Abstract

Grammatical class is a fundamental property of language, and all natural languages distinguish between nouns and verbs. Brain activation studies have provided conflicting evidence concerning the neural substrates of noun and verb processing. A major limitation of many previous imaging studies is that they did not disentangle the impact of grammatical class from the differences in semantic correlates. In order to tease apart the role of semantic and grammatical factors, we performed a functional magnetic resonance imaging study presenting Italian speakers with pictures of events and asked them to name them as 1) Infinitive Verb (e.g., mangiare [to eat]); 2) Inflected Verb (e.g., mangia [she/he eats]); and 3) Action Noun (e.g., mangiata [the eating]). We did not find any verb-specific activation. However, reliable left inferior frontal gyrus (IFG) activations were found when contrasting the Action Noun with the Infinitive Verb condition. A second-level analysis indicated then that activation in left IFG was greatest for Action Nouns, intermediate for Inflected Verbs, and least for Infinitive Verbs. We conclude that, when all other factors are controlled, nouns and verbs are processed by a common neural system. In the present case, differences in left IFG activation emerge as a consequence of increasing linguistic and/or general processing demands.

Introduction

The neural substrate of naming events has received substantially less attention than the neural network recruited in naming objects. However, the investigation of the neural network engaged in naming events is of interest not only because it can reveal differences between these domains of knowledge but also because it bears the potential to answer questions concerning the representation of grammatical class information.

It is well know that grammatical class plays a crucial role in language processing. For example, during language production, grammatical class affects all types of errors made by normal speakers (Garrett 1984). Event-related potential (ERP) experiments during language comprehension tasks showed that grammatical class violations in sentences produce specific responses as shortly as 200 ms after presentation (Friederici et al. 1996). In a study with nouns, verbs, and class-ambiguous words, verbs elicited a left-lateralized, frontal positivity between 200 and 400 ms after stimulus onset, when used in predictable verb context (Federmeier et al. 2000). ERP differences between verbs and nouns were also found between 300 and 800 ms after stimulus onset during a semantic priming task (Khader et al. 2003).

There is also evidence that grammatical class affects the performance of language disorder patients. A double dissociation between noun and verb processing is well documented in the neuropsychological literature, with some patients showing a disproportionate deficit with verbs compared with nouns (Miceli et al. 1988; Daniele et al. 1994; Miozzo et al. 1994; Rapp and Caramazza 1997) and others showing the reverse pattern (Zingeser and Berndt 1990; Damasio and Tranel 1993; Silveri et al. 2003). This double dissociation has been linked to lesions involving different anatomical sites. Patients with verb-specific impairments often have anterior lesions, usually involving the left prefrontal cortex (e.g., Daniele et al. 1994), whereas patients with noun-specific impairments generally show left temporal lesions (Damasio et al. 1993). Moreover, it has been reported that patients with frontotemporal dementia showed a greater action naming impairment compared with patients suffering from Alzheimer disease (Cappa et al. 1998).

Neuroimaging studies in normal subjects, however, have provided only limited support to these lesion-based hypotheses proposing a neural separation between noun and verb processing. The discrepancies among the studies may be attributed to a number of factors. Most important, the vast majority of previous patient and imaging studies investigating differences in noun and verb processing did not distinguish between grammatical class and the semantic correlates of grammatical class. Across languages, the distinction between nouns and verbs is rooted in the distinction between objects and actions (for a discussion, see Vinson and Vigliocco 2002; Vigliocco et al. 2004). Most previous studies used as stimuli nouns referring to objects and verbs referring to actions, introducing therefore a confound between grammatical and semantic class. In the imaging literature, this is the case not only in early studies using verb generation (Petersen et al. 1988, 1989; Martin et al. 1995; Warburton et al. 1996) but also in more recent studies using picture naming (Tranel et al. 2005).

Studies that attempted to lessen the semantic confound by using both concrete and abstract nouns and verbs provided mixed results. In a lexical decision study by Perani et al. (1999) in Italian using concrete and abstract nouns and verbs, verb-specific activations were reported in left inferior frontal gyrus (IFG) but no noun-specific activations were observed. This finding was replicated in English in a semantic decision study using inflected nouns and verbs (Tyler et al. 2004) but not in a lexical decision study using uninflected words (Tyler et al. 2001). Greater left IFG activations for verbs than nouns in these previous studies might have not reflected grammatical class differences, but morphosyntactic processes that may be more demanding for verbs than nouns, if not simply greater processing demands imposed by verbs than nouns (see Thompson-Schill et al. 1997; Binder et al. 2004; Vigliocco et al. 2006). These findings arise because previous studies showing verb-specific left IFG activation used explicit tasks, such as lexical (Perani et al. 1999) or semantic (Tyler et al. 2004) decision. It should also be noted that most languages have more morphologically Inflected Verb than noun forms (e.g., in English there are at least 4 different forms for verbs and only 2 for nouns), so task performance on verbs may place greater demands on selection and decision processes attributed to left IFG (Thompson-Schill et al. 1997; Gold and Buckner 2002; Binder et al. 2004), in a way that verb-specific activation may result from an interaction between grammatical class and task demands. This possibility has been tackled in a study by Longe et al. (2006). Greater activations of left IFG for verbs than nouns were observed when English speakers made semantic judgments on inflected words but not when they made judgments on the same words presented in an uninflected form. However, as discussed in Shapiro et al. (2006), it can be argued that this specific task does not engage grammatical class information for uninflected words.

In an attempt to control for the semantic correlates of noun–verb differences, in a functional magnetic resonance imaging (fMRI) study, Shapiro et al. (2006) considered only areas of significant activations emerging both when speakers were producing phrases in response to real words (not only object nouns and action verbs but also more abstract nouns and verbs) and when speakers were producing phrases in response to pseudowords. Moreover, in order to control for morphophonological correlates of grammatical class differences, only areas of significant activation when speakers were producing both regularly and irregularly inflected nouns and verbs were considered. Across 3 experiments, participants were presented with words (either a noun or a verb) or pseudowords (to be used in either a noun or verb context) and their task was to produce short phrases such as “many doors,” or “he sweeps.” Significant greater activations for nouns across experiments were observed within the left fusiform gyrus (Brodmann area [BA] 20). Significant greater activations for verbs were found in left prefrontal cortex (BA 9) and in left superior parietal cortex (BA 7). However, it can be argued that the stimuli used in this study still conflate semantics and grammatical class. It has been proposed, in fact, that the mental representation for nouns and verbs develops around an initial distinction between prototypical nouns, that are objects and prototypical verbs, that are actions (Bates and MacWhinney 1982; O'Grady 1997; Gauthier et al. 1999). In this view, when presented with novel words, speakers infer prototypical semantic properties of objects, if these words are used in “noun” contexts, and prototypical semantic properties of actions, if they are used in “verb” contexts. The finding of fusiform activations just anterior to the areas involved in object recognition for nouns and the frontoparietal activations for verbs in the Shapiro et al. (2006) may thus still reflect semantic differences between nouns and verbs.

A different approach to controlling for the semantic correlates of grammatical class was used in a positron emission tomography study by Vigliocco et al. (2006) in Italian. In this study, only words referring to events, either nouns (e.g., corsa [run]) or verbs (e.g., galoppa [she/he gallops]), and referring to either sensation (e.g., solletico [tickle]) or motion (e.g., giravolta [twirl]) were used. Participants were presented auditorily with blocks of sensory or motor nouns or verbs and asked to simply listen to the words. Whereas significant activations related to differences between sensory and motion words were found in anterior temporal cortex, for sensory words and in primary motor cortex, for motion words, no specific activations for nouns or verbs were observed. However, again, one may argue that just listening to single words (even if inflected) may not engage grammatical class information.

The fMRI experiment we report here provides a further test of grammatical class differences when semantics is controlled in an attempt to better understand the conflicting results reported in the literature. Using a mini-block design, we asked native Italian speakers to name the same picture of an event as follows:

  1. Infinitive Verb (e.g., mangiare [to eat]),

  2. Inflected Verb (e.g., mangia [she/he eats]), and

  3. Action Noun (e.g., mangiata [the eating]).

Using the same picture in the different naming conditions, we eliminate the semantic confound because participants name the same event using either a noun or a verb. Moreover, as the same picture is always presented, any difference cannot be attributed to visual processing demands. Using a picture-naming task, we fully engage the language production system from conceptual retrieval to word retrieval, thus, if there are any differences in processing nouns and verbs these should emerge, taking in account that noun–verb dissociations in the neuropsychological literature have been established using picture naming.

A further remark is that producing the infinitive form of a verb (example 1) engages lexical retrieval processes, whereas producing the inflected form (example 2) further engages syntactic and morphophonological processes. This is because Italian is a “pro-drop” language in which the subject of a sentence does not need to be produced (as it is the case, instead, in English). Thus, an Inflected Verb is either a full sentence (for intransitive verbs, e.g., cammina [she/he walks]) or a sentence fragment (for transitive verbs, e.g., mangia [she/he eats]). Therefore, if activations in left IFG previously reported during verb processing are due to engagement of morphosyntactic processes, these activations should be greater in the Inflected Verb than in the Infinitive Verb condition. Importantly, by explicitly instructing participants to name the picture using a noun or a verb, we further ensure that grammatical class information is relevant to the task (in contrast to previous studies that used words as stimuli and tasks, such as lexical decision, in which grammatical class is neither implicitly engaged nor relevant to the task, e.g., Perani et al. 1999; Tyler et al. 2001). Finally, as an element of novelty, we examine nouns that are morphologically complex as they are morphologically derived from verbs (example 3).

If noun- and verb-specific activations in previous studies emerged as a consequence of semantic differences between stimuli, we should not find any specific activation in the contrast between Infinitive Verb and Action Noun (examples 1 and 3). If previously observed verb-specific activations in left IFG were due to greater processing demands of the stimuli rather than by verb processing per se, we may observe left IFG activations for Action Nouns (example 3). This is because these stimuli are not only more linguistically complex (the nouns are morphologically derived from the verb), but moreover, in order to produce the noun response, speakers have to suppress the more natural verb response to the materials (i.e., given a picture of an action, if no specific naming instructions are given, speakers name it using a verb rather than the corresponding noun).

Methods

Subjects

Twelve Italian native speakers (7 males, 5 females; mean age 26), recruited within the student population of Vita-Salute University participated in the study. All participants had no history of neurological or psychiatric disorders and were right handed. Right handedness was verified using the Edinburgh Inventory (Oldfield 1971). Prior to the experiment, all participants gave written consent to participate.

Materials

Seventy-two black and white line drawings depicting actions were used. These pictures were taken from the object and action naming battery of Druks and Materson (2000) . Fifty-two other pictures from this set were excluded because 1) the actions depicted could not be named as Action Nouns in Italian (e.g., crawling) and 2) name agreement in Italian, as established in a norming phase, was lower than 70%.

Experimental Design

Subjects viewed black and white pictures one at a time, and they were asked to name them overtly. Written instructions appearing before picture presentation indicated whether they had to produce: 1) Infinitive Verb (e.g., correre [to run]); 2) Inflected Verb (e.g., corre [he/she runs]); 3) Action Noun (e.g., corsa [the run]) (see Fig. 1 for examples of pictures and naming responses). Each picture was presented 3 times, once in each of the experimental runs. Each picture was available on the screen for 2.5 s. Instructions remained on the screen for 1.5 s. Type of response was changed every 4, 5 items, introducing a mini-block design. Each picture was presented 3 times, one time per run, in a different condition in each run (e.g., RUN1 – Infinitive Verb, RUN2 – Action Noun, RUN3 – Inflected Verb). Order of pictures within each mini block was randomized, and the order of experimental runs was randomized across subjects. A mini-block design was chosen instead of a completely event-related paradigm on the basis of the results of behavioral testing that we carried out first. Fourteen subjects were tested outside the scanner using the same materials and timing. In this behavioral study, we established that a mini-block design was preferable to a fully random design because this latter produced high levels of variability in the reaction times (RTs), likely because of the increased difficulty of the task that continuously required subjects to switch response type. The mini-block design was combined with jittered inter stimulus interval (= 1, 2, or 3 s in order to optimize statistical power (Dale 1999).

Figure 1.

Examples of pictures used in the study along with naming responses required in the 3 different conditions.

Figure 1.

Examples of pictures used in the study along with naming responses required in the 3 different conditions.

Procedures

Pictures were presented to participants via a portable computer placed outside the magnet room equipped with Presentation 0.91 (Neurobehavioral Systems, Albany, CA). A projector connected to the laptop delivered stimuli on a translucent screen placed at the foot of the magnet bore. Participants viewed the screen through a mirror system attached to the top of the head coil. Before entering in the scanner, all participants underwent a training session on a different set of pictures in order to familiarize them with the task and to optimize their performance during the experimental session. Moreover, participants were trained to minimize jaw and tongue movements while naming to reduce movement artifacts. During the scanning, participants were asked to speak into a plastic tube attached to the head coil, and they were required to avoid head movements as much as they could. The plastic tube conveyed the sound outside the scanner room where there was a small microphone connected to a computer (for recording response times) and a tape recorder (for recording naming responses in order to score accuracy).

Imaging Data Acquisition and Analysis

Brain imaging was conducted on a 1.5-T Sigma scanner (General Electric, Milwaukee, WI) equipped with a standard head coil. Twenty contiguous, gradient-echo echo-planar images, sensitive to blood oxygen level–dependent (BOLD) contrast, parallel to the anterior commisure-posterior commisure plane (AC–PC) were acquired using a T2*-weighted gradient-echo echo -planar imaging sequence (time repetition [TR] 2.5 s, time echo [TE] 60 ms, field of view [FOV] 280 × 280, 64 × 64 matrix, flip angle 90°, slice thickness 4 mm, 4.375 × 4.375 mm in-plane resolution). High-resolution T1-weighted images were also acquired (3 -dimensional spoiled gradient-echo sequence, TR 4.50 s, TE 9 ms, FOV 24 × 24 cm, matrix 2562, slice thickness 1.5 mm).

Data processing and statistical analysis were performed with SPM2 (www.fil.ion.ucl.ac.uk). The entire volume set for each subjects was corrected for slice acquisition order and realigned to the first volume of the first sequence. Functional images were then spatially normalized to the Montreal Neurological Institute standard space to allow for group analysis. Prior to statistical analysis, all images were smoothed using an isotropic Gaussian kernel (full-width half-maximum = 8 mm).

The data for each subject were modeled using the general linear model (Friston et al. 1995). Three sessions and 3 variables were entered into the model (Infinitive Verb, Inflected Verb, Action Noun). The BOLD response for each event was modeled with the canonical hemodynamic response function along with the time derivative. The first 10 volumes of each run were discarded to allow attainance of a T2 equilibrium. Movement parameter estimates produced by the realignment procedure were entered as confound regressors in the first-level single-subject design matrices, in order to correct for potential movement artifacts caused by the overt naming task. Contrast images for the 3 individual conditions of all subjects were combined into a second-level random effects group analysis (basic models, one-way analysis of variance [ANOVA]) (Penny and Holmes 2003). Within the one-way ANOVA design, we computed Student t-tests, including 1) Main effect (Infinitive Verb + Inflected Verb + Action Noun) and 2) Direct comparisons (Action Noun – Infinitive Verb; Action Noun – Inflected Verb; Inflected Verb – Infinitive Verb).

We assessed the cytoarchitectonic probability of activations in the left fronto-lateral cortex, using the probability maps (Amunts et al. 1999) for BA 44 and BA 45 available with the SPM Anatomy toolbox (http://www.fz-juelich.de/ime/spm_anatomy_toolbox).

Results

Behavioral Data

Response times and error rates were recorded for a group of 14 participants (Group 1, from the same population as the participants to the fMRI study) outside the scanner and for the 12 participants (Group 2) who performed the naming task during functional scanning. Data collected outside the scanner are closely comparable to data collected during fMRI and are reported here to point out that scanning did not interfere with participants' performance on the naming task and that our apparatus for recording RTs during scanning produced reliable and comparable data (see Fig. 2). For both groups, statistical analyses on RTs showed that subjects were significantly slower in the Action Noun condition (Group 1: mean = 1633 ms, standard deviation [SD] = 601; Group 2: mean = 1745 ms, SD = 563) compared with both other conditions (Tukey–Kramer HSD test P < 0.001, for both groups), whereas there was no significant difference between Infinitive Verb (Group 1: mean = 1212 ms, SD = 458; Group 2: mean = 1409 ms, SD = 364) and Inflected Verb (Group 1: mean = 1395 ms, SD = 533; Group 2: mean = 1387 ms, SD = 389) (See Fig. 2). No significant difference in accuracy was found across the 3 conditions and across both groups outside and inside the scanner. (Group 1: Action Nouns 88%, Infinitive Verbs 97%, Inflected Verbs 96%; Group 2: Action Nouns 88%, Infinitive Verbs 97%, Inflected Verbs 96%).

Figure 2.

Mean naming latencies (in ms) for the 3 experimental conditions for 14 participants (behavioral, Group 1) who performed the naming task outside the scanner and for the 12 participants (fMRI, Group 2) performing the task during scanning.

Figure 2.

Mean naming latencies (in ms) for the 3 experimental conditions for 14 participants (behavioral, Group 1) who performed the naming task outside the scanner and for the 12 participants (fMRI, Group 2) performing the task during scanning.

Functional Data

First, we performed a main effect analysis including all 3 naming conditions (P < 0.001, family wise error type correction). We found activations in the left IFG, the left superior temporal gyrus, the temporal poles bilaterally, and the right hippocampus (Table 1, Fig. 3).

Figure 3.

Standard SPM2 3-dimensional rendering showing the significant activations for the main effect including all 3 naming conditions (P < 0.001, family wise error correction).

Figure 3.

Standard SPM2 3-dimensional rendering showing the significant activations for the main effect including all 3 naming conditions (P < 0.001, family wise error correction).

Table 1

Montreal Neurological Institute stereotactic coordinates (in mm) for the significant activations in the main effect including all 3 experimental conditions (P < 0.001, family wise error corrected)

 x y z Z-score CP 
Left IFG −46 10 24 6.30 40% BA 44 
Left ventral premotor cortex −54 38 Inf  
Left superior temporal gyrus −52 −14 −6 6.48  
Left temporal pole −44 −10 6.48  
Right temporal pole 44 −14 6.64  
Right hippocampus 26 −34 6.77  
 x y z Z-score CP 
Left IFG −46 10 24 6.30 40% BA 44 
Left ventral premotor cortex −54 38 Inf  
Left superior temporal gyrus −52 −14 −6 6.48  
Left temporal pole −44 −10 6.48  
Right temporal pole 44 −14 6.64  
Right hippocampus 26 −34 6.77  

CP, cytoarchitectonic probabilities (Amunts et al. 1999); Inf, inferior.

Second, we performed the direct comparisons between the 3 naming conditions. We found significant effects (P < 0.05, false discovery rate type correction) only when we performed the subtraction Action Noun minus Infinitive Verb; these effects were exclusively located in the left IFG, both ventrally and dorsocaudally (pars opercularis, 60% probability for BA 44, according to cytoarchitectonic probability maps; Amunts et al. 1999, Fig. 4). The subtraction Action Noun minus Inflected Verb provided significant effects only when lowering the significance threshold (P < 0.001, uncorrected); again, the effects were exclusively located in the left IFG. No significant differences were found in the subtraction Inflected Verb minus Infinitive Verb even at an uncorrected significance threshold, except for an effect in the left ventral IFG (P < 0.05, uncorrected) (Table 2, Fig. 5). Thus, signal change within left IFG, particularly in the ventral aspect, was somewhat modulated by the type of processes required in the naming of the same event: greatest for the Action Noun, intermediate for the Inflected Verb, and least for the Infinitive Verb.

Figure 4.

Standard SPM2 3-dimensional rendering showing the significant activations for the direct subtraction Actions Noun minus Infinitive Verb (P < 0.05, false discovery rate correction).

Figure 4.

Standard SPM2 3-dimensional rendering showing the significant activations for the direct subtraction Actions Noun minus Infinitive Verb (P < 0.05, false discovery rate correction).

Figure 5.

Histograms representing percent signal change and 90% confidence intervals in the 3 experimental conditions (Infinitive Verb: white; Inflected Verb: gray; Action Noun: black) in 2 different portions (dorsal vs. ventral) of the left IFG (upper part: x = −46, y = 16, z = 26; lower part: x = −48, y = 24, z = −4).

Figure 5.

Histograms representing percent signal change and 90% confidence intervals in the 3 experimental conditions (Infinitive Verb: white; Inflected Verb: gray; Action Noun: black) in 2 different portions (dorsal vs. ventral) of the left IFG (upper part: x = −46, y = 16, z = 26; lower part: x = −48, y = 24, z = −4).

Table 2

Montreal Neurological Institute stereotactic coordinates (in mm) for the significant activations in the direct comparisons

 x y z Z-score CP 
Action Noun − Infinitive Verb 
    Left IFG −46 16 26 5.02 60% BA 44 
    Left IFG −48 24 −4 4.67  
Action Noun − Inflected Verb 
    Left IFG −46 16 26 4.24 60% BA 44 
    Left IFG −48 24 −4 3.47  
Inflected Verb − Infinitive Verb 
    Left IFG −48 24 −4 1.73  
 x y z Z-score CP 
Action Noun − Infinitive Verb 
    Left IFG −46 16 26 5.02 60% BA 44 
    Left IFG −48 24 −4 4.67  
Action Noun − Inflected Verb 
    Left IFG −46 16 26 4.24 60% BA 44 
    Left IFG −48 24 −4 3.47  
Inflected Verb − Infinitive Verb 
    Left IFG −48 24 −4 1.73  

CP, cytoarchitectonic probabilities (Amunts et al. 1999); Inf, inferior.

Discussion

The results of our fMRI study provide evidence that naming an event relies on the activation of the left perisylvian cortex, the right temporal pole, and the right hippocampus and that, within this system, the activity of the left IFG is greatest when the same event is named by Action Nouns, intermediate when it is named by Inflected Verbs, and least when it is named by Infinitive Verbs. In what follows, we will concentrate on the interpretation of differences between the 3 naming conditions, as revealed by the direct comparisons.

The dissociation observed in neuropsychological studies between noun and verb processing has been taken to support hypotheses that words from different grammatical classes engage different neural networks, by virtue of their grammatical properties (e.g., Caramazza and Hillis 1991). However, the distinction between nouns and verbs largely cooccurs with semantic differences between these 2 types of words. Semantic differences between nouns referring to objects and verbs referring to actions have been linked to different neural substrates, with nouns preferentially engaging left inferior temporal cortices (e.g., Damasio et al. 2004), involved in perceptual object recognition and verbs preferentially engaging left premotor cortex, left IFG, and left middle temporal, responsible for action processing (e.g., Perani et al. 1999; Tranel et al. 2005).

Previous studies investigating noun–verb processing have attempted to reduce the impact of the semantic confound by using abstract words (Perani et al. 1999; Tyler et al. 2001, 2004; Bedny and Thompson-Schill 2006; Shapiro et al. 2006) or by comparing real words to pseudowords (Shapiro et al. 2006). However, as discussed in the Introduction, these strategies set up to control for semantic correlates may not be ideal, because it cannot be excluded that, for example, a pseudoword is processed as object-like, when presented in a noun context, and as event-like, when presented in a verb context.

In our study, we eliminated the semantic confound by using only words referring to events. More precisely, we asked speakers to name the same event using either a noun or a verb. Grammatical class information was directly relevant to the task as we instructed participants to name the picture either as a noun or a verb. Thus, although speakers were producing single words, grammatical class information was relevant to the task, and therefore this study is not susceptible to this criticism that can be, instead, raised to some previous studies that failed to observe noun- or verb-specific activations (e.g., Tyler et al. 2001; Vigliocco et al. 2006).

In contrast to some previous studies (Perani et al. 1999; Tyler et al. 2004), we failed to observe any verb-specific activations. We found, instead, noun-specific activations in left IFG, specifically in BA 44 and more ventrally in BA 45/47. Importantly, these noun-specific activations fall in close proximity to the verb-specific activations reported in the aforementioned studies. In our follow-up analysis, we further showed that the signal change in left ventral IFG show a trend toward being modulated by the processing differences across the 3 conditions. In particular, we found strongest activation for the Action Nouns, intermediate for Inflected Verbs, and weakest for Infinitive Verbs. This finding clearly indicates that left IFG engagement is not selectively linked to verb processing. Rather, our finding suggests that left IFG is engaged in the combinatorial and selection processes required for producing an inflected form of a verb (that, as we described in the Introduction can be considered as syntactic in addition to morphological processes) and, to a greater extent, in the processes engaged in producing derived word forms.

A number of previous studies have established a role for left IFG in syntactic and morphological processing during comprehension (e.g., Gabrieli et al. 1998; Dapretto and Bookheimer 1999; Caplan et al. 2000; Moro et al. 2001; Marangolo et al. 2004; Cooke et al. 2006). The few studies that have investigated phrasal production have also reported activations related to phrasal production in left IFG. In experiments by Indefrey et al. (2004), speakers of German were presented with displays showing animated scenes and they were asked to describe them using full sentences (e.g., “The red square launches the blue ellipse”), sequences of phrases with local, but not sentence-level syntactic structure (e.g., “red square,” “blue ellipse,” etc.), or sequences of words without any syntactic structure (e.g., “square,” “red,” “ellipse,” etc.). It was found that a left IFG region, overlapping in part with BA 44, although mainly posterior to Broca's area (most likely in BA 6) was most activated in the sentence condition, less active in the phrasal condition, and least active in the word condition. Activations in left IFG (greater, however, in BA 45 than in BA 44) were also reported in the study by Haller et al. (2005) in which German speakers were asked to produce sentences aloud (e.g., “The child throws the ball”) in response to syntactically incomplete stimuli (e.g.,“throws,” “ball”; “child”), relative to the condition in which they were asked to read aloud the same words or in which they were asked to read aloud corresponding sentences. What is common between these studies and our study is that the conditions in which left IFG activations are observed as ones where combinatorial and selection processes are engaged. A role for left IFG in integrating different types of linguistic information has been argued by researchers investigating sentence comprehension (Price 1998; Martin and Chao 2001; Hagoort 2005).

It is noteworthy that 2 separate activation foci were found in left IFG. One could be localized on the basis of probability mapping in BA 44, the other was approximately localized more ventrally in BA 45/47. There is extensive evidence for the fractionation between these components of the classical Broca's region, on the basis of cytoarchitectonic and hodological differences (Petrides and Pandya 2002; Anwarder et al. 2006). This structural distinction has been proposed to reflect a functional differentiation. It has been argued that the dorsal, 44/6 area is involved in the computation of local syntactic dependencies in word production and comprehension, whereas the ventral 45/47 area is linked to lexical/semantic processing (Paulesu et al. 1997; Poldrack et al. 1999; Indefrey et al. 2004; Friederici 2006). Our study was not designed to test any specific hypothesis of differentiation within left IFG. It is, however, possible that whereas the graded effect we observed in dorsal IFG reflects to a greater extent the engagement of combinatorial and syntactic processes, the graded effect we found in ventral IFG might be linked to processes of selecting and selecting the lexical status of inflected and derived forms to-be-produced, these processes could be more demanding for derived than inflected words.

The processes recruited in producing Action Nouns may not be strictly linguistic (i.e., they may be domain general). In our study, when presented with pictures depicting actions, the most automatic naming response is to name the picture using a verb and specifically an Infinitive Verb in Italian. This can be established by simply presenting pictures to speakers and asking them to name them without giving them any more specific instructions regarding the form of the naming response. In order to name the same picture with an Action Noun, speakers have to select a less favorite response instead. Thompson-Schill has proposed that BA 44 plays a crucial role in conditions characterized by “high selectional demands” among competing alternatives (Thompson-Schill et al. 1997). In our experiment, the Action Noun was the less favored response to the pictures as indicated by the fact that subjects were slower in producing Action Nouns than in all the other conditions (see Fig. 2). Thus, it is plausible that Action Nouns posed the highest selection demands among competing alternatives in our experiment. Conspiring in rendering Action Nouns the least favorite response among our conditions, these words are also less frequent and are acquired later than the verbs. It has been shown in previous fMRI studies greater left IFG activations for low-frequency than high-frequency words (Chee et al. 2002; Fiebach et al. 2003).

In conclusion, our study demonstrates that words belonging to different grammatical classes (nouns and verbs) are not represented in segregated neural networks. Rather, neural segregation emerges as the result of other differences between nouns and verbs that are usually confounded with the grammatical class difference across languages. First, in most previous patient and imaging work, differences between nouns and verbs may have come about as a semantic difference between objects and actions. In previous studies that did not suffer of this confound, and in which morphological processes were not highly engaged, no verb-specific activations in left IFG was found (Saccuman et al. 2006; Vigliocco et al. 2006). In the present study, we showed that if we control for semantic differences (asking participants to name only events and more specifically the same events as either noun or verb) and manipulate the extent of morphological processing across the grammatical class of verbs and nouns, left IFG activations are shown to be modulated by the complexity of the morphological processes involved and/or by the selection demands of the morphological task rather than being associate to verb-specific processing.

Thus, our results do call into question the view that grammatical class per se drives neural segregation within the language network at the lexical level restricting the role of this information in driving neural organization to its use in connected speech.

The work reported here was supported by a Biotechnology and Biological Science Research Council grant (BBSRC 31/S18749) to GV and an MIUR grant to SFC. Conflict of Interest: None declared.

References

Amunts
K
Schleicher
A
Buergel
U
Mohlberg
H
Uylings
HBM
Zilles
K
Broca's region revisited: cytoarchitecture and intersubject variability
J Comp Neurol
 , 
1999
, vol. 
412
 (pg. 
319
-
341
)
Anwander
A
Tittgemeyer
M
von Cramon
DY
Friederici
AD
Knosche
TR
Connectivity-Based Parcellation of Broca's Area
Cereb Cortex
 , 
2006
, vol. 
17
 (pg. 
816
-
825
)
Bates
EA
MacWhinney
B
Wanner
E
Gleitman
LR
Functionalist approaches to grammar
Language acquisition: the state of the art
 , 
1982
Cambridge (MA)
Cambridge University Press
(pg. 
173
-
218
)
Bedny
M
Thompson-Schill
SL
Neuroanatomically separable effects of imageability and grammatical class during single-word comprehension
Brain Lang
 , 
2006
, vol. 
98
 (pg. 
127
-
139
)
Binder
JR
Liebenthal
E
Possing
ET
Medler
DA
Ward
BD
Neural correlates of sensory and decision processes in auditory object identification
Nat Neurosci
 , 
2004
, vol. 
7
 (pg. 
295
-
301
)
Caplan
D
Alpert
N
Walters
G
Olivieri
A
Activation of Broca's area by syntactic processing under conditions of concurrent articulation
Hum Brain Mapp
 , 
2000
, vol. 
9
 (pg. 
65
-
71
)
Cappa
SF
Binetti
G
Pezzini
A
Padovani
A
Rozzini
L
Trabucchi
M
Object and action naming in Alzheimer's disease and frontotemporal dementia
Neurology
 , 
1998
, vol. 
50
 (pg. 
351
-
355
)
Caramazza
A
Hillis
AE
Lexical organization of nouns and verbs in the brain
Nature
 , 
1991
, vol. 
349
 (pg. 
788
-
790
)
Chee
MWL
Hon
NHH
Caplan
D
Lee
HL
Goth
J
Frequency of concrete words modulates prefrontal activation during semantic judgments
Neuroimage
 , 
2002
, vol. 
16
 (pg. 
259
-
268
)
Cooke
A
Grossman
M
DeVita
C
Gonzales-Atavales
J
Moore
P
Chen
W
Gee
J
Detre
J
Large-scale neural network for sentence processing
Brain Lang
 , 
2006
, vol. 
96
 (pg. 
14
-
36
)
Dale
AM
Optimal experimental design for event-related fMRI
Hum Brain Mapp
 , 
1999
, vol. 
8
 (pg. 
109
-
114
)
Damasio
AR
Tranel
D
Nouns and verbs are retrieved with differently distributed neural systems
Proc Natl Acad Sci USA
 , 
1993
, vol. 
90
 (pg. 
4957
-
4960
)
Damasio
H
Tranel
D
Grabowsky
TJ
Adolphs
R
Damasio
HR
Neural systems behind word and concept retrieval
Cognition
 , 
2004
, vol. 
92
 (pg. 
179
-
229
)
Daniele
A
Giustolisi
L
Silveri
MC
Colosimo
C
Gainotti
G
Evidence for a possible neuroanatomical basis for lexical processing of nouns and verbs
Neuropsychologia
 , 
1994
, vol. 
32
 (pg. 
1325
-
1341
)
Dapretto
M
Bookheimer
SY
Form and content: dissociating syntax and semantics in sentence comprehension
Neuron
 , 
1999
, vol. 
24
 (pg. 
427
-
432
)
Druks
J
Masterson
J
An object and action naming battery
2000
London (UK)
Psychology Press
Federmeier
KD
Segal
JB
Lombrozo
T
Kutas
M
Brain responses to nouns, verbs and class-ambiguous words in context
Brain
 , 
2000
, vol. 
123
 (pg. 
2552
-
2566
)
Fiebach
CJ
Friederici
AD
Muller
K
von Cramon
DY
Hernandez
AE
Distinct brain representations for early and late learned words
Neuroimage
 , 
2003
, vol. 
19
 (pg. 
1627
-
1637
)
Friederici
AD
Broca's area and the ventral premotor cortex in language: functional differentiation and specificity
Cortex
 , 
2006
, vol. 
42
 
4
(pg. 
472
-
475
)
Friederici
AD
Hahne
A
Mecklinger
A
Temporal structure of syntactic parsing: early and late event-related brain potential effects
J Exp Psychol Learn Mem Cogn
 , 
1996
, vol. 
22
 (pg. 
1219
-
1248
)
Friston
KJ
Holmes
AP
Poline
JB
Grasby
PJ
Williams
SC
Frackoviak
RS
Turner
R
Analysis of fMRI time-series revisited
Neuroimage
 , 
1995
, vol. 
2
 (pg. 
45
-
53
)
Gabrieli
JDE
Poldrack
RA
Desmond
JE
The role of left prefrontal cortex in language and memory
Proc Natl Acad Sci USA
 , 
1998
, vol. 
95
 (pg. 
906
-
913
)
Garrett
MF
Caplan
D
Lecours
AR
Smith
A
The organization of processing structure for language production: applications to aphasic speech
Biological perspectives on language
 , 
1984
London (UK)
MIT Press
(pg. 
172
-
193
)
Gauthier
I
Tarr
MJ
Anderson
AW
Skudlarski
P
Gore
JC
Activation of the middle fusiform “face area” increases with expertise in recognizing novel objects
Nat Neurosci
 , 
1999
, vol. 
2
 
6
(pg. 
568
-
573
)
Gold
BT
Buckner
RL
Common prefrontal regions coactivate with dissociable posterior regions during controlled semantic and phenological tasks
Neuron
 , 
2002
, vol. 
35
 (pg. 
803
-
812
)
Hagoort
P
On Broca, brain, and binding: a new framework
Trends Cogn Sci
 , 
2005
, vol. 
9
 (pg. 
416
-
423
)
Haller
S
Radue
EW
Erb
M
Grodd
W
Kircher
T
Overt sentence production in event-related fMRI
Neuropsychologia
 , 
2005
, vol. 
43
 (pg. 
807
-
814
)
Indefrey
P
Hellwig
F
Herzog
H
Seitz
RJ
Hagoort
P
Neural responses to the production and comprehension of syntax in identical utterances
Brain Lang
 , 
2004
, vol. 
89
 (pg. 
312
-
319
)
Khader
P
Scherag
A
Streb
J
Rosler
F
Differences between noun and verb processing in a minimal phrase context: a semantic priming study using event-related brain potentials
Brain Res Cogn Brain Res
 , 
2003
, vol. 
17
 (pg. 
293
-
313
)
Longe
O
Randall
B
Stamatakis
EA
Tyler
LK
Grammatical categories in the brain: the role of morphological structure
Cereb Cortex.
 , 
2007
 
doi: 10.1093/cercor/bhl099
Marangolo
P
Piras
F
Galati
G
Burati
C
The neural substrates of derivational morphological processing: an FRMI study
Cortex
 , 
2004
, vol. 
40
 (pg. 
185
-
186
)
Martin
A
Chao
LL
Semantic memory and the brain: structure and processes
Curr Opin Neurobiol
 , 
2001
, vol. 
11
 (pg. 
194
-
201
)
Martin
A
Haxby
JV
Lalonde
FM
Wiggs
CL
Ungerleider
LG
Discrete cortical regions associated with knowledge of color and knowledge of action
Science
 , 
1995
, vol. 
270
 (pg. 
102
-
105
)
Miceli
G
Silveri
MC
Nocentini
U
Caramazza
A
Patterns of dissociation in comprehension and production of nouns and verbs
Aphasiology
 , 
1988
, vol. 
2
 (pg. 
351
-
358
)
Miozzo
A
Soardi
M
Cappa
SF
Pure anomia with spared action naming due to a left temporal lesion
Neuropsychologia
 , 
1994
, vol. 
32
 (pg. 
71101
-
1109
)
Moro
A
Tettamanti
M
Perani
D
Donati
C
Cappa
SF
Fazio
F
Syntax and the brain: disentangling grammar by selective anomalies
Neuroimage
 , 
2001
, vol. 
13
 (pg. 
110
-
118
)
O'Grady
W
Syntactic development
1997
Chicago (IL)
U Chicago Press
Oldfield
RC
The assessment and analysis of headedress: The Edinburgh inventory
Neuropsychologia
 , 
1971
, vol. 
9
 (pg. 
97
-
113
)
Paulesu
E
Goldacre
B
Scifo
P
Cappa
SF
Gilardi
MC
Castiglioni
I
Perani
D
Fazio
F
Functional heterogeneity of left inferior frontal cortex as revealed by fMRI
Neuroreport
 , 
1997
, vol. 
8
 
8
(pg. 
2011
-
2017
)
Penny
WD
Holmes
AP
Frackowiak
RSJ
Friston
KJ
Frith
CD
Dolan
R
Price
CJ
Zeki
S
Ashburner
J
Penny
WD
Random effects analysis
Human brain function
 , 
2003
San Diego (CA)
Academic Press
(pg. 
843
-
850
)
Perani
D
Cappa
SF
Schnur
T
Tettamanti
M
Collina
S
Rosa
MM
Fazio
F
The neural correlates of verb and noun processing: a PET study
Brain
 , 
1999
, vol. 
122
 (pg. 
2337
-
2344
)
Petersen
SE
Fox
PT
Posner
MI
Mintum
M
Raichle
ME
Position emission tomographic studies of the cortical anatomy of single-word processing
Nature
 , 
1988
, vol. 
331
 (pg. 
585
-
589
)
Petersen
SE
Fox
PT
Posner
MI
Mintum
M
Raichle
ME
Position emission tomographic studies of the processing of single words
J Cogn Neurosci
 , 
1989
, vol. 
1
 (pg. 
153
-
170
)
Petrides
M
Pandya
DN
Comparative cytoarchitectonic analysis of the human and the macaque ventrolateral prefrontal cortex and corticocortical connection patterns in the monkey
Eur J Neurosci
 , 
2002
, vol. 
16
 
2
(pg. 
291
-
310
)
Poldrack
RA
Wagner
AD
Prull
MW
Desmond
JE
Glover
GH
Gabrieli
JD
Functional specialization for semantic and phonological processing in the left inferior prefrontal cortex
Neuroimage
 , 
1999
, vol. 
10
 
1
(pg. 
15
-
35
)
Price
CJ
The functional anatomy of word comprehension and production
Trends Cogn Sci
 , 
1998
, vol. 
2
 (pg. 
281
-
288
)
Rapp
B
Caramazza
A
The modality-specific organization of grammatical categories: evidence from impaired spoken and written sentence production
Brain Lang
 , 
1997
, vol. 
56
 (pg. 
248
-
286
)
Saccuman
MC
Cappa
SF
Bates
EA
Arevalo
A
Della Rosa
P
Danna
M
Perani
D
The impact of semantic reference on word class: an fMRI study of action and object naming
Neuroimage
 , 
2006
, vol. 
32
 (pg. 
1865
-
1878
)
Shapiro
KA
Moo
LR
Caramazza
A
Cortical signatures of noun and verb production
Proc Natl Acad Sci USA
 , 
2006
, vol. 
103
 (pg. 
1644
-
1649
)
Silveri
MC
Perri
R
Cappa
A
Grammatical class effects in brain-damaged patients: functional locus of noun and verb deficit
Brain Lang
 , 
2003
, vol. 
85
 (pg. 
49
-
66
)
Thompson-Schill
SL
D'Esposito
M
Aguirre
GK
Farah
MJ
Role of left inferior prefrontal cortex in retrieval of semantic knowledge: a reevaluation
Proc Natl Acad Sci USA
 , 
1997
, vol. 
23
 (pg. 
14792
-
14797
)
Tranel
D
Martin
C
Damasio
H
Grabowski
TJ
Hichwa
R
Effects of noun-verb homonymy on the neural correlates of naming concrete entities and actions
Brain Lang
 , 
2005
, vol. 
92
 (pg. 
288
-
299
)
Tyler
LK
Bright
P
Fletcher
P
Stamatakis
EA
Neural processing of nouns and verbs: the role of inflectional morphology
Neuropsychologia
 , 
2004
, vol. 
42
 (pg. 
512
-
523
)
Tyler
LK
Russell
R
Fadili
J
Moss
HE
The neural representation of nouns and verbs: PET studies
Brain
 , 
2001
, vol. 
124
 (pg. 
1619
-
1634
)
Vigliocco
G
Vinson
DP
Lewis
W
Garrett
MF
Representing the meanings of object and action words: the featural and unitary semantic space hypothesis
Cognit Psychol
 , 
2004
, vol. 
48
 (pg. 
422
-
488
)
Vigliocco
G
Warren
J
Siri
S
Arcuili
J
Scott
S
Wise
R
The role of semantics and grammatical class in the neural representation of words
Cereb Cortex
 , 
2006
, vol. 
16
 (pg. 
1790
-
1796
)
Vinson
DP
Vigliocco
G
A semantic analysis of noun-verb dissociation in aphasia
J Neurolinguistics
 , 
2002
, vol. 
15
 (pg. 
317
-
351
)
Warburton
E
Wise
RJ
Price
CJ
Weiller
C
Hadar
U
Ramsay
S
Noun and verb retrieval by normal subjects. Studies with PET
Brain
 , 
1996
, vol. 
119
 (pg. 
1803
-
1817
)
Zingeser
LB
Berndt
RS
Retrieval of nouns and verbs in agrammatism and anomia
Brain Lang
 , 
1990
, vol. 
39
 (pg. 
14
-
32
)