striatum

Neuropsychologia 46 (2008) 174185

The role of the striatum in sentence procpriming study in early stages of Hu

oux-Lerventi, Franreteil,es Co

e Laboratoire de Sciences Cognitives et Psycholinguistique, UMR8554, EHESS-ENS-CNRS, Paris, FranceReceived 5 May 2007; received in revised form 24 July 2007; accepted 31 July 2007

Available online 3 August 2007

Abstract

The roleprocessing ihave difficulM., Hickok,of declarativHere we bol

Both proccompared idthat containcrushed the(e.g., bail)

HD patienword changwhereas it ishypothesis b 2007 Else

Keywords: St

1. Introdu

While testablishedstriatum anstill controaged patien

CorresponTel.: +33 6 66

E-mail ad

0028-3932/$doi:10.1016/jof sub-cortical structures such as the striatum in language remains a controversial issue. Based on linguistic claims that languagemplies both recovery of lexical information and application of combinatorial rules it has been shown that striatal damaged patientsties applying conjugation rules while lexical recovery of irregular forms is broadly spared (e.g., Ullman, M. T., Corkin, S., Coppola,G., Growdon, J. H., Koroshetz, W. J., et al. (1997). A neural dissociation within language: Evidence that the mental dictionary is parte memory, and that grammatical rules are processed by the procedural system. Journal of Cognitive Neuroscience, 9(2), 266276).stered the striatum-rule hypothesis by investigating lexical abilities and rule application at the phrasal level.essing aspects were assessed in a model of striatal dysfunction, namely Huntingtons disease (HD). Using a semantic priming task weiomatic prime sentences involving lexical access to whole phrases (e.g., Paul has kicked the bucket) with idiom-derived sentencesed passivation changes involving syntactic movement rules (e.g., Paul was kicked by the bucket), word changes (e.g., Paul hasbucket) or either. Target words that were either idiom-related (e.g., death) reflecting lexical access to idiom meanings, word-relatedreflecting lexical access to single words, or unrelated (e.g., table).ts displayed selective abnormalities with passivated sentences whereas priming was normal with idioms and sentences containing only

es. We argue that the role of the striatum in sentence processing specifically pertains to the application of syntactic movement rulesnot involved in canonical rules required for active structures or in lexical processing aspects. Our findings support the striatum-rule

ut suggest that it should be refined by tracking the particular kind of language rules depending on striatal computations.vier Ltd. All rights reserved.

riatum; Syntax; Rules; Huntingtons disease

ction

he linguistic role of cortical areas is relatively well, the role of sub-cortical structures such as thed the way they impact on language processing isversial. Indeed, studies investigating striatal dam-ts have shown various kinds of language disorders

ding author at: AP-HP, Hopital Henri Mondor, Creteil, France.263919/1 49812309.dress: [email protected] (M. Teichmann).

ranging from semantic disorganisation to syntactic impairmentin both neurodegenerative diseases such as Huntingtons andParkinsons disease (e.g., Copland, 2003; Frank, McDade, &Scott, 1996; Illes, 1989) and in patients with vascular dam-age (e.g., Damasio, Damasio, Rizzo, Varney, & Gersh, 1982;Wallesch & Papagno, 1988). This might be related to the differ-ent lesion patterns regarding the different aetiologies: vasculardisorders affect various parts of the striatum and often involvesurrounding white matter fibre tracks, Parkinsons disease (PD)is due to neural degeneration in the substantia nigra resultingin dopamine-related dysfunction of the striatum, whereas Hunt-ingtons disease (HD) is characterised by neuronal death that

see front matter 2007 Elsevier Ltd. All rights reserved..neuropsychologia.2007.07.022Marc Teichmann a,b,c,d,, Emmanuel DupAnne-Catherine Bachoud

a INSERM U841, Equipe 1, Neuropsychologie Inteb Universite Paris XII, Creteil

c AP-HP, Hopital Henri Mondor, Cd Ecole Normale Superieure, Departement dEtudessing: Evidence from antingtons disease

c,d,e, Pierre Cesaro a,b,c,

vi a,b,c,donnelle, Creteil, Francece

Francegnitives, Paris, France

M. Teichmann et al. / Neuropsychologia 46 (2008) 174185 175

specifically originates in the neostriatum comprising the cau-date head and the putamen (see Peschanski, Cesaro, & Hantraye,1995 for asent a reliabal., 1982; Madults, by mthe picturewere drawnogy (Tettam& Lewis, 2et al., 2001led to two cthat the strcific processupervise aet al., 2000& Tyler, 20the striatumthat may cuogy, morph2001). Accto impact othat apply tproposal ishuman langa mental lecontains alphemes, wthe compuare applied

The Ullspecificallyassessed inthe conjuge.g., walk/w(lexical-baatal damagyielded cothese patieregular veirregular vtum subserwhich wasTeichmannwith HD, rLongworthand irregultheory. Surules assesand/or toThus, to fstriatum-rudomain extsentence pr

In this vthe striatumnot compa

way. With respect to rule application, most authors used non-canonical sentences like passives (e.g., the boy was kissed

girl), inomptheAcc1977tion

n-can

[2: actctur

rulesdiffeg (

hat een rel trans ofnon

tienthercturMcNouloof Hassition

sponcan

irl wwateat de.g., ). Ht picallynsatontr

tic remen

as

l conInde

men

on (euttel callowied bancyETand

to neinine hebseqreview). Thus, only early stages of HD might repre-le model of near exclusive striatal disorders (Kuhl etazziotta et al., 1987). Yet, investigations of healthyeans of functional brain imaging, have not clarified

either. They showed striatal activation with tasks thatfrom various language domains including phonol-anti et al., 2005), word morphology (Vannest, Polk,

005) and syntax (e.g., Friederici & Kotz, 2003; Moro). Indeed, this functionally unspecified picture hasompeting views. On the one hand, it has been arguediatum merely underpins general non-language spe-ses such as executive functioning, which is thought tond modulate linguistic information (e.g., Grossman, 2002; Longworth, Keenan, Barker, Marslen-Wilson,05). Conversely, several authors have proposed thatsubserves particular kinds of language computationst across different linguistic levels, including phonol-ology and syntax (e.g., Ullman et al., 1997; Ullman,ording to this latter view, the striatum was claimedn the computation of combinatorial language ruleso various kinds of lexically stored information. Thisderived from psycholinguistic models that divide theuage faculty into two core components comprisingxicon and a computational grammar. The lexicon

l linguistic idiosyncrasies such as phonemes, mor-ords and even whole phrases (e.g., idioms), whereastational grammar holds the combinatorial rules thatto the lexical input (Chomsky, 1965; Pinker, 1999).man proposal, according to which the striatum is

involved in rule computations has been mainlythe domain of word morphology by comparing

ation of regular verbs or non-verbs (rule-based;alk-ed or splush/splush-ed) and of irregular verbs

sed; e.g., go/went). However, studies using stri-ed patients such as PD, HD and stroke patients

nflicting views. Ullman et al. (1997) showed thatnts are mainly impaired on the conjugation of

rbs and non-verbs whereas their performance onerbs is near-normal. They concluded that the stria-ves the application of the English past tense rule,thought to substantiate the striatum-rule theory.

et al. (2005), using the richer morphology of Frenchestricted this view to conjugation sub-rules, whileet al. (2005) using a priming study with regular

ar prime-target pairs fully rejected the striatum-rulech conflicting data might be due to the differentsed in different languages, to the different tasksthe differences of the respective lesion patterns.urther clarify this issue, the general claim of thele hypothesis should be tested in a second languageending evidence from the word level to the level ofocessing.ein, several studies have investigated the link between

and syntactic processing aspects. However, they didre syntactic rules and lexical operations in a direct

by thekissedverb-ckissedgirl).1965,applicathe nokissedcal struTheseto themeaninNote thas bematicamodelversus

PD paorder wcal stru1999;Natsopstageswere m

applicarate reand thethe ggirl isture thone (ea girlcorrec

especicompe

In csyntacinvolvof themphrasaadults.in thesentatiPearlmmaticathe fofollowexpectin a Pminersso as

maintacaudatand su) or object-relatives (e.g., the boy that the girlwhich the usual agent-action-theme order (subject-

lement) is inverted as compared to actives (the boygirl) or subject-relatives (the boy that kissed theording to several linguistic accounts (e.g., Chomsky,, 1986) such non-canonical sentences involve theof syntactic movement rules that allow mapping ofonical surface structure (e.g., the boy[3: theme] was

ion] by the girl[1: agent]) on the underlying canoni-e (the girl[1: agent] kissed[2: action] the boy[3: theme]).are thought to be critical for assigning thematic rolesrent words and thus to convey the correct sentencewho [agent] does what [action] to whom [theme]).ven if the initial concept of grammatical movementvised in more recent theories, the notion of gram-sformation always holds a central role in almost all

generative grammar. When studying this canonical-canonical contrast several studies have shown thats have difficulties interpreting non-canonical wordeas they have near normal performance on canoni-es, such as actives and subject-relatives (Kemmerer,amara, Krueger, OQuin, Clark, & Durso, 1996;s et al., 1993). Similar results were reported in early

D. Teichmann et al. (2005) showed that these patientsvely impaired on sentence comprehension when theof syntactic movement rules was required for accu-

se. The authors independently varied the plausibilityonicity of simple sentences (e.g., canonical plausible:aters the flower, non-canonical non-plausible: thered by the flower), which were presented with a pic-picted either the corresponding scene or the inverteda girl that waters a flower or a flower that watersD patients had severe difficulties in choosing theture when confronted with non-canonical sentenceswhen non-plausible sentence meanings impeded theory use of pragmatic reasoning.ast to these previous studies exclusively assessingule application, several authors evaluated striatalt in lexical aspects of sentence processing. Most

sessed the processing of grammatical categories intexts using functional brain imagery with healthyed, grammatical categories are claimed to be stored

tal lexicon together with the respective word repre-.g., cat

-noun, eat-verb, the-determiner, . . .) (MacDonald,r, & Seidenberg, 1994). Furthermore, certain gram-tegories such as determiners place constraints onng sentence word (e.g., determiners are alwaysy nouns), which defines a specific grammatical. The kind of grammatical expectancy was usedstudy of Moro et al. (2001) who inverted deter-nouns in sentences that consisted of pseudowordsutralise the access to semantic components while

g function words. They reported activation of thead when participants covertly read such sentencesuently made acceptability judgements. Likewise, in

176 M. Teichmann et al. / Neuropsychologia 46 (2008) 174185

a functional MRI study Friederici, Ruschemeyer, Hahne, &Fiebach (2003) showed that violations of the expectancy ofgrammaticresults in tmaterials wing ERP reshowed thaindex late(Friederici,Frisch, vonpatients (Fings led topost-linguioutput of e(Friedericiby Longwocompetingand, possibing (Longwstriatal damically storein turn, migever, intrigstriatum secategories.ily depend othat striatala marker ofthat containwas eatenKotz et al.,

In summplays a roltroversial alexical opeprocesses wunclear whaspects oflanguage srules and th

The curatal disorddirectly coof idiomatilinguistic athe bucketchunks (Caccess to thnot dependconvey a lbucket). Sadults usinprime senteceded targemeaning (e(e.g., bailThey found

faster on both word-related and idiom-related target words ascompared to unrelated targets (e.g., lie), showing that idiom

sing involves lexical access to both the component wordse figurative idiom meaning. In the present study we usedlar priming task manipulating idioms (e.g., Paul hasthe bucket) as well as idiom-derived sentences that

ned passivation changes involving movement rules (e.g.,as kicked by the bucket), word changes (e.g., Paul has

d the bucket) or either (e.g., Paul was crushed by the). This allowed assessing both lexical accesses to idiomword meanings as well as the ability to apply syntac-

vement rules. HD patients were predicted to have normalmance with idioms and with idiom-derived sentences con-

onlyces,lar,

nearHEnterpterprs of Hguis

l as merall

hods

rtici

nty-fostag

nd 24d from

the sondo

tory opeatswere

l Fs 35). Healthy controls had no neurological or psychiatric disor-matched to the HD patients according to their age and educational1). All participants gave informed consent. Demographic data aren Table 1.

l assessment

ts were evaluated using the unified Huntingtons disease ratingS; Huntington Study Group, 1996) and the Mattis dementiaMDRS; Mattis, 1976). The UHDRS comprises the evaluationability (total functional capacities; TFC), of motor functioningtor part) and of executive parameters such as literal fluency,st and the digit symbol code (UHDRS cognitive part). Generalsummarised in Table 2.

data of HD patients and controls

HD Controls

24 2413 M/11 F 12 M/12 F44.4 10.9 41.7 7.9

ation 12.7 3.1 13.9 2.322 R/2 L 23 R/1 L

ation (years) 5.5 3.4 44.4 3.3


Table 2Results for the general assessment in HD patients

Total functionUHDRS motoMDRS

Executive funStroop coloFluency PRSymbol dig

MDRS: Mattithe Stroop tes

a Norms areb Norms arec Norms are

3. Primingsentences

We usedsentencesParticipantversus non

tic primingfacilitates tbail or gets (e.g., advantagethey do nomaterials. Tand allow asentence pr

Prime sebucket) aword orderwas changboth wereidioms; Pin which acrushed thand contenthe bucketrelated targan unrelateneurre).

We reasshould resusuch primitences as thof syntactiof lexical pFurthermorpassivatedshould hinSchriefers,diminished

esis that HD patients display rule disorders but intact lexicalabilities, we predicted them to be insensitive to passivation but

ve toto d

ivesere b-deped psivehe worm

andtivelrderr the

ple aPauwithountredithatence

ater

cons

ioms that). Bmai

rivedng thsivem-reas w

f 16d ons 13s (es

8 (ed fiHD Normal publishedrange

al capacity (TFC) 12.0 0.9 13r score 27.0 16.1 0

131.2 9.6 >136ctionsur/words 28.6 11.0 >35aV (2 min) 47.2 24.0 >56bit code 28.9 12.8 >37c

s dementia rating scale; Stroop colour/words: interference part oft; fluency PRV: mean number of words P, R, V in 2 min.issued from Golden (1978).issued from Cardebat, Doyon, Puel, Goulet, & Joanette (1990).issued from Wechsler (1981).

task with idioms and idiom-derived

a semantic priming task, in which auditory primewere followed by visually presented target words.s performed lexical decisions on the targets (word-word decision). The main characteristic of seman-

is that a prime (e.g., He has kicked the bucket)he recognition of semantically related targets (e.g.,death) as compared to semantically unrelated tar-table). Such priming paradigms have a theoretical

over classical sentence comprehension tasks in thatt depend on explicit comprehension of the sentencehus, they diminish the impact of controlled processesssessing more implicit language-specific aspects ofocessing.ntences included idioms (e.g., Paul has kicked the

nd different idiom-derived sentences, in which thewas maintained, but either the syntactic structure

ed (passivation), a content word was changed ormodified. This yielded idioms with passivation (P-aul was kicked by the bucket), active non-idioms,content word was changed (NI-actives; Paul has

e bucket) and non-idioms with both passivationt word change (NI-passives; Paul was crushed by

sensitidictedNI-actthey win ruleminishNI-pasboth tyield nidiomsrespec

In owhethein simversus

pairedthe amwere ptencesto sent

3.1. M

Weand idclause(sub-se

Thewe deinvertiNI-pasan idiotargettotal omatche(idiom1393 mphonemDA: 1contain). Each prime sentence was paired with an idiom-et (e.g., death), a word-related target (e.g., bail),d target (e.g., table) and a non-word target (e.g.,

oned that idioms convey a figurative meaning, whichlt in priming for idiom-related targets. Conversely,

ng effects should be disrupted in idiom-derived sen-ese convey literal sentence meanings either because

c rule application in passivated versions or becauserocessing in sentences with content word change.e, we assumed that semantic restriction violations inversions (a bucket can neither kick nor crush)der the integration of the verbs arguments (seeFriederici, & Rose, 1998; Tyler, 1985) and thus yieldpriming for word-related targets. Under the hypoth-

six words iwords wernumber of(2), unreof phonemeunrelated:more, idiomtarget wordtheir log-trIQUE 2 da(idiom-relagets 28.3 p(89.2)).

The subtained semacontent word changes. Hence HD patients were pre-isrupt priming for idiom-related targets only withand NI-passives but not with P-idioms. Furthermore,elieved not to detect semantic restriction violations

endent passivated versions and thus to display undi-riming for word-related targets with P-idioms ands. On the contrary, intact lexical abilities in HD atord level and the sentence level were thought toal priming magnitudes for word-related targets withNI-actives and for idiom-related targets with idioms,

y.to further explore lexical abilities in HD we verifiedse patients have intact access to semantic restrictionsctive clauses (e.g., *Paul has eaten his car (SR)l has driven his car (SR+). These clauses were onlyword-related targets (e.g., truck) in order to limitof the sentence materials. Like controls, HD patientscted to display increased reaction times with sen-contain semantic restriction violations as compareds that do not.

ials

tructed two experimental sets that contained idioms-derived sentences (main-set) and simple activet did or did not contain semantic restriction violationsoth sets are summarised in Table 3.

n-set was composed of 10 French idioms, from which10 P-idioms (by passifying the source idiom ande agent and the theme role), 10 NI-actives and 10

s. Each of the 40 prime sentences were paired withlated target, a word-related target and an unrelatedell as with a target non-word. This resulted in a

0 prime-target pairs. The four sentence types weree with another (all Ps > 0.1) for auditory duration

15 ms (114), P-idioms 1314 ms (133), NI-actives118), NI-passives 1338 ms (79)) and number of

(idioms: 18.2 (2), PI-idioms: 18.2 (2), idioms-2), idioms-DP: 18 (2)). Idioms and NI-activesve words while P-idioms and NI-passives contained

ncluding the French function word par (by). Targete matched one with another (all Ps > 0.1) for theirletters (idiom-related: 7.6 (3), word-related: 6.6

lated: 7.6 (2.4), non-word targets: 6.6 (1.2)) ands (idiom-related: 6.2 (2.9), word-related: 5 (1.9),

5.9 (2.2), non-word targets: 4.7 (0.8)). Further--related targets, word-related targets and unrelated

s were matched one with another (all Ps > 0.1) foransformed token frequency according to the LEX-tabase (New, Pallier, Brysbaert, & Ferrand, 2004)ted targets 50 per million (107.2), word-related tar-er million (38.6), unrelated targets 36 per million

-set contained 30 active clauses. Half of them con-ntic restriction violations (SR) while the other half


Table 3The different types of prime sentences and their respective targets

Set s

ed

Main-set la

Sub-set uck

SR+: no semaa The Frenc

did not (SRtarget and60 sentenc(P > 0.1) fo1396 (99and numbeFurthermor(P > 0.1) fand 6.2 ((2.8)).

Finally,to obtain awithin thewith unrelaAll non-wochanging thdistinct froand letter s

The foutive targetsNI-actives,5 passiveyielding aidiom blocthe Sub-secontainedwithin eacsentence typassives) wfrom idiomblock 69%the last senwas no se

We choserevealed thhigh expecundifferentsentence ty

roce

h sti+) inby aoneset oaseosit

ide aletters the

was a

ant hget orticipespo(idi

ordetherandPrime types Examples N Target type

Word-relat

Idioms Paul a casse sa pipe 10

Cigare, bai

Paul has kicked the bucketP-idioms Paul est casse par sa pipe 10

Paul was kicked by the bucketNI-actives Paul a brise sa pipe 10

Paul has crushed the bucketNI-passives Paul est brise par sa pipe 10

Paul was crushed by the bucket

SR+ Paul a conduit sa voiture 15

Camion, trPaul has driven his carSR Paul a mange sa voiture 15Paul has eaten his car

ntic restriction violation; SR: semantic restriction violation.h word cigare is semantically related to pipe just as bail is related to bucket.

+). Each sentence was paired with a word-relatedwith a target non-word. This resulted in a total ofe-target pairs. Both sentence types were matchedr their auditory duration (SR+: 1354 ms (81), SR:)), number of words (SR+: 5 (0), SR: 5 (0))r of phonemes (SR+: 18.6 (3), SR: 19.4 (3.4)).e, word targets and non-word targets were matched

or their number of letters (respectively 6.3 (0)0)) and phonemes (respectively 5.3 (3.4) and 5.5

we added 20 passive filler prime sentences in ordern equivalent number of passive and active sentencesstimuli materials. Ten filler sentences were pairedted targets and 10 were paired with target non-words.rd stimuli of the experiment were constructed bye first phoneme of existing French nouns, which werem the target words, and consisted of legal phonemetrings.r sentence types of the main-set and their respec-were separated into four blocks (idioms, P-idioms,NI-passives). Fifteen stimuli of the sub-set and

3.2. P

Eaccross (lowedheadphthe offuppercsame pto dectargetto prestargetdominthe tarthe paafter rblockstationacross

sented

filler stimuli were added to each of the 4 blockstotal of 60 stimuli per block. The idiom and the P-k contained eight SR+ and seven SR stimuli oft while the NI-actives and the NI-passives blockseach seven SR+ and eight SR stimuli. Overall,h block 66% of the stimuli belonged to the samepe (respectively idioms, P-idioms, NI-actives and NI-hile the other 33% of the stimuli were not deriveds (fillers, active clauses). Furthermore, within eachof the targets were semantically related to eithertences word or the idiom meaning, whereas there

mantic link for the remaining 31% of the targets.a blocked presentation because a pilot study hadat mixing idiom and idiom-derived sentences led totancy of idiomatic meaning, which was reflected byiated priming effects for idiom-related targets with allpes.

was followinitiated wwere recor

for binaurasoftware. Tcomputer u

4. Results

Analyseipants (F1)dependentexperimentmissing dacalculatedparticipantstandard dIdiom-related Unrelated Non-word

Mort, death Table, table Neurre, neurre

Neval, neval

dure and apparatus

mulus trial consisted of the presentation of a fixationthe middle of the computer screen for 1000 ms fol-

n auditory prime sentence that was displayed overs. The visual target word was presented 200 ms afterf the prime sentence on the computer screen (blackletters, white background) and was centred in theion as the fixation cross. Participants were instructeds accurately and as quickly as possible whether thestring was a real French word or not. They were toldbutton corresponding to their dominant hand if thereal word and the button corresponding to their non-

and if it was not. Reaction times were measured fromnset. Targets remained on the computer screen untilant responded. A new trial was initiated 1000 ms

nse. The stimuli were presented in four successiveoms, P-idioms, NI-actives, NI-passives) the presen-r of which was permuted across the 24 controls and24 HD patients. Within the blocks stimuli were pre-omly for each participant. Each of the four blocks

ed by a short pause and the subsequent block washen the participant felt ready. All prime sentencesded by a native French male speaker and digitisedl presentation over headphones using COOL EDIThe experiment was administered on an IBM laptopsing E-PRIME experimental software.

s of variance (ANOVAs) were conducted by partic-and by items (F2) with reaction time (RT) as the

variable. They were conducted separately for the twoal sets. For each participant, incorrect responses andta were excluded from the analyses. Furthermore, wemean reaction times and standard deviations for eachand excluded data that deviated by more than two

eviations from the mean values for a given partici-


Fig. 1. Exper nt prime and target types in Controls and HD patients. *P is significantat the 0.05 lev

pant. This y(HD: 7.1%(HD: 5.8%

4.1. Main-

We usedtrols), primand targetsummarisecontrols an

RT wetrols (596108) = 3760(word relunrelated108) = 44.6741 ms 2passives 77P = 0.02)138) = 2.09interactionP = 0.002)interactionP = 0.006).and word-ring effectswere similafaster for boP < 0.001; FP < 0.001;gets (contrP < 0.001.P = 0.006).targets (coP > 0.1. Hgets (contrP < 0.001. HP < 0.001).idiom-relat

.55,targ

2.561.96trols

in23)rd-r6.521.29er, ciomand

ll senrdertedtar

(RTriesg w). Prnd Hs 0.1). There was a prime type target type(F1(6, 276) = 7.80, P < 0.001; F2(6, 108) = 3.68,

as well as a triple group prime type target type(F1(6, 276) = 2.85, P = 0.010; F2(6, 108) = 3.24,Restricted analyses compared RT for idiom-related

18) = 1related18) = 118) = 1in consimilar(F1(1,for wo23) = 223) = 2togethwith ididiomswith a

In ocalcularelatedtargetsond sepriminFs < 1trols a(both FF2(1,P = 0.0elated targets with RT for unrelated targets (prim-). Results with idioms, NI-actives and NI-passivesr in HD patients and controls. With idioms, RT wereth idiom-related targets (controls: F1(1, 23) = 99.04,2(1, 18) = 32.86, P = 0.001. HD: (F1(1, 23) = 47.00,

F2(1, 18) = 10.33, P < 0.005) and word-related tar-ols: F1(1, 23) = 91.16, P < 0.001; F2(1, 18) = 49.45,HD: (F1(1, 23) = 55.30, P < 0.001; F2(1, 18) = 9.54,With NI-actives, RT were similar for idiom-related

ntrols: (F1(1, 23) = 2.20, P > 0.1; F2(1, 18) = 2.11,D: both Fs < 1) but faster for word-related tar-ols: F1(1, 23) = 93.66, P < 0.001; F2(1, 18) = 36.68,

D: (F1(1, 23) = 30.90, P < 0.001; F2(1, 18) = 24.37,Likewise, with NI-passives, RT were similar for

ed targets (controls: F1(1, 23) = 2.86, P > 0.1; F2(1,

fact that p

Table 4Amount of prtargets in idio

IdiomsP-idiomsNI-activesNI-passives

Asterisks indlated). Boldcontrols: undminished primP > 0.1. HD: both Fs < 1) but faster for word-ets (controls: F1(1, 23) = 27.95, P < 0.001; F2(1,

, P = 0.002. HD: F1(1, 23) = 34.28, P < 0.001; F2(1,, P = 0.003). Results with P-idioms were differentand HD patients: RT for idiom-related targets werecontrols (both Fs < 1) but faster in HD patients

= 21.80, P < 0.001; F2(1, 18) = 10.28, P = 0.005). RTelated targets were faster in both controls (F1(1,,P < 0.001;F2(1, 18) = 6.29,P = 0.02) and HD (F1(1,, P < 0.001; F2(1, 18) = 16.58, P = 0.001). Takenontrols displayed priming for idiom-related targetss alone, whereas HD patients displayed priming withP-idioms. Priming for word-related targets existedtence types in both controls and HD patients.to compare the magnitude of the priming effects wethe amount of priming in milliseconds for idiom-gets (RTunrelated RTidiom-related) and word-relatedunrelated RTidiom-related) and entered them in a sec-of ANOVAs (see Table 4). For idiom-related targetsith idioms was similar in controls and HD (bothiming for word-related targets was similar in con-D patients with idioms (both Fs < 1) and NI-actives1) but not with P-idioms (F1(1, 46) = 9.57, P = 0.003;6.36, P = 0.033) and NI-passives (F1(1, 46) = 7.38,F2(1, 9) = 10.85, P = 0.009). This was related to the

riming in controls was smaller with P-idioms than

iming in milliseconds for idiom-related targets and word-relatedms, P-idioms, NI-actives and NI-passives

Unrelated idiom-related Unrelated word-related

Controls HD Controls HD

95* 97* 102* 108*8 99* 44* 122*18 3 108* 122*15 22 47* 129*

icate significant priming (idiom-related/word-related vs. unre-characters indicate priming effects that differentiate HD fromisrupted priming for idiom-related targets in P-idioms and undi-ing for word-related targets in P-idioms and NI-passives.


Fig. 2. Sub-s ith si(SR) seman ing thcontexts.

with idiomP = 0.017)23) = 20.75there was n

Finally,might haveing sentencbehaviour waware thatThus, we raorder comtences (firsindicated thand that in(both Fs

0.1) or were also norder with(all Fs < 1ent behaviowhen thepost-hoc Awith the vablock, thirthat there wF < 1; NI-aMoreover,F < 1; P-idpassives: FF < 1; P-idipassives: Fblock ordeidioms: F(6F < 1).

Performtrasts. Perfanalysis b96.87% 1in the anThere was

(wo 5

9.49.09,

alysi 3

es 97t in twas

1.02 pr; F2

ub-se

usednd

sultstien46)ontrenten1) w.70,lyseet: mean reaction times and standard deviations in controls and HD patients wtic restrictions. HD patients are sensitive to semantic restriction violations show

s (F1(1, 23) = 16.67, P < 0.001; F2(1, 18) = 6.91,and with NI-passives than with NI-actives (F1(1,, P < 0.001; F2(1, 18) = 6.90, P = 0.017), whereaso difference in HD (all Fs < 1).we checked whether the blocked design of the study

resulted in the use of conscious strategies dur-e processing. First, participants might have changedithin the same block when becoming progressively

66% of the stimuli belong to the same sentence type.n a post-hoc ANOVA with the variable intra-blockparing the first half of idioms/idiom-derived sen-

t seen) with the second half (second seen). Resultsat there was no intra-block order effect (both Fs < 1)

tra-block order did not interact either with group1) or with prime type (F1 < 1; F2(3, 108) = 1.58,target type (F1(2, 92) = 1.97,P > 0.1;F2 < 1). Thereo triple or quadruple interactions of the intra-block

the variables group, prime type and target type). Second, participants might have displayed differ-ur as a function of the overall block order especially

idiom block was presented first. We therefore ranNOVAs by subjects for each of the four prime typesriable block order (first seen block, second seen

d seen block, forth seen block). Analyses showedas no block order effect (idioms: F < 1; P-idioms:

items98.49%108) =92) = 2the an98.89%passivbut noThere276) =groupP > 0.1

4.2. S

Wetrols) athe reHD pa(F1(1,both cSR+ sP = 0.046) = 2

Ana

ctives: F < 1; NI-passives: F(3, 40) = 1.03, P > 0.1).block order neither interact with group (idioms:ioms: F(3, 40) = 1.19, P > 0.1; NI-actives: F < 1; NI-(3, 40) = 2.16, P > 0.1) nor with target type (idioms:oms: F(6, 80) = 2.14, P = 0.06; NI-actives: F < 1; NI-(6, 80) = 1.65, P > 0.1). There was also no triple

r group target type interaction (idioms: F < 1; P-, 80) = 1.82, P > 0.1; NI-actives: F < 1; NI-passives:

ance accuracy was analysed along the same con-ormance in HD and controls was similar in the

y subjects (controls: 99.34% 2.62 correct, HD:0.02 correct; F1(1, 46) = 3.74, P = 0.06) but not

alysis by items (F2(1, 108) = 54.47, P < 0.001).a target type effect only in the analysis by

to RT. Perfthan in coP < 0.001;Fand SR+ seno group P > 0.1; F2

5. Correla

The expwith the Mof the UHtest, word(total functic rule appmple active clause sentences that either respect (SR+) or violateat they have lexical access to verb-specific information in phrasal

rd related 98.80% 3.56 correct, idiom-related.99 correct, unrelated 97.03% 10.73 correct; F2(2,, P < 0.001) but not in the analysis by subjects (F1(2,P > 0.1). There was a prime type effect only in

s by items (idioms 98.68% 4.30 correct, P-idioms.94 correct, NI-actives 97.64% 8.93 correct, NI-.22% 10.27 correct; F2(3, 108) = 5.16, P = 0.002)

he analysis by subjects (F1(3, 138) = 2.49, P = 0.06).no prime type target type interaction (F1(6,

, P > 0.1; F2(6, 108) = 1.23, P > 0.1) nor a tripleime type target type interaction (F1(6, 276) = 1.55,(6, 108) = 2.10, P > 0.1).

t

the two independent variables group (HD, con-semantic restriction (SR+, SR). Fig. 2 summarisesin controls and HD patients. RT were slower in

ts (869 ms 154) than in controls (559 ms 101)= 78.94, P < 0.001; F2(1, 28) = 226.26, P < 0.001). Inols and HD patients SR yielded slower RT thances (F1(1, 46) = 56.80, P < 0.001; F2(1, 28) = 7.18,ith no group semantic restriction interaction (F1(1,P > 0.1; F2 < 1).s on performance accuracy were conducted similar

ormance was lower in HD (94.17% 5.26 correct)ntrols (98.75% 2.63 correct) (F1(1, 46) = 26.01,2(1, 28) = 8.47,P < 0.01) but did not differ with SR

ntences (F1(1, 46) = 1.55, P > 0.1; F2 < 1). There wassemantic restriction interaction (F1(1, 46) = 2.56,

< 1).

tion analyses

erimental results of HD patients were correlatedDRS and different disease progression parametersDRS including executive function scores (Stroopfluency, symbol digit code), motor scores and TFCtional capacity). Correlation analyses with syntac-lication used the amount of priming with P-idioms

M.T

eichmann

etal./Neuropsychologia

46(2008)174185181

Table 5Correlation analyses in HD patients

Disease progression Sentence processing

TFC UHDRS motor Executive functions Rule application Lexical processes

Word fluency Symbol digit test Stroop W/C P-idioms idiom-rel P-idioms word-rel NI-passives word-rel Idioms idiom-rel

MDRSR 0.577* 0.532* 0.450* 0.638* 0.644* 0.029 0.555* 0.042 0.296P 0.004 0.009 0.027 0.001 0.001 0.896 0.006 0.851 0.160

TFCR 0.420* 0.408* 0.636* 0.567* 0.240 0.201 0.303 0.353P 0.041 0.048 0.001 0.004 0.258 0.347 0.150 0.091

UHDRS motorR 0.427* 0.708* 0.666* 0.085 0.105 0.140 0.363P 0.037 0.000 0.000 0.693 0.825 0.515 0.081

Word fluencyR 0.485* 0.652* 0.325 0.247 0.011 0.369P 0.016 0.001 0.121 0.244 0.958 0.076

Symbol digit codeR 0.740* 0.037 0.444* 0.110 0.353P 0.000 0.862 0.030 0.610 0.090

Stroop W/CR 0.204 0.244 0.081 0.383P 0.340 0.251 0.708 0.065

R, correlation coefficient; P * is significant at the 0.05 level (two-tailed). TFC: total functional capacity; UHDRS motor: UHDRS motor scores; word fluency: word fluency (PRV) in 2 min; Stroop W/C: word-coloursub-part of the Stroop test; idiom-rel: idiom-related targets; word-rel: word-related targets.


and NI-passives: (1) P-idioms (RTunrelated RTidiomrelated),(2) P-idioms (RTunrelated RTwordrelated), (3) NI-passives(RTunrelatedabilities usmeanings:

Diseaseexecutive(MDRS) codid not corrule applic(RTunrelateding lexicalTable 5.

6. Discuss

We invetion by testsyntactic ruof sentenceorders, namwe tested blexical acceidiom mea

Controlidiomatic srelated targtargets wittic restrictisentencescontrols apas reflectedwith P-idiowith P-idiotrols withdid not appreflected byidioms andP-idioms a

Taken tphrasal strguage reprHD, lexicasyntactic mserve the abut are impspecificallynon-canonia linear detences (firsrole). Thisencountereposition wsentation ois reminisction showiFrench non

that respond to less frequent sub-regularities (Teichmann et al.,2005).

res

ompent

s. Atinvoexte200res.ent996

t on m997

ent gtic mm apngundenompll stus thetic moulo

l subatedx, Kropoleadi

tolikeIn li

inhibn theA swhoof int to mdeci

cessepactsing ey. Tateduse

thare cr

ore

nctiprog

l/met be

tic meners inf a gions. RTword-related). Correlation analyses with lexicaled the amount of priming reflecting access to idiomidioms (RTunrelated RTidiom-related).progression scores (TFC, UHDRS motor scores,

function scores) and global intellectual capacitiesrrelated with one another. Executive function scores

relate with either priming effects reflecting syntactication (except for the symbol digit code and P-idioms RTword-related)) nor with priming effects reflect-access to idiom meanings. Results are detailed in

ion

stigated the role of the striatum in sentence percep-ing the claim that the striatum specifically subservesle use as opposed to lexical processes. Both aspectsprocessing were assessed in a model of striatal dis-ely early stages of HD. Using a priming paradigm

oth the application of syntactic movement rules andss to word meanings, to semantic restrictions and to

nings.s had lexical access to word meanings and to wholeentences as shown by priming respectively for word-ets with all sentence types and for idiom-related

h idioms. In addition, they had access to seman-ons as reflected by increased reaction times withcontaining semantic restriction violations. Finally,plied syntactic movement rules to passive structuresby the disruption of priming for idiom-related targetsms and decreased priming for word-related targetsms and NI-passives. HD patients behaved like con-respect to idioms and NI-actives. In contrast, theyly movement rules to passive sentence structures aspersisting priming for idiom-related targets with P-undecreased priming for word-related targets with

nd NI-passives.ogether, our results indicate that the processing ofuctures relies on both lexical access to stored lan-esentations and the application of syntactic rules. Inl abilities are preserved whereas the application ofovement rules is deficient. Indeed, HD patients con-bility to process idioms and simple active sentencesaired on the processing of passive structures. More, our results suggest that HD patients process thecal structure of passives along what could be calledfault rule, which permits the parsing of active sen-

t noun of the sentence = agent role; last noun = themekind of default rule applied whenever HD patientsd passive structures independently from the sentenceithin a given block and independently from the pre-rder of the different sentence blocks. This findingent of results provided in the domain of conjuga-ng that HD patients apply a basic default rule to-verbs while they are impaired with non-verb stimuli

Ourtence cmovem

aspecttum is2001)et al.,structumovem

et al., 1impacet al., 1represlinguisstriatutinct laindepeboth csevera

impairsyntacNatsoplogicaautomDupoubeen prules,shownforms1997).ties inresult iations.(2005)cesses

respecdo notin proily imclarifyphologconjugin thegestingthat awere m

is a futo theventrait mighsyntacral degabilitieview ooperatults suggest that the striatum holds a role in sen-rehension that pertains to the application of syntacticrules whereas it is not involved in lexical processingfirst view, this supports the broad claim that the stria-lved in the computation of language rules (Ullman,nding evidence from the word domain (Teichmann5; Ullman et al., 1997) to the domain of phrasalHowever, the striatum seems to subserve complexrules in syntax (see also Kemmerer, 1999; McNamara; Natsopoulos et al., 1993), whereas it was claimed to

ore basic default rules in word morphology (Ullman). Thus, assuming that both complex and default rulesenuine language rules in that they intervene whenever

emory fails (Jackendoff, 2002; Pinker, 1999), thepears to underpin distinct rule computations in dis-

age domains. On the contrary, it may be argued that,t from the language domain, the striatum impacts onex and default rules, yet at different degrees. Indeed,dies have shown that striatal damage prominentlycomputation of complex and infrequent rules, such asovement (Kemmerer, 1999; McNamara et al., 1996;s et al., 1993; Teichmann et al., 2005) and morpho--regularities (Teichmann et al., 2005), while moredefault rules are more mildly affected (Teichmann,ouider, & Bachoud-Levi, 2006). Alternatively, it hassed that striatal damage mainly impacts on defaultng to excessive rule use in HD patients, who wereover-regularise during verb conjugation producingdig-ed (from to dig) instead of dug (Ullman et al.,ne with this, HD patients may basically have difficul-iting the over-active default rule, which in turn, mightdiminished application of more infrequent rule oper-imilar view has been adopted by Longworth et al.

stated that the striatum is involved in general pro-hibition that may account for language deficits withorphology and syntax. Yet, even if the present data

sively demonstrate whether the striatum is involveds of default rule inhibition or whether it primar-the application of complex language rules, some

vidence is available from the domain of word mor-eichmann et al. (2006) assessing the perception ofverb forms showed that HD patients were impaired

of both complex sub-rules and the default rule sug-t the striatum comprises computational propertiesitical for rule application as such. Yet, sub-rulesaffected than the default rule indicating that there

onal gradient within the striatum that might relateression of neural degeneration following a dorso-

dio-lateral gradient (Vonsattel et al., 1985). Likewise,that syntactic rule impairments in HD extend fromovement to canonical word order processing as neu-ation proceeds. Assessing the different syntactic ruledifferent stages of HD should help to confirm theenuine, but probably graduated, impairment on rule


In each case, our findings are at odds with the view statingthat the striatum impacts the processing of lexical information.Friedericiposed that tthat checkthe syntactwere claimrather thanproposal wthat the strcation in gfunctions.language saccess to lsentationsmanipulatistrategies. Yand the crione cannotexplicit linprocesses ctask and mthis, we ranindependenand the preseries of coparameter arule applicdamage ev

In summmovementat the phratum is invoFurthermorabsence ofas the nearparametersvia languagpossibilitytional circu2000), alsoas impliedKotz et al. (ent with finretrieval delexicon invOjemann, Otrolled prodamaged pbal fluencyand object nsemantic/leto conciliatmight holdcontrolledareas and imcircuits. Th

diffusion tensor imaging (Lehericy et al., 2004) showing that thehuman striatum is connected to both the dorsal-lateral prefrontal

, knot, 200act ort, 2ulatig (Landthe sn, Shs, &In pe strettereirann

cisiown toflect

ere

requeat sttion

ogethemen

gatingestge-st notl lexexec

rifiednt pon thethe orule

d by, its

hat gg ther stutionserv

wled

s woallo

nt HtionWeandu anre w

Mons. Wand Kotz (2003) and Longworth et al. (2005) pro-he striatum is involved in late integrational processeswhether lexicalised information is compatible withic frame. Such kinds of lexical checking processesed to relate to controlled post-linguistic operationsto implicit language-specific processes. A similar

as formulated by Grossman et al. (2002) who claimediatum impacts on sentence processing via its impli-eneral control processes that are tied to executive

However, our results challenge this lexical and non-pecific view showing that HD patients had normalexical information, such as word and idiom repre-and semantic restrictions, by means of an impliciton task that minimised executive-controlled response

et, despite the short interval between the target onsettical word of the prime sentence, namely 200 ms,exclude that the participants might have established

ks between the primes and the targets. Such explicitould have modified the response strategy along theight have induced performance biases. To check forpost-hoc analyses showing that priming effects weret from both the sentence position within a given blocksentation order of the blocks. Moreover, we ran arrelation analyses showing that only one executivemong nine correlated with performance on syntactic

ation suggesting that rule impairment and executiveolve independently in HD.ary, our results show that the application of syntacticrules is deficient in HD whereas lexical processingsal level is not. Thus it is suggested that the stria-lved in the former processes but not in the latter.e, the use of an implicit manipulation task and theboth prime order and block order effects as wellabsence of correlations between executive and rulesuggest that the striatum subserves rule applicatione-specific operations. Yet, we do not preclude the

that the striatum, which is known to encompass func-its of executive functioning (e.g., Middleton & Strick,controls lexical output at some post-linguistic level

by the data of Friederici et al. (1999) and Friederici,2003) and Kotz et al. (2003). This view is also coher-dings showing that lexical operations such as wordpend on controlled research strategies in the mentalolving prefrontal and striatal structures (e.g., Rosen,

llinger, & Petersen, 2000). The same kinds of con-cesses might also account for the fact that striatalatients have difficulties in explicit tasks assessing ver-(e.g., Butters, Wolfe, Granholm, & Martone, 1986)aming (e.g., Frank et al., 1996) and that they producexical paraphasias (Damasio et al., 1982). In an efforte both points of view, we propose that the striatuma twofold role in language processing comprising

processes of lexical access via the link to executiveplicit rule computations via more language specific

is proposal is compatible with recent findings from

cortexKnighto impHagoomaniplearnin2004)tied toSalmoButter1987).that thfrom linto thTeichmcal deunknoular inrules mhigh fing thapplica

Altinvolvinvestiwe suglanguation bucontrotied tother vedifferement i

Ontum inclaimeIndeedrules tdefininFurthedistincare sub

Ackno

Thi(2001)AssistaFondamann.

scores

DolbeathermoHenripatientwn to subserve executive functions (e.g., Stuss &2), and to portions of Brocas area, which is claimed

n combinatorial aspects of language processing (e.g.,005). Likewise, it is compatible with data of impliciton tasks showing that rule use in artificial grammarieberman, Chang, Chiao, Bookheimer, & Knowlton,in word morphology (Teichmann et al., 2006) aretriatum whereas lexical operations are not (Heindel,ults, Walicke, & Butters, 1989; Salmon, Shimamura,

Smith, 1988; Shimamura, Salmon, Squire, & Butters,articular, Lieberman et al. (2004) using fMRI showediatum is activated when healthy adults abstract rulesstrings even though they have no conscious insightinherent rule structure. Similarly, in morphology,

et al. (2006) assessed HD patients with a lexi-n task (word versus non-word decision), in which,

the participants, the frequency of regular and irreg-ed verb forms was manipulated. HD patients appliedly to low frequency regulars while they accessedncy forms directly in the mental lexicon, suggest-

riatal disorders hamper implicit processes of rule.er, this study provided novel arguments for thet of the striatum in language rule application throughg the domain of sentence processing. In particular,that certain portions of the striatum are involved in

pecific computations that are related to rule applica-to lexical operations. Conversely, the striatum might

ical operations via other functional circuits that areutive functioning. Such a twofold role should be fur-by providing more detailed correlations between thertions of the striatum and their respective involve-different language components.ther hand, our data showed that the role of the stria-processing might be more restricted than initiallythe declarative-procedural model (Ullman, 2001).

computational function seems to pertain to linguisticovern syntactic movement whereas more basic rulesprocessing of simple active sentences are preserved.

dies are necessary to evaluate this more fine-grainedand to track the specific kind of language rules thated by striatal structures.

gments

rk was conducted with the help of an Avenir grantcated to Bachoud-Levi by the INSERM and anospitalier de Recherche (AHR AP-HP/INSERM) etpour la Recherche Medicale grants to Marc Teich-thank Marie-Francoise Boisse for assessing MDRS

cognitive scales of the UHDRS and Guillaumed Amandine Rialland for transmitting the data. Fur-e wish to thank the Neurology Department of the

dor Hospital and the CIC for providing access to thee are grateful to the Reseau Huntington de Langue


Francaise who supports this work. We wish also to thank HelenBates and Chantal Vachon for the corrections of the manuscript.

Reference

Butters, N., Wverbal recdisease. C

Cacciari, C.,Memory a

Cardebat, D.,semantic lproductioBelgica, 9

Chomsky, N.Chomsky, N.

AkmajianChomsky, N. (

Press.Chomsky, N.Copland, D.

insights flesions, PaNeuropsyc

Damasio, A. Rsia with nArchives o

Frank, E. M.secondaryof Commu

Friederici, A.Functiona

Friederici, A.Syntacticmatic andNeuropsyc

Friederici, A.of left infesion: Loca7077.

Friederici, A.brain potelesions. B

Golden, C. J.Grossman, M

H. I. (200in Parkins

Grossman, Mal. (2002)Parkinson

Heindel, W. C(1989). Ntems: A cpatients. Tfor Neuro

Hagoort, P. (2Cognitive

Huntington SReliabilityMovemen

Illes, J. (1989dissociatetons, and

Jackendoff, Revolution.

Kemmerer, Dtions in Pa

Kotz, S. A., Frisch, S., von Cramon, D. Y., & Friederici, A. D. (2003). Syntacticlanguage processing: ERP lesion data on the role of the basal ganglia. Journalof the International Neuropsychological Society, 9, 10531060.

. E., P982).ed by254

y, S.,pon, Cicostrian, M004).nced438.rth, C. (200vasc

ald, Mntact

S. (19rly pa771ta, J.t al. (1at ris357ara, Pcalitympar

nce, 8on, F.or and250.., Te

1). Syroimaulos,u, Z.,preheropsyc., Pallch lexrs, 36ski, Mting once, 6S. (19c Boo. J., O

of brI. Br

, D. Psema

of Clirnatiors, H

d reconitionura, Adysfuroscien, I. (1neur

., & Kord U, D. Aessions

olfe, J., Granholm, E., & Martone, M. (1986). An assessment ofall, recognition and fluency abilities in patients with Huntingtonsortex, 22, 1132.& Tabossi, P. (1988). The comprehension of idioms. Journal ofnd Language, 27, 668683.Doyon, B., Puel, M., Goulet, P., & Joanette, Y. (1990). Formal andexical evocation in normal subjects. Performance and dynamics ofn as a function of sex, age and educational level. Acta Neurologica0, 207217.(1965). Aspects of the theory of syntax. Cambridge: MIT Press.(1977). On Wh movement. In P. Culicover, T. Wasow, & A.(Eds.), Formal syntax. New York: Academic Press.1980).Rules and representations. New York: Columbia University

(1986). Barriers. Cambridge, MA: MIT Press.(2003). The basal ganglia and semantic engagement: Potentialrom semantic priming in individuals with subcortical vascularrkinsons disease, and cortical lesions. Journal of the Internationalhological Society, 9, 10411052.., Damasio, H., Rizzo, M., Varney, N., & Gersh, F. (1982). Apha-

on-hemorrhagic lesions in the basal ganglia and internal capsule.f Neurology, 39, 1524.

, McDade, H. L., & Scott, W. K. (1996). Naming in dementiato Parkinsons, Huntingtons, and Alzheimers diseases. Journalnication Disorders, 29, 183197., & Kotz, S. A. (2003). The brain basis of syntactic processes:l imaging and lesion studies. Neuroimage, 20(Suppl 1), 817.D., Kotz, S. A., Werheid, K., Hein, G., & von Cramon, D. Y. (2003).comprehension in Parkinsons disease: Investigating early auto-late integrational processes using event-related brain potentials.hology, 17, 133142.

, Ruschemeyer, S. A., Hahne, A., & Fiebach, C. J. (2003). The rolerior frontal and superior temporal cortex in sentence comprehen-lising syntactic and semantic processes. Cerebral Cortex, 13(2),

D., von Cramon, D. Y., & Kotz, S. A. (1999). Language relatedntials in patients with cortical and subcortical left hemisphererain, 122, 10331047.(1978). Stroop color and word test. Chicago: Stoelting.., Kalmanson, J., Bernhardt, N., Morris, J., Stern, M. B., & Hurtig,0). Cognitive resource limitations during sentence comprehensionons disease. Brain and Language, 73, 116.., Zurif, E., Lee, C., Prather, P., Kalmanson, J., Stern, M. B., et. Information processing speed and sentence comprehension ins disease. Neuropsychology, 16, 174181.., Salmon, D. P., Shults, C. W., Walicke, P. A., & Butters, N.

europsychological evidence for multiple implicit memory sys-omparison of Alzheimers, Huntingtons, and Parkinsons diseasehe Journal of Neuroscience: The Ofcial Journal of the Societyscience, 9(2), 582587.005). On Broca, brain, and binding: A new framework. Trends insciences, 9(9), 416423.tudy Group. (1996). Unified Huntingtons disease rating scale:

and consistency. Movement Disorders: Ofcial Journal of thet Disorder Society, 11, 136142.). Neurolinguistic features of spontaneous language productionthree forms of neurodegenerative disease: Alzheimers, Hunting-Parkinsons. Brain and Language, 37, 628642.. (2002). Foundations of language: Brain, meaning, grammar,New York: Oxford University Press.

. (1999). Impaired comprehension of raising-to-subject construc-rkinsons disease. Brain and Language, 66, 311328.

Kuhl, DJ. (1min12, 4

LehericPoucort

LiebermJ. (2bala427

LongwoL. Kfrom

Macdonof sy

Mattis,elde(pp.

MazziotH., ejects316,

McNammatiA coscie

MiddletMot236

Moro, A(200Neu

Natsoposaro

com

NeuNew, B

Frenpute

Peschangrafscie

Pinker,Basi

Rosen, HisonFMR

SalmonandnalInte

Schriefewor

CogShimam

oryNeu

Shoulsowith

Stuss, DOxf

Swinneyexpr534.helps, M. E., Markham, C. H., Metter, E. J., Riege, W. H., & Winter,Cerebral metabolism and atrophy in Huntingtons disease deter-18FDG and computed tomographic scan. Annals of Neurology,34.Ducros, M., Van de Moortele, P. F., Francois, C., Thivard, L.,., et al. (2004). Diffusion tensor fiber tracking shows distinct

atal circuits in humans. Annals of Neurology, 55, 522529.. D., Chang, G. Y., Chiao, J., Bookheimer, S. Y., & Knowlton, B.An event-related fMRI study of artificial grammar learning in a

chunk strength design. Journal of Cognitive Neuroscience, 16(3),

. E., Keenan, S. E., Barker, R. A., Marslen-Wilson, W. D., & Tyler,5). The basal ganglia and rule-governed language use: Evidence

ular and degenerative conditions. Brain, 128, 584596.. C., Pearlmutter, N. J., & Seidenberg, M. S. (1994). Lexical nature

ic ambiguity resolution. Psychological Review, 101, 676703.76). Mental status examination for organic mental syndrome intients. In L. Bellak & T. B. Karasu (Eds.), Geriatric psychiatry21). New York: Grune & Straton.C., Phelps, M. E., Pahl, J. J., Huang, S. C., Baxter, L. R., Riege, W.987). Reduced cerebral glucose metabolism in asymptomatic sub-

k for Huntingtons disease. The NewEngland Journal ofMedicine,362.., Krueger, M., OQuin, K., Clark, J., & Durso, R. (1996). Gram-judgments and sentence comprehension in Parkinsons disease:

ison with Brocas aphasia. The International Journal of Neuro-6, 151166.A., & Strick, P. L. (2000). Basal ganglia and cerebellar loops:cognitive circuits. Brain Research. Brain Research Reviews, 31,

ttamanti, M., Perani, D., Donati, C., Cappa, S. F., & Fazio, F.ntax and the brain: Disentangling grammar by selective anomalies.ge, 13, 110118.D., Grouios, G., Bostantzopoulou, S., Mentenopoulos, G., Kat-& Logothetis, J. (1993). Algorithmic and heuristic strategies innsion of complement clauses by patients with Parkinsons disease.hologia, 31, 951964.

ier, C., Brysbaert, M., & Ferrand, L. (2004). Lexique 2: A newical database. Behavior Research Methods, Instruments & Com-

(3), 516524.., Cesaro, P., & Hantraye, P. (1995). Rationale for intrastriatal

f striatal neuroblasts in patients with Huntingtons disease. Neuro-8, 273285.99). Words and rules: The ingredients of language. New York:ks.jemann, J. G., Ollinger, J. M., & Petersen, S. E. (2000). Compar-

ain activation during word retrieval done silently and aloud usingain and Cognition, 42, 201217.., Shimamura, A. P., Butters, N., & Smith, S. (1988). Lexicalntic priming deficits in patients with Alzheimers disease. Jour-nical and Experimental Neuropsychology: Ofcial Journal of thenal Neuropsychological Society, 10, 477494.., Friederici, A. D., & Rose, U. (1998). Context effects in visualgnition: Lexical relatedness and syntactic context. Memory &, 26, 12921303.. P., Salmon, D. P., Squire, L. R., & Butters, N. (1987). Mem-

nction and word priming in dementia and amnesia. Behavioralnce, 101(3), 347351.981). Huntington disease: Functional capacities in patients treated

oleptic and antidepressant drugs. Neurology, 31, 13331335.night, R. T. (2002). Principles of frontal lobe function. New York:

niversity Press.., & Cutler, A. (1979). The access and processing of idiomatics. Journal of Verbal Learning and Verbal Behavior, 18, 523


Teichmann, M., Dupoux, E., Kouider, S., Brugie`res, P., Boisse, M. F., Baudic,S., et al. (2005). The role of the striatum in rule application: The model ofHuntingtons disease at early stage. Brain, 128, 11551167.

Teichmann, M., Dupoux, E., Kouider, S., & Bachoud-Levi, A. C. (2006). Therole of the striatum in processing language rules: Evidence from word per-ception in Huntingtons disease. Journal of Cognitive Neuroscience, 18(9),115.

Tettamanti, M., Moro, A., Messa, C., Moresco, R. M., Rizzo, G., Carpinelli, A., etal. (2005). Basal ganglia and language: Phonology modulates dopaminergicrelease. Neuroreport, 16(4), 397401.

Titone, D. A., & Connine, C. M. (1994). The comprehension of idiomaticexpressions: Effects of predictability and literality. Journal of ExperimentalPsychology. Learning, Memory, and Cognition, 20, 11261138.

Tyler, L. K. (1985). Real-time comprehension processes in agrammatism: Acase study. Brain and Language, 26, 259275.

Ullman, M. T. (2001). A neurocognitive perspective on language: The declara-tive/procedural model. Nature Reviews. Neuroscience, 2(10), 717726.

Ullman, M. T., Corkin, S., Coppola, M., Hickok, G., Growdon, J. H., Koroshetz,W. J., et al. (1997). A neural dissociation within language: Evidence that themental dictionary is part of declarative memory, and that grammatical rulesare processed by the procedural system. Journal of Cognitive Neuroscience,9(2), 266276.

Vannest, J., Polk, T. A., & Lewis, R. L. (2005). Dual-route processing of com-plex words: New fMRI evidence from derivational suffixation. Cognitive,Affective & Behavioral Neuroscience, 5(1), 6776.

Vonsattel, J. P., Myers, R. H., Stevens, T. J., Ferrante, R. J., Bird, E. D., &Richardson, E. P., Jr. (1985). Neuropathological classification of Hunting-tons disease. Journal of Neuropathology and Experimental Neurology, 44,559577.

Wallesch, C. W., & Papagno, C. (1988). Subcortical aphasia. In F. C. Rose,R. Whurr, & M. Wyke (Eds.), Aphasia (pp. 256287). London: WhurrPublishers.

Wechsler, D. (1981). Wechsler adult intelligence scale-revised manual. NewYork: Psychological Corporation.

The role of the striatum in sentence processing: Evidence from a priming study in early stages of Huntingtons diseaseIntroductionMethodsParticipantsGeneral assessment

Priming task with idioms and idiom-derived sentencesMaterialsProcedure and apparatus

ResultsMain-setSub-set

Correlation analysesDiscussionAcknowledgmentsReferences

striatum

Documents