doi:10.1093/brain/awn036
Brain (2008), 131, 1046 ^1056
Language processing within the striatum: evidence
from a PET correlation study in Huntington’s disease
Marc Teichmann,1,2,3,4 Ve¤ronique Gaura,5 Jean-Franc ois De¤monet,6 Fre¤de¤ric Supiot,7 Marie Delliaux,8
Christophe Verny,9 Pierre Renou,10 Philippe Remy2,5 and Anne-Catherine Bachoud-Le¤vi1,2,3
1
INSERM U841, Equipe 1, Neuropsychologie Interventionnelle, IM3/Paris XII, Cre¤teil, 2AP-HP, Ho“pital Henri Mondor, Service
de neurologie, Cre¤teil, 3Ecole Normale Supe¤rieure, De¤partement d’Etudes Cognitives, Paris, 4Laboratoire de Sciences
Cognitives et Psycholinguistique, UMR8554, EHESS-ENS-CNRS, Paris, 5URA CEA-CNRS 2210, Service Hospitalier Fre¤de¤ric
Joliot Orsay, 6INSERM U825, Po“le Neurosciences, Ho“pital Purpan, Toulouse, France, 7Ho“pital Erasme. Service de neurologie,
Bruxelles, Belgium, 8CHRU de Lilles, Ho“pital Roger Salengro, Service de Neurologie, Lille, 9CHU d’Angers, Service de
neurologie, Angers and 10CHU de Nantes. Ho“pital Guillaume et Rene¤ Laennec, Service de neurologie, Nantes, France
Correspondence to: Anne-Catherine Bachoud-Le¤vi, Ho“pital Henri Mondor. Service de neurologie. 54 avenue du Mare¤chal
de Lattre de Tassigny. 94000 Cre¤teil, France
E-mail: [email protected]
The role of sub-cortical structures in language processing, and more specifically of the striatum, remains
controversial. In line with psycholinguistic models stating that language processing implies both the recovery of
lexical information and the application of combinatorial rules, the striatum has been claimed to be involved
either in the former component or in the latter. The present study reconciles these conflicting views by showing
the striatum’s involvement in both language processes, depending on distinct striatal sub-regions. Using PET
scanning in a model of striatal disorders, namely Huntington’s disease (HD), we correlated metabolic data of
31 early stage HD patients regarding different striatal sub-regions with behavioural scores on three rule/lexicon
tasks drawn from word morphology, syntax and from a non-linguistic domain, namely arithmetic. Behavioural
results reflected impairment on both processing aspects, while deficits predominated on rule application. Both
correlated with the left striatum but involved distinct striatal sub-regions. We suggest that the left striatum
encompasses linguistic and arithmetic circuits, which differ with respect to their anatomical and functional
specification, comprising ventrally located regions dedicated to rule computations and more dorsal portions
pertaining to lexical devices.
Keywords: striatum; language processing; PET imaging; Huntington’s disease
Abbreviations: HD = Huntington’s disease; PD = Parkinson’s disease; NV = non-verbs
Received October 18, 2007. Revised January 14, 2008. Accepted February 11, 2008. Advance Access publication March 11, 2008
Introduction
While the role of cortical areas in linguistic processing is
relatively well established, the role of sub-cortical structures
such as the striatum, and the way they impact language
processing, is still controversial. Indeed, studies using both
patients sustaining striatal damage and functional brain
imaging with healthy subjects have suggested that the
striatum is involved in various aspects of language comprising
phonology (e.g. Démonet et al., 1991; Tettamanti et al., 2005),
word morphology (e.g. Ullman et al., 1997; Teichmann et al.,
2005; Vannest et al., 2005; Teichmann et al., 2006) and syntax
(e.g. Illes, 1989; Moro et al., 2001; Friederici and Kotz, 2003;
Teichmann et al., 2005). To clarify this functionally
unspecified picture it has been proposed that the striatum
impacts computational processes, which may cut across the
different language domains. In line with this proposal, several
authors have adopted a non-language specific view, positing
that the striatum modulates language via general processes
such as executive functioning or working memory
(Lieberman et al., 1992; Grossman et al., 1993, 2000, 2002).
However, two more language-specific accounts were built on
the observation that patients sustaining damage to the
striatum either demonstrate difficulties with lexical processing such as word retrieval (Crosson, 1985; Wallesch and
Papagno, 1988; Copland, 2003; Longworth et al., 2005) or
with grammatical rule application as required in tense
agreement and sentence comprehension (Ullman, 2001;
Teichmann et al., 2005). This controversy, which may be
ß The Author (2008). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For Permissions, please email: [email protected]
Language processing within the striatum
referred to as the lexicon/rule conflict, is grounded on
psycholinguistic models stating that the human language
faculty is organized along a dual processing architecture
comprising both a mental lexicon and a computational
grammar. The lexicon is thought to contain all linguistic
idiosyncrasies such as phonemes, morphemes and words,
whereas the computational grammar holds the combinatorial
rules which respectively apply to the lexical input (Chomsky,
1965; Pinker, 1999). The present study attempted to resolve
the lexicon/rule conflict by testing the hypothesis that striatal
structures are involved in both language aspects depending on
the anatomical subcomponents of the striatum.
In morphology, the lexicon/rule contrast has primarily
been assessed by comparing the conjugation of regular
verbs or non-verbs (NV) (rule-based; e.g. walk/walk-ed or
splush/splush-ed) and of irregular verbs (lexical-based; e.g.
go/went). Several studies have shown that striatal damage,
due to stroke or to neuro-degenerative diseases such as
Huntington’s disease (HD), characterized by primary neural
death in the striatum (Vonsattel et al., 1985; Peschanski
et al., 1995) and Parkinson’s disease (PD), related to
dopaminergic deficit in the striatum, are specifically
impaired in the conjugation of regular verbs and NV,
whereas their performance in irregular verbs is largely
preserved (Ullman et al., 1997; Teichmann et al., 2005).
Similar results were obtained during verb perception,
showing that HD patients are impaired in judging
conjugated NV, whereas they demonstrate relatively good
performance with irregular forms (Teichmann et al., 2006).
However, Longworth et al. (2005) reported data which
were at odds with the striatum-rule hypothesis whereas
favouring the lexical view. As in the previous studies
(Ullman et al., 1997; Teichmann et al., 2005) the authors
used a conjugation task which required the generation of
the past tense form upon the presentation of a verb
infinitive. Assessing PD, HD and stroke patients with
striatal damage, Longworth et al. (2005) did not find any
difference between regular and irregular forms, whereas
they reported a tendency for lexical intrusions substituting
semantically related verbs (e.g. bang-ed instead of slamm-ed).
Similarly, using a priming task, the authors showed that
regular past tense forms (e.g. ‘jump-ed’) primed their
respective base forms (e.g. ‘jump’) suggesting that rulebased processes of stem-suffix decomposition are intact
in these patients. In contrast, unlike control subjects PD
patients displayed priming also for similar sounding words
(e.g. captive—captain) suggesting that striatal damage
results in difficulty inhibiting inappropriate lexical items.
Furthermore, such a lexical view of the striatum appeared
to be coherent with data showing that striatal damage
results in verbal paraphasias (e.g. Cambier et al., 1979;
Damasio et al., 1982; Puel et al., 1984) and in naming
deficits (e.g. Damasio et al., 1982; Cappa et al., 1983;
Wallesch, 1985; Alexander et al., 1987; Démonet et al.,
1991; Frank et al., 1996). In brief, the previous studies
provided conflicting findings with respect to the functional
Brain (2008), 131, 1046 ^1056
1047
specification of the striatum, which may be due to various
factors such as the differing demands of the tasks and/or
distinct lesion patterns of the different patient populations.
However, the investigation of another language domain,
namely syntax, did not clarify this controversial picture,
presumably because rule application and lexical aspects
were not tested simultaneously. Several authors tested the
ability of patients with striatal damage to understand noncanonical sentences like passives, in which the usual word
order is inverted (e.g. ‘The girl was observed by the boy’) as
compared to canonical structures (e.g. ‘The boy observed
the girl’). According to a number of linguistic accounts,
such non-canonical structures critically depend on the
application of syntactic movement rules that allow for word
order re-mapping and thus for the usual canonical sentence
interpretation (e.g. Chomsky, 1965, 1977, 1986). Assessing
PD patients with non-canonical and canonical sentences
showed that these patients experience difficulty interpreting
the former clauses, whereas they have near-normal
performance with the latter (Natsopoulos et al., 1993;
McNamara et al., 1996; Kemmerer, 1999). Similar results
were reported in HD. Teichmann et al. (2005) varied the
plausibility and the canonicity of simple French clauses,
showing that HD patients were massively impaired in
sentence comprehension with non-canonical non-plausible
sentences, in which only the application of syntactic
movement rules allows for accurate responses. Moreover,
they found significant correlations between the performance
in these sentences and the atrophy of the caudate head, as
measured by the bicaudate ratio on MRI. Yet, the striatum
may also impact lexical aspects of phrasal processing such
as access to grammatical categories (Moro et al., 2001;
Friederici and Kotz, 2003). Indeed, grammatical categories
are stored in the mental lexicon together with the respective
word representation (e.g. cat-noun, eat-verb, the-determiner . . .)
(MacDonald et al., 1994). Furthermore, certain grammatical
categories, such as determiners, define a specific grammatical expectancy by placing rule-based constraints on the
kinds of lexical units that can follow (e.g. ‘determiners’ are
part of noun phrases, and thus, are always followed by
either ‘nouns’ or ‘adjectives’). In a PET study, Moro et al.
(2001) used the inversion of determiners and nouns in
sentences which consisted of pseudo-words so as to
neutralize the access to semantic components while
maintaining function words. They reported activations of
the left caudate head when participants covertly read such
sentences and subsequently made acceptability judgements.
Likewise, in a functional MRI study, Friederici et al. (2003)
reported that violations of the expectancy of grammatical
categories (e.g. ‘the ice cream was in the eaten’) result in
the activation of the left putamen. Similar sentence
materials were also used with patients sustaining striatal
damage during ERP recollection. Several authors showed
that the P600 component, which is hypothesized to index
late stages of grammatical integration, is either absent
(Friederici et al., 2003a; Kotz et al., 2003) or reduced in
1048
Brain (2008), 131, 1046 ^1056
these patients (Friederici et al., 1999). Hence in light of
these studies, it may be hypothesized that the striatum is
involved in both lexical access to stored grammatical
categories and in rule-based expectation of such categories.
Although language studies in morphology and syntax did
not allow for the functional specification of the striatum,
there are some additional data from another cognitive
domain which is also characterized by a dual lexicon/rule
architecture, namely arithmetic. Indeed, arithmetic processing is claimed to be subdivided into a lexical memory
component, containing arithmetic facts such as number
representations or multiplication tables, and a rule device
allowing exact calculation as required for example in
subtractions (e.g. Warrington, 1982; McCloskey and
Caramazza, 1985; McCloskey et al., 1991). Assessing HD
patients’ performance on simple multiplications and
subtractions, Teichmann et al. (2005) reported that these
patients were more impaired in the latter operations,
suggesting predominant impairment on arithmetic rule
application. However, some case studies provided evidence
for the inverse pattern, showing that patients with striatal
damage were unable to handle basic multiplication tables,
whereas they succeeded in solving simple addition and
subtraction problems (Hittmair-Delazer et al., 1994;
Dehaene and Cohen, 1997).
In brief, the lexicon/rule conflict remains unresolved,
despite its investigation in both language and arithmetic.
However, this controversy may be resolved by viewing the
striatum not as a simple anatomical and functional entity,
but rather as a highly complex structure encompassing
various neural circuits which originate from distinct cortical
areas. These circuits project from the striatal input nuclei,
namely the caudate, to the pallidal output nuclei, conveying
information via the thalamus back to the cortical areas
where the information was initiated. As evidenced in
primate models, these circuits are organized along a
segregated architecture, which is dedicated to different
functions, comprising motor and cognitive processes
(Alexander et al., 1986; Hoover and Strick, 1993;
Middleton and Strick, 2000). This multi-channel view has
recently been substantiated in humans by means of
diffusion tensor imaging showing that the striatum is
connected with cortical regions, which are related to motor
functions, executive abilities and, interestingly, to language
processing such as Broca’s area (Lehéricy et al., 2004a, b).
Likewise, there may be also connections with more
posterior language areas given that the primate striatum
is connected with several temporal regions (Middleton
and Strick, 1996). Intriguingly, Broca’s area is thought to
contribute to linguistic rule application (Grodzinsky, 2000;
Tyler et al., 2005; Hagoort, 2005; Santi and Grodzinsky,
2007), whereas posterior temporal cortices are known to
subserve lexical operations such as word retrieval (Damasio,
1992; Goodglass, 1993; Indefrey and Levelt, 2002).
However, the respective striatal sub-regions receiving
input from these cortical language areas have not yet
Marc Teichmann et al.
been investigated. Indeed, the previous language studies
hardly considered the different sub-regions of the striatum,
which was rather viewed as one single anatomical and
functional entity. Hence, the authors did not distinguish
between the different lesion patterns within vascular
patients, which were merely subdivided into disorders of
the putamen and/or of the caudate (Damasio et al., 1982;
Cappa et al., 1983; Wallesch, 1985; Alexander et al., 1987;
Ullman et al., 1997; Friederici et al., 1999; Longworth et al.,
2005). Furthermore, it is difficult to define an exact
anatomical lesion pattern in neuro-degenerative diseases
such as HD and PD. In brief, there are hardly any data
about the functional role of the distinct sub-portions of the
striatum, either with respect to rule application or to lexical
processing. The only available data are those drawn from
two functional imagery studies suggesting that lexical
aspects of phrasal processing may be linked to dorsal
portions either of the putamen (Friederici et al., 2003) or of
the caudate (Moro et al., 2001).
In the present study, we hypothesized that the striatum
comprises distinct language and possibly arithmetic circuits,
which are thought to be anatomically and functionally
segregated. Such circuits and their neural integration in
distinct portions of the striatum may subserve lexical
memory processes on the one hand, and combinatorial rule
application on the other hand. Yet the paucity of data in
this domain did not allow for formulating a specific
prediction with respect to the intra-striatal localization
of such sub-portions. In order to test our hypothesis we
evaluated 31 HD patients with tasks assessing the rule/
lexicon dichotomy in the domain of morphology, syntax
and arithmetic. To ensure that their performance is
specifically linked to rule application and lexical capacities
we compared them to a control group of 20 age-matched
healthy adults from Teichmann et al. (2005). We then
scanned the HD patients with PET using 18FDG and
applying a mask, which allowed for the metabolic
specification of the different striatal portions as well as of
the pallidum. Finally, we ran correlation analyses which
linked the different portions of the striatum with the
lexicon/rule abilities of the patients.
Methods
Participants
Thirty-one patients with HD at an early stage participated in this
study (stages I and II according to the ‘Total Functional Capacity
scale’; Shoulson, 1981). HD patients were recruited from the
out-clinic patients within the follow-up of the Multicentric
Intracerebral Grafting in Huntington’s Disease (MIG-HD) trial
which was approved by the ethics committee of the Henri Mondor
Hospital. All patients were assessed with the rule/lexicon tasks
before grafting within each centre participating in the trial
(Créteil, Toulouse, Bruxelles, Lille, Angers and Nantes). The HD
diagnosis was genetically confirmed (CAG repeats 435). Patients
had no previous neurological or psychiatric history other than
HD. We compared them to 20 healthy volunteers that had already
Language processing within the striatum
Brain (2008), 131, 1046 ^1056
Table 1 Demographic data of HD patients and healthy
controls
N
Sex
Age (years)
Years of education
Laterality
Disease duration (years)
CAG repeats
HD
Controls
31
13F/18M
45.4 9.0
12.5 3.5
27R/4L
5.7 2.6
44.8 4.1
20
14F/6M
46.1 6.6
13.2 4.3
19R/1L
^
^
Table 2 Clinical performance of HD patients (mean values
and standard deviations)
Total functional capacity (TFC)
UHDRS motor score
MDRS
Stroop colour/words
Fluency for PRV in 2 min
Symbol digit code
HD
Normal values
11.0 1.3
35.5 15.0
128.8 6.9
27.4 10.4
40.3 16.1
25.0 9.3
13
0
4136a
435b
445c
437d
MDRS, Mattis dementia rating scale. Normal values are issued
from: aMattis (1976), bGolden (1978), cCardebat et al. (1990),
d
Wechsler (1981).
participated in our previous study using the same rule/lexicon
tasks (Teichmann et al., 2005). Healthy controls had no
neurological or psychiatric disorders and were paired to the HD
patients according to their age and educational level (all Fs 51).
All participants gave informed consent. Demographic data are
summarized in Table 1.
General assessment
All patients were evaluated using the UHDRS (Huntington Study
Group, 1996) and the Mattis Dementia Rating Scale (MDRS;
Mattis, 1976). Data are summarized in Table 2.
Behavioural tasks
Patients were asked to complete three tasks related to morphology,
syntax and arithmetic. The three tasks drawn from morphology,
syntax and arithmetic were previously used with HD patients and
extensively described in Teichmann et al. (2005).
Morphology
We used verb inflection to assess morphological processes. Indeed,
inflection is the primary domain for which online rule-based
decomposition has been argued and most models of word
processing regard inflected words as the only complex forms
for which rule-based decomposition is likely to take place
(e.g. Caramazza et al., 1988; Niemi et al., 1994). Furthermore,
inflection is fully productive and semantically and grammatically
consistent (e.g. to govern—he governs—he governed), whereas
derivation can change the meaning and the grammatical category of
the base word (e.g. to govern—government) which may complicate
the interpretation of experimental results. In a conjugation task we
used high frequency regular and irregular verbs to assess lexical
1049
abilities. Although low-frequency regulars are assumed to depend on
rule application (e.g. Pinker, 1999), high-frequency forms of both
regular and irregular verbs have been shown to be stored in the
mental lexicon (Schreuder and Baayen, 1995; Pinker and Ullman,
2002). Conversely, rule application was assessed through the
conjugation of NV which, by definition, do not have any lexical
representation and thus specifically depend on rule application.
NV were constructed following two types of French conjugation
rules, providing two levels of rule processing: the main rule,
which captures regularities pertaining to verbs ending in ‘-er’
(e.g. arriver—il arrive—il arrivera (to arrive—he arrives—he will
arrive) and sub-rules which capture less-frequent regularities such as
those pertaining to verbs in ‘-ir’ or in ‘-oire’ (e.g. finir—il finit—il
finira (to finish—he finishes—he will finish), croire—il croit—il
croira (to believe—he believes—he will believe). The two types of
NV are respectively referred to as ‘regular NV’ (e.g. ‘garouster’) and
‘subregular NV’ (e.g. ‘saurentir’, ‘olissoire’).
The materials contained 24 regular verbs, 23 irregular verbs, 24
regular NV and 18 sub-regular NV which were to be conjugated in
the present and in the future tense (third person singular). Regular
and irregular verbs were matched for the number of syllables
[F(1,45) = 2.22, P40.1] and phonemes (F51) and for their logtransformed frequencies (F51) according to the LEXIQUE 2
database (New et al., 2004). Furthermore, regular and sub-regular
NV were matched for the number of syllables [F(1,40) = 1.56,
P40.1] and phonemes (F51). All NV consisted of orthographically and phonotactically legal letter strings. The stimuli were
randomized within the two verb and NV conditions. The stimuli
order was the same for each participant.
Syntax
Sentence comprehension was assessed with a sentence–picturematching task using canonical (actives, subject-relatives) and noncanonical structures in which the usual subject–verb–complement
order is inversed (passives, object-relatives). We reasoned that the
processing of non-canonical structures critically depends on the
application of syntactic movement rules, whereas canonical
structures can be processed using word-order information to
reconstruct the thematic roles (first noun = agent, verb = action,
second noun = theme). This kind of word-order mapping was
claimed to involve more basic rules generally referred to as
syntactic formation rules (Jackendoff, 2002). Note that the
processing of both canonical and non-canonical structures
requires intact access to the lexical representations of the words
used in the sentences. Furthermore, in each of the two sentence
types we manipulated the plausibility of the clauses so as to vary
pragmatic factors which may contribute to sentence comprehension. This yielded four types of sentences referred to as canonical
plausible [e.g. ‘La fille arrose la fleur qui est blanche’ (the girl
waters the flower which is white); N = 4], canonical non-plausible
[e.g. ‘La fleur arrose la fille qui est blanche’ (the flower waters the
girl who is white); N = 4], non-canonical plausible [e.g. ‘La fleur
est arroseé par la fille qui est blanche’ (the flower is watered by the
girl who is white); N = 4] and non-canonical non-plausible
sentences [e.g. ‘La fille est arrosée par la fleur qui est blanche’
(the girl is watered by the flower which is white); N = 4]. Each
sentence was paired with one picture that depicted either the
plausible version of the sentence or the non-plausible version,
yielding 32 sentence–picture pairs (Fig. 1). Participants were
asked whether the orally presented sentence and the picture
1050
Brain (2008), 131, 1046 ^1056
Marc Teichmann et al.
The girl waters the flower which is white
(can+pl+)
The flower waters the girl who is white
(can+pl−)
The flower is watered by the girl who is white
(can−pl+)
The girl is watered by the flower which is
white (can−pl−)
Fig. 1 The different sentence types and the plausible and non-plausible pictures they were matched with. can+pl+ = canonical plausible;
can+pl = canonical non-plausible; can pl+ = non-canonical plausible; can pl = non-canonical non-plausible.
were correctly matched. Canonical and non-canonical sentences
were randomized. The presentation order of the sentence–picture
pairs was the same for each participant.
Arithmetic
statistical power, and the threshold of the correlation analysis was
set at P50.01. The striatum was subdivided into three main
components referred to as the ventral striatum and as the dorsal
striatum which itself comprises the caudate nucleus and the
putamen. The distinction between dorsal and ventral regions
is founded on numerous observations that these regions are
anatomically distinct (e.g. Heimer and Wilson, 1975), connect to
different regions of the cortex (Yeterian et al., 1991; Leh et al.,
2007) and serve distinct behavioural functions (e.g. O’Doherty
et al., 2004). This structural and functional distinction may also
have implications for the different aspects of language processing.
Figure 2 shows a 3D reconstruction of the different striatal
portions.
We assessed arithmetic rule use and lexical processing, respectively, with simple subtractions with carry-over (e.g. 12
7 = 5;
N = 20) and multiplications (e.g. 3 7 = 21; N = 20). Fifty percent
of the multiplications and subtractions were correct, whereas the
other 50% contained errors. Multiplications and subtractions were
matched for the number of digits they contained (F51).
Participants were instructed to check whether the visually
presented operations were correct or not. The multiplication
and subtraction problems were randomized. The stimuli order was
the same for each participant.
Behavioural results
PET scanning
Statistical analyses used analyses of variance (ANOVAs) which
were run by participants (F1) and by items (F2).
As described elsewhere (Gaura et al., 2004), PET examinations
were performed with a high-resolution EXACT HR+ tomograph
(CTI/Siemens) using a 3D acquisition. The subject’s head was
maintained using an individually moulded head holder. All studies
were carried out in a quiet, dark environment while the patients
were in a resting state with their eyes closed. Metabolic images
were acquired 30–50 min after intravenous injection of 118–280
MBq of [18F]fluoro-2-deoxy-d-glucose (18FDG).
Image analysis
A detailed description of image analysis has been reported
previously (Gaura et al., 2004). Images were analysed using the
statistical parametric mapping software (SPM99; Welcome
Department of Cognitive Neurology, London, UK, Friston,
1996). Briefly, images were transformed into Talairach’s standard
stereotaxic space (Talairach and Tournoux, 1988). The images
were then smoothed with an 8 8 8 mm3 Gaussian filter
to compensate for inter-subject variability of brain anatomy
(Friston, 1996).
A correlation analysis using a multiple regression model was
performed to investigate the relationships between behavioural
scores and normalized values of metabolic activity in the striatum.
Statistical analyses were restricted to the striatum and the
pallidum, using a mask which excluded other brain regions and
allowing for voxel-by-voxel comparisons in this area of interest
(small volume correction). This mask provides an increase in
Conjugation task
We used ‘accuracy’ as dependent variable and ‘stimulus type’
(irregular verbs, regular verbs, regular NV, sub-regular NV) and
‘groupe’ (HD, controls) as independent variables. Performance
was better for controls (99.07% 2.59 correct) than for patients
[84.70% 21.28 correct; F1(1,49) = 44.15, P50.001; F2(1,85) =
149.17, P50.001]. There was an effect of stimulus type
[F1(3,147) = 73.94, P50.001; F2(3,85) = 97.91, P50.001] and a
significant group stimulus-type interaction [F1(3,147) = 40.44,
P50.001; F2(3,85) = 61.50, P50.001]. This interaction was due to
the fact that controls had similar performance with the four
stimulus types [F1(3,57) = 2.59, P = 0.06; F2(3,85) = 1.96, P40.1],
whereas there were differences for HD patients [irregular verbs
91.16% 14.66 correct, regular verbs 99.19% 2.26 correct,
regular
NV
90.73% 10.19
correct,
sub-regular
NV
57.71% 21.97 correct; F1(3,90) = 71.97, P50.001; F2(3,85) =
86.14, P50.001]. Post hoc analyses showed that HD patients
performed poorer with NV than with verbs [F1 (1,30) = 78.43,
P50.001; F2(1,87) = 36.52, P50.001] suggesting rule disorders.
Indeed, this difference cannot be attributed to the difference
between natural and non-natural stimuli because performance was
similar with regular NV and irregular verbs (F151; F251).
Furthermore, performance differed within NV showing poorer
performance with sub-regular NV than with regular NV
[F1(1,30) = 88.15,
P50.001;
F2(1,40) = 87.19,
P50.001].
Language processing within the striatum
Brain (2008), 131, 1046 ^1056
1051
100
body
90
head
% correct
caudate nucleus
posterior putamen
ventral striatum
can+pl+
80
can+pl−
can−pl+
70
can−pl−
60
anterior putamen
50
Fig. 2 Antero-lateral view of the left striatum comprising the
dorsal striatum (caudate nucleus and putamen) and the ventral
striatum. Three-dimensional reconstruction of the striatum from
Douaud et al. (2006).
controls
Fig. 4 Sentence comprehension in controls and HD patients
with the four sentence types.
100
100
90
regular verbs
irregular verbs
regular NV
sub-regular NV
80
70
% correct
90
% correct
HD
80
Multiplication
70
Subtraction
60
60
50
50
controls
HD
Fig. 3 Conjugation performance in controls and HD patients
with the four stimulus types.
This dissociation confirms that HD patients do not have general
difficulties with non-words but that their performance is related to
specific rule disorders: relative conservation of the main rule but
break-down of sub-rules. Finally, performance with verbs was
poorer in HD than in controls suggesting that HD patients may
also have some lexical disorders [F1(1,49) = 5.55, P = 0.02;
F2(1,46) = 18.20, P50.001]. Results are displayed in Fig. 3.
Sentence^picture matching task
We used ‘accuracy’ as dependent variable and ‘canonicity’
(canonical/non-canonical sentences), ‘plausibility’ (plausible/nonplausible sentences) and ‘groupe’ (HD, controls) as independent
variables. Performance was better for controls (96.56% 5.62
correct) than for patients [81.55% 18.21 correct; F1(1,49) =
43.42, P50.001; F2(1,28) = 16.62, P50.001]. There was an effect
of canonicity [F1(1,49) = 84.29, P50.001; F2(1,28) = 11.13,
P = 0.002] but not of plausibility [F1(1,49) = 3.70, P = 0.06; F251]
and a significant groupe canonicity interaction [F1(1,49) = 45.16,
P50.001; F2(1,28) = 8.23, P50.008]. This interaction was due
to the fact that controls had similar performances with
canonical (97.19% 5.29 correct) and non-canonical sentences
(95.94% 5.93 correct; F151; F251) and with plausible
(97.50% 5.06
correct)
and
non-plausible
sentences
[95.63% 6.04 correct; F1(1,19) = 3.35, P = 0.08, F2(1,28) = 1.64,
P40.1]. In contrast, performance in HD was poorer with noncanonical (70.36% 17.42 correct) than with canonical sentences
(92.74% 10.51 correct) reflecting specific impairment of movement rules [F1(1,30) = 90.64, P50.001, F2(1,28) = 10.01, P = 0.004],
whereas it was similar with plausible (82.86% 16.51 correct) and
non-plausible sentences [80.24% 19.80 correct; F1(1,30) = 1.96,
controls
HD
Fig. 5 Calculation performance in controls and HD patients
with multiplications and subtractions.
P40.1, F251]. Finally, performance with canonical sentences was
poorer in HD than in controls suggesting that HD patients may also
have some disorders with formation rules and/or with lexical access
to the words of the sentences [F1(1,49) = 4.36, P = 0.04;
F2(1,15) = 4.69, P = 0.047]. The results are summarized in Fig. 4.
Arithmetic task
Twenty-seven HD patients performed this task. We used
‘accuracy’ as dependent variable and ‘operation type’ (subtraction,
multiplication) and groupe (HD, controls) as independent
variables. Performance was better for controls (97.87% 2.97
correct) than for patients [89.72% 11.01 correct; F1(1,45) =
12.13, P = 0.001; F2(1,38) = 26.15, P50.001] but there was no
effect of operation type [F1(1,45) = 4.02, P = 0.05; F2(1,38) = 1.09,
P40.1] and no groupe operation interaction [F1(1,45) = 2.25,
P40.1; F251]. Post hoc analyses revealed however that performance in HD tended to be better with multiplications
(91.30% 9.96 correct) than with subtractions (88.15% 11.94
correct) in the analysis by participants [F1(1,26) = 3.75, P = 0.06]
but not in the analysis by items [F2(1,38) = 1.07, P40.1].
In contrast, performance with multiplications (98.00% 2.99
correct) and subtractions (97.75% 3.02 correct) was similar in
controls (F151; F251). Results are displayed in Fig. 5.
PerformanceçPET correlation analyses
Performance of controls was too similar and too homogeneous to
provide any correlation with metabolic activity of the striatum.
Thus correlation analyses were only run with behavioural data of
HD patients which showed the required variability between
1052
Brain (2008), 131, 1046 ^1056
Marc Teichmann et al.
Table 3 Significant correlations between behavioural scores and metabolic PET data in HD (P50.01)
Cognitive domain
Morphology
Regular NV
Sub-regular NV
Irregular verbs
Syntax
can
can pl+
can+pl
Arithmetic
Subtractions
Multiplications
Anatomical region
Lateralization
Talairach coordinates (x, y, z)
Z score
P
Number of voxels
Caudate head
Ventral striatum
Putamen
Caudate head
Ventral striatum
Caudate head
Left
Left
Left
Left
Left
Left
16,
18,
18,
16,
16,
16,
20,
20,
18,
24,
20,
24,
2
8
6
4
8
2
2.65
2.52
2.57
3.72
2.86
2.46
0.004
0.006
0.005
0.000
0.002
0.007
102
18
21
75
12
133
Ventral striatum
Ventral striatum
Ventral striatum
Putamen
Pallidum
Caudate head
Right
Left
Right
Left
Left
Left
16,
16,
14,
28,
26,
16,
4, 10
10, 12
8, 8
10, 4
10, 6
22, 2
2.44
2.51
2.39
2.50
2.69
2.57
0.007
0.006
0.009
0.006
0.004
0.005
53
84
69
44
43
106
Putamen
Pallidum
Ventral striatum
Ventral striatum
Left
Left
Left
Left
30,
26,
18,
18,
12, 6
12, 4
16, 10
14, 8
2.82
2.37
2.39
2.43
0.002
0.009
0.009
0.008
88
27
44
49
One voxel = 8 mm3.
conditions and across patients. The significant results (P50.01)
reflecting positive correlations between low behavioural scores and
low metabolic activity in the different striatal and pallidal areas are
listed in Table 3.
In morphology, performance with both regular NV (main rule)
and sub-regular NV (sub-rule) correlated with the left ventral
striatum and with ventrally situated portions of the left caudate
head. Furthermore, performance on regular NV also involved
ventral regions of the left putamen. In contrast, performance with
irregular verbs (lexicon) correlated with more dorsal portions of
the left caudate head. Dissociations between rule and lexical
processing within the caudate head are shown in Fig. 6. In syntax,
performance with non-canonical sentences (movement rules)
correlated with the ventral striatum bilaterally, with ventrally
situated portions of the left putamen and with the left pallidum.
Conversely, performance with canonical sentences (formation
rules, lexicon) correlated with more dorsally localized regions of
the left caudate head. In arithmetic, performance with subtractions
correlated with the left ventral striatum, with the left putamen and
with the left pallidum, whereas performance with multiplications
only correlated with the left ventral striatum.
Discussion
In this study, we tested the hypothesis that the striatum is
involved in language and arithmetic abilities with respect to
both rule application and lexical processing. Our findings
support this hypothesis, showing that HD patients were
impaired in both processing aspects which correlated with
distinct portions of the striatum.
In morphology, performance in controls was similar with
all four stimulus types, whereas HD patients displayed a
different behavioural pattern. Patients were impaired in
both lexical processes with irregular verbs and rule
application with NV. Yet lexical processing and the
application of the conjugation main rule were only slightly
affected, whereas there was massive impairment of sub-rule
application. The dissociation between regular NV and subregular NV furthermore shows that the processing deficits
in HD cannot be attributed to the non-natural character of
NV but that they are related to the different rule processes
captured by the different NV types. The correlation data
showed that performance on lexical processing and on rule
application were dissociated, involving dorsal portions of
the caudate for lexical operations and more ventral portions
of the caudate and the putamen for rule application. In
syntax, performance in controls was similar for the different
sentence types, whereas HD patients were specifically
impaired with non-canonical sentences suggesting damage
of syntactic movement rules. In contrast, performance was
largely preserved with canonical sentences, suggesting that
formation rules and access to lexical information are only
slightly hampered in HD. Importantly, performance was
not influenced by the plausibility status of the sentences,
indicating that sentence processing was independent of
pragmatic factors and solely depended on language-related
processes. As in morphology, correlation data revealed a
ventral-dorsal dissociation. Syntactic movement rules correlated with ventrally situated portions of the striatum and
with the pallidum, whereas lexical processing and syntactic
formation rules correlated with regions which are more
dorsally located, such as the left caudate head. Yet,
intriguingly, we did not find a significant correlation with
non-canonical non-plausible sentences, although the processing of such sentences is assumed to rely exclusively on
the application of movement rules. Likewise, this does not
fit with previous findings of ours showing that the
performance in such sentences correlates with the atrophy
Language processing within the striatum
Brain (2008), 131, 1046 ^1056
1053
Irregular verbs
Sub-regular NV
z scores
x = −16, y = 24, z = −4
z scores
x = −16, y = 24, z = 2
Fig. 6 Anatomical dissociation between rule and lexical processing in the caudate head: correlation analyses with sub-rule application
(sub-regular NV) yielded a ventrally situated cluster of the caudate, whereas lexical processing (irregular verbs) involved more dorsal
portions localized in the same axial plane. This dissociation is reflected by the differences between z-coordinates while x/y-coordinates
are identical in the two conditions.
of the caudate head as measured by the bicaudate ratio on
MRI (Teichmann et al., 2005). Yet a plausible explanation
could be that the HD patients of the present study were
quite homogeneously impaired in language processing,
which may have prevented certain correlations by yielding
floor effects. Nonetheless, the correlation data from both
the domains of word morphology and sentence processing
reveal a rule/lexicon dissociation within the striatum,
showing that rule application with NV and non-canonical
clauses correlate with ventrally situated portions of the
striatum, whereas lexical processes, as assessed with
irregular verbs and canonical clauses, seem to involve
more dorsal sub-regions. Finally, in arithmetic, we found
processing impairments in both multiplications and subtractions with a tendency towards better performance in
multiplications. Furthermore, as in language, arithmetic
performance was also dissociated, since the performance in
multiplications correlated with the left ventral striatum
only, whereas the performance in subtractions correlated
with the left ventral striatum but also with the left putamen
and pallidum.
In sum, the behavioural results of the present study are
coherent with the data of our previous study (Teichmann
et al., 2005) reflecting predominant impairment on
language rule application as compared to lexical processing
aspects. Thus, at first view, our findings favour the claim
that the striatum holds a role in language processing
pertaining to linguistic rule application (e.g. Ullman, 2001).
However, we also found evidence of slight impairment on
lexical operations in agreement with concurrent proposals
reporting that damage to the striatum hampers lexical
processing aspects such as word retrieval and the inhibition
of inappropriate items (e.g. Crosson, 1985; Wallesch and
Papagno, 1988; Copland, 2003; Longworth et al., 2005).
Indeed, the lexicon/rule conflict remained unresolved
because the previous findings were exclusively based on
behavioural data from patient studies comprising lesion
patterns which may have involved different sub-portions of
the striatum and which may have extended beyond striatal
structures. In HD, for example, neural degeneration spreads
within the striatum and progressively invades cortical
regions (Vonsattel et al., 1985; Andrews et al., 1999).
Likewise, vascular lesions affect various portions of the
striatum and are rarely restricted to the caudate or the
putamen. The present study, correlating metabolic data of
the striatum and rule/lexicon scores, provides more direct
evidence for the linguistic involvement of striatal structures.
In accordance with the wide consensus that language
abilities are homed in the left hemisphere, we showed that
primarily left-lateralized portions of the striatum correlate
with the processing of morphological and syntactic
information. This finding suggests that the striatum impacts
on language via its implication in language-specific
processes (Ullman, 2001), whereas it appears to be less
compatible with concurrent claims positing that the
striatum modulates language through more general operations such as executive functioning and/or working memory
(Lieberman et al., 1992; Grossman et al., 1993, 2000, 2002).
Our correlation data further specify this linguistic function,
showing that distinct left striatal structures are involved in
rule application and lexical processing, relating respectively
to ventral portions and to more dorsal regions of the
striatum. Finally, the computational function of the left
striatum seems to apply also to arithmetic aspects of lexical
and rule processing. Interestingly, the left-lateralization of
exact calculation is supported by several studies using
functional imagery with healthy participants. Dehaene et al.
(1999) reported that exact arithmetic, which depends on
1054
Brain (2008), 131, 1046 ^1056
calculation rules, leads to left-lateralized activation in the
inferior frontal cortex, whereas approximation yielded
bilateral activation in the parietal lobes. Moreover,
Stanescu-Cosson et al. (2000) showed that exact arithmetic
involves in addition to cortical areas the left putamen.
Taken together, our results suggest that the left striatum
holds both lexical and rule devices which seem to be
dedicated to language and arithmetic. Furthermore, lexical
processes and rule application appear to be implemented by
distinct striatal circuits encompassing distinct sub-portions
of the striatum which presumably hold distinct neural and
computational properties. This finding is crucial in that it
contributes to resolve the lexicon/rule conflict by conciliating two extreme views of striatal functioning. Indeed, a
2-fold lexicon/rule view allows us to account for a number of
apparently conflicting findings. Damage to lexicon-related
portions of the striatum may induce naming impairments
(e.g. Damasio et al., 1982; Cappa et al., 1983; Wallesch, 1985;
Alexander et al., 1987; Frank et al., 1996), the production of
paraphasias (e.g. Cambier et al., 1979; Damasio et al., 1982),
difficulties to suppress infrequent meanings of homophones
in semantic priming (Copland, 2003) and deficient handling
of multiplications (Dehaene and Cohen, 1997). Conversely,
damage to rule-related portions may result in conjugation
disorders of regular verbs and non-verbs (Ullman et al., 1997;
Teichmann et al., 2005, 2006), incorrect comprehension
of non-canonical sentences (Natsopoulos et al., 1993;
McNamara et al., 1996; Kemmerer, 1999; Grossman et al.,
2000, 2002; Teichmann et al., 2005), and difficulties to
manipulate subtractions (Teichmann et al., 2005). In fact,
vascular damage, neural degeneration in HD and dopaminerelated dysfunction in PD may affect distinct sub-portions
of the striatum. The present study helps to separate these
sub-portions and furthermore suggests a specific localization
pattern. In particular, our correlation data indicate that
ventrally situated portions of the striatum may be involved
in linguistic rule aspects, whereas more dorsal portions of
the caudate and the putamen may impact on lexical
processing. This ventral/dorsal dissociation does not hold
for arithmetic suggesting that arithmetic capacities involve
distinct circuits of the striatum. This is also coherent
with modular theories of cognition claiming that different
cognitive domains are functionally and anatomically
separated (Fodor, 1983). Yet, like for language, the striatum
seems to implement both rule and lexical processing aspects
in distinct sub-portions which, for arithmetic, involve
ventral regions of the striatum as well as the putamen
and the pallidum. The behavioural pattern of our HD
patients provides some additional insight in the functional
localization of language processes. Indeed, the intra-striatal
lesion pattern in HD has been extensively investigated.
Unfortunately, the data are conflicting. On one hand,
neuropathological studies (Vonsattel et al., 1985; Vonsattel
and DiFiglia, 1998) and several investigations using
imaging-based morphometry (Kassubek et al., 2004;
Douaud et al., 2006) reported that neural degeneration
Marc Teichmann et al.
follows a dorso-ventral and a caudo-rostral, medio-lateral
gradient (Vonsattel et al., 1985; Vonsattel and DiFiglia,
1998; Douaud et al., 2006). On the other hand, Thieben
et al. (2002), using imaging-based morphometry with
non-symptomatic gene careers, showed the inverse lesion
pattern, namely a ventro-dorsal gradient. Interestingly, the
latter study provides evidence for the initial lesion
distribution in HD, suggesting that neural degeneration
specifically originates in ventrally situated portions regarding the left striatum. Our behavioural pattern in early stages
of HD, showing mild impairment on lexical processing but
massive disorders with rule application, fit with both the
lesion gradient reported by Thieben et al. (2002) and our
correlation data. Indeed, neural degeneration in HD seems
to originate in the left ventral striatum, leading to prominent
rule disorders which coherently correlate with the metabolic
decline of ventrally and left-situated portions of the striatum.
In contrast, lexical processes are lesser affected and correlate
with more dorsal portions of the striatum. However, some
caution is warranted when considering such attempts of
functional localization. Indeed, as mentioned before, the
morphometric data, reflecting the intra-striatal lesion pattern
in HD, are far from being clear-cut. Furthermore, the
relatively small number of participants and the existence of
some structural overlap in the present correlations make it
difficult to clarify the exact functional localizations. Finally,
localizing rule and lexical devises in the striatum necessarily
remains speculative because of the insufficient number
of studies in this domain. We provided the first study
investigating such functional specifications within striatal
structures, and corroborative arguments are still needed.
Conclusion
Here we provided novel functional and structural evidence
showing that the striatum is involved in both rule and
lexical processing in language as well as in arithmetic. Our
findings suggest that the respective neural substrates for
rule computation and lexical operations involve distinct
striatal sub-portions which may encompass distinct corticostriatal circuits. However, despite the anatomical distinctness, the individualization of discrete functional circuits
remains speculative. Our data nonetheless suggest that
ventral portions of the striatum may subserve linguistic rule
computations, whereas more dorsally situated portions
underpin lexical operations. Such sub-portions of the
striatum are supposed to receive input from cortical areas
which have been incriminated in rule application and
lexical processing such as Broca’s area (e.g. Grodzinsky,
2000; Tyler et al., 2005; Hagoort, 2005) and posterior
temporal cortices, respectively (e.g. Goodglass, 1993;
Indefrey and Levelt, 2002). Further work is needed to
confirm these latter assumptions and to clarify the
functional segregation within the cortico-striatal circuitry,
which presumably accounts for its involvement in various
cognitive processing aspects and domains.
Language processing within the striatum
Acknowledgements
The results of this work have been obtained through an
ancillary study of the MIG-HD clinical trial (Multicentric
Intracerebral Grafting in Huntington’s Disease, PHRC grant
NCT00190450) sponsored by the ‘Assistance Publique/
Hôptitaux de Paris’. The study was conducted with
the help of an Avenir Grant (2001) allocated to AC
Bachoud-Lévi by the INSERM, an Assistant Hospitalier de
Recherche Grant (AHR AP-HP/INSERM) awarded to Marc
Teichmann, and a Gis-Maladies rares grant (A04159JS). We
would like to thank Marie-Françoise Boissé for assessing
MDRS scores and cognitive scales of the UHDRS, and
Guillaume Dolbeau and Amandine Rialland for transmitting the data. Furthermore we wish to thank the Center of
Clinical Investigations (CIC) as well as the Huntington
French-Speaking Group (Réseau Huntington de Langue
Française) which supports this research. Thank you also to
Claire Sanderson and Laura Robotham for the English
correction of the manuscript.
References
Alexander GE, DeLong MR, Strick PL. Parallel organization of functionally
segregated circuits linking basal ganglia and cortex. Annu Rev Neurosci
1986; 9: 357–81.
Alexander MP, Naeser MA, Palumbo C. Correlations of subcortical CT
lesion sites and aphasia profiles. Brain 1987; 110: 961–91.
Andrews TC, Weeks RA, Turjanski N, Gunn RN, Watkins LH, Sahakian B,
et al. Huntington’s disease progression. PET and clinical observations.
Brain 1999; 122 (12): 2353–63.
Cambier J, Elghozi D, Strube E. Haemorrhage of the head of the left
caudate nucleus: disorganization of speech and graphic expression, and
disturbances in gestures (author’s translation). Rev Neurol 1979; 135:
763–74.
Cappa SF, Cavallotti G, Guidotti M, Papagno C, Vignolo LA. Subcortical
aphasia: two clinical-CT scan correlation studies. Cortex 1983; 19:
227–41.
Caramazza A, Laudanna A, Romani C. Lexical access and inflectional
morphology. Cognition 1988; 28: 297–332.
Cardebat D, Doyon B, Puel M, Goulet P, Joanette Y. Formal and semantic
lexical evocation in normal subjects. Performance and dynamics of
production as a function of sex, age and educational level. Acta Neurol
Belg 1990; 90: 207–17.
Chomsky N. Aspects of the theory of syntax. Cambridge, MA: MIT Press;
1965.
Chomsky N. On Wh movement. In: Culicover P, Wasow T, Akmajian A.,
editors. Formal Syntax. New York: Academic Press; 1977.
Chomsky N. Barriers. Cambridge MA: MIT Press; 1986.
Copland D. The basal ganglia and semantic engagement: potential insights
from semantic priming in individuals with subcortical vascular lesions,
Parkinson’s disease, and cortical lesions. J Int Neuropsychol Soc 2003; 9:
1041–52.
Crosson B. Subcortical functions in language: a working model. Brain
Lang 1985; 25: 257–92.
Dehaene S, Cohen L. Cerebral pathways for calculation: double dissociation between rote verbal and quantitative knowledge of arithmetic.
Cortex 1997; 33: 219–50.
Dehaene S, Spelke E, Pinel P, Stanescu R, Tsivkin S. Sources of
mathematical thinking: behavioral and brain-imaging evidence. Science
1999; 284: 970–4.
Damasio AR, Damasio H, Rizzo M, Varney N, Gersh F. Aphasia with nonhemorrhagic lesions in the basal ganglia and internal capsule.
Arch Neurol 1982; 39: 15–24.
Brain (2008), 131, 1046 ^1056
1055
Damasio AR. Aphasia. N Engl J Med 1992; 326: 531–9.
Démonet J-F, Puel M, Celsis P, Cardebat D. Subcortical aphasia: some
proposed pathophysiological mechanisms and their rCBF correlates
revealed by SPECT. J Neurolinguistics 1991; 6: 319–44.
Douaud G, Gaura V, Ribeiro M-J, Lethimonnier F, Maroy R, Verny C,
et al. Distribution of grey matter atrophy in Huntington’s disease
patients: a combined ROI-based and voxel-based morphometric study.
Neuroimage 2006; 32: 1562–75.
Fodor JA. The modularity of mind. Cambridge, MA: MIT Press; 1983.
Frank EM, McDade HL, Scott WK. Naming in dementia secondary to
Parkinson’s, Huntington’s, and Alzheimer’s diseases. J Commun Disord
1996; 29: 183–97.
Friederici AD, von Cramon DY, Kotz SA. Language related brain
potentials in patients with cortical and subcortical left hemisphere
lesions. Brain 1999; 122: 1033–47.
Friederici A, Kotz SA. The brain basis of syntactic processes: functional
imaging and lesion studies. Neuroimage 2003; 20 (Suppl 1): 8–17.
Friederici AD, Kotz SA, Werheid K, Hein G, von Cramon DY. Syntactic
comprehension in Parkinson’s disease: investigating early automatic and
late integrational processes using event-related brain potentials.
Neuropsychology 2003a; 17: 133–42.
Friederici A, Ruschemeyer SA, Hahne A, Fiebach CJ. The role of left inferior
frontal and superior temporal cortex in sentence comprehension:
localising syntactic and semantic processes. Cereb Cortex 2003; 13: 70–7.
Friston KJ. Statistical parametric mapping and other analyses of functional
imaging data. In: Toga AW, Mazziotta JC, editors. Brain mapping. The
methods. San Diego: Academic Press; 1996. pp. 363–86.
Gaura V, Bachoud-Lévi A-C, Ribeiro MJ, Nguyen J-P, Frouin V, Baudic S,
et al. Striatal neural grafting improves cortical metabolism in
Huntington’s disease patients. Brain 2004; 127: 65–72.
Golden CJ. Stroop color and word test. Chicago: Stoelting; 1978.
Goodglass H. Understanding aphasia. San Diego: Academic Press; 1993.
Grodzinsky Y. The neurology of syntax: language use without Broca’s area.
Behav Brain Sci 2000; 23: 1–71.
Grossman M, Carvell S, Gollomp S, Stern MB, Reivich M, Morrison D,
et al. Cognitive and physiological substrates of impaired sentence
processing in Parkinson’s disease. J Cogn Neurosci 1993; 5: 480–98.
Grossman M, Kalmanson J, Bernhardt N, Morris J, Stern MB, Hurtig HI.
Cognitive resource limitations during sentence comprehension in
Parkinson’s disease. Brain Lang 2000; 73: 1–16.
Grossman M, Lee C, Morris J, Stern MB, Hurtig HI. Assessing resource
demands during sentence processing in Parkinson’s disease. Brain Lang
2002; 80: 603–16.
Hagoort P. On Broca, brain, and binding: a new framework. Trends Cogn
Sci 2005; 9: 416–23.
Heimer L, Wilson RD. The subcortical projections of the allocortex:
similarities in the neural associations of the hippocampus, the piriform
cortex, and the neocortex. In: Santini M, editor. Perspectives in
neurobiology. Golgi Centennial Symposium. New York, NY: Raven
Press; 1975. pp. 177–93.
Hittmair-Delazer M, Semenza C, Denes G. Concepts and facts in
calculation. Brain 1994; 117: 715–28.
Hoover JE, Strick PL. Multiple output channels in the basal ganglia.
Science 1993; 259: 819–21.
Huntington Study Group. Unified Huntington’s disease rating scale:
reliability and consistency. Mov Disord 1996; 11: 136–42.
Illes J. Neurolinguistic features of spontaneous language production
dissociate three forms of neurodegenerative disease: Alzheimer’s,
Huntington’s, and Parkinson’s. Brain Lang 1989; 37: 628–42.
Indefrey P, Levelt WJM. The spatial and temporal signatures of word
production components. Cognition 2004; 92: 101–44.
Jackendoff R. Foundations of language: brain, meaning, grammar,
evolution. New York: Oxford University Press; 2002.
Kassubek J, Juengling FD, Kioschies T, Henkel K, Karitzky J, Kramer B,
et al. Topography of cerebral atrophy in early Huntington’s disease: a
voxel based morphometric MRI study. J Neurol Neurosurg Psychiatry
2004; 75: 213–20.
1056
Brain (2008), 131, 1046 ^1056
Kemmerer D. Impaired comprehension of raising-to-subject constructions
in Parkinson’s disease. Brain Lang 1999; 66: 311–28.
Kotz SA, Frisch S, von Cramon DY, Friederici AD. Syntactic language
processing: ERP lesion data on the role of the basal ganglia. J Int
Neuropsychol Soc 2003; 9: 1053–60.
Leh SE, Ptito A, Chakravarty MM, Strafella AP. Fronto-striatal connections
in the human brain: a probabilistic diffusion tractography study.
Neurosci Lett 2007; 419: 113–8.
Lehéricy S, Ducros M, Van de Moortele PF, Francois C, Thivard L,
Poupon C, et al. Diffusion tensor fiber tracking shows distinct
corticostriatal circuits in humans. Ann Neurol 2004a; 55: 522–9.
Lehéricy S, Ducros M, Krainik A, Francois C, Van de Moortele P-F,
Ugurbil K, et al. 3-D diffusion tensor axonal tracking shows distinct
SMA and Pre-SMA projections to the human striatum. Cereb Cortex
2004b; 14: 1302–9.
Lieberman P, Kako E, Friedman J, Tajchman G, Feldman LS, Jiminez EB.
Speech production, syntax comprehension, and cognitive deficits in
Parkinson’s disease. Brain Lang 1992; 43: 169–89.
Longworth CE, Keenan SE, Barker RA, Marslen-Wilson WD, Tyler LK.
The basal ganglia and rule-governed language use: evidence from
vascular and degenerative conditions. Brain 2005; 128: 584–96.
Macdonald MC, Pearlmutter NJ, Seidenberg MS. Lexical nature of
syntactic ambiguity resolution. Psychol Rev 1994; 101: 676–703.
Mattis S. Mental status examination for organic mental syndrome in
elderly patients. In: Bellak L, Karasu TB, editors. Geriatric psychiatry.
New York: Grune & Straton; 1976. p. 77–21.
McCloskey M, Caramazza A. Cognitive mechanisms in number processing and calculation: evidence from dyscalculia. Brain Cogn 1985; 4:
171–96.
McCloskey M, Alimosa D, Sokol SM. Facts, rules, and procedures in
normal calculation: evidence from multiple single-patient studies of
impaired arithmetic fact retrieval. Brain Cogn 1991; 17: 154–203.
McNamara P, Krueger M, O’Quin K, Clark J, Durso R. Grammaticality
judgments and sentence comprehension in Parkinson’s disease: a
comparison with Broca’s aphasia. Int J Neurosci 1996; 86: 151–66.
Middleton FA, Strick PL. The temporal lobe is a target of output from the
basal ganglia. Proc Natl Acad Sci USA 1996; 93: 8683–7.
Middleton FA, Strick PL. Basal ganglia and cerebellar loops: motor and
cognitive circuits. Brain Res Brain Res Rev 2000; 31: 236–50.
Moro A, Tettamanti M, Perani D, Donati C, Cappa SF, Fazio F. Syntax
and the brain: disentangling grammar by selective anomalies.
Neuroimage 2001; 13: 110–8.
Natsopoulos D, Grouios G, Bostantzopoulou S, Mentenopoulos G,
Katsarou Z, Logothetis J. Algorithmic and heuristic strategies in
comprehension of complement clauses by patients with Parkinson’s
disease. Neuropsychologia 1993; 31: 951–64.
New B, Pallier C, Brysbaert M, Ferrand L. Lexique 2: a new French lexical
database. Behav Res Methods Instrum Comput 2004; 36: 516–24.
Niemi J, Laine M, Tuominen J. Cognitive morphology in Finnish:
foundations of a new model. Lang Cogn Process 1994; 9: 423–46.
O’Doherty J, Dayan P, Schultz J, Deichmann R, Friston K, Dolan RJ.
Dissociable roles of ventral and dorsal striatum in instrumental
conditioning. Science 2004; 304: 452–54.
Pinker S. Words and rules: the ingredients of language. New York: Basic
Books; 1999.
Pinker S, Ullman M. The past and future of the past tense. Trends Cogn
Sci 2002; 6: 456–63.
Peschanski M, Cesaro P, Hantraye P. Rationale for intrastriatal grafting of
striatal neuroblasts in patients with Huntington’s disease. Neuroscience
1995; 68: 273–85.
Marc Teichmann et al.
Puel M, Demonet J-F, Cardebat D, Bonafé A, Gazounaud Y,
Guiraud-Chaumeil B, et al. Subcortical aphasia. Neurolinguistic and
X-ray computed tomography studies of 25 cases. Rev Neurol 1984; 140:
695–710.
Santi A, Grodzinsky Y. Working memory and syntax interact in Broca’s
area. Neuroimage 2007; 37: 8–17.
Schreuder R, Baayen H. Modelling morphological processing. In:
Feldman L, editor. Morphological aspects of language processing.
Hillsdale, NJ: Erlbaum Press; 1995. p. 131–54.
Shoulson I. Huntington disease: functional capacities in patients
treated with neuroleptic and antidepressant drugs. Neurology 1981;
31: 1333–5.
Stanescu-Cosson R, Pinel P, van der Moortele PF, Le Bihan D, Cohen L,
Dehaene S. Understanding dissociations in dyscalculia. A brain
imaging study of the impact of number size on the cerebral
networks for exact and approximate calculation. Brain 2000; 123:
2240–55.
Talairach J, Tournoux P. Co-planar stereotaxic atlas of the human brain.
Stuttgart: Thieme; 1988.
Teichmann M, Dupoux E, Kouider S, Bachoud-Lévi A-C. The role of the
striatum in processing language rules: evidence from word perception in
Huntington’s disease. J Cogn Neurosci 2006; 18: 1–15.
Teichmann M, Dupoux E, Kouider S, Brugières P, Boissé M-F, Baudic S,
et al. The role of the striatum in rule application. The model of
Huntington’s disease at early stage. Brain 2005; 128: 1155–67.
Tettamanti M, Moro A, Messa C, Moresco RM, Rizzo G, Carpinelli A,
et al. Basal ganglia and language: phonology modulates dopaminergic
release. Neuroreport 2005; 16: 397–401.
Thieben MJ, Duggins AJ, Good CD, Gomes L, Mahant N, Richards F,
et al. The distribution of structural neuropathology in pre-clinical
Huntington’s disease. Brain 2002; 125: 1815–28.
Tyler LK, Marslen-Wilson WD, Stamatakis EA. Differentiating lexical
form, meaning, and structure in the neural language system. Proc Natl
Acad Sci USA 2005; 102: 8375–80.
Ullman MT, Corkin S, Coppola M, Hickok G, Growdon JH,
Koroshetz WJ, et al. A neural dissociation within language: evidence
that the mental dictionary is part of declarative memory, and that
grammatical rules are processed by the procedural system. J Cogn
Neurosci 1997; 9: 266–76.
Ullman MT. A neurocognitive perspective on language: the declarative/
procedural model. Nat Rev Neurosci 2001; 2: 717–26.
Vannest J, Polk TA, Lewis RL. Dual-route processing of complex words:
new fMRI evidence from derivational suffixation. Cogn Affect Behav
Neurosci 2005; 5: 67–76.
Vonsattel JP, Myers RH, Stevens TJ, Ferrante RJ, Bird ED,
Richardson EP Jr. Neuropathological classification of Huntington’s
disease. J Neuropathol Exp Neurol 1985; 44: 559–77.
Vonsattel JP, DiFiglia M. Huntington disease. J Neuropathol Exp Neurol
1998; 57: 369–84.
Yeterian EH, Pandya DN. Prefrontostriatal connections in relation to
cortical architectonic organization in rhesus monkeys. J Comp Neurol
1991; 312: 43–67.
Wallesch CW. Two syndromes of aphasia occurring with ischemic lesions
involving the left basal ganglia. Brain Lang 1985; 25: 357–61.
Wallesch CW, Papagno C. Subcortical aphasia. In: Rose FC, Whurr R,
Wyke M, editors. Aphasia. London: Whurr Publishers; 1988. pp. 256–87.
Warrington EK. The fractionation of arithmetical skills: a single case study.
Q J Exp Psychol 1982; 34: 31–51.
Wechsler D. Wechsler adult intelligence scale-revised manual. New York:
Psychological Corporation; 1981.