Skip to main content
NCBI home page
Human Brain Mapping logoLink to Human Brain Mapping
. 2009 Jun 24;31(1):65–79. doi: 10.1002/hbm.20845

The effect of presentation paradigm on syntactic processing: An event‐related fMRI study

Donghoon Lee 1, Sharlene D Newman 1,✉
PMCID: PMC6870720  PMID: 19554559

Abstract

An event‐related fMRI study was conducted to investigate the effect of two different sentence presentation paradigms—rapid serial visual presentation (RSVP) and whole sentence presentation—on syntactic processing. During scanning, sentences were presented using one of the two presentation paradigms and were followed by a short delay and a probe to verify sentence comprehension. The delay was included in an attempt to separate sentence‐related activity from probe‐related activity. The behavioral data showed a main effect of syntactic complexity for reaction time and accuracy, and accuracy revealed an interaction between complexity and the presentation paradigm employed—RSVP produced many more errors for syntactically complex sentences than did whole sentence presentation. The imaging data revealed a syntactic complexity effect during the sentence phase in left BA 44 and during the probe phase in left BA 44 and the left posterior MTG. In addition, timecourse analysis revealed that these two regions also showed an interaction between complexity and presentation paradigm such that there was no complexity effect during RSVP but a significant effect during whole sentence presentation. In addition to finding that these two presentation paradigms differentially affected syntactic processing, there were main effects within the visual pathway (V1/V2 vs. V5) and the hippocampus that revealed significant differences in activation between the paradigms. Hum Brain Mapp, 2010. © 2009 Wiley‐Liss, Inc.

Keywords: RSVP, event‐related fMRI, sentence comprehension, syntactic complexity

INTRODUCTION

In recent neuroimaging studies, the syntactic complexity effect has been localized to distinct brain areas such as the left inferior frontal gyrus (LIFG), particularly BA 44, and the left posterior superior/middle temporal gyrus (S/MTG) [e.g., Just et al., 1996b] and it appears to be independent of input modality [Constable et al., 2004; Michael et al., 2001]. Although the activation of both regions has been found to be modulated by syntactic complexity, the functional roles assigned to each have been quite different. For example, Caplan and Waters [1999] have suggested that the posterior region of LIFG (Broca's area) is specialized for syntactic working memory processes necessary for syntactic analysis of sentences. There have been a number of studies presented by Caplan et al., as well as others, that have provided supportive evidence of this claim [Caplan et al., 2008; Fiebach et al., 2005; Hashimoto and Sakai, 2002]. Although Wernicke's area has traditionally been regarded as a semantic processing center, many studies have found its activation is modulated by syntactic complexity [Michael et al., 2001; Newman et al., in press]. In an attempt to reconcile these two ideas, a number of hypotheses have been generated. For example, it has been suggested that a more specific function of the posterior MTG is in processing phrase‐level meaning [Cutting et al., 2006; Grossman et al., 2002; Vigneau et al., 2006] and integration of lexico‐syntactic information [see Grodzinsky and Friederici, 2006]. In this vein, syntactic manipulations not only affect syntactic processing but also semantic level processing that explains the involvement of Wernicke's area.

Although many studies have found the involvement of the left IFG and S/MTG during syntactic processing tasks, there has been some inconsistency in the findings across studies. One possible source of such inconsistency is in the variation in the sentence presentation paradigms used across studies; ranging from whole sentence presentation to rapid serial visual presentation (RSVP) to a moving window presentation paradigm. It may very well be that these different presentation paradigms require different cognitive demands which may interact with syntactic processing. For example, Cooke et al. [2002] examined the effect of syntactic complexity using serial visual presentation and found an effect of complexity in the inferior frontal gyrus, but not in temporal cortex. Keller et al. [2001] also presented sentences one word at a time, but using a moving window paradigm, and found no syntactic effects for sentences containing high‐frequency words in either the inferior frontal gyrus or temporal cortex. In a study using whole sentence presentation, syntactic complexity effects were observed in both the inferior frontal gyrus and temporal cortex [Michael et al., 2001]. Although these studies used different presentation paradigms, they may not be easily comparable because of possible important methodological or stimulus differences. However, studies from the same lab using the same stimulus materials also show different patterns. For example, in an fMRI study using RSVP, Caplan et al. [2002] failed to observe a syntactic complexity effect in either the LIFG or the posterior S/MTG but instead found an effect in the right angular gyrus. However, in earlier studies using the same sentence materials but different sentence presentation paradigms—auditory [Caplan et al., 1999] and whole sentence [Caplan et al., 1998]—an effect in the LIFG was observed. Caplan [2001] explained that the lack of a syntactic complexity effect in the LIFG was due to the cognitive demands associated with RSVP. Although these discrepancies are present in the literature, to our knowledge, there is no neuroimaging study that examines whether different presentation paradigms have a differential effect on syntactic processing; therefore, the current study attempts to fill this gap.

RSVP, initially developed by Forster [1970], has been a staple of psycholinguistic research because it provides for control of stimulus presentation time [Potter, 1984]. In RSVP reading, each word is serially presented in the center of the screen and the speed of reading is under the experimenter's control. Because all the words are presented in the same location, eye movements are minimized. It has been demonstrated that with the use of RSVP reading speed can be accelerated without severe decrements in comprehension. For this reason, it has been considered an educational tool for speed reading and as a potential technique to present text messages in small electric devices such as cell phones, PDAs, and even wrist watches [Chien and Chen, 2007; Goldstein et al., 2003; Muter, 1996]. In addition, it is used as a reading aid for those with visual limitations such as patients with bilateral macular disease [Beccue and Vila, 2004; Rubin and Turano, 1992]. RSVP has also led to the development of various experimental reading methods such as the self‐paced reading and the moving window presentation paradigm [Just et al., 1982; Kieras and Just, 1984].

Although RSVP was originally developed for use in behavioral studies, it has been adopted for use in event‐related potential (ERP) studies because it makes it possible to time lock to a particular word within a sentence [Ditman et al., 2007; Hagoort and Brown, 2000; King and Kutas, 1995; Kutas and Hillyard, 1984; Neville et al., 1991]. Recent neuroimaging studies have also employed this paradigm for language comprehension and sometimes it allows for synchronization between ERP and fMRI data [Caplan et al., 2002; Cooke et al., 2002; Cutting et al., 2006; Fiebach et al., 2005; Hashimoto and Sakai, 2002; Kiehl et al., 2002]. Thus, RSVP has a number of advantages forresearch as well as commercial and educational purposes.

Beyond its practical uses, a critical question is whether underlying cognitive processing (e.g., language and memory processing) during RSVP is similar to whole sentence presentation. The physical characteristics of the two are quite different in that gaze time on each word is not under the reader's control and regression is not allowed in RSVP. Because of these differences, RSVP reading can be uncomfortable and may increase cognitive demands such as attention and working memory. Nevertheless, RSVP is believed to not be very different from whole sentence presentation, at least in terms of language processing (for a review, see Potter [1984]). However, the evidence that has been used to support this idea is based on the comprehension of short, simple sentences. During the processing of these sentences no severe decrements in comprehension have been observed, not even at much faster presentation rates (e.g., 12 words per second (wps)) than normal reading speed (e.g. 6 wps), [Beccue and Vila, 2004; Juola et al., 1982; Potter, 1984; Rubin and Turano, 1992; Ward and Juola, 1982].

Although no comprehension decrements have been observed for short sentences, this is not necessarily true for longer reading materials such as passages. Potter et al. [1980] compared RSVP and whole sentence/text presentation for a paragraph with the approximately same total presentation time, using rates of 4, 8, or 12 wps. At a slow presentation rate, 4 wps, they reported that the percentage of idea units recalled in the first half of the paragraph in the RSVP condition was significantly lower in the whole sentence/text presentation condition. They also examined the effect of topic sentence on memory by placing the topic sentence at the beginning, middle, or end of the paragraph. Interestingly, recall rates during RSVP were significantly affected by this factor, but it was not a factor during whole sentence/text reading. When a topic sentence was placed at the beginning, better performance was observed in the RSVP condition, although performance was still worse than that observed in whole sentence/text presentation. Also, Rusell and Chaparro [2001] presented 18 passages using RSVP at 4 wps and 11 wps with various font sizes and found reading comprehension at 11 wps was significantly lowered. Recently, similar results were found with Chinese languages [Chien and Chen, 2007; Lin and Shieh, 2006]. Although Potter [1984] suggested that the decrement for RSVP for longer passages is due to lack of memory consolidation or deeper meaning processes [Potter, 1984; pp. 110–111], a comprehensive explanation has not been presented.

In addition to the length of a text, linguistic complexity (e.g., syntactic complexity) of reading materials also affects RSVP reading and it has been found to interact with reading speed. In fact, initial findings by Forster [1970] revealed this effect. Linguistically complex sentences (such as passive sentences, sentences with an embedded clause) produced higher recall errors compared to simple active sentences when the number of words was equated; and the recall errors were worse at faster presentation rates [Forster, 1970; Forster and Olbrei, 1972; French, 1981; Holmes and Forster, 1972]. This finding has important implications for studies examining syntactic complexity. As a matter of fact, syntactically complex sentences typically increase online processing demands including working memory [Caplan and Waters, 1999; Just et al., 1996a, b]. Although the characteristic of the working memory demand is still under debate (domain‐general vs. domain‐specific [Caplan and Waters, 1999; Fedorenko et al., 2006; Fiebach et al., 2001, 2005; Just et al., 1996a, b; King and Kutas, 1995; MacDonald et al., 1992; Waters and Caplan, 1996], the fact that syntactically complex sentences require more computational resources as does RSVP may suggest that RSVP may interfere with syntactic processing.

In the current event‐related fMRI study, we aimed to explore the effect that the presentation paradigm employed has on syntactic processing. To accomplish this goal, conjoined active and object‐relative sentences were compared and two presentation paradigms were examined—RSVP using a moderate reading speed (2.5 wps) and whole sentence presentation. Each sentence was followed by a true/false comprehension probe. All probes were presented using the whole sentence presentation paradigm and they all were simple active sentences. To examine the activation related to sentence processing separately from that associated with the comprehension probe a 6‐s delay was placed between the sentence and probe. On the basis of previous literature, our focus was on two regions in particular, BA 44 and the posterior S/MTG. Both of these regions have been implicated in syntactic processing, and the goal here was to determine whether their activation is modulated by the differing demands of the presentation mode.

METHODS

Participants

Twenty participants took part in the experiment. They were all Indiana University students without any history of neurological disorders. Before scanning, all participants gave written informed consent that was approved by the Indiana University Institutional Review Board. Data from two participants were discarded due to poor behavioral performance during the fMRI session. Data from 18 participants (9 male, 9 female, age = 22 ± 1.87) were used for the current data analysis. In a training session, participants were also introduced to the sentence comprehension task and underwent several practice trials to familiarize them with the experimental procedure. After the experiment, they completed a debriefing questionnaire.

Experimental Design

The current fMRI experiment used a single trial event‐related design in which each trial was treated as an event block [Postle et al., 2000; Zarahn, 2000; Zarahn et al., 1997]. A trial consisted of a sentence, a 6‐s delay, and a comprehension probe. The study was a 2 × 2 design with sentence presentation mode and syntactic complexity as within‐subject factors. Two presentation modes; the whole sentence presentation and RSVP were used to display sentence materials. Sentences were presented for 5 s in their entirety in the middle of the screen for whole sentence presentation. During RSVP, sentences were presented one word at a time in the middle of the screen; each sentence was presented in 5 s with an average presentation rate of 2.4 wps. This presentation rate was slightly slower than normal reading speed. The second factor, syntactic complexity, was manipulated by using two sentence types, conjoined active and object‐relative sentences (see later). Sentence materials were adapted from Keller et al. [2001] in which all the sentences consisted of 12 words, and frequency and length of words in sentences were equated across conditions. There were 80 sentences presented (40 conjoined active sentences and 40 object‐relative sentences) and half of sentences in each type were presented using each of the two presentation modes.

  • Conjoined active
    • Sentence: The pilot scared the escort and broke the mirror on the closet.
    • Probe: The pilot broke the mirror.
  • Object‐relative
    • Sentence: The pilot that the escort scared broke the mirror on the closet.
    • Probe: The pilot broke the mirror.

To verify reading comprehension of the sentence, a short probe followed the sentence after a 6‐s delay. The probe sentence had a simple SVO structure with one verb and two nouns, all taken from the sentence and the probe described one of the events in the sentence. One third of the probes were false (e.g. “the escort broke the mirror”). Probes were always displayed using the whole sentence presentation paradigm regardless of sentence presentation; along with the probe response cue (F | T) was presented in the middle of screen. Participants were to press a button with their left index finger to denote false and a button with the right index finger to denote true. So, neither syntactic complexity nor presentation mode was manipulated in the comprehension probe.

One of the important elements in the current fMRI design is the 6‐s delay. The 6‐s delay was inserted to obtain separate estimations of the blood oxygen level dependent (BOLD) response from the sentence reading phase and the probe phase, somewhat separately. Because we manipulated syntactic complexity and presentation mode in only the sentences and not the probe, we wanted to investigate these effects on the brain activation during sentence reading separately from the effects of responding to the comprehension probe. Thus, we tested two time delays (3 and 6 s) in a couple of pilot studies and verified that the 6‐s delay allowed for the better separation of the hemodynamic response peaks for each phase (you can see this activation pattern in Fig. 4). With the inserted delay, the present experimental design became similar to the fMRI design for a delayed matching task [e.g. Postle et al., 2000; Zarahn et al., 1997]. Zarahn et al. [1997] described an fMRI data analysis procedure for estimating separate hemodynamic responses for each component in a trial. Thus, the current fMRI data analysis followed the method in Zarahn et al. [1997]. After the probe, a 12‐s rest denoted by an asterisk (*) was inserted between trials to allow the BOLD signal to return to the baseline. In Figure 1, the experimental design is summarized.

Figure 4.

Figure 4

Open in a new tab

The extracted timecourses for each of the conditions for the two regions of interest, BA44 and MTG. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Figure 1.

Figure 1

Open in a new tab

A single event‐related fMRI design and examples of sentences. Slash(/) represents the unit of RSVP.

The fMRI scan consisted of four functional sessions with 20 trials (five trials in four conditions) presented in a pseudorandom order. In each session, three 28‐s fixation periods were included in the beginning, middle, and end of the session to estimate a baseline BOLD signal. To remind participants how to perform the task, two additional practice trials were inserted in the beginning of the first session, but the scans from the practice trials were discarded from the data analysis.

fMRI Data Acquisition and Analysis

fMRI scanning was conducted on a 3T Siemens TRIO scanner with an 8‐channel radio frequency head coil located in the Imaging Research Facility at Indiana University. Functional images were obtained in eighteen oblique axial slices with 5 mm thickness and a 1 mm gap (TR = 1,000 ms, TE = 25 ms, flip angle = 60°, matrix size = 64 × 64, FOV = 240 × 240 mm2) using a gradient echo planar imaging (EPI) sequence. Before statistical analysis, for all of the functional images, slice timing correction, head motion correction by realignment, and spatial normalization were conducted using the SPM99 software (Wellcome Department of Imaging Neuroscience; http://www.fil.ion.ucl.ac.uk/spm). In the spatial normalization step, all of the functional images were warped directly to the Montreal Neurological Institute (MNI) EPI template and resampled to the 2 × 2 × 2 voxel dimensions. The spatially normalized images were entered for the statistical analysis based on the general linear model (GLM) and the Gaussian random field theory, which were implemented in the SPM package [Friston et al., 1995]. Zarahn et al.'s [1997] method was applied for the statistical analysis. A regressor was built for the sentence reading, the delay, and the probe phase, respectively. Each regressor was constructed by convolving a canonical hemodynamic response function (HRF) with a stimulus delta function. The stimulus delta function was created with stimulus onset time and duration. For the sentence reading, the beginning time of each sentence presentation was set for the onset and the total presentation time, 5 s, was entered for the duration. A brief event HRF was setup in the middle of the delay for the delay period activation. For the probe phase, the onset of probes and 5 s were entered for duration because the probe was also presented for 5 s. Six realignment parameters and one additional regressor for incorrect response trials, if there were any, were entered in the model to remove contaminating effects from head motion and careless responses. Therefore, only correct trials were used in the analysis.

Individual activation maps for the sentence reading and probe phases in each condition were achieved by contrasting each phase to the fixation condition. On the basis of those contrasted images, in the second level, conjunction analysis and within participant ANOVAs were performed. The conjunction analysis was conducted to examine the common activation for sentence reading in both presentation modes [Friston et al., 1999, 2005; Nichols et al., 2005]. The main effects of syntactic complexity and presentation and the interaction between the two factors were examined for the sentence reading and the probe phase. The main effect of presentation was examined using a corrected threshold of P < 0.05 using family wise error correction. To examine the syntactic complexity effect, the entire brain was searched using an uncorrected threshold of P < 0.001 with 25 voxel extent threshold. Because the interaction is a relatively small effect, the entire brain was not searched. Instead the syntactic complexity maps were used as an inclusive mask (the mask threshold was an uncorrected P < 0.0001). Because the cluster size found from this analysis was small, a region of interest (ROI) analysis was performed. For the ROI analysis, a 10 mm radius, spherical ROI whose center was the activation peak from the conjunction analysis map for sentence reading was constructed. The two ROIs of primary interest were the LIFG and LMTG and we ensured that they were included in the ROIs generated. Then, timecourse data were extracted from individual datasets by using the Marsbar toolbox [Brett et al., 2002]. Individual timecourse data were truncated by each trial and sorted by conditions and averaged across trials in for each condition. Timecourse data from incorrect trials were excluded. The averaged timecourse data at each time point was converted into a percent signal change (PSC) value using the formula, (signal − baseline/baseline) × 100, where the baseline was the mean signal of the fixation period. Finally, the PSC was baseline corrected. The timecourse analysis was also applied to other regions in which the main effect of presentation was found (see Fig. 6).

Figure 6.

Figure 6

Open in a new tab

Presentation effects in the sentence reading phase. Right: presentation effects overlaid in an axial slice (z = −2) of the canonical brain. Left: time course of percent signal change (PSC) in (A) the right MT(V5) (x = 44, y = −70, z = 0), (B) the left hippocampus (x = −22, y = −30, z = −4), (C) calcarine sulcus (V1) (x = 4, y = −68, z = 6).

RESULTS

Behavioral Results

Behavioral responses to comprehension probes were collected during the scan. The average response time (RT) for correct trials (incorrect trials were excluded for the RT analysis) was used to calculate the RT, and the number of errors was used to calculate accuracy. The data were analyzed using a 2 × 2 (presentation mode × syntactic complexity) within participant ANOVA for RT and accuracy. As for RT, a main effect of syntactic complexity [F(1,17) = 22.99, P < 0.0001], was found but neither a main effect of presentation nor an interaction was found. The main effect of syntactic complexity indicated that the RT was significantly faster for conjoined active than the object relative sentences (see Fig. 2). However, when examining accuracy, the main effect of syntactic complexity [F(1,17) = 11.83, P < 0.003], the main effect of presentation mode [F(1,17) = 37.7 P < 0.0001], and the interaction between the two factors [F(1,17) = 7.52, P < 0.014] were all found to be significant. The RSVP condition resulted in more errors on average and showed a larger syntactic complexity effect than the whole sentence condition, see Figure 2. In fact, most of participants stated in their debriefings that RSVP reading was more difficult compared to whole sentence reading.

Figure 2.

Figure 2

Open in a new tab

Behavioral responses to comprehension probes (error bars represent standard errors).

fMRI Results

The current fMRI experimental design allowed us to estimate BOLD responses for the sentence reading phase and those for the comprehension probe phase separately (see Methods). We first examined the syntactic complexity effect during each phase as well as how this effect is modulated by presentation mode, with a focus on BA 44 and posterior S/MTG. Generally, we found a syntactic complexity effect and an interaction during the sentence phase in left BA 44 (see Table I and Fig. 3). We also found a similar effect during the probe phase in left posterior S/MTG (see Table II and Fig. 4). We then examined the common and differential activation during the sentence reading phase for the two presentation paradigms. There were regions showing differential activation between presentation modes (Table III and Fig. 5) as well as a number of regions that revealed common activation during sentence reading for both presentation modes (Table IV and Fig. 6). Below, we summarized each of these effects in detail.

Table I.

Main effect of syntactic complexity (object‐relative minus conjoined active) and interaction between presentation mode and syntactic complexity in the sentence reading phase (uncorrected P < 0.001)

Area R/L Size t‐value Coordinate
Main effect of syntactic complexity
 Inferior frontal gyrus (BA44) L 139 4.77 −42, 4, 28
Interaction(More syntactic effect in whole)
 Inferior frontal gyrus (BA44) L 8 2.46 −48, 16, 30
Open in a new tab

Figure 3.

Figure 3

Open in a new tab

Syntactic complexity effects in the sentence reading phase. (A) Main effect of syntactic complexity. (B) Interaction: more syntactic effect during whole sentence presentation than RSVP in the IFG. (C) Simple main effect of syntactic complexity during whole sentence presentation. (D) Simple main effect of syntactic complexity (no effect) during RSVP.

Table II.

Main effect of syntactic complexity and interaction between presentation mode and syntactic complexity in the probe phase (uncorrected P < 0.001)

Area R/L Size t‐value Coordinate
Main effect of syntactic complexity
 Inferior frontal gyrus (BA45) L 103 4.69 −46, 30, −6
R 82 4.62 46, 15, 34
 Precentral gyrus (including BA44) L 1458 6.51 −38, 2, 60
 Insula L 133 4.65 −32, 20, −14
 Supplementary motor area L/R 313 5.51 −2, 8, 60
 Posterior middle temporal gyrus L 199 5.11 −62, −34, 2
 Inferior parietal lobule L 771 5.61 −28, −56, 40
R 207 4.58 38, −58, 38
 Precuneus (including superior parietal lobule) L/R 289 5.39 0, −70, 42
 Cerebellum R 212 6.55 8, −74, −30
93 4.70 40, −64, −36
Interaction I (more syntactic effect in whole)
 Posterior middle temporal gyrus L 42 3.01 −58, −44, 2
Interaction II (More syntactic effect in RSVP)
 Superior parietal lobule R 20 2.83 −24, −74, 52
 Cerebellum R 24 3.16 36, −60, −32
Open in a new tab

Table III.

Presentation mode effects in the sentence reading phase (FWE corrected P < 0.05)

Area R/L Size t‐value Coordinate
Whole‐RSVP
 Hippocampus L 220 22.66 −22, −30, −4
R 209 17.17 22, −32, 4
 Calcarine sulcus (V1/V2) L/R 11085 20.77 4, −68, 6
 Superior parietal lobule R 123 8.79 22, −62, 54
RSVP‐whole
 Middle temporal gyrus (V5) R 749 12.81 44, −70, 0
L 170 8.97 −46, −74, −2
 Superior temporal gyrus R 43 7.30 56, −38, 20
Open in a new tab

Figure 5.

Figure 5

Open in a new tab

Syntactic complexity effects in the probe phase. (A) main effects of syntactic complexity. (B) Simple main effects of syntactic complexity during whole sentence presentation. (C) Simple main effects of syntactic complexity during RSVP. (D) Interaction (E) averaged time course data in the left posterior MTG which showed the interaction effect.

Table IV.

Common activation during sentence reading in the whole and RSVP modes (FWE corrected P < 0.05)

Area R/L Size z‐value Coordinate
Inferior frontal gyrus (BA45/47) L 70 7.22 −34, 28, −4
Inferior frontal gyrus (BA44) L 64 6.48 −48, 16, 24
Precentral gyrus L 6 5.56 −46, −2, 54
L 7 5.51 −46, −10, 58
L 6 5.13 −42, −4, 56
Insula R 13 5.73 38, 18, 0
Supplementary motor area L 118 6.53 −2, 0, 60
R 10 5.41 8, 10, 48
Anterior middle temporal gyrus L 10 5.76 −56, −12, −14
Posterior middle temporal gyrus L 66 7 −56, −44, 4
Inferior temporal gyrus L 12 5.9 −40, −48, −18
Angular gyrus R 7 5.53 28, −64, 46
Inferior occipital gyrus L 217 6.85 −30, −90, −10
Calcarine sulcus R 165 6.48 20, −98, −6
Putamen L 103 6.16 −18, 8, 0
Open in a new tab

Syntactic complexity effects during the sentence reading phase

During the sentence reading phase, the only region to show a significant effect of syntactic complexity was the left inferior frontal gyrus (LIFG). The region also revealed an interaction between syntactic complexity and presentation mode (Table I and Fig. 3A). To take a closer look at the processing taking place in this region, the percent signal change (PSC) values were retrieved for this region (see Fig. 3B). The bar graph in Figure 3 presents PSC values at the activation peak (peaked at −48, 16, 30 in MNI coordinates) and Figure 4 presents the timecourse in the ROI across conditions. These figures reveal that the lack of syntactic complexity effect in RSVP reading was due to relatively more activation in the conjoined active condition for RSVP compared to whole sentence presentation. This interaction can be observed clearly when the simple main effect of syntactic complexity is calculated in each presentation mode (compare Fig. 3C,D). Although the left inferior frontal regions (BA 44 and BA45) showed the syntactic complexity effect during whole sentence reading, no region showed a significant effect in RSVP.

Syntactic complexity effects during the comprehension probe phase

All the comprehension probes were constructed using a simple SVO structure and presented using the whole sentence mode. As a result, all of the effects observed during the probe phase presumably resulted from factors manipulated in the sentence phase (i.e., complexity and presentation mode). Here, there was no main effect of presentation mode, but there were significant syntactic complexity effects and a significant interaction with presentation mode (Table II) in many regions. The main effect of syntactic complexity during the probe phase was found at the LIFG (BA 44) again, but it was also found in other regions including the left precentral gyrus, supplementary motor area, bilateral inferior parietal lobules, the left posterior MTG, precuneus, and right cerebellum (Table II and Fig. 5A).

An interaction was observed in the left posterior middle temporal gyrus and was due to a larger syntactic effect for whole sentence presentation compared to RSVP (Fig. 5D). The syntactic complexity effect was only significant for whole sentence presentation but not for RSVP (compare Fig. 5B,C). For further investigation, the averaged PSC values were extracted from a spherical ROI in this region (peaked at −58, −44, 2) and plotted in Figure 5E and in Figure 4. This plot revealed the relatively greater activation for the conjoined active sentences after RSVP reading, which was a very similar pattern observed in the LIFG during reading (see Figs. 3B and 4)

Other regions such the right precuneus and the right cerebellum also revealed an interaction, but the interaction was due to a larger syntactic complexity effect for RSVP compared to whole sentence presentation (see Table II).

Presentation mode effects during the sentence reading phase

Presentation mode effects were revealed by subtracting estimated hemodynamic responses for RSVP from those associated with whole sentence presentation and vice versa (Table III). One region that showed a critical difference between the presentation paradigms was the hippocampus. The disparate response in this region suggests there is differential memory processing associated with these two presentation paradigms. The bilateral hippocampi revealed little activation during RSVP but was strongly activated during whole sentence reading. In addition, the hippocampus was also activated during the probe phase in which probe statements were presented using the whole sentence presentation mode (Fig. 6B). Although early visual cortex (V1/V2) showed more activation for whole sentence than RSVP reading, bilateral MT (V5) was uniquely activated during RSVP (Fig. 6A,C). The activation in MT (V5) indicated that the rapidly presented words in the same position engaged motion‐related processes [He et al., 1998]. The differential activation in these two visual processing areas appears to be due to the different visual stimulus properties or differences in the amount of visual processing required for the presentation paradigms.

Common activation for sentence reading during whole sentence and RSVP reading

Conjunction analysis revealed that most of language regions were commonly activated during sentence reading (Table IV). The commonly activated regions included left dominant language areas such as the inferior frontal and precentral regions, the left anterior/posterior middle temporal region, and the inferior temporal region including the fusiform gyrus. The supplementary motor area, primary visual cortex, left putamen, insula, and right angular gyrus were also involved during both presentation paradigms (see Fig. 7).

Figure 7.

Figure 7

Open in a new tab

Common activation between the RSVP and the whole sentence presentation.

DISCUSSION

The current fMRI study aimed to determine whether syntactic processing was modulated by the sentence presentation paradigm used. Behaviorally, an expected syntactic complexity effect was found in response time and accuracy to comprehension probes for both RSVP and whole sentence presentation. However, the interaction observed in the accuracy data indicated that RSVP elicited more errors, particularly for the syntactically complex sentences. Differential syntactic complexity effects were found between the two presentation paradigms. The syntactic complexity effect in Broca's area was only significant during whole sentence presentation. A similar interaction was observed in Wernicke's area when processing the probes. Conversely, a larger syntactic effect was found in the precuneus and right cerebellum when responding to the comprehension probe after RSVP, seemingly corresponding to the higher error rates elicited by this presentation paradigm. In addition, the imaging data revealed both common and differential activation for RSVP and whole sentence presentation. Both presentation paradigms elicited activation in the left dominant language areas including Broca's and Wernicke's areas. However, differential activation was found in the hippocampus and the visual pathway. For example, the hippocampus revealed very little involvement during RSVP but was significantly activated during whole sentence presentation.

Sentence Processing Phase

Most psycholinguistic research has agreed that language processing is fast and autonomous. The common activation across these two presentation paradigms in language processing areas such as Broca's and Wernicke's area support this claim. However, RSVP and whole sentence presentation appears to differentially affect syntactic processes. Syntactic complexity interacted with the sentence presentation paradigm in several regions either during the reading of the sentence or when answering the comprehension probes. The interaction during sentence reading occurred in the LIFG. This region has been strongly associated with syntactic processing, and it has consistently showed syntactic complexity effects in many neuroimaging studies [Caplan et al., 1998, 1999; Caplan and Waters, 1999; Carpenter et al., 1999; Cooke et al., 2006; Fiebach et al., 2001, 2005; Friederici et al., 2006; Just et al., 1996b; Keller et al., 2001; Newman et al., 2003]. In particular, the region has been thought to be involved in the processing of word order information and is more activated when the word order within a sentence is not the typical order (e.g., SVO in English) [Grodzinsky and Friederici, 2006; Grewe et al., 2005]. Here, we observed the syntactic complexity effect in the LIFG during whole sentence reading but not during RSVP. The lack of a syntactic complexity effect during RSVP was consistent with some previous results [Caplan et al., 2002; Cooke et al., 2002]. When the signal change measures were retrieved from the region, the signal for the conjoined active sentences during RSVP was larger than the signal in the whole sentence condition. This suggests that the computational demands associated with RSVP may have interfered with the readers' ability to take full computational advantage of the simpler syntactic structure. One possible explanation is that because the sentence is presented serially during RSVP, the words and their order must be maintained in a buffer. Therefore, the serial order processing linked with RSVP may interfere with the word order processing required during the processing of syntactically complex (e.g., sentences not in the typical SVO word order) sentences.

As implied earlier, one possible source of the increased computational demand associated with RSVP is memory processing. Almost 30 years ago, Potter et al. [1980] observed memory decrements when the stimuli was presented using RSVP (i.e., a substantial drop in recall and an interaction with the location of the topic within a paragraph). She later proposed that there may be an impairment of memory consolidation during RSVP resulting in the meaning of the stimuli not being deeply processed [Potter, 1984]. However, to our knowledge, there is no subsequent research on this topic. This study in some way replicated the interaction with memory in that here we found an interaction between presentation paradigm and syntactic complexity. In addition, differential hippocampal activation for RSVP and whole sentence presentation was observed such that the whole sentence condition elicited a larger response from the hippocampus than did RSVP. The hippocampus has long been implicated in memory consolidation and episodic memory formation [Eichenbaum, 2004; McGaugh, 2000; Squire and Alvarez, 1995; Tulving and Markowitsch, 1998]. Eichenbaum [2004] defined three elemental cognitive processes linked to the hippocampus: associative representation, sequential organization, and relational networking. Davachi and Wagner [2002] found that the hippocampus was engaged during relational and item‐based learning, and it seemed to be involved in the relational binding of items into an integrated memory system. It may be that this binding process is limited during RSVP. Because the words are presented serially, requiring the use of a computationally, resource demanding internal working memory buffer to perform the relational binding process, this binding process may be more difficult to perform or in this case may not be performed at all during RSVP. During whole sentence presentation, on the other hand, the display is an external buffer, freeing up resources to allow for this binding process to occur. Thus, the lack of hippocampal activation in RSVP may coincide with an impairment of memory consolidation and may result in poor performance, particularly for the more computationally demanding object‐relative constructions.

All together it seems that RSVP is more computationally, resource demanding and interferes with syntactic processing. Again, BA 44 is thought to be involved in processing linear word order so the word order processing required for RSVP may be interfering with the word order processing required to comprehend more syntactically complex sentences. This syntax related order processing may interact with the binding processes associated with the hippocampus that is required to obtain the deeper semantic meaning of the sentence. Therefore, if the syntax‐based word ordering is not performed, then no binding is performed and no activation of the hippocampus is elicited. Although this explanation fits the current data, it is speculative and requires further research to validate it.

Thus far, we have argued that the lack of a syntactic complexity effect observed during RSVP in Broca's area is due to an interaction with memory processing. However, there is an alternative explanation. Both of the critical sentence types used are locally ambiguous: at the position of “that,” the relative clause constructions could continue either as an object relative clause (“The pilot that the escort scared broke the mirror on the closet,” as in this study) or as a subject relative clause (“The pilot that scared the escort and broke the mirror on the closet.”), whereas, at the position of “and,” the conjoined actives permit either a VP‐coordination (“The pilot scared the escort and broke the mirror on the closet,” as in this study), an NP‐coordination (“The pilot scared the escort and the admiral.”) or S‐coordination (“The pilot scared the escort and the admiral raised the alarm.”). In both sentence types, the ambiguous region (“that” or “and”) is a short word and, therefore, relatively unlikely to be fixated during natural reading [cf. Rayner, 1998]. Hence, the disambiguating information, which is provided by the immediately following word, will have been available to the reader much more rapidly during whole sentence presentation than during RSVP. As a result, RSVP may have enhanced the effects of local ambiguity during the comprehension process and elicited precisely the activation pattern observed for the LIFG during the reading phase: a general, ambiguity‐related activation increase for both sentence types in comparison to whole sentence reading. For whole sentence reading, by contrast, there is no (or only very little) ambiguity, allowing the activation pattern to reflect the general complexity difference between object relative and conjoined active sentences. Furthermore, this alternative explanation is consistent with previous reports of ambiguity‐related activation in the LIFG [e.g. Bornkessel et al., 2005; Fiebach et al., 2004; Stowe et al., 2004].

Probe Processing Phase

Interestingly, the syntactic complexity effect was observed, and more widespread, during the processing of the probes. This effect is quite interesting because the probe statements were constructed with a simple SVO structure so the effect must be the result of reprocessing the sentence via the mental representation that was generated during sentence reading. This syntactic reanalysis taking place during the probe appears to not be affected by how the sentences were presented because the effect was consistent for both presentation paradigms. In a recent study, we reported that syntactic complexity effects during the probe phase interacted with probe type [Newman et al., in press]. There, probe sentences all had the same structure, but the distance (in the preceding sentence) between the verb and noun was manipulated. For example, for a conjoined active sentence such as “the pilot scared the escort and broke the mirror on the closet,” true probe sentences could be either “the pilot scared the escort” or “the pilot broke the mirror.” It was found that the first type of probe was easier to answer and the brain activity in several regions including LIFG revealed lower activation levels compared to the second probe type. Although the distance was not manipulated here, the finding of syntactic complexity effects off‐line (during the probe) in language processing regions has been replicated here and provides further support for the idea that Broca's area is indeed associated with syntactic processing, both during sentence reading as well as when responding to the comprehension probe.

While Broca's area revealed a consistent effect of complexity during the probe for both sentence presentation modes, other regions revealed an interaction between complexity and presentation during the probe phase. For example, the left posterior MTG revealed a complexity effect during whole sentence presentation but not during RSVP. This interaction pattern during the probe phase in the MTG was similar to the interaction observed in the LIFG during the sentence reading phase—a complexity effect for whole sentence but not RSVP. Actually, the current activation patterns in the left posterior MTG is very interesting for several reasons. First, this region is commonly activated for both RSVP and whole sentence presentation indicating this region is essential for sentence processing. Second, the syntactic complexity effect in this region was not found during the reading phase (i.e. online sentence processing) but was observed during the probe phase (i.e., off‐line processing). This supports the claim made by Caplan [2001] in which he argues that the syntactic complexity effect in Wernicke's area is not due to the online syntactic processing but may be due to off‐line task demands. Third, the syntactic complexity effect in the probe phase was only significant when sentences were presented using the whole sentence presentation paradigm. According to recent neuroimaging studies, the posterior temporal region is involved in the integration of lexical and syntactic information and structured knowledge such as propositional or thematic concepts [Grodzinsky and Friederici, 2006; Grossman et al., 2002; Hart and Kraut, 2007; Kable et al., 2005]. On the basis of this, we may consider again the syntactic complexity effect observed in this region. The comprehension probes used here queried the thematic knowledge structure of the previous sentence (e.g., who did what to whom). To answer the probes, the mental representation of the previously read sentence should be retrieved. Given that complexity effects are observed during the probe, this suggests that the representation generated contains not just sentence meaning information but also structural information that may make determining those thematic relationships a more demanding process in object‐relative compared to conjoined active sentences. However, this was not the case when sentences were presented using RSVP; there was no complexity effect when processing the probe in temporal cortex. As stated earlier, because of the nature of RSVP it may have impeded both the word order processing required during non‐canonical structures in BA 44, but also the binding or integration processes thought to be associated with the hippocampus. This may mean that responding to the probe was based almost completely on sentence word order information instead of the “deeper” meaning of the sentence. If the sentence word order is being used to respond to the probe, then more errors would be expected for the object‐relative compared to the conjoined active sentences because the conjoined active sentences can be answered based solely on word order.

Limitations of Current Research and Implications for Future Research

This study would be the first neuroimaging study to compare RSVP and whole sentence presentation paradigms. The results indicate a critical difference between the two presentation methods in terms of memory processing and how they interact with syntactic complexity. Although the study does account for the lack of a syntactic complexity effect observed in some studies, it does appear to conflict with others that do show a complexity effect using RSVP [e.g., Cooke et al., 2002]. However, when one looks more closely at these studies the presentation paradigm used is just one methodological difference across these studies. For example, the Cooke study required a judgment regarding the gender of the agent and allowed participants to respond as soon as they knew the answer causing the termination of the trial and the advancement to the next trial. Therefore, it is likely that many participants did not read the entire sentence. To obtain a clear picture of the processing taking place during sentence processing and to compare across studies, it will be necessary to understand how all of these methodological differences interact with processing.

The study provides important messages for both research and commercial applications of RSVP, there are some limitations of the study. One such limitation is with the use of the 6‐s delay between sentence and probe. This was done so that we could separate the sentence‐related and probe‐related activation. Although we were able to do that we also introduced a possible confound. It may be that the processing taking place during sentence reading is affected by the need to retain the sentence in memory for 6 s before being presented with the comprehension probe. This additional memory processing may have contaminated the results. However, given that the need to retain the information is the same for both RSVP and whole sentence presentation it is our claim that it would not have a major effect when examining condition effect size differences.

Another limitation to this study is that we only examined two sentence presentation paradigms that are routinely used in psycholinguistic research. There are others such as the moving window paradigm in which stimuli are presented serially but in the space to the right of the previous word instead of the same location as the previous word. There it would be expected that the working memory load would be less than with RSVP due to the ability to make regressions back to the location of a previously presented word. But it is unclear how this paradigm, or others such as phrase by phrase instead of word by word presentation, may interact with syntactic complexity. Further study is needed to help provide a more complete characterization of the influence that the presentation paradigm employed has on language processing.

A final limitation of the study is related to the hippocampus activation—the hippocampus lies incredibly close to the LGN (lateral geniculate nucleus), part of the visual pathway. The activity of the LGN has been shown to be modulated by saccadic eye movements [Reppas et al., 2002; Thilo et al., 2004]. Given that RSVP limits eye movements and whole sentence presentation allows for both forward and regressive eye movements, activation differences in the LGN are possible. That being said, the coordinates for the activation maxima (and the ROI center) are clearly within the hippocampus. However, we cannot rule out the possibility that the activation attributed to the hippocampus did not extend into the LGN.

The results presented here may have several important implications on research and commercial usage of RSVP. First, examining more closely how the memory load engendered by RSVP affects language processing more generally requires further investigation. RSVP has been frequently used in many research areas and has also been considered for use in commercial products. If it has a serious impact on memory, it must be very carefully considered before being commercially used. Second, while RSVP and whole sentence presentation showed common activation in the typical language areas, the current results also showed that they interacted with syntactic complexity in different ways. One major implication is that the syntactic complexity effect could be reduced when using RSVP. This is important for psycholinguistic research, particularly for ERP and fMRI studies that use this presentation paradigm routinely. In sum, although RSVP is a convenient tool for research and holds additional benefits (e.g. fast reading), the advantages and disadvantages of using RSVP should be carefully evaluated.

Acknowledgements

We would like to thank Thomas Burns, Tara Muratore, Kristen Ratliff, Andrea Sampson, and Benjamin Pruce for all of their help with data collection. We would also like to thank those who have provided wonderful insight on previous versions of this manuscript.

REFERENCES

  1. Beccue B, Vila J ( 2004): Assessing the impact of rapid serial visual presentation (RSVP): A reading technique Adv Distributed Syst. vol. 3061, Springer Berlin: Heidelberg, 42–53. [Google Scholar]
  2. Bornkessel I, Zysset S, Friederici AD, von Cramon Y, Schlesewsky M ( 2005): Who did what to whom? The neural basis of argument hierarchies during language comprehension. NeuroImage 26: 221–233. [DOI] [PubMed] [Google Scholar]
  3. Brett M, Anton JL, Valabregue R, Poline JB ( 2002): Region of interest analysis using an SPM toolbox (Abstract). Presented at the 8th International Conferance on Functional Mapping of the Human Brain, Sendai, Japan.
  4. Caplan D ( 2001): Functional neuroimaging studies of syntactic processing. J Psycholing Res 30: 297–320. [DOI] [PubMed] [Google Scholar]
  5. Caplan D, Waters GS ( 1999): Verbal working memory and sentence comprehension. Behav Brain Sci 22: 77–94; discussion 95–126. [DOI] [PubMed] [Google Scholar]
  6. Caplan D, Alpert N, Waters G ( 1998): Effects of syntactic structure and propositional number on patterns of regional cerebral blood flow. J Cogn Neurosci 10: 541–552. [DOI] [PubMed] [Google Scholar]
  7. Caplan D, Alpert N, Waters G ( 1999): PET studies of syntactic processing with auditory sentence presentation. Neuroimage 9: 343–351. [DOI] [PubMed] [Google Scholar]
  8. Caplan D, Vijayan S, Kuperberg G, West C, Waters G, Greve D, Dale AM ( 2002): Vascular responses to syntactic processing: Event‐related fMRI study of relative clauses. Hum Brain Mapp 15: 26–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Caplan D, Chen E, Waters G ( 2008): Task‐dependent and task‐independent neurovascular responses to syntactic processing. Cortex 44: 257–275. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Carpenter PA, Just MA, Keller TA, Eddy WF, Thulborn KR ( 1999): Time course of fMRI‐activation in language and spatial networks during sentence comprehension. Neuroimage 10: 216–224. [DOI] [PubMed] [Google Scholar]
  11. Chien Y‐H, Chen C‐H ( 2007): The use of dynamic display to improve reading comprehension for the small screen of a wrist watch In: Human Interface and the Management of Information. Methods, Techniques and Tools in Information Design; vol. 4557, Springer Berlin: Heidelberg, pp 814–823. [Google Scholar]
  12. Constable RT, Pugh KR, Berroya E, Mencl WE, Westerveld M, Ni W, Shankweiler D ( 2004): Sentence complexity and input modality effects in sentence comprehension: An fMRI study. NeuroImage 22: 11–21. [DOI] [PubMed] [Google Scholar]
  13. Cooke A, Zurif EB, DeVita C, Alsop D, Koenig P, Detre J, Gee J, Pinango M, Balogh J, Grossman M ( 2002): Neural basis for sentence comprehension: Grammatical and short‐term memory components. Hum Brain Mapp 15: 80–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Cooke A, Grossman M, DeVita C, Gonzalez‐Atavales J, Moore P, Chen W, Gee J, Detre J ( 2006): Large‐scale neural network for sentence processing. Brain and Language 96: 14–36. [DOI] [PubMed] [Google Scholar]
  15. Cutting LE, Clements AM, Courtney S, Rimrodt SL, Schafer JGB, Bisesi J, Pekar JJ, Pugh KR ( 2006): Differential components of sentence comprehension: Beyond single word reading and memory. NeuroImage 29: 429–438. [DOI] [PubMed] [Google Scholar]
  16. Davachi L, Wagner AD ( 2002): Hippocampal contributions to episodic encoding: Insights from relational and item‐based learning. J Neurophysiol 88: 982–990. [DOI] [PubMed] [Google Scholar]
  17. Ditman T, Holcomb PJ, Kuperberg GR ( 2007): An investigation of concurrent ERP and self‐paced reading methodologies. Psychophysiology 44: 927–935. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Eichenbaum H ( 2004): Hippocampus: Cognitive processes and neural representations that underlie declarative memory. Neuron 44: 109–120. [DOI] [PubMed] [Google Scholar]
  19. Fedorenko E, Gibson E, Rohde D ( 2006): The nature of working memory capacity in sentence comprehension: Evidence against domain‐specific working memory resources. J Mem Lang 54: 541–553. [Google Scholar]
  20. Fiebach CJ, Schlesewsky M, Friederici AD ( 2001): Syntactic working memory and the establishment of filler‐gap dependencies: Insights from ERPs and fMRI. J Psycholinguist Res 30: 321–338. [DOI] [PubMed] [Google Scholar]
  21. Fiebach CJ, Schlesewsky M, Bornkessel I, Friederici AD ( 2004): Distinct neural correlates of legal and illegal word order variations in German: How can fMRI inform cognitive models of sentence processing In: Carreiras V, Clifton C, Jr, editors. The On‐Line Study of Sentence Comprehension. New York: Psychology Press; pp 357–370. [Google Scholar]
  22. Fiebach CJ, Schlesewsky M, Lohmann G, von Cramon DY, Friederici AD ( 2005): Revisiting the role of Broca's area in sentence processing: Syntactic integration versus syntactic working memory. Hum Brain Mapp 24: 79–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Forster KI ( 1970): Visual perception of rapidly presented word sequences of varying complexity. Perception Psychophys 8: 215–221. [Google Scholar]
  24. Forster KI, Olbrei I ( 1972): Semantic heuristics and syntactic analysis. Cognition 2: 319–347. [Google Scholar]
  25. French P ( 1981): Semantic and syntactic factors in the perception of rapidly presented sentences. J Psycholing Res 10: 581–591. [Google Scholar]
  26. Friederici AD, Fiebach CJ, Schlesewsky M, Bornkessel ID, von Cramon DY ( 2006): Processing linguistic complexity and grammaticality in the left frontal cortex. Cereb Cortex 16: 1709–1717. [DOI] [PubMed] [Google Scholar]
  27. Friston KJ, Frith CD, Turner R, Frackowiak RS ( 1995): Characterizing evoked hemodynamics with fMRI. Neuroimage 2: 157–165. [DOI] [PubMed] [Google Scholar]
  28. Friston KJ, Holmes AP, Price CJ, Buchel C, Worsley KJ ( 1999): Multisubject fMRI studies and conjunction analyses. Neuroimage 10: 385–396. [DOI] [PubMed] [Google Scholar]
  29. Friston KJ, Penny WD, Glaser DE ( 2005): Conjunction revisited. Neuroimage 25: 661–667. [DOI] [PubMed] [Google Scholar]
  30. Goldstein M, Öquist G, Björk S ( 2003): Evaluating sonified rapid serial visual presentation: An immersive reading experience on a mobile device In: Universal Access Theoretical Perspectives, Practice, and Experience vol. 2615, Springer Berlin: Heidelberg, pp 508–523. [Google Scholar]
  31. Grewe T, Bornkessel I, Zysset S, Wiese R, von Cramon DY, Schlesewsky M ( 2005): The emergence of the unmarked: A new perspective on the language‐specific function of Broca's area. Hum Brain Mapp 26: 178–190. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Grodzinsky Y, Friederici AD ( 2006): Neuroimaging of syntax and syntactic processing. Curr Opin Neurobiol 16: 240–246. [DOI] [PubMed] [Google Scholar]
  33. Grossman M, Koenig P, DeVita C, Glosser G, Alsop D, Detre J, Gee J ( 2002): Neural representation of verb meaning: An fMRI study. Hum Brain Mapp 15: 124–134. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Hagoort P, Brown CM ( 2000): ERP effects of listening to speech compared to reading: The P600/SPS to syntactic violations in spoken sentences and rapid serial visual presentation. Neuropsychologia 38: 1531–1549. [DOI] [PubMed] [Google Scholar]
  35. Hart J, Kraut M ( 2007): Neural Basis of Semantic Memory. Cambridge: Cambridge University Press. [Google Scholar]
  36. Hashimoto R, Sakai KL ( 2002): Specialization in the left prefrontal cortex for sentence comprehension. Neuron 35: 589–597. [DOI] [PubMed] [Google Scholar]
  37. He S, Cohen ER, Hu X ( 1998): Close correlation between activity in brain area MT/V5 and the perception of a visual motion aftereffect. Curr Biol 8: 1215–1218. [DOI] [PubMed] [Google Scholar]
  38. Holmes VM, Forster KI ( 1972): Perceptual complexity and underlying sentence structure. J Verbal Learn Verbal Behav 11: 148–156. [Google Scholar]
  39. Juola JF, Ward NJ, McNamara T ( 1982): Visual search and reading of rapid serial presentations of letter strings, words, and text. J Exp Psychol Gen 111: 208–227. [Google Scholar]
  40. Just MA, Carpenter PA, Woolley JD ( 1982): Paradigms and processes in reading comprehension. J Exp Psychol Gen 111: 228–238. [DOI] [PubMed] [Google Scholar]
  41. Just MA, Carpenter PA, Keller TA ( 1996a): The capacity theory of comprehension: new frontiers of evidence and arguments. Psychol Rev 103: 773–780. [DOI] [PubMed] [Google Scholar]
  42. Just MA, Carpenter PA, Keller TA, Eddy WF, Thulborn KR ( 1996b): Brain activation modulated by sentence comprehension. Science 274: 114–116. [DOI] [PubMed] [Google Scholar]
  43. Kable JW, Kan IP, Wilson A, Thompson‐Schill SL, Chatterjee A ( 2005): Conceptual representations of action in the lateral temporal cortex. J Cogn Neurosci 17: 1855–1870. [DOI] [PubMed] [Google Scholar]
  44. Keller TA, Carpenter PA, Just MA ( 2001): The neural bases of sentence comprehension: A fMRI examination of syntactic and lexical processing. Cereb Cortex 11: 223–237. [DOI] [PubMed] [Google Scholar]
  45. Kiehl KA, Laurens KR, Liddle PF ( 2002): Reading anomalous sentences: An event‐related fMRI study of semantic processing. NeuroImage 17: 842–850. [PubMed] [Google Scholar]
  46. Kieras DE, Just MA ( 1984): New Methods in Reading Comprehension Research. Hillsdale, N.J.: L. Erlbaum Associates. [Google Scholar]
  47. King JW, Kutas M ( 1995): Who did what and when? Using word‐ and clause‐level erps to monitor working memory usage in reading. J Cogn Neurosci 7: 376–395. [DOI] [PubMed] [Google Scholar]
  48. Kutas M, Hillyard SA ( 1984): Brain potentials during reading reflect word expectancy and semantic association. Nature 307: 161–163. [DOI] [PubMed] [Google Scholar]
  49. Lin Y‐C, Shieh K‐K ( 2006): Reading a dynamic presentation of Chinese text on a single‐line display. Displays 27: 145–152. [DOI] [PubMed] [Google Scholar]
  50. MacDonald MC, Just MA, Carpenter PA ( 1992): Working memory constraints on the processing of syntactic ambiguity. Cogn Psychol 24: 56–98. [DOI] [PubMed] [Google Scholar]
  51. McGaugh JL ( 2000): Memory—A century of consolidation. Science 287: 248–251. [DOI] [PubMed] [Google Scholar]
  52. Michael EB, Keller TA, Carpenter PA, Just MA ( 2001): fMRI investigation of sentence comprehension by eye and by ear: Modality fingerprints on cognitive processes. Hum Brain Mapp 13: 239–252. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Muter P ( 1996): Interface design and optimization of reading of continuous text In: van Oostendorp H, de Mul A, editors. Cognitive aspects of electronic text processing. Norwood, NJ: Ablex; pp 161–180. [Google Scholar]
  54. Neville H, Nicol JL, Barss A, Forster KI, Garrett MF ( 1991): Syntactically based sentence processing classes: Evidence from event‐related brain potentials. J Cogn Neurosci 3: 151–165. [DOI] [PubMed] [Google Scholar]
  55. Newman SD, Just MA, Keller TA, Roth J, Carpenter PA ( 2003): Differential effects of syntactic and semantic processing on the subregions of Broca's area. Brain Res Cogn Brain Res 16: 297–307. [DOI] [PubMed] [Google Scholar]
  56. Newman SD, Lee D, Ratliff KL: Off‐line sentence processing: What is involved in answering a comprehension probe? Hum Brain Mapp (in press). [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Nichols T, Brett M, Andersson J, Wager T, Poline J‐B ( 2005): Valid conjunction inference with the minimum statistic. NeuroImage 25: 653–660. [DOI] [PubMed] [Google Scholar]
  58. Postle BR, Zarahn E, D'Esposito M ( 2000): Using event‐related fMRI to assess delay‐period activity during performance of spatial and nonspatial working memory tasks. Brain Res Protocols 5: 57–66. [DOI] [PubMed] [Google Scholar]
  59. Potter MC ( 1984): Rapid serial visual presentation (RSVP): A method for studying language processing In: Kieras DE, Just MA, editors. New Methods in Reading Comprehension Research. Hillsdale, NJ: L. Erlbaum Associates; pp 91–118. [Google Scholar]
  60. Potter MC, Kroll JF, Harris C ( 1980): Comprehension and memory in rapid sequential reading In: Nickerson R, editor.Attention and Performance, Vol. VIII Hillsdale, NJ: Erlbaum; pp 395–418. [Google Scholar]
  61. Rayner K ( 1998): Eye movements in reading and information processing: 20 years of research. Psychol Bull 124: 372–422. [DOI] [PubMed] [Google Scholar]
  62. Reppas JB, Usrey WM, Reid RC ( 2002): Saccadic eye movements modulate visual responses in the lateral geniculate nucleus. Neuron 35: 961–974. [DOI] [PubMed] [Google Scholar]
  63. Rubin GS, Turano K ( 1992): Reading without saccadic eye movements. Vision Res 32: 895–902. [DOI] [PubMed] [Google Scholar]
  64. Rusell MC, Chaparro BS ( 2001): Exploring effects of speed and font size with RSVP. Paper Presented at Human Factors and Ergonomics Society 45th Annual Meeting, Minneapolis, Minnesota.
  65. Squire LR, Alvarez P ( 1995): Retrograde amnesia and memory consolidation: A neurobiological perspective. Curr Opin Neurobiol 5: 169–177. [DOI] [PubMed] [Google Scholar]
  66. Stowe LA, Paans AMJ, Wijers AA, Zwarts F ( 2004): Activations of “motor” and other non‐language structures during sentence comprehension. Brain Lang 89: 290–299. [DOI] [PubMed] [Google Scholar]
  67. Thilo KV, Santoro L, Walsh V, Blakemore C ( 2004): The site of saccadic suppression. Nat Neurosci 7: 13–14. [DOI] [PubMed] [Google Scholar]
  68. Tulving E, Markowitsch HJ ( 1998): Episodic and declarative memory: Role of the hippocampus. Hippocampus 8: 198–204. [DOI] [PubMed] [Google Scholar]
  69. Vigneau M, Beaucousin V, Herv PY, Duffau H, Crivello F, Houd O, Mazoyer B, Tzourio‐Mazoyer N ( 2006): Meta‐analyzing left hemisphere language areas: Phonology, semantics, and sentence processing. NeuroImage 30: 1414–1432. [DOI] [PubMed] [Google Scholar]
  70. Ward NJ, Juola JF ( 1982): Reading with and without eye movements: Reply to Just, Carpenter, and Woolley. J Exp Psychol Gen 111: 239–241. [PubMed] [Google Scholar]
  71. Waters GS, Caplan D ( 1996): The capacity theory of sentence comprehension: Critique of Just and Carpenter (1992): Psychol Rev 103: 761–772. [DOI] [PubMed] [Google Scholar]
  72. Zarahn E ( 2000): Testing for neural responses during temporal components of trials with BOLD fMRI. NeuroImage 11: 783–796. [DOI] [PubMed] [Google Scholar]
  73. Zarahn E, Aguirre G, D'Esposito M ( 1997): A trial‐based experimental design for fMRI. NeuroImage 6: 122–138. [DOI] [PubMed] [Google Scholar]

Articles from Human Brain Mapping are provided here courtesy of Wiley

RESOURCES