Brain Sci. 2013, 3, 1198-1214; doi:10.3390/brainsci3031198
OPEN ACCESS
brain sciences
ISSN 2076-3425
www.mdpi.com/journal/brainsci/
Article
Neural Correlates of Processing Passive Sentences
Jennifer E. Mack 1, Aya Meltzer-Asscher 2,3, Elena Barbieri 1,4 and Cynthia K. Thompson 1,4,5,*
1
2
3
4
5
Department of Communication Sciences and Disorders, Center for the Neurobiology of Language,
Northwestern University, Francis Searle Building, 2240 Campus Drive, Evanston, IL 60208, USA;
E-Mails: jennifer-mack-0@northwestern.edu (J.E.M.); elena.barbieri83@gmail.com (E.B.)
Department of Linguistics, Tel Aviv University, Webb Building, Ramat Aviv, Tel Aviv 69978,
Israel; E-Mail: ameltzer@post.tau.ac.il
Sagol School of Neuroscience, Webb Building, Tel Aviv University, Ramat Aviv,
Tel Aviv 69978, Israel
Cognitive Neurology and Alzheimer’s Disease Center, Feinberg School of Medicine, Northwestern
University, 320 E. Superior, Searle 11-453, Chicago, IL 60611, USA
Department of Neurology, Feinberg School of Medicine, Northwestern University, Abbott Hall,
11th Floor, 710 North Lake Shore Drive, Chicago, IL 60611, USA
* Author to whom correspondence should be addressed; E-Mail: ckthom@northwestern.edu;
Tel.: +1-847-467-7591; Fax: +1-847-467-7377.
Received: 27 May 2013; in revised form: 16 July 2013 / Accepted: 19 July 2013 /
Published: 2 August 2013
Abstract: Previous research has shown that comprehension of complex sentences
involving wh-movement (e.g., object-relative clauses) elicits activation in the left inferior
frontal gyrus (IFG) and left posterior temporal cortex. However, relatively little is known
about the neural correlates of processing passive sentences, which differ from other complex
sentences in terms of representation (i.e., noun phrase (NP)-movement) and processing
(i.e., the time course of syntactic reanalysis). In the present study, 27 adults (14 younger
and 13 older) listened to passive and active sentences and performed a sentence-picture
verification task using functional Magnetic Resonance Imaging (fMRI). Passive sentences,
relative to active sentences, elicited greater activation in bilateral IFG and left
temporo-occipital regions. Participant age did not significantly affect patterns of activation.
Consistent with previous research, activation in left temporo-occipital cortex likely reflects
thematic reanalysis processes, whereas, activation in the left IFG supports processing of
complex syntax (i.e., NP-movement). Right IFG activation may reflect syntactic reanalysis
processing demands associated with the sentence-picture verification task.
Brain Sci. 2013, 3
1199
Keywords: fMRI; sentence processing; syntactic processing; thematic processing
1. Introduction
A considerable body of research has investigated the neural basis of sentence comprehension by
comparing complex to simple sentences. One dimension of sentence complexity is the presence vs.
absence of wh-movement, which influences the mapping of verb arguments onto the surface
representation of sentences. Compare a simple active sentence (1a) to an object wh-question (1b):
1. a. The boy is hugging the girl.
b. Whoi is the boy hugging ti?
c. The girli was hugged ti by the boy.
Both sentences require the listener to build a syntactic structure that guides the integration of the
verb hug with its arguments (i.e., event participants that receive a thematic role from the verb): the
agent (performer of the action) and the theme (undergoer of the action). However, these processes are
more difficult for (1b) than (1a). Some linguistic theories (e.g., Government and Binding Theory; [1])
claim that in object wh-questions, the theme argument originates in the post-verbal position, as it does
in its active counterpart, and is displaced to the beginning of the sentence, leaving behind a “trace” of
the movement operation, co-indexed with the moved wh-word. As a result, object wh-questions have a
noncanonical verb-argument structure in which the theme precedes the agent, in contrast with
canonical active sentences in which the agent is mapped to the subject position, preceding the theme,
which is mapped to the object position. Thus, object wh-questions are more complex than simple
active sentences. Previous psycholinguistic studies have found that the moved wh-word (the filler) is
reactivated immediately after the verb (the gap, i.e., the hypothesized trace site), lending support to the
psychological reality of wh-movement [2–7]. For example, using a visual world paradigm,
wh-movement structures (including object-extracted wh-questions and object cleft structures) elicit
automatic eye movements to the filler at the gap site in healthy adults and listeners with aphasia [2,3].
Additional evidence comes from event-related potential (ERP) studies [8], indicating that wh-embedded
questions—as compared to whether-questions that do not entail gap-filling—elicit a positivity at the
gap position around 400–700 ms.
Passive sentences (1c) also are noncanonical and, hence, are more complex than canonical, active
forms (1a). In passive sentences, the theme is the grammatical subject, and the agent is mapped to an
adjunct prepositional phrase. According to some theorists, passive sentences also involve syntactic
movement [1], i.e., NP-movement (due to the movement of a noun phrase), in contrast with
wh-movement, in which a wh-word is displaced. In movement-based accounts of passive sentences,
the theme originates post-verbally in the direct object position and moves to the grammatical subject
position. However, in other linguistic theories (e.g., Head-Driven Phrase Structure Grammar [9] and
Lexical-Functional Grammar [10]), passive sentences do not involve movement; instead, passive
sentences are distinguished from actives solely on the basis of lexical/thematic structure.
These accounts make different predictions about the processing of passive sentences.
Movement-based accounts predict that identification of a passive structure during language
Brain Sci. 2013, 3
1200
comprehension triggers syntactic reanalysis, in which a trace is constructed and co-indexed with the
grammatical subject, whereas lexical/thematic accounts do not. Psycholinguistic studies have found
that passive sentences elicit delayed or absent reactivation of the filler at the hypothesized gap site, in
contrast with wh-movement structures. One cross-modal priming study [11] found evidence of delayed
reactivation of the grammatical subject, with effects reaching significance only 1000 ms after the verb.
Similarly, delayed reactivation effects have been reported for unaccusative structures (e.g., The leafi
fellti) that are also hypothesized to involve NP-movement [12–14]. These findings are consistent with
the claim that syntactic reanalysis does take place in structures with NP-movement, though on a
slowed time course. Other studies, however, have failed to show gap-filling effects for passive
structures with either healthy or aphasic listeners either at the gap site or downstream from it [3]. Thus,
it remains an open question whether passive sentences require syntactic reanalysis.
Both accounts predict that passive sentences, like other noncanonical sentences, should trigger a
process of thematic reanalysis, i.e., revision of an initial mapping of thematic roles. Psycholinguistic
studies have shown that healthy listeners tend to interpret sentence-initial noun phrases as agents
unless there are additional cues to thematic mapping in the linguistic representation, such as
case-marking [15], or in the discourse context [16]. This “agent-first bias” has been demonstrated
through visual-world eye tracking studies in which listeners tend to direct initial looks to scenes in
which the first noun phrase is the agent ([15,17–20], although cf. [21]). In the case of active
(canonical) sentences, the agent-first bias corresponds to the true structure of the sentence, and no
reanalysis is required. However, passive (noncanonical) sentences require thematic reanalysis as soon
as the structure is identified as passive (in English, upon encountering the past participial morphology
characteristic of the passive voice). At this point, the thematic role assignment of the first noun phrase
must be revised from agent to theme. Some studies have found longer reaction times for passive as
compared to active sentences [22,23], which may be due to the processing costs of thematic reanalysis
and/or syntactic reanalysis.
Studies examining the neural mechanisms of sentence processing have primarily focused on
wh-movement structures by comparing neural activation patterns elicited by complex versus simple
sentences, such as object and subject relative structures [24–28]. These studies consistently report
activation in left inferior frontal gyrus (IFG) (see [29], for review), with some researchers suggesting
that this activation reflects syntactic movement operations [30–34]. Although few neuroimaging studies
examining NP-movement structures have been reported, all studies contrasting passive and active
sentences have also found activation in the left IFG ([26,35–38]; cf. [39], who report activation in the
left frontal operculum). Shetreet, Friedmann and Hadar [34] also found left IFG activation in an fMRI
study of Hebrew unaccusative sentences, which like English passive structures, involve
NP-movement (but, see [40], who did not find IFG activation for a different NP-movement structure in
German). Notably, the majority of passive sentence processing studies have been conducted with
Japanese-speaking participants and, as pointed out by Yokoyama et al. [37], in Japanese, passive verbs
are morphologically marked (i.e., by the morpheme rare), whereas active verbs are uninflected. Thus,
the IFG activation for passive sentences may be attributable to morphological complexity, rather than
movement operations. Japanese also is a verb-final language and, therefore, the thematic and syntactic
reanalysis processes take place after both noun phrases are presented (see [38] for discussion). In
Brain Sci. 2013, 3
1201
addition, Japanese has three types of passive sentences [41]. For these reason, it is an open question
whether English passive sentences elicit patterns like those for Japanese passives.
Several neuroimaging studies of complex wh-movement sentences also have reported left posterior
perisylvian activation; however, fewer NP-movement structure studies find this pattern (see [35–38],
who reported no posterior activation). These regions include the posterior middle and superior
temporal gyri (pMTG, pSTG) and the inferior parietal cortex, (i.e., angular gyrus (AG)). Notably, these
regions have been found to reflect verb-argument structure processing [42–44], as well as the
integration of verbs with their arguments ([45,46]; see [47] for review). These findings suggest that left
posterior perisylvian regions may support thematic reanalysis, but this raises the question of why
activation in these regions has not emerged consistently for passive sentences. One possible
explanation is that task demands influence the likelihood of observing activation in this region, with
activation more likely when the task places significant demands on thematic reanalysis processes.
Hirotani et al. [26] aimed to disentangle the neural correlates of thematic and syntactic reanalysis by
comparing active sentences (which require neither thematic nor syntactic reanalysis), passive sentences
(which are assumed by the authors to require both) and causative sentences (which are claimed to
require thematic reanalysis, but not syntactic reanalysis, because they do not involve syntactic
movement). Relative to active sentences, both passive and causative sentences elicited activation in
regions, including the left IFG (pars triangularis) and the left posterior superior temporal gyrus
(pSTG), whereas direct comparison of passive to causative sentences showed no differential activation.
However, the time course of activation differed for the two sentence types in the left IFG (greater
activation for passive than causative sentences approximately 8 s after the critical point of the
sentence), but not in the pSTG. The authors argue that both the left IFG and pSTG support thematic
reanalysis, whereas the left IFG additionally supports syntactic reanalysis.
Studies of aphasia also provide insight into the neural basis of complex sentence processing.
Individuals with agrammatic (Broca’s) aphasia, resulting from stroke or head-injury, typically have
impaired production and comprehension of noncanonical sentences, including both wh- and
NP-movement structures [48–51]. Because agrammatic aphasia often is associated with damage to
Broca’s area, this pattern suggests that Broca’s area plays a crucial role in complex sentence
processing. Other research, however, shows that stroke-induced lesions in agrammatic aphasia often
extend well beyond Broca’s area and include cortical (and subcortical) tissue in temporoparietal
regions, as well (see [52,53], for lesions associated with stroke-induced agrammatic aphasia). In
addition, studies directly investigating the relationship between complex sentence processing and
lesion location have reported mixed results: damage to anterior and/or posterior perisylvian regions can
result in deficits in complex sentence processing [54–56]. Two recent studies examining cortical
atrophy in patients with the agrammatic variant of primary progressive aphasia (PPA-G) found
correlations between cortical atrophy in the left IFG and impaired complex sentence processing in
these patients [57,58]. However, Caplan and colleagues [55] argued for a primary role of posterior
perisylvian regions in syntactic processing in stroke-induced aphasia. The authors found
that stroke-induced lesions in both Wernicke’s area, and the anterior inferior temporal lobe were
associated with accuracy on a sentence-to-picture matching task, whereas lesions in the inferior and
superior parietal lobe predicted performance on a task tapping thematic role assignment and
co-indexation for syntactically complex sentences, as compared to simple sentences. In addition,
Brain Sci. 2013, 3
1202
Thothathiri, Kimberg, and Schwartz [59], in a voxel-based lesion mapping (VLSM) study in
stroke-induced aphasia, found correlations between noncanonical sentence comprehension deficits and
lesions in the left temporo-parietal cortex, as well as other regions, but not in Broca’s area. These
mixed results may be due in part to variable deficits in thematic and/or syntactic reanalysis processes
across individuals with aphasia.
Collectively, the results of both neuroimaging studies with healthy participants and lesion-deficit
correlation studies suggest that both anterior and posterior regions are engaged for complex sentence
processing. However, because of the dearth of studies examining NP-movement structures in English,
as well as the mixed findings derived from Japanese studies focused on these sentences, further
research is needed. In addition, further evidence is necessary to test the hypothesis that the left IFG
supports syntactic reanalysis whereas, left posterior perisylvian regions support thematic reanalysis
during complex sentence processing (cf. [26]).
The present study investigated the neural correlates of processing passive and active sentences in
English, using a sentence-picture verification task to probe comprehension. This task was selected in
order to maximize demands on thematic mapping and reanalysis processes by requiring the participant
to integrate the linguistic representation with an external representation of the event (the visual scene).
Following previous studies, we expected to find activation in the left IFG for passive as compared to
active sentences. Despite mixed findings with respect to left posterior perisylvian activation, we
expected activation in these regions, as well, due to our use of a thematically-demanding task. If
passive sentences entail syntactic reanalysis, as suggested based on linguistic descriptions of passives
as NP-movement structures, we expected greater IFG activation for passive as compared to active
sentences. Additionally, if passive sentence computation engages thematic reanalysis, posterior
perisylvian activation would be expected.
2. Results and Discussion
2.1. Behavioral Results
Participants performed overall very well on both active (97.7%) and passive (98%) sentences.
Accuracy was equally good when the sentence matched the picture displayed on the screen and when
sentence and picture did not match (98.1% and 97.5%, respectively). Reaction times (RT) obtained on
correct trials were faster for active sentences (3519 ± 362 ms) as compared to passive sentences
(3578 ± 375 ms) and for matched as compared to mismatched sentences (3497 ± 335, 3601 ± 395). A
mixed-effects regression analysis was conducted by introducing one predictor at a time: sentence type
(active, passive), condition (match, mismatch) and age, to evaluate the contribution of each to the
variance. ANOVAs between consecutive models were run with subjects and items introduced as
random factors for all. Random slopes for significant predictors were also introduced in the model one
at a time, and their contribution to the explanation of variance was evaluated in the same way as the
fixed factors. Prior to the analysis, reaction times were log-transformed in order to render a normal
distribution. Results indicated a significant main effect of sentence type (t = 2.2, p = 0.031), with
longer RT for passive as compared to active sentences and of condition, where sentences that did not
match the picture elicited longer RT than sentences matching the picture (t = 3.7, p < 0.001). The
Brain Sci. 2013, 3
1203
introduction of random slopes for both sentence type and condition significantly improved the model
(χ2 = 6.8, p = 0.033 and χ2 = 26.4, p < 0.001 respectively). No significant interaction between sentence
type and condition emerged (t = −0.4, p = ns). Age did not reach significance as a predictor (t = 1.15,
p = ns).
2.2. fMRI Results
Passive sentences, as compared to active sentences, elicited significant clusters of activation in the
bilateral inferior frontal gyrus (pars opercularis and pars triangularis, overlapping with BA’s 44 and
45), as well as in the left temporo-occipital junction (middle occipital gyrus and posterior middle
temporal gyrus); see Table 1 and Figure 1. Additional clusters that did not survive correction for
multiple comparisons were observed in the bilateral supplementary motor area, the left superior
parietal lobule, and the left precentral gyrus. The reverse contrast, of active over passive sentences, did
not reveal any significant areas of activation. In addition, no significant effects of age were observed
for either contrast (passive > active, active > passive).
Table 1. Areas of differential activation for passive and active sentences. Peak Montreal
Neurological Institute (MNI) coordinates, cluster sizes (k), maximal t-values, and
cluster-corrected (family-wise error rate) p-values are reported (voxel-wise threshold of
p < 0.001, uncorrected, k ≥ 85). Notes: LH, left hemisphere; RH, right hemisphere; IFG,
inferior frontal gyrus; pMTG, posterior middle temporal gyrus; SMA, supplementary
motor area; SPL, superior parietal lobule.
Contrast
Passive > Active
Active > Passive
Age
(Passive > Active,
Active > Passive)
Region
RH IFG (pars opercularis,
pars triangularis)
LH IFG (pars opercularis,
pars triangularis)
LH middle occipital gyrus,
pMTG
Bilateral SMA
L SPL
L precentral gyrus
None
None
Peak Coordinates
k
t
p
x
y
z
56
24
28
497
6.09
0.005
−36
4
32
349
4.52
0.019
−46
−76
4
265
5.21
0.048
−2
−40
−38
22
−46
−2
54
52
54
217
136
117
4.77
3.60
4.2
0.084
0.227
0.289
Brain Sci. 2013, 3
1204
Figure 1. Regions of differential activation for passive as compared to active sentences
(voxel-wise threshold of p < 0.001; k ≥ 85).
2.3. Discussion
The present study investigated the neural correlates of processing passive sentences in English.
Consistent with previous studies, we found activation for passive sentences in the left IFG, including
tissue in both the pars opercularis and pars triangularis (cf. [26,36–39]); additionally, activation was
found in homologous regions in the right hemisphere. We also found activation in the left
temporo-occipital junction, including the pMTG. As pointed out earlier, some previous studies have
reported posterior perisylvian activation for passive as compared to active sentences [26,39], whereas
others have not found activation in this region [36–38]. These results contribute to our understanding
of the neural basis of complex sentence processing, in particular, the functions of the left and right IFG
and left posterior temporal regions in supporting noncanonical sentence comprehension.
2.3.1. Roles of the Left and Right IFG in Passive Sentence Comprehension
Several explanations have been proposed for the observation that complex sentence comprehension
consistently elicits activation in the left IFG. It has been argued by some researchers that the left IFG
supports the processing of complex syntactic representations [24,25,53,60–62] or, more specifically,
syntactic movement [30–34], whereas other accounts suggest that the IFG supports morphological
processing [37], semantic/pragmatic aspects of argument processing [63–67] or working memory [68].
Results of the present study are not consistent with morphological, semantic or working memory
accounts of left IFG activation. First, the experimental sentences were controlled for morphological
complexity, through the use of auxiliary verbs combined with present participles in active sentences
and past participles in passive sentences. Second, the experimental sentences were semantically
reversible, with the same arguments used in passive and active sentences, and therefore, left IFG
activation is not likely due to processing the intrinsic semantic features of arguments. Third, working
memory demands were low in the present study, due to the use of short sentences; furthermore, in
Brain Sci. 2013, 3
1205
contrast with wh-structures, passive sentences are unlikely to place enhanced demands on working
memory, because the subject NP is not identifiable as a filler until the verb is processed. Thus, the IFG
activation found here is most likely associated with the greater syntactic and/or verb-argument structure
complexity of passive as compared to active sentences. The present results, however, do not directly
address whether the source of left IFG activation is syntactic movement and/or noncanonical
verb-argument structure mapping. Previous research suggests that the left IFG is involved in both
aspects of syntactic complexity, supporting both thematic and syntactic reanalysis processes [26].
Studies also show that these processes may be supported by different subregions of the IFG, although
research examining the neural substrates of syntactic movement report peak activation in the pars
opercularis (roughly BA 44) [25,27], as well as in the pars triangularis (roughly BA 45) [26,35,36,38];
see discussion in [29]. Similarly, both the pars opercularis and pars triangularis have been linked to
noncanonical argument mapping (see discussion in [64]).
Thus, the present results are consistent with the view that the left IFG supports syntactic
movement [30–34]. However, they are incompatible with accounts claiming that the IFG supports only
certain types of movement: that is, wh-movement, but not NP-movement found in passive sentences.
For example, Christensen [69] argues that movement to the complementizer phrase (CP) domain
(i.e., movement to the left of the grammatical subject position), which takes place in wh-movement
structures, should elicit activation in the left IFG and posterior perisylvian regions, whereas movement
within the inflectional phrase (IP) domain (i.e., movement to or to the right of the grammatical subject
position), which occurs in NP-movement structures, such as passive sentences (under standard
assumptions), should elicit activation in the left anterior temporal cortex instead. Santi and
Grodzinsky [28] propose that left IFG activation is elicited only by movement that is predictable at the
point of the antecedent (i.e., wh-movement structures in which the wh-word is immediately identifiable
as a filler). However, consistent with previous studies on passive sentence processing in Japanese and
Chinese [26,35–39], our findings suggest that if indeed, the left IFG supports syntactic movement; it
does so not only for wh-movement structures, but also for NP-movement structures. Furthermore, despite
differences between Japanese and English passives with respect to linguistic representation [41] and
processing costs [39], the present results indicate that English passives, like Japanese passives, elicit
left IFG activation, possibly because passive sentences in both languages require syntactic reanalysis.
In addition, passive sentences elicited greater activation in the right IFG. This finding contrasts with
previous studies of passive sentence comprehension, which did not find activation in the right
IFG ([26,35–39]), and in general, effects of syntactic complexity in the IFG tend to be strongly
left-lateralized (for a review, see [29]). However, some evidence suggests that right IFG supports
syntactic reanalysis of complex sentences, particularly in the context of integration of a linguistic
representation with a visual scene. In an fMRI study, Meltzer, McArdle, Schafer and Braun [62]
investigated the neural correlates of syntactic reanalysis demands during comprehension of complex
(noncanonical object-relative) and simple (canonical subject-relative) sentences. In 50% of the trials,
participants simply listened to sentences (sentence-alone condition); in the other 50% of the trials,
participants performed a sentence-picture matching task. Syntactic reanalysis demands were higher in
the sentence-picture matching condition, because participants were required to hold the linguistic
representation of the sentence in memory, compare it to the picture probes, then reanalyze it, if
necessary. The authors reported main effects of syntactic complexity (object-relatives > subject-relatives)
Brain Sci. 2013, 3
1206
in the left IFG across tasks; however, in sentence-picture matching trials, effects of syntactic
complexity were observed in bilateral IFG. The authors propose that the right IFG specifically supports
conscious and effortful syntactic reanalysis. Similarly, right IFG activation has been found in previous
studies that required participants to generate linguistic judgments about syntactic structures [70,71].
The present study also placed high demands on syntactic reanalysis processes, as participants were
required to build a linguistic representation of the sentence and compare it to a visual scene. Thus, high
syntactic reanalysis demands are one plausible explanation for the right IFG activation observed for
passive sentences. However, we note that the design of the present study (passive/active blocks) makes
it difficult to distinguish the neural correlates of reanalysis processes that are due to sentence type
(passive vs. active sentences) from those that are due to the correspondence between the sentence and
picture (mismatch vs. match trials).
A less likely explanation for the presence of bilateral IFG activation is the inclusion of both young
and older adult participants. Some previous research has suggested that healthy aging is associated
with increased bilateral language activation, especially in frontal regions [72–74]. Therefore, one
might hypothesize that bilateral IFG activation for passive sentences in the present study was driven by
the older participants. However, we did not observe any significant effects of age on the processing of
passive (or active) sentences. For this reason, it is more likely that bilateral IFG activation is due to the
high syntactic reanalysis demands imposed by the sentence-picture verification task.
2.3.2. The Role of Left Posterior Temporal Cortex in Passive Sentence Comprehension
Previous research has shown that left posterior temporal cortex supports thematic reanalysis in
passive sentences [26], and more generally, the integration of verbs with their arguments [29,43–45,47].
Therefore, the left temporo-occipital activation observed for passive sentences in the present study is
likely due to thematic reanalysis in the context of a sentence-picture verification task. However, it is an
open question why previous studies of passive sentence processing have yielded mixed results with
respect to activation in this region. One possible explanation is that posterior perisylvian activation is
more sensitive to the choice of task than is left IFG activation, such that tasks that place greater
demands on thematic mapping and reanalysis processes are more likely to elicit posterior perisylvian
activation. Consistent with this hypothesis, Caplan, Chen and Waters [61] found left IFG activation for
noncanonical as compared to canonical sentence comprehension across three tasks (sentence
verification, plausibility judgment and non-word detection), whereas posterior perisylvian activation
was found only for the sentence verification and plausibility judgment tasks, both of which target
thematic role mapping. Studies of passive sentence processing have reported nonhomogeneous
findings that nonetheless provide some support for this hypothesis. For example, Yokoyama et al. [37]
used a lexical decision task that did not target thematic role mapping and did not elicit posterior
perisylvian activation. Studies that targeted thematic role mapping within the sentence (e.g.,
plausibility judgments, comprehension questions about thematic role assignment) have yielded mixed
results, with some studies [26,39] eliciting temporo-parietal activations and others not [36,38].
In addition, one study [35] used a sentence-picture verification task, as in the present study. The
authors did not find posterior perisylvian activation for passive as compared to active sentences, but
did find posterior temporal (pMTG/pSTG) activation for another type of noncanonical structure
Brain Sci. 2013, 3
1207
(scrambled sentences) as compared to active sentences. Thus, there is some evidence that activation for
passive sentences may emerge in left posterior perisylvian regions only with tasks that place explicit
demands on verb-argument integration.
The left posterior temporo-occipital activation found in the present study is located inferiorly and
posteriorly to the pSTG activation reported by Hirotani et al. [26] and in studies of verb-argument
integration [29,47]. However, we note that some studies of complex sentence processing have elicited
activation in the left temporo-occipital junction [62,75,76]. Meltzer and colleagues [62] also found left
pMTG/pSTG activation for semantically reversible relative to non-reversible sentences in both
sentence-alone and sentence-picture matching conditions. However, this activation was shifted
posteriorly in the sentence-picture matching condition (peak MNI coordinates: (−49 −73 4); cf. peak
coordinates in the present study: (−46 −76 4)) relative to the sentence-alone condition (peak
coordinates: (−53 −46 11)). This suggests that integration of auditorily-presented linguistic stimuli
with visually-presented scenes may result in a shift posteriorly to the temporo-occipital junction for
thematic processing (see, also, the discussion of task effects on activation patterns in [29]).
3. Experimental Section
3.1. Participants
Fourteen healthy young adults (mean age: 24.9; range = 19–38; two males) and thirteen healthy
older adults (mean age: 61.2; range = 54–70; seven males) participated in the study. All were
right-handed native speakers of English with normal vision and hearing and no history of
speech/language, learning or neurological disorders. The study was approved by the Institutional
Review Board at Northwestern University, and all participants gave informed consent.
3.2. Materials
Twenty verbs were selected for inclusion in the experiment. All were semantically reversible,
frequently-occurring (M log frequency = 4.33; Corpus of Contemporary American English
(COCA); [77]) and had a regular passive form (-ed). Each verb was embedded in four sentences, all
including the same noun phrase participants: two active sentences (e.g., The brother was pushing the
sister; The sister was pushing the brother) and two passive sentences (e.g., The brother was pushed by
the sister; The sister was pushed by the brother). All sentences contained past-tense verb forms; the
passive sentences included a past-tense auxiliary combined with the past participle (was V-ed),
whereas the active sentences contained a past-tense auxiliary with a progressive main verb (was
V-ing), in order to control for morphological complexity across the two conditions. The two sentence
types were also controlled for length in syllables (active M = 6.15; passive M = 6.3, p =0.42). The
nouns referring to participants were all frequently-occurring (M log frequency = 5.00, COCA) and
referred to humans or animals (e.g., brother, cat, mouse, woman).
Sentences were recorded by a female native English speaker in a sound proof booth, using Audacity.
Maximum amplitude of the sound files was normalized to −3 dB. All sentences were between 2.5 and
3 s, with 100 ms of silence added at the offset of each sentence. An additional variable period of silence
was added at the beginning of each sentence, so that all sound files were 3500 ms long.
Brain Sci. 2013, 3
1208
Twenty black and white drawings were prepared, one for each verb. All drawings depicted two
animate participants engaged in an event, such that the agent acted upon the theme (see, e.g., Figure 1).
The same line drawing was used for all four sentences constructed with the same verb.
3.3. Procedures
In each trial, a picture and a sentence were presented simultaneously, with the picture remaining on
the screen for 6000 ms, followed by 1000 ms fixation. Participants held a response box in their left
hand and were asked to press with their index finger if the picture matched the sentence and with their
middle finger if it did not. Participants could respond at any point during the trial, and reaction times
were measured relative to the onset of stimulus presentation. The experiment was presented using
E-Prime, and participant responses were recorded using Cedrus RB610 response boxes.
The experiment used a block design, such that each block included four trials of the same syntactic
structure (active or passive), with matched and mismatched trials pseudorandomized within and across
blocks. Each block ended with a 12 s fixation cross, for a total of 40 s per block. The experiment
included 20 blocks, which were pseudorandomized and assigned to two runs of 6:40 min each. The
order of presentation of the two blocks was counterbalanced across participants.
3.4. Data Acquisition
MRI data were acquired using a Siemens 3T Tim Trio scanner with a 32-channel head coil. At the
beginning of each scan, a T1-weighted anatomical image was acquired, with the following
parameters: time to repeat (TR) = 2300 ms, time to echo (TE) = 2.91 ms, flip angle = 9 degrees; matrix
size = 256 × 256; field of view (FOV) = 256 mm; voxel size = 1 × 1 × 1 mm; 176 slices). During the
experimental task, blood oxygen level dependent (BOLD) contrast images were acquired using the
following parameters: TR= 2000 ms; TE = 30 ms, flip angle = 80 degrees; matrix size = 64 × 64;
FOV = 220.16 mm; voxel size = 3.44 × 3.44 × 3 mm; 32 slices.
3.5. Data Analysis
3.5.1. Behavioral Data
Accuracy and reaction time (RT) data were analyzed by means of logistic (for accuracy) or linear
(for RT) regression with mixed-effects, following the approach described by Jaeger [78] and Baayen,
Davidson and Bates [79]. Subject and item were introduced as random effects in the regression
analysis, and the contribution of each predictor to the explanation of the variance was evaluated by
performing ANOVA comparisons between models.
3.5.2. Neuroimaging Data
Preprocessing and statistical analysis of the MRI data were performed using SPM8. Preprocessing
consisted of slice-timing correction of the functional scans, realignment of the functional scans to a
mean functional volume, normalization of the anatomical and functional scans to the MNI 152-subject
Brain Sci. 2013, 3
1209
template brain, reslicing of functional and anatomical scans to a 2 × 2 × 2 mm voxel and smoothing of
the functional images using a 9 mm Gaussian kernel.
In order to eliminate scanner drift, a high-pass filter of 128 s was used in the first-level statistical
analysis. In addition to the two experimental conditions (active, passive), a parameter for run and six
motion parameters were entered into this analysis. In second-level analyses, contrast maps from each
participant (passive > active sentences; active > passive sentences) were entered into ANCOVA
analyses. Participant age was entered as a covariate to control for and evaluate the potential effects of
age on passive and active sentence processing. The results were evaluated using a
voxel-level threshold of p < 0.001 with a minimum cluster size (k) of 85 (680 mm3), applying
family-wise error rate correction at the cluster level to identify clusters of significant activation. This is
equivalent to, or more stringent than, the threshold used in some previous studies testing subtle
linguistic contrasts (e.g., [42] and references therein).
4. Conclusions
The results of the present study suggest that in both younger and older adults, comprehension of
passive sentences is supported by bilateral IFG and left posterior temporo-occipital regions. These
findings are largely consistent with previous research on the comprehension of complex sentences,
which has linked left IFG activation to syntactic complexity and left posterior temporal activation to
verb-argument structure integration (see, e.g., [29,80]). Thus, despite the linguistic and
psycholinguistic differences between passive sentences and other complex structures, they may be
largely supported by the same brain regions. The right IFG activation found in the present study may
reflect effortful reanalysis processes required by the sentence-picture verification task. Together with
previous studies [62,70,71], this suggests that the right IFG may play a role in controlled syntactic
reanalysis and the generation of linguistic judgments.
Acknowledgments
This research was supported by NIH grant R01DC01948-19 to C.K. Thompson. The authors would
like to thank Chien-Ju Hsu and Julia Schuchard for help with stimuli construction, Ellyn Riley, Ellen
Fitzmorris and Shezena Samsair for assistance with data analysis and Caitlin Radnis for assistance
with manuscript preparation.
Conflict of Interest
The authors declare no conflict of interest.
References
1.
2.
3.
Chomsky, N. Lectures on Government and Binding; Foris: Dordrecht, The Netherlands, 1981.
Dickey, M.W.; Choy, J.J.; Thompson, C.K. Real-time comprehension of wh-movement in aphasia:
Evidence from eyetracking while listening. Brain Lang. 2007, 100, 1–22.
Dickey, M.W.; Thompson, C.K. Automatic processing of wh- and NP-movement in agrammatic
aphasia: Evidence from eyetracking. J. Neurolinguist. 2009, 22, 563–583.
Brain Sci. 2013, 3
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
1210
Love, T.; Swinney, D.; Walenski, M.; Zurif, E. How left inferior frontal cortex participates in
syntactic processing: Evidence from aphasia. Brain Lang. 2008, 107, 203–219.
Nicol, J.; Fodor, J.D.; Swinney, D. Using Cross-Modal lexical decision tasks to investigate
sentence processing. J. Exp. Psychol. Learn. 1994, 20, 1229–1238.
Sussman, R.S.; Sedivy, J.C. The time-course of processing syntactic dependencies: Evidence from
eye movements. Lang. Cogn. Process. 2003, 18, 143–161.
Zurif, E.; Swinney, D.; Prather, P.; Solomon, J.; Bushell, C. An on-line analysis of syntactic
processing in Broca’s and Wernicke’s aphasia. Brain Lang. 1993, 45, 448–464.
Fiebach, C.J.; Schlesewsky, M.; Friederici, A.D. Separating syntactic memory costs and syntactic
integration during parsing: The processing of German WH-questions. J. Mem. Lang. 2002, 47,
250–272.
Pollard, C.; Sag, I.A. Head-Driven Phrase Structure Grammar; The University of Chicago Press:
Chicago, IL, USA, 1994.
Bresnan, J. Lexical-Functional Syntax; Blackwell: Oxford, UK, 2000.
Osterhout, L.; Swinney, D.A. On the temporal course of gap-filling during comprehension of
verbal passives. J. Psycholinguist. Res. 1993, 22, 273–286.
Burkhardt, P.; Piñango, M.M.; Wong, K. The role of the anterior left hemisphere in real-time
sentence comprehension: Evidence from split intransitivity. Brain Lang. 2003, 86, 9–22.
Friedmann, N.; Taranto, G.; Shapiro, L.P.; Swinney, D. The leaf fell (the leaf): The online
processing of unaccusatives. Linguist. Inq. 2008, 39, 355–377.
Koring, L.; Mak, P.; Reuland, E. The time course of argument reactivation revealed: Using the
visual world paradigm. Cognition 2012, 123, 361–379.
Kamide, Y.; Scheepers, C.; Altmann, G.T. Integration of syntactic and semantic information in
predictive processing: Cross-linguistic evidence from German and English. J. Psycholinguist. Res.
2003, 32, 37–55.
Olson, D.R.; Filby, N. On the comprehension of passive and active sentences. Cogn. Psychol.
1972, 3, 361–381.
Knoeferle, P. Comparing the Time-course of Processing Initially Ambiguous and Unambiguous
German SVO/OVS Sentences in Depicted Events. In Eye Movement Research: A Window on
Mind and Brain; van Gompel, R., Fischer, M., Murray, W., Hill, R., Eds.; Elsevier: Oxford, UK,
2007; pp. 517–531.
Knoeferle, P.; Crocker, M.W.; Scheepers, C.; Pickering, M.J. The influence of the immediate
visual context on incremental thematic role-assignment: Evidence from eye-movements in
depicted events. Cognition 2005, 95, 95–127.
Meyer, A.M.; Mack, J.E.; Thompson, C.K. Tracking passive sentence comprehension in
agrammatic aphasia. J. Neurolinguist. 2012, 25, 31–43.
Stromswold, K.; Eisenband, J.; Norland, E.; Ratzan, J. Tracking the Acquisition and Processing of
English Passives: Using Acoustic cues to Disambiguate Actives and Passives. Paper Presented at
the CUNY Conference on Sentence Processing, New York, NY, USA, 21–23 March 2002.
Hanne, S.; Sekerina, I.A.; Vasishth, S.; Burchert, F.; de Bleser, R. Chance in agrammatic sentence
comprehension: What does it really mean? Evidence from eye movements of German agrammatic
aphasic patients. Aphasiology 2011, 25, 221–244.
Brain Sci. 2013, 3
1211
22. Brookshire, R.H.; Nicholas, L.E. Verification of active and passive sentences by aphasic and
nonaphasic subjects. J. Speech Hear. Res. 1980, 23, 878–893.
23. Ferreira, F. The misinterpretation of noncanonical sentences. Cogn. Psychol. 2003, 47, 164–203.
24. Caplan, D.; Alpert, N.; Waters, G. Effects of syntactic structure and propositional number on
patterns of regional cerebral blood flow. J. Cogn. Neurosci. 1998, 10, 541–552.
25. Friederici, A.D.; Fiebach, C.J.; Schlesewsky, M.; Bornkessel, I.D.; von Cramon, D.Y. Processing
linguistic complexity and grammaticality in the left frontal cortex. Cereb. Cortex 2006, 16,
1709–1717.
26. Hirotani, M.; Makuuchi, M.; Rüschemeyer, S.A.; Friederici, A.D. Who was the agent? The neural
correlates of reanalysis processes during sentence comprehension. Hum. Brain Mapp. 2011, 32,
1775–1787.
27. Makuuchi, M.; Bahlmann, J.; Anwander, A.; Friederici, A.D. Segregating the core computational
faculty of human language from working memory. Proc. Natl. Acad. Sci. USA 2009, 106,
8362–8367.
28. Santi, A.; Grodzinsky, Y. Broca’s area and sentence comprehension: A relationship parasitic on
dependency, displacement or predictability? Neuropsychologia 2012, 50, 821–832.
29. Friederici, A.D. The brain basis of language processing: From structure to function. Physiol. Rev.
2011, 91, 1357–1392.
30. Ben-Shachar, M.; Hendler, T.; Kahn, I.; Ben-Bashat, D.; Grodzinsky, Y. The neural reality of
syntactic transformations: Evidence from functional magnetic resonance imaging. Psychol. Sci.
2003, 14, 433–440.
31. Ben-Shachar, M.; Palti, D.; Grodzinsky, Y. Neural correlates of syntactic movement: Converging
evidence from two fMRI experiments. Neuroimage 2004, 21, 1320–1336.
32. Grodzinsky, Y. The neurology of syntax: Language use without Broca’s area. Behav. Brain Sci.
2000, 23, 1–21.
33. Santi, A.; Grodzinsky, Y. Working memory and syntax interact in Broca’s area. Neuroimage 2007,
37, 8–17.
34. Shetreet, E.; Friedmann, N.; Hadar, U. The neural correlates of linguistic distinctions:
Unaccusative and unergative verbs. J. Cogn. Neurosci. 2010, 22, 2306–2315.
35. Kinno, R.; Kawamura, M.; Shioda, S.; Sakai, K.L. Neural correlates of noncanonical syntactic
processing revealed by a picture-sentence matching task. Hum. Brain Mapp. 2008, 29, 1015–1027.
36. Ye, Z.; Zhou, X. Conflict control during sentence comprehension: fMRI evidence. Neuroimage
2009, 48, 280–290.
37. Yokoyama, S.; Miyamoto, T.; Riera, J.; Kim, J.; Akitsuki, Y.; Iwata, K.; Yoshimoto, K.; Horie, K.;
Sato, S.; Kawashima, R. Cortical mechanisms involved in the processing of verbs: An fMRI study.
J. Cogn. Neurosci. 2006, 18, 1304–1313.
38. Yokoyama, S.; Okamoto, H.; Miyamoto, T.; Yoshimoto, K.; Kim, J.; Iwata, K.; Hyeonjeong, J.;
Uchida, S.; Ikuta, N.; Sassa, Y.; et al. Cortical activation in the processing of passive sentences in
L1 and L2: An fMRI study. Neuroimage 2006, 30, 570–579.
39. Yokoyama, S.; Watanabe, J.; Iwata, K.; Ikuta, N.; Haji, T.; Usui, N.; Taira, M.; Miyamoto, T.;
Nakamura, W.; Sato, S.; et al. Is Broca’s area involved in the processing of passive sentences? An
event-related fMRI study. Neuropsychologia 2007, 45, 989–996.
Brain Sci. 2013, 3
1212
40. Wartenburger, I.; Heekeren, H.R.; Burchert, F.; de Bleser, R.; Villringer, A. Grammaticality
judgments on sentences with and without movement of phrasal constituents: An event-related
fMRI study. J. Neurolinguist. 2003, 16, 301–314.
41. Hoshi, H. Theta-role assignment, passivization, and excorporation. J. East Asian Ling. 1994, 3,
147–178.
42. Meltzer-Asscher, A.; Schuchard, J.; den Ouden, D.B.; Thompson, C.K. The neural substrates of
complex argument structure representations: Processing “alternating transitivity” verbs. Lang.
Cogn. Process. 2012, doi:10.1080/01690965.2012.672754.
43. Thompson, C.K.; Bonakdarpour, B.; Fix, S.; Blumenfeld, H.K.; Parrish, T.B.; Gitelman, D.R.;
Mesulam, M.-M. Neural correlates of verb argument structure processing. J. Cogn. Neurosci.
2007, 19, 1753–1767.
44. Thompson, C.K.; Bonakdarpour, B.; Fix, S. Neural mechanisms of verb argument structure
processing in agrammatic aphasic and healthy age-matched listeners. J. Cogn. Neurosci. 2010, 22,
1993–2011.
45. Bornkessel, I.; Zysset, S.; Friederici, A.D.; von Cramon, D.Y.; Schlesewsky, M. Who did what to
whom? The neural basis of argument hierarchies during language comprehension. Neuroimage
2005, 26, 221–233.
46. Friederici, A.D.; Rüschemeyer, S.A.; Hahne, A.; Fiebach, C.J. The role of left inferior frontal and
superior temporal cortex in sentence comprehension: Localizing syntactic and semantic processes.
Cereb. Cortex 2003, 13, 170–177.
47. Thompson, C.K.; Meltzer-Asscher, A. Neurocognitive Mechanisms of Verb Argument Structure
Processing. In Structuring the Argument; Bachrach, A., Roy, I., Stockall, L., Eds.; John
Benjamins: Amsterdam, The Netherlands, 2013, in press.
48. Bastiaanse, R.; Edwards, S. Word order and finiteness in Dutch and English Broca’s and
Wernicke’s aphasia. Brain Lang. 2004, 89, 91–107.
49. Grodzinsky, Y.; Piñango, M.M.; Zurif, E.; Drai, D. The critical role of group studies in
neuropsychology: Comprehension regularities in Broca’s aphasia. Brain Lang. 1999, 67, 134–147.
50. Linebarger, M.C.; Schwartz, M.F.; Saffran, E.M. Sensitivity to grammatical structure in so-called
agrammatic aphasics. Cognition 1983, 13, 361–392.
51. Luzzatti, C.; Toraldo, A.; Guasti, M.T.; Ghirardi, G.; Lorenzi, L.; Guarnaschelli, C.
Comprehension of reversible active and passive sentences in agrammatism. Aphasiology 2001, 15,
419–441.
52. Kielar, A.; Meltzer-Asscher, A.; Thompson, C.K. Electrophysiological responses to argument
structure violations in healthy adults and individuals with agrammatic aphasia. Neuropsychologia
2012, 50, 3320–3337.
53. Thompson, C.K.; den Ouden, D.B.; Bonakdarpour, B.; Garibaldi, K.; Parrish, T.B. Neural
plasticity and treatment-induced recovery of sentence processing in agrammatism. Neuropsychologia
2010, 48, 3211–3227.
54. Caplan, D.; Hildebrandt, N.; Makris, N. Location of lesions in stroke patients with deficits in
syntactic processing in sentence comprehension. Brain 1996, 119, 933–949.
Brain Sci. 2013, 3
1213
55. Caplan, D.; Waters, G.; Kennedy, D.; Alpert, N.; Makris, N.; Dede, G.; Michaud, J.; Reddy, A.
A study of syntactic processing in aphasia II: Neurological aspects. Brain Lang. 2007, 101,
151–177.
56. Dronkers, N.F.; Wilkins, D.P.; van Valin, R.D.; Redfern, B.B.; Jaeger, J.J. Lesion analysis of the
brain areas involved in language comprehension. Cognition 2004, 92, 145–177.
57. Rogalski, E.; Cobia, D.; Harrison, T.M.; Wienecke, C.; Thompson, C.K.; Weintraub, S.;
Mesulam, M.-M. Anatomy of language impairments in primary progressive aphasia. J. Neurosci.
2011, 31, 3344–3350.
58. Wilson, S.M.; Dronkers, N.F.; Ogar, J.M.; Jang, J.; Growdon, M.E.; Agosta, F.; Henry, M.L.;
Miller, B.-.L.; Gorno-Tempini, M.L. Neural correlates of syntactic processing in the nonfluent
variant of primary progressive aphasia. J. Neurosci. 2010, 15, 16845–16854.
59. Thothathiri, M.; Kimberg, D.Y.; Schwartz, M.F. The neural basis of reversible sentence
comprehension: Evidence from voxel-based lesion symptom mapping in aphasia. J. Cogn.
Neurosci. 2012, 24, 212–222.
60. Just, M.A.; Carpenter, P.A.; Keller, T.A.; Eddy, W.F.; Thulborn, K.R. Brain activation modulated
by sentence comprehension. Science 1996, 274, 114–116.
61. Caplan, D.; Chen, E.; Waters, G. Task-dependent and task-independent neurovascular responses
to syntactic processing. Cortex 2008, 44, 257–275.
62. Meltzer, J.A.; McArdle, J.J.; Schafer, R.J.; Braun, A.R. Neural aspects of sentence comprehension:
Syntactic complexity, reversibility, and reanalysis. Cereb. Cortex 2010, 20, 1853–1864.
63. Bornkessel-Schlesewsky, I.; Schlesewsky, M.; von Cramon, D.Y. Word order and Broca’s region:
Evidence for a supra-syntactic perspective. Brain Lang. 2009, 111, 125–139.
64. Bornkessel-Schlesewsky, I.; Grewe, T.; Schlesewsky, M. Prominence vs. aboutness in sequencing:
A functional distinction within the left inferior frontal gyrus. Brain Lang. 2012, 120, 96–107.
65. Bornkessel-Schlesewsky, I.; Schlesewsky, M. The role of prominence information in the real-time
comprehension of transitive constructions: A cross-linguistic approach. Lang. Linguist. Compass
2009, 3, 19–58.
66. Grewe, T.; Bornkessel, I.; Zysset, S.; Wiese, R.; von Cramon, D.Y.; Schlesewsky, M. The
emergence of the unmarked: A new perspective on the language-specific function of Broca’s area.
Hum. Brain Mapp. 2005, 26, 178–190.
67. Grewe, T.; Bornkessel, I.; Zysset, S.; Wiese, R.; von Cramon, D.Y.; Schlesewsky, M. Linguistic
prominence and Broca’s area: The influence of animacy as a linearization principle. Neuroimage
2006, 32, 1395–1402.
68. Fiebach, C.J.; Schlesewsky, M.; Lohmann, G.; von Cramon, D.Y.; Friederici, A.D. Revisiting the
role of Broca’s area in sentence processing: Syntactic integration versus syntactic working
memory. Hum. Brain Mapp. 2005, 24, 79–91.
69. Christensen, K.R. Interfaces, syntactic movement, and neural activation: A new perspective on the
implementation of language in the brain. J. Neurolinguist. 2008, 21, 71–103.
70. Moro, A.; Tettamanti, M.; Perani, D.; Donati, C.; Cappa, S.F.; Fazio, F. Syntax and the brain:
Disentangling grammar by selective anomalies. Neuroimage 2001, 13, 110–118.
Brain Sci. 2013, 3
1214
71. Tettamanti, M.; Rotondi, I.; Perani, D.; Scotti, G.; Fazio, F.; Cappa, S.F.; Moro, A. Syntax
without language: Neurobiological evidence for cross-domain syntactic computations. Cortex
2009, 45, 825–838.
72. Davis, S.W.; Kragel, J.E.; Madden, D.J.; Cabeza, R. The architecture of cross-hemispheric
communication in the aging brain: Linking behavior to functional and structural connectivity.
Cereb. Cortex 2012, 22, 232–242.
73. Park, D.; Reuter-Lorenz, P. The Adaptive Brain: Aging and Neurocognitive Scaffolding. Annu.
Rev. Psychol. 2009, 60, 173–196.
74. Tyler, L.K.; Shafto, M.A.; Randall, B.; Wright, P.; Marslen-Wilson, W.D.; Stamatakis, E.A.
Preserving syntactic processing across the adult life span: The modulation of the frontotemporal
language system in the context of age-related atrophy. Cereb. Cortex 2010, 20, 352–364.
75. Caplan, D. Task effects on BOLD signal correlates of implicit syntactic processing. Lang. Cogn.
Process. 2010, 25, 866–901.
76. Cooke, A.; Zurif, E.B.; DeVita, C.; Alsop, D.; Koenig, P.; Detre, J.; Gee, J.; Pinãngo, M.;
Balogh, J.; Grossman, M. Neural basis for sentence comprehension: Grammatical and short-term
memory components. Hum. Brain Mapp. 2002, 15, 80–94.
77. Davies, M. The Corpus of Contemporary American English, 2008. Available online:
http://corpus.byu.edu/coca/ (accessed on 1 August 2011).
78. Jaeger, T.F. Categorical data analysis: Away from ANOVAs (transformation or not) and towards
logit mixed models. J. Mem. Lang. 2008, 59, 434–446.
79. Baayen, R.H.; Davidson, D.J.; Bates, D.M. Mixed-effects modeling with crossed random effects
for subjects and items. J. Mem. Lang. 2008, 49, 390–412.
80. Thompson, C.K.; Kielar, A. Neural bases of sentence processing: Evidence from neurolinguistic
and neuroimaging studies. In The Oxford Handbook of Language Production; Goldrick, M.,
Ferreira, V., Miozzo, M., Eds.; Oxford University Press: New York, NY, USA, 2014.
© 2013 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).