review of 30 years of research. Journal ofMemory and Language, 46(3), 441–517.

Yonelinas, A. P., Kroll, N. E., Quamme, J. R., Lazzara,M. M., Sauvé, M. J., Widaman, K. F., & Knight, R. T.(2002). Effects of extensive temporal lobe damage or mildhypoxia on recollection and familiarity. NatureNeuroscience, 5, 1236–1241.

Yovel, G., & Paller, K. A. (2004). The neural basis of thebutcher-on-the-bus phenomenon: When a faceseems familiar but is not remembered. NeuroImage, 21,789–800.

226 REFERENCES

Dow

nloa

ded

by [

Nor

thw

este

rn U

nive

rsity

] at

08:

25 2

8 Ju

ly 2

012