Abstract

Patients with anterior temporal lobe (ATL) lesions show semantic and lexical retrieval deficits, and the differential role of this area in the 2 processes is debated. Functional neuroimaging in healthy individuals has not clarified the matter because semantic and lexical processes usually occur simultaneously and automatically. Furthermore, the ATL is a region challenging for functional magnetic resonance imaging (fMRI) due to susceptibility artifacts, especially at high fields. In this study, we established an optimized ATL-sensitive fMRI acquisition protocol at 4 T and applied an event-related paradigm to study the identification (i.e., association of semantic biographical information) of celebrities, with and without the ability to retrieve their proper names. While semantic processing reliably activated the ATL, only more posterior areas in the left temporal and temporal–parietal junction were significantly modulated by covert lexical retrieval. These results suggest that within a temporoparietal network, the ATL is relatively more important for semantic processing, and posterior language regions are relatively more important for lexical retrieval.

Introduction

The occasional failure to name well-known people is a common experience in healthy individuals and suggests that naming is a process somehow independent from the identification of a person. In aphasia and in normal aging, this difficulty in retrieving names can become pathological and is called anomia. Anomic subjects can show preserved semantic knowledge of items they cannot name, thus suggesting that, even in pathological situations, the processes of lexical and conceptual knowledge retrieval can dissociate. In these cases, anomia can be caused by lexical and phonological deficits (Howard 1995; Lambon Ralph et al. 2000; Howard and Gatehouse 2006). The dissociation between semantic knowledge and naming is not a double dissociation though, since patients who have semantic deficits invariably show lexical retrieval impairments as well (Gainotti et al. 1981, 1986; Butterworth et al. 1984; Hillis et al. 1990; Hodges et al. 1992; Nickels and Howard 1994), a finding consistent with a serial, although interacting, naming model in which name retrieval follows semantic processing (Bruce and Young 1986; Burton et al. 1990; Bredart et al. 1995; Valentine et al. 1996).

Despite the behavioral distinction between semantic and naming processes, the identification of the anatomical correlates of the 2 processes has been difficult and is still debated. Single case studies in which semantic and lexical retrieval processes have been studied in detail suggest that left temporal and temporoparietal areas are crucial for naming but the precise anatomical location of the lesion was usually not detailed in these reports (Howard 1995; Lambon Ralph et al. 2000; Howard and Gatehouse 2006). Group studies (Damasio et al. 1996, 2004; Tranel 2006) in a large population of patients with focal lesions suggest that the left anterior temporal lobe (ATL) is crucially involved in naming faces, while the right ATL is crucial for recognizing famous faces (Tranel et al. 1997). The presumptive role of the left ATL in naming was then explained by Damasio et al. (2004) in their “convergence zone” account. According to this account, the left ATL would hold the “dispositions for naming.” Dispositions are the potentiality to produce the explicit mental representation of the word or its written and spoken patterns. Taken together, this evidence suggests that the left temporal lobe and the left inferior parietal region are involved in semantics and naming, but the specific role of each region is still not clarified.

Recent evidence from patients with primary progressive aphasia (PPA) (Mesulam 1982, 2007; Mesulam and Weintraub 1992; Grossman et al. 1996; Gorno-Tempini et al. 2004, 2011) has suggested a functional distinction between posterior temporoparietal areas on the one hand and the ATL on the other. Patients with left posterior temporal and parietal damage have logopenic PPA and anomia but not a multimodal semantic deficit (Henry and Gorno-Tempini 2010; Gorno-Tempini et al. 2011), whereas patients with ATL atrophy due to semantic variant PPA typically have naming problems but also a multimodal semantic deficit (Warrington 1975; Schwartz et al. 1979; Patterson et al. 2007).

While patient studies suffer from uncertainty regarding the precise anatomical location of the lesion responsible for the cognitive impairment, functional imaging studies on semantics and naming have to meet 2 different challenges. First, semantics and name retrieval occur usually simultaneously and automatically and are difficult to dissociate in cognitive tasks (Vandenberghe et al. 1996; Gorno-Tempini et al. 1998, 2000; Mummery et al. 1998). However, as initially mentioned, the failure to retrieve proper names is relatively frequent in healthy individuals (Brown 1991). In this study, we therefore used a paradigm involving famous people. This gave us the opportunity to dissociate semantics and lexical processing, an opportunity we would not have had with categories of nonunique objects (e.g., animals, tools, and vehicles) and common names.

Secondly, the ATL is a region of the brain that is difficult to investigate with functional magnetic resonance imaging (fMRI). The proximity of bone and air-filled cavities with very different magnetic susceptibilities leads to geometric distortions and signal loss, well-recognized limitations of echo planar imaging (EPI), particularly with high-field MRI (Ojemann et al. 1997; Devlin et al. 2000; Gorno-Tempini et al. 2002; Robinson et al. 2004; for review, see Visser et al. 2010). Therefore, the use of standard EPI may preclude the detection of task-related activity in the ATL, especially when using a higher field magnet. Studies using more sophisticated image acquisition techniques (Binney et al. 2010; Simmons et al. 2010; Visser et al. 2010) succeeded in finding ATL activation for semantic representations.

The main aim of this study was to characterize the role of the left temporal and inferior parietal regions in semantics and name retrieval using blood oxygen level–dependent (BOLD) fMRI. To address this, we first optimized BOLD sensitivity of 4-T gradient-echo EPI in ATL areas, considering slice thickness, echo time, polarity of the phase-encode gradient, slice angle, and shimming. We then used the optimized fMRI protocol to study semantic biographical and proper name retrieval in a group of 21 healthy subjects. Subjects' performance and attention to semantic information were ensured by asking them to perform a semantic (profession) same–different matching task in the scanner. The ability to identify and name the famous faces that were shown was assessed individually in a postscanning behavioral test, presenting all famous faces once again. Based on this postscanning assessment, we were able to compare the BOLD response during trials in which celebrities could be correctly identified and named, to trials in which faces were correctly identified without the name being recalled. Based on previous findings in PPA (Patterson et al. 2007; Gorno-Tempini et al. 2008, 2011), we predicted that a network of regions including bilateral ATL, left posterior temporal, and the inferior parietal regions would be activated by the semantic matching task but that the more posterior left lexical and phonological regions would show greater response for name retrieval.

Materials and Methods

We first optimized and evaluated a single-shot gradient-echo EPI protocol for 4-T fMRI in the ATL. This optimized protocol (for parameters, see below) was used to characterize semantic and lexical retrieval in the temporal lobes.

All MRI data was acquired with a 4-T MRI scanner (Bruker Medical, Ettlingen, Germany) using a birdcage transmit, 8-channel receive head RF coil. Structural images were acquired using a 3D magnetization prepared rapid gradient echo optimized for gray–white matter contrast, with time echo [TE]/time repetition [TR]/time to inversion [TI] = 4.18/2700/1020 ms, flip angle = 7°, 1 × 1 × 1 mm3, Generalized Autocalibrating Partially Parallel Acquisition acceleration factor 2 (Papinutto and Jovicich 2008).

All participants were right handed, had normal or corrected-to-normal vision, and none reported a history of head injury or other neurological problems. Specific demographics are indicated below, separately for the 2 groups of subjects used in the optimization of the EPI protocol and the semantic and lexical retrieval experiment. All participants gave written informed consent for their participation in the study. The experimental procedures were approved by the ethical committee for experiments involving humans at the University of Trento.

Optimized EPI Protocol for ATL at 4 T

The following parameters were investigated to minimize susceptibility-loss effects in the ATL and to increase time-series signal-to-noise ratio (tSNR), a good index for BOLD sensitivity (Triantafyllou et al. 2005). We tested different TE (Bandettini et al. 1994; Gati et al. 1997; Kruger et al. 2001), slice thickness and orientation (Deichmann et al. 2003), polarity of the phase-encoding gradient and shimming (De Panfilis and Schwarzbauer 2005), following previous evidence at lower field strength (Robinson et al. 2004; Weiskopf et al. 2006). Voxelwise tSNR was assessed in 10 healthy volunteers (mean age: 32.9 years, range: 24–45 years) in our standard EPI protocol (TE = 33 ms, 3-mm isotropic voxels, TR = 2000 ms, flip angle = 750, 37 axial AC-PC–oriented slices, slice gap = 0.45 mm) and the optimized EPI protocol (TE = 21 ms, 3 × 3 mm2 in-plane voxels, 2 mm slice thickness, 43 axial slices oriented approximately −200 relative to the AC–PC plane [approximately parallel to the longitudinal axis of the temporal lobes], slice gap = 0.3 mm). Each volunteer underwent a 10-min resting-state scan with each EPI protocol. Full-brain coverage was not possible with the optimized EPI protocol. Approximately the upper 2 cm of the brain were not included, while the main areas of interest were covered, including the entire temporal lobes, the inferior parietal regions as well as the occipital and most of the frontal lobes.

Images were preprocessed in SPM5 using standard methods (see below). The tSNR was used as a metric of BOLD sensitivity and was calculated as follows. Low-frequency signal changes (such as drift) were removed by subtracting a second-order polynomial fit to total slice signal. tSNR was calculated by dividing the voxelwise detrended signal mean by the standard deviations. Comparison between the optimized and the standard EPI protocols using paired t-test and thresholding with a false discovery rate of 0.05 revealed significant increases in tSNR with the optimized EPI protocol in bilateral ATL. Furthermore, tSNR distributions within the bilateral ATL were calculated over all subjects. The ATL was defined as the volume of the temporal lobes anterior to the limen insula (approximately defined as the anteroposterior position of y = 4 mm in the Montreal Neurological Institute [MNI] template space; Insausti et al. 1998) excluding the parahippocampal formation and amygdalae. The distributions showed higher tSNR with optimized EPI (mean tSNR = 156) compared with the standard protocol (mean tSNR = 111). The mean tSNR improvement in the ATL was 41%. Results are illustrated in Supplementary Figure 1.

Semantic and Lexical Retrieval Experiment

Subjects and Procedure

Twenty-one native Italian-speaking volunteers took part in the study (7 males; mean age: 28.4 years, range: 19–49 years). All participants underwent 2 functional scanning runs with the task, each of 14.2 min duration. After the scanning session, subjects were presented with each famous face to assess identification and naming scores.

Task and Design

A mixed blocked/event-related design was used. There were 3 different conditions, which were presented in blocks. At the start of each block, a written instruction was shown for 2 s to inform subjects of the upcoming task (“famous faces” or “unknown faces”). Within each block, trials were jittered to allow analysis based on different responses. Each condition involved the presentation of pairs of pictures. In the first condition, 2 famous faces were presented, and subjects were asked to perform a semantic task, deciding whether the people shown had the same profession (condition FF). In the second condition, 2 unknown faces were presented, and subjects were asked to perform a same–different visual matching task, deciding whether the images were of the same person (condition UF). In the third condition, 2 scrambled faces were presented, and subjects were asked to perform a perceptual task, deciding whether the 2 images were identical (condition SF). Subjects were instructed to press a key with their right index finger to indicate a “Yes” response and to press a key with the left index finger to indicate a “No” response. Response times were collected using in-house software ‘‘ASF” (Schwarzbach 2011), based on the MATLAB Psychtoolbox-3 (Brainard 1997) for Windows.

Of the 21 subjects, 13 were scanned with 80, 80, and 40 trials, respectively, in the conditions FF, UF, and SF. The remaining 8 subjects were presented with 130, 40, and 40 picture pairs in the same conditions. In this second group, the larger number of picture pairs was used for the FF condition because trials in this condition were to be subdivided in the analysis according to the naming and identification abilities of each individual subject, assessed after the scanning session (for details, see below).

Each picture pair was presented at the center of the screen on a black background, for 3.5 s. Intertrial intervals were jittered in a range of 2–7 s (mean = 4.5 s). A black screen with a fixation cross in the center was shown during these periods. Each scanning run contained 15 blocks: 5 in each condition. Each block of a given condition type contained the same number of trials.

Subjects were familiarized with the task before the experiment using a separate set of stimuli. They were instructed to respond as quickly as possible and to respond even if they were unsure about their decision. For famous faces, they were asked to concentrate on the semantic task without thinking about the name of the person.

Postscanning Behavioral Assessment of Identification and Naming Ability

After the scanner session, subjects were again presented with all the famous faces they had been shown in the fMRI experiment. Each face was presented on a computer screen for a maximum of 5 s, and subjects were asked to state the proper name and the profession/category of the person shown. A face was considered as identified correctly if the profession/category was stated correctly. These data were used to categorize trials of condition FF individually for each subject, depending on their ability to name and identify the faces shown in the corresponding picture pair. This categorization was critical for testing our main hypothesis on semantic and lexical processing and is explained in detail below.

Stimuli

Three types of stimuli were used: famous faces, unknown faces, and scrambled faces. Black and white photographs of 105 famous faces of Italian and internationally known celebrities were selected. Their names are listed in Supplementary Table 1. Thirty-six healthy controls (ages ranging from 25 to 70) were asked to identify, name, and rate the faces for familiarity. All of the celebrities belong to one of the following categories: politics, entertainment, sports, clergy, royal family, journalism, and business. The famous faces were then assembled in pairs of celebrities belonging to the same category (65 picture pairs: 15 pairs of women and 50 pairs of men) or belonging to different categories (65 picture pairs: 16 pairs of women and 49 pairs of men). We selected pictures in order to maximize attention and FF semantic processing. For example, pairs, see Figure 1A,B and for a complete list of pairs, see Supplementary Table 2.

Figure 1.

Examples of stimulus pairs for the 3 conditions FF (A,B), UF (C,D), and SF (E,F). Subjects had to do a same/different judgment regarding the persons' profession (condition FF), the identity of the faces (condition UF), or the identity of the images (condition SF). Respectively, one matched pair (A,C,D) and one unmatched pair (B,D,F) are shown for each condition. For details, see text.

Figure 1.

Examples of stimulus pairs for the 3 conditions FF (A,B), UF (C,D), and SF (E,F). Subjects had to do a same/different judgment regarding the persons' profession (condition FF), the identity of the faces (condition UF), or the identity of the images (condition SF). Respectively, one matched pair (A,C,D) and one unmatched pair (B,D,F) are shown for each condition. For details, see text.

In order to create picture pairs of unknown faces, we chose 150 grayscale pictures of unfamiliar faces (74 females and 76 males) from the “Multiracial Faces” database created by the Tarrlab at Brown University (“Stimulus images courtesy of Michael J. Tarr, Center for the Neural Basis of Cognition, Carnegie Mellon University, http://www.tarrlab.org/”). Four types of picture pairs were created: same females, same males, different females, and different males. Picture pairs of the same person were taken from slightly different perspectives and showed the person with slightly different facial expressions. In contrast, picture pairs of different persons were selected to be as similar as possible (e.g., pairs, see Fig. 1C,D). This allowed us to increase task difficulty, matching it as closely as possible to the level of difficulty in the condition FF. Unfamiliar faces were matched with famous faces for age, nationality, and confounding factors such as position of the face, expressions, luminosity, and the presence of glasses or earrings. We selected 80 picture pairs (i.e., 20 of each type), which were most consistently perceived as the same or different person in tests with a sample of 18 healthy subjects.

The stimuli for the control condition SF were created using custom written Matlab scripts, scrambling both types of faces, that is, famous faces (FF) and unknown faces (UF) (see Fig. 1E,F). To maintain a constant spatial frequency power density spectrum in these scrambled faces, the manipulation was performed on the phases of each spatial frequency in the image. The phase of each lower frequency component, starting from the lowest frequency, was swapped with the phase of a corresponding higher frequency component, starting with the highest. A pattern was obtained that was no longer recognizable as a face. The scrambled faces were arranged in 20 pairs of different pictures and 20 pairs of identical pictures.

All pictures were scaled to 315 × 260 pixels (visual angle: 6.05° × 4.85°). Pairs of pictures were displayed next to each other, in the center of the visual field and on a black background.

Visual Stimulation

Stimuli were back-projected onto a screen with a liquid-crystal projector at a frame rate of 60 Hz and a screen resolution of 1280 × 1024 pixels. Participants viewed the stimuli binocularly via an adjustable mirror mounted on the head coil. Stimulation was programmed using the in-house software ASF (Schwarzbach 2011), based on the MATLAB Psychtoolbox-3 (Brainard 1997) for Windows.

Behavioral Data Collection and Analysis

Subject responses were collected with fMRI-compatible response pads for the left and right hand (Lumina LU400-PAIR, Cedrus, United States). Reaction time and accuracy of response were calculated for the different conditions and compared among conditions using one-way repeated measures analysis of variance. Post hoc analysis was performed using t-tests and Bonferroni correction.

fMRI Data Acquisition and Preprocessing

Functional and structural images were acquired with the parameters listed in the methods description of the optimization study. Only optimized EPI was used for functional scanning runs. For both functional runs, 405 volumes were acquired. A point-spread function (PSF) scan was acquired prior to each functional run for distortion correction (Zeng and Constable 2002; Zaitsev et al. 2004). The first 5 volumes of each run were discarded to allow T1 equilibrium to be established. Further preprocessing was performed with SPM5 (http://www.fil.ion.ucl.ac.uk/spm/software/spm5; Friston et al. 2007), including slice time correction and motion correction. The mean functional image was coregistered with the structural image using a rigid body transformation. Structural images were segmented, bias corrected, and spatially normalized to MNI space using a unified segmentation procedure (Ashburner and Friston 2005). Functional images were normalized to MNI space, using the same parameters, and spatially smoothed with a Gaussian kernel of 8-mm full-width at half-maximum (FWHM).

Trial Splitting and Hypothesis Testing

The aim of our study was to reveal, in a first analysis, brain areas involved in the processing of famous faces in general. In a second analysis, we attempted to distinguish areas contributing to lexical retrieval. Both these analyses required the classification of trials in the FF condition depending on the subjects' ability to correctly identify and/or name the famous faces in the postscanning behavioral assessment. Although we collected behavioral responses to the profession-matching task in the scanner, we considered the postscanning explicit description as a more specific index of semantic knowledge. Furthermore, naming could only be assessed postscanning. The schemes of trial splitting for our 2 major analyses are described below:

Trial splitting for Analysis 1 (famous faces network).

In the first analysis, we were interested in revealing the overall effect of semantic processing, independent of lexical retrieval. To do so, we isolated trials in which subjects knew both famous faces and therefore had access to the related semantic information. This was done by splitting trials in condition FF in 2 groups: trials in which both faces were correctly identified in the postscanning assessment (FF-known) and trials in which subjects could not identify both faces correctly (FF-unknown). Contrasting FF-known against condition SF would then reveal the overall network involved in processing famous faces. The range of processes captured by this contrast would include high level visual processing specific to faces as well as semantic and lexical retrieval processes. A conjunction of this contrast, that is, FF-unknown versus SF, with the contrast of condition UF versus SF would allow isolation of high-level visual processing common to both tasks. A third contrast, between FF-known and condition UF, could finally reveal all semantic and lexical processes which go beyond the pure visual processing of faces.

Trial splitting for Analysis 2 (naming effect).

In the second analysis, we were interested in identifying networks contributing to lexical retrieval processes. In order to study lexical retrieval, without confounding it with different levels of semantic processing or the task performed in the scanner, we included only trials in which both faces could be correctly identified (i.e., FF-known) but split these trials further in 2 subgroups: trials in which both faces were correctly named in the postscanning testing (FF-named) and trials in which subjects could not name both faces (FF-unnamed). The ability to retrieve proper names could then be captured by contrasting FF-named against FF-unnamed. In order to avoid effects being compromised by noise, we conducted this second analysis only for subjects who had at least 16 trials of each type. There were 12 subjects who met this criterion and were therefore included in the second analysis.

Since FF trials were classified as known, unknown, named, or unnamed based on the postscanning session we included in the analyses also trials in which subjects gave an “incorrect” performance in the scanner. Since postscanning performance showed that subjects actually knew the celebrities, these incorrect responses on the semantic matching task in the scanner were really atypical categorizations since celebrities can have more than one profession (e.g., actor and singer). Similarly, FF trials in which subjects did not identify both faces postscanning were classified as FF-unknown and excluded from the analyses, regardless of performance in the scanner.

fMRI Statistical Analysis

Effects at the individual subject level were estimated by fitting a general linear model for each voxel using SPM5. The 2 functional runs for each subject were concatenated. The design matrix consisted of one explanatory variable (EV) per experimental condition and run. The number of EVs was different for our 2 analyses, depending on the scheme of trial splitting in condition FF (see above). For the first analysis, 4 EVs were used, corresponding to condition FF-known, FF-unknown, UF, and SF. The EV for condition FF-unknown was included in the model as an effect of no interest. For the second analysis, 5 EVs were used, corresponding to the conditions FF-named, FF-unnamed, FF-unknown, UF, and SF. Here, only the first 2 EVs were of interest for the experimental hypothesis. All of these EVs were created by convolving a boxcar function (corresponding in duration to the stimulus presentation) with a canonical hemodynamic response function. For each run, 6 additional regressors were included, corresponding to the head motion parameters estimated during the realignment step, and one variable encoding the mean of the run. Model parameters were estimated using restricted maximum likelihood (ReML) using an autoregressive AR(1) model to correct for nonsphericity arising from serial correlations. The data and model were high-pass filtered with a cutoff frequency of 1/128 Hz.

Contrast images calculated in the first level of analysis were entered in a random effects analysis, to infer effects on the population level. This second level of analysis was conducted using the flexible-factorial design implemented in SPM5. Contrasts at the second level were calculated at the single voxel level and corrected for multiple comparisons. For the Analysis 1 (i.e., semantic contrast), we corrected for familywise error (FWE) at P < 0.05, taking advantage of the ability of FWE to detect small clusters that are reliably activated. For the Analysis 2 (i.e., lexical contrast), we did cluster size correction because we did not anticipate strong effects given that naming was an implicit process and given the lower number of trials. Following a whole brain uncorrected voxelwise thresholding at P < 0.01, we accepted as significant only activations surviving at P < 0.05 (FWE) at the cluster level. We also performed a small volume correction to reduce the risk of false negative results in the left temporal lobe. The ATL volume included the temporal pole (TP) and extended posteriorly to the −10 mm MNI coordinate, in order to include also the MTG cluster found in the semantic contrast between conditions FF versus UF (Fig. 3D).

All results were displayed with MRIcron (Version 7 July 2009, Chris Rorden, http://www.mricro.com), overlaying functional data on the provided single subject T1 template. Anatomical labels were determined based on visual inspection of the data with reference to the atlas of Duvernoy (1999).

Results

Behavioral Results

Based on the postscanning performance, we first sorted FF trials in FF-known and FF-unknown (Analysis 1, overall famous faces network). For the lexical retrieval analysis (Analysis 2, naming effect), we split the FF-known trials in FF-named and FF-unnamed (for more details, see Materials and Methods). The number of FF-known trials ranged from 79 to 130. FF-unknown trials were generally fewer and even absent for some subjects and were not included in the analyses. After splitting FF-known further, we had enough FF-named (range from 16 to 97) and FF-unnamed (range from 16 to 63) trials for 12 subjects. The proportions of FF-named and FF-unnamed trials are shown for these 12 subjects in Figure 2B.

Figure 2.

Reaction times and accuracy in the scanner is shown in panel A for all 21 subjects and in all conditions (i.e., conditions FF, UF, and SF). According to the postscanning assessment, trials of condition FF were split individually for each subject into a first group in which subjects knew the pair of presented famous faces (FF-known) and a second group in which they did not know both faces (FF-unknown). The second group of trials was considered “real” errors, since subjects could do the semantic matching task inside the scanner only on guessing. Consequently, these trials were excluded from the analysis of functional data. Analysis 1 investigated the overall famous faces network. For Analysis 2, which investigated the naming effect, the trials with known famous faces were further split into 2 subgroups (i.e., FF-named, FF-unnamed) according to subjects' ability to name them in the postscanning assessment. Panel B shows the percentages of FF trials falling in these subgroups for the 12 subjects included in Analysis 2. Panel C shows for the same subjects the performance in the semantic matching task done during the scanning session. For details, see text.

Figure 2.

Reaction times and accuracy in the scanner is shown in panel A for all 21 subjects and in all conditions (i.e., conditions FF, UF, and SF). According to the postscanning assessment, trials of condition FF were split individually for each subject into a first group in which subjects knew the pair of presented famous faces (FF-known) and a second group in which they did not know both faces (FF-unknown). The second group of trials was considered “real” errors, since subjects could do the semantic matching task inside the scanner only on guessing. Consequently, these trials were excluded from the analysis of functional data. Analysis 1 investigated the overall famous faces network. For Analysis 2, which investigated the naming effect, the trials with known famous faces were further split into 2 subgroups (i.e., FF-named, FF-unnamed) according to subjects' ability to name them in the postscanning assessment. Panel B shows the percentages of FF trials falling in these subgroups for the 12 subjects included in Analysis 2. Panel C shows for the same subjects the performance in the semantic matching task done during the scanning session. For details, see text.

Performance in the scanner was analyzed for trials split based on the postscanning results described above. Subjects responded faster and more accurately during trials of type FF-known compared with trials of type FF-unknown (Fig. 2A). This finding was to be expected since subjects' performance on the trials in which they do not know the celebrities should depend mainly on guessing, that is, “real errors.” Compared with trials of type FF-known, subjects reacted faster in conditions UF (t20 = −12.14, P < 0.001) and SF (t20 = −11.21, P < 0.001). Accuracy was also higher in UF (t20 = 4.82, P < 0.01) and SF (t20 = 3.28, P = 0.02) when compared with FF-known. Since FF-known were correctly recognized postscanning, “errors” in this condition are likely due to “atypical” categorization of celebrities with multiple professions (see Materials and Methods). These findings nevertheless indicate that the UF and SF conditions were less effortful. No significant difference was found between conditions UF and SF.

For Analysis 2 (Fig. 2C), there was no significant difference in accuracy during scanning for the FF-named and FF-unnamed (t11 = 2.07, P = 0.38), but reaction times were significantly faster (t11 = −3.83; P = 0.016) for FF-named. This finding suggests greater effort for the FF-unnamed trials. As a consequence, we argue that any positive effect of naming on the functional data (i.e., FF-named > FF-unnamed) cannot simply be explained by task difficulty, which has the opposite sign.

Functional Data

Results of Analysis 1 (Famous Faces Network)

To identify the overall effect of famous face processing (e.g., perceptual processing, structural encoding, face recognition, semantic, lexical and phonological retrieval, and emotional processing), we contrasted FF-known versus SF (Fig. 3A and Table 1). The areas revealed by this contrast were bilateral fusiform and inferior occipital gyrus, left occipitoparietal junction, left precuneus, bilateral amygdala and hippocampus, bilateral caudate, bilateral inferior frontal gyrus (IFG), right medial frontal gyrus, bilateral middle temporal gyrus (MTG), and bilateral TP.

Table 1

Regions activated in the different contrasts performed for Analysis 1

Contrast Brain area MNI coordinates Extent (mm3P Max T 
x y z  
FF versus SF (including 21 subjects) Right fusiform 42 −51 −24 15 282 15.56 
Right inferior occipital 42 −78 −12   13.78 
Right middle temporal 45 −53 15   9.17 
Left precuneus −3 −54 15 54 756 <0.001 12.98 
Left amygdala −21 −6 −12   10.25 
Left thalamus −6 −9   9.88 
Right amygdala 24 −6 −15   8.97 
Right hippocampus 33 −12 −18   8.72 
Left caudate −12   8.72 
Left hippocampus −30 −15 −15   7.99 
Right caudate 12 12   7.08 
Left inferior frontal (pars triangularis) −45 24 21 12 150 <0.001 11.84 
Left fusiform −42 −72 −18 7749 <0.001 10.92 
Left inferior occipital −42 −81 −15   10.68 
Left occipitoparietal junction −36 −75 42 4644 <0.001 9.67 
Right inferior frontal (pars orbitalis) 33 33 −12 1296 <0.001 9.05 
Left TP −39 12 −33 1215 <0.001 9.04 
Right middle temporal 54 −9 −21 1701 <0.001 8.2 
Left middle temporal −57 −6 −18 1269 <0.001 7.65 
Right medial frontal 42 −18 1944 <0.001 7.43 
Right TP 36 12 −33 405 0.001 6.87 
Right inferior frontal (pars triangularis) 45 24 21 648 0.003 6.42 
Conjunction (FF vs. SF; UF vs. SF) (including 21 subjects) Right fusiform 42 −51 −24 9450 <0.001 15.56 
Right inferior occipital 42 −78 −12   13.78 
Right middle temporal 48 −66 12   6.83 
Left fusiform −42 −72 −18 5940 <0.001 10.92 
Left inferior occipital −42 −81 −15   10.68 
Right amygdala 24 −6 −15 2889 <0.001 8.97 
Left amygdala −21 −6 −15 1728 <0.001 8.96 
Right inferior frontal (pars orbitalis) 36 33 −15 324 <0.001 6.97 
Right inferior frontal (pars triangularis) 45 27 18 405 0.004 6.29 
FF versus UF (including 21 subjects) Left precuneus −3 −54 12 128 466 <0.001 15.91 
Left caudate −12   12.27 
Left thalamus −9 −6   11.87 
Left posterior cingulum −3 −36 30   9.48 
Left inferior frontal (pars triangularis) −39 27   9.3 
Right caudate 18 21 −3   9.18 
Left inferior frontal (pars orbitalis) −36 33 −12   7.86 
Left TP −39 15 −33   7.53 
Left temporoparietooccipital junction −33 −72 39 9450 <0.001 11.48 
Left middle temporal −60 −6 −18 2403 <0.001 8.56 
Right middle temporal 60 −3 −15 1809 <0.001 7.54 
Left medial orbitofrontal −3 60 −9 2025 <0.001 7.18 
Right temporoparietooccipital junction 45 −66 30 2457 <0.001 7.08 
Left superior frontal −21 57 270 <0.001 7.01 
Left middle temporal −54 −39 −6 918 0.004 6.25 
Right lingual gyrus 18 −45 −9 324 0.005 6.23 
Right hippocampus 36 −12 −18 81 0.02 5.72 
Right middle temporal 48 −27 54 0.026 5.63 
FF versus UF (including 12 subjects) Left cuneus −6 −66 27 38 664 <0.001 12.81 
Left precuneus −6 −57 12   12.61 
Left posterior cingulum −3 −39 30   7.63 
Left inferior frontal (pars triangularis) −42 24 24 2160 <0.001 8.6 
Left cingulum −9 −18 27 351 <0.001 8.16 
Right middle temporal 60 −15 729 <0.001 8.04 
Left parietooccipital junction −33 −72 42 2889 <0.001 7.65 
Left TP −42 15 −33 540 0.001 7.08 
Right angular 51 −66 27 729 0.004 6.57 
Left middle temporal −57 −6 −18 405 0.005 6.47 
Left superior frontal −24 54 108 0.005 6.46 
Right putamen 15 15 −3 567 0.008 6.27 
Left medial orbitofrontal −3 54 −12 837 0.01 6.18 
Left inferior frontal (pars triangularis) −39 27 81 0.019 5.93 
Left inferior frontal (pars orbitalis) −39 33 −12 54 0.026 5.81 
Left caudate −6 297 0.029 5.77 
Contrast Brain area MNI coordinates Extent (mm3P Max T 
x y z  
FF versus SF (including 21 subjects) Right fusiform 42 −51 −24 15 282 15.56 
Right inferior occipital 42 −78 −12   13.78 
Right middle temporal 45 −53 15   9.17 
Left precuneus −3 −54 15 54 756 <0.001 12.98 
Left amygdala −21 −6 −12   10.25 
Left thalamus −6 −9   9.88 
Right amygdala 24 −6 −15   8.97 
Right hippocampus 33 −12 −18   8.72 
Left caudate −12   8.72 
Left hippocampus −30 −15 −15   7.99 
Right caudate 12 12   7.08 
Left inferior frontal (pars triangularis) −45 24 21 12 150 <0.001 11.84 
Left fusiform −42 −72 −18 7749 <0.001 10.92 
Left inferior occipital −42 −81 −15   10.68 
Left occipitoparietal junction −36 −75 42 4644 <0.001 9.67 
Right inferior frontal (pars orbitalis) 33 33 −12 1296 <0.001 9.05 
Left TP −39 12 −33 1215 <0.001 9.04 
Right middle temporal 54 −9 −21 1701 <0.001 8.2 
Left middle temporal −57 −6 −18 1269 <0.001 7.65 
Right medial frontal 42 −18 1944 <0.001 7.43 
Right TP 36 12 −33 405 0.001 6.87 
Right inferior frontal (pars triangularis) 45 24 21 648 0.003 6.42 
Conjunction (FF vs. SF; UF vs. SF) (including 21 subjects) Right fusiform 42 −51 −24 9450 <0.001 15.56 
Right inferior occipital 42 −78 −12   13.78 
Right middle temporal 48 −66 12   6.83 
Left fusiform −42 −72 −18 5940 <0.001 10.92 
Left inferior occipital −42 −81 −15   10.68 
Right amygdala 24 −6 −15 2889 <0.001 8.97 
Left amygdala −21 −6 −15 1728 <0.001 8.96 
Right inferior frontal (pars orbitalis) 36 33 −15 324 <0.001 6.97 
Right inferior frontal (pars triangularis) 45 27 18 405 0.004 6.29 
FF versus UF (including 21 subjects) Left precuneus −3 −54 12 128 466 <0.001 15.91 
Left caudate −12   12.27 
Left thalamus −9 −6   11.87 
Left posterior cingulum −3 −36 30   9.48 
Left inferior frontal (pars triangularis) −39 27   9.3 
Right caudate 18 21 −3   9.18 
Left inferior frontal (pars orbitalis) −36 33 −12   7.86 
Left TP −39 15 −33   7.53 
Left temporoparietooccipital junction −33 −72 39 9450 <0.001 11.48 
Left middle temporal −60 −6 −18 2403 <0.001 8.56 
Right middle temporal 60 −3 −15 1809 <0.001 7.54 
Left medial orbitofrontal −3 60 −9 2025 <0.001 7.18 
Right temporoparietooccipital junction 45 −66 30 2457 <0.001 7.08 
Left superior frontal −21 57 270 <0.001 7.01 
Left middle temporal −54 −39 −6 918 0.004 6.25 
Right lingual gyrus 18 −45 −9 324 0.005 6.23 
Right hippocampus 36 −12 −18 81 0.02 5.72 
Right middle temporal 48 −27 54 0.026 5.63 
FF versus UF (including 12 subjects) Left cuneus −6 −66 27 38 664 <0.001 12.81 
Left precuneus −6 −57 12   12.61 
Left posterior cingulum −3 −39 30   7.63 
Left inferior frontal (pars triangularis) −42 24 24 2160 <0.001 8.6 
Left cingulum −9 −18 27 351 <0.001 8.16 
Right middle temporal 60 −15 729 <0.001 8.04 
Left parietooccipital junction −33 −72 42 2889 <0.001 7.65 
Left TP −42 15 −33 540 0.001 7.08 
Right angular 51 −66 27 729 0.004 6.57 
Left middle temporal −57 −6 −18 405 0.005 6.47 
Left superior frontal −24 54 108 0.005 6.46 
Right putamen 15 15 −3 567 0.008 6.27 
Left medial orbitofrontal −3 54 −12 837 0.01 6.18 
Left inferior frontal (pars triangularis) −39 27 81 0.019 5.93 
Left inferior frontal (pars orbitalis) −39 33 −12 54 0.026 5.81 
Left caudate −6 297 0.029 5.77 

Note: P values (P) and maximum T statistics (Max T) are reported for the local maximum of each cluster. P values were controlled for FWE (in the whole group of 21 subjects: FWHM = 12.1, 12.5, and 11.6 mm; Volume = 42 101, voxels = 566.7 resels; in the subgroup of 12 subjects: FWHM = 13.0, 13.4, and 12.3 mm; Volume = 43 152, voxels = 478.1 resels). For single clusters, which clearly extended into several areas of the brain, the local maxima in these additional areas are indicated in italics.

To isolate further the effect of visual processing of faces, we calculated the conjunction of contrasts FF-known versus SF and UF versus SF (Fig. 3B and Table 1). Since semantic processing of UF is not possible, this conjunction should identify perceptual areas. Common activations were present in bilateral fusiform and inferior occipital gyrus, right MTG, bilateral amygdala, and right IFG.

Finally, to identify the effect of semantic and covert lexical processes, we contrasted FF-known versus UF (Fig. 3C and Table 1). Areas revealed by this contrast were left IFG, left TP, bilateral temporoparietal junction (TPJ), bilateral MTG, left precuneus, thalamus, posterior cingulum, bilateral caudate, left medial orbitofrontal gyrus, left superior frontal gyrus, right lingual gyrus, and hippocampus.

Consistent with previous studies, perceptual processing of faces involved mainly the fusiform and occipital cortex (Kanwisher et al. 1997; McCarthy et al. 1997), while semantic and lexical processing went well beyond these visual association regions, including our temporal and parietal regions of interest (ROIs) (Perani et al. 1999; Gorno-Tempini and Price 2001).

As stated above, we included only 12 subjects in our second analysis on lexical retrieval. However, we first wanted to show that the subgroup was a representative sample of the whole group. For this reason, we calculated the contrast between trials of type FF-known and condition UF again for the subgroup (Table 1). The overall pattern of activation in this contrast was similar, although some clusters were activated to a lesser extent. Most importantly, stable activation clusters were still present in the left TP, the bilateral anterior MTG, and the bilateral TPJ. These were areas predicted to be involved in famous face processing (Gorno-Tempini et al. 1998), and our particular aim was to disambiguate the role of these areas in semantic and lexical processing, respectively.

Figure 3.

Surface rendering of the parametric maps of t-statistic for Analysis 1. Overall effect of processing famous faces in 21 subjects (A, contrast FF-known vs. SF); effects due to high-level visual processing of faces in 21 subjects (B, conjunction between FF-known vs. SF and UF vs. SF); and effects due to semantic and/or lexical processes in 21 subjects (C, contrast FF-known vs. UF). Height threshold and scale of t-statistic is indicated in C. For details, see text.

Figure 3.

Surface rendering of the parametric maps of t-statistic for Analysis 1. Overall effect of processing famous faces in 21 subjects (A, contrast FF-known vs. SF); effects due to high-level visual processing of faces in 21 subjects (B, conjunction between FF-known vs. SF and UF vs. SF); and effects due to semantic and/or lexical processes in 21 subjects (C, contrast FF-known vs. UF). Height threshold and scale of t-statistic is indicated in C. For details, see text.

Results of Analysis 2 (Naming Effect)

The aim of the second analysis was the identification of brain areas that would show greater activation for implicit naming. For this purpose, we compared the BOLD signal for FF-known trials that were named in the postscanning session (FF-named) versus the ones that were correctly identified but not named (FF-unnamed) postscanning. Data from 12 subjects were included in this analysis (for explanation, see Behavioral Results). Using cluster size correction (see Materials and Methods), this analysis revealed one large cluster (size = 269 voxels = 7263 mm3), comprising areas in the left inferior parietal (TPJ) and in the left posterior MTG (Fig. 4). To reduce the risk of a false negative result in the left ATL caused by signal noise, we used also a small volume correction including only the left ATL. This analysis confirmed the absence of any effect in that region.

Figure 4.

Effect of naming (Analysis 2). Surface rendering of the significant cluster revealed by cluster thresholding at P = 0.01 (A). Blue lines indicate the anterior–posterior position of coronal sections shown in panels B and C. The local maxima in the MTG (B) and in the TPJ (C) are indicated by blue crosshairs. Percent signal change is shown for masked 10 mm spheres in both local maxima.

Figure 4.

Effect of naming (Analysis 2). Surface rendering of the significant cluster revealed by cluster thresholding at P = 0.01 (A). Blue lines indicate the anterior–posterior position of coronal sections shown in panels B and C. The local maxima in the MTG (B) and in the TPJ (C) are indicated by blue crosshairs. Percent signal change is shown for masked 10 mm spheres in both local maxima.

In order to visualize the size of the lexical retrieval effect, BOLD signal was calculated within 2 ROIs centered at the local maxima of the activation cluster in the posterior MTG (Fig. 4B, MNI coordinates: x −63; y −54; z 6) and in the TPJ (Fig. 4C; MNI coordinates: x −42; y −60; z 48). ROIs were defined as all voxels within a 10 mm sphere around the local activation maximum and being located within the overall activation cluster.

Post Hoc Analyses

A supplementary analysis was carried out to investigate how FF familiarity could contribute to the naming effect. Familiarity ratings were collected during stimulus assembly (see above). The average familiarity rating of the 2 FF shown in each trial was covaried out at the single subject level by adding an additional EV. The clusters in the MTG and TPJ were still activated for FF-named versus FF-unnamed, though their volumes were reduced (4023 and 891 mm3, respectively). Only the MTG survived correction for multiple comparison at the cluster level. So, familiarity might have contributed to the effect found in the MTG and TPJ but could not explain it entirely.

We measured BOLD sensitivity in the ATL (because of EPI protocol optimization), MTG, and in the TPJ to investigate whether the lack of a significant ATL effect in naming could be explained by lower overall signal in the ATL. This is actually unlikely since ATL optimization is expected to reduce BOLD signal in areas where there is no susceptibility artifact, such as the MTG and TPJ. We nevertheless investigated this possibility by comparing tSNR in 3 ROIs along the left temporal–parietal lobes. The first ROI was the left ATL, defined anatomically as for the optimization study. The other 2 ROIs were the MTG and TPJ clusters. The tSNR was calculated from resting-state data collected during the optimization study. The average tSNR value in the ATL ROI was significantly higher than that in the other 2 ROIs (see Supplementary Fig. 2). Thus, the lack of significant lexical retrieval effect in the ATL cannot be explained by lower BOLD sensitivity in this area.

Discussion

The main goal of our study was to identify the differential role of the ATL and posterior temporoparietal regions in processing semantic (biographical information) and lexical (proper names) information. We applied an ATL-optimized fMRI protocol and showed that a network of regions in the bilateral temporal lobes is involved in recognizing, identifying, and naming famous people. The ATL bilaterally was mainly involved in semantic processing, while more posterior left temporoparietal regions were modulated by lexical retrieval processes. Here, we discuss the implications of our results for understanding the functional neuroanatomy of semantic processing and lexical retrieval.

We found that the ATL was involved in semantic processing irrespective of whether or not names could be retrieved for the identified famous faces. Both the left and the right ATL responded to processing semantic information, regardless of naming ability, suggesting a major role of both ATLs in person-related semantic processing. Patients with semantic variant PPA (or semantic dementia) and ATL atrophy indeed have severe problems identifying objects, including people. While most patients with semantic variant PPA have bilateral or left greater than right ATL atrophy and show deficits for objects and people (Patterson et al. 2007), some patients with greater right ATL atrophy show greater difficulty in processing biographical information regarding people (Evans et al. 1995; Gainotti et al. 2003). Our results support the view that both hemispheres play an important role in retrieving person-specific semantic information, although they do not exclude that different types of information are processed by each hemisphere (for related literature, see Snowden et al. 2004; Rankin et al. 2006; Gainotti 2007; Butler et al. 2009; Brambati et al. 2010). Connectivity of the ATL with visuospatial and emotional networks in the right and language areas in the left hemisphere might determine a preferential role of this region in processing visual, verbal, and social information (Snowden et al. 2004; Gainotti 2007). Furthermore, our results can be accommodated in relation to recent cognitive models of ATL function. One prominent model states that the ATL acts as a semantic hub, forming amodal semantic representations, which would enable semantic generalization on the basis of conceptual structure rather than modality-specific features (Patterson et al. 2007; Lambon Ralph and Patterson 2008). Another prominent account claims that the ATL supports social conceptual knowledge in general (Simmons and Martin 2009; Simmons et al. 2010). Our study supports a central role of the bilateral ATL in semantics processing. Differences in emotional valence between famous and nonfamous faces might have contributed to the activation in the ATL. However, our study investigated only social stimuli (i.e., faces) and was designed to investigate lexical and not category-specific effects.

We found that covert naming modulated activation in more posterior parts of the left temporal lobe and TPJ. These findings are consistent with a role of these regions in the retrieval and encoding of phonological forms of lexical items. A view suggested by Benson (1979) and Geschwind (1967) in their seminal descriptions of different types of anomia and their neural correlates. Also, the finding that these areas are often damaged in patients with Wernicke's aphasia or transcortical sensory aphasic, who evolve to anomic aphasia (Albert et al. 1981), is consistent with this idea. There is no detailed study of semantic memory in the old cases, but patients with aphasia due to vascular posterior left perisylvian damage do not usually report object or face identification deficits in everyday life. Similarly, patients with logopenic PPA show impaired naming but relatively spared nonverbal semantic association abilities (Gorno-Tempini et al. 2004; Henry and Gorno-Tempini 2010). Given their atrophy being most prominent in posterior temporal and inferior parietal areas, the symptoms of these patients give further support for the role of these areas in lexical–phonological processing. However, many functional neuroimaging studies, including ours, have shown activations in the left inferior parietal regions in semantic tasks and a role of this region in semantics has been postulated (Price et al. 1999; Binder et al. 2009,Seghier et al. 2010). Most of these studies were not designed to differentiate regions that would respond preferentially to naming and semantic categorization. When we performed this specific contrast the left TPJ was most involved in naming. Our results therefore suggest that the left inferior parietal region (together with the ATL, the IFG, medial frontal, and subcortical areas) is part of the semantic network but that, within this network, it is particularly involved in name retrieval.

Additional roles in language processing have been suggested for other regions of the inferior parietal lobe. For example, the ventral supramarginal gyrus might have a role in articulation and higher phonological processing (for review, see Price 2010). This area was not activated in our experiment, possibly because the covert retrieval of proper names did not reach the phonological/articulatory level.

One of the strengths of our study was the idea to utilize famous people as stimuli. Famous faces indeed allowed us to dissociate semantics and naming in healthy subjects, as lexical retrieval failure for proper names is common. This would not have been possible with other object categories (e.g., animals, tools, etc.) and common names. However, dissociations between semantic and name retrieval processes for common and proper names have been reported (for review, see Semenza 2006). Whether our findings generalize to all lexical items remains to be established.

A limitation of our study might be that naming abilities could be tested only after the scanning session. The naming ability outside the scanner might have been slightly better due to the repeated presentation of all stimuli or a bit worse due to fatigue. Misclassification of faces as either named or unnamed might have slightly weakened the statistical contrast between these trials. It cannot be excluded that increasing sensitivity could reveal an effect for naming also in the ATL. However, we suggest that in this case, the effect in the posterior temporal and parietal areas would increase as well. The main conclusion that the posterior temporal and parietal areas play the predominant role in the retrieval of proper names would then remain valid.

We want to emphasize, however, that our results do not imply that lexical processing is exclusively accomplished by these posterior areas. For instance, earlier stages of lexical processing involving intermediate representations between semantic and phonological levels (termed “lemmas” by some researchers) may depend on more anterior temporal regions (Damasio et al. 1996, 2004; Indefrey and Levelt 2004; Tsukiura et al. 2008). In a recent study using voxel-based lesion-symptom mapping, Schwartz et al. (2009) (Walker et al. 2011) found that damage to anterior and midtemporal regions was predictive of semantic naming errors (e.g., naming a cat as “dog”), suggesting a role in lemma retrieval (though cf. Tsapkini et al. 2011) who did not find an anterior temporal locus for semantic errors in acute stroke patients. If anterior and midtemporal regions are involved in intermediate stages of lexical access, such regions would likely be undetected by our paradigm because they may be activated even when the phonological form of a name cannot be retrieved.

In summary our data suggest that the ATL is mainly involved in semantic processing, while lexical retrieval is attributed mainly to areas in the posterior–temporal lobe and the TPJ. One can therefore speculate a cascade of processes in the temporal lobe network, starting with semantic integration in the ATL, and leading further to the activation of lexical representations in the posterior portion of the MTG and phonological assembly in the posterior superior and TPJ.

Conclusion

Using an imaging sequence optimized for the ATL and considering the participants' ability to identify and name famous faces, we were able to study the neural basis of semantic memory and lexical retrieval and in particular the differential roles of anterior and posterior temporal regions in these processes. Our findings indicate that the ATL is involved in semantic processing, while more posterior left temporal and temporoparietal regions are involved in lexical retrieval processes.

Supplementary Material

Supplementary material can be found at: http://www.cercor.oxfordjournals.org/

Funding

Marie Curie International Reintegration Grants (MIRG-CT-2007-046512 to M.L.G.T.); Provincia Autonoma di Trento; Fondazione Cassa di Risparmio di Trento e Rovereto, Italy.

We thank Maxim Zaitsev for providing the PSF EPI sequence. Conflict of Interest : None declared.

References

Albert
ML
Goodglass
H
Helm
NA
Rubens
AB
Alexander
MP
Clinical aspects of dysphasia
 , 
1981
Wien (Austria)
Springer-Verlag
Ashburner
J
Friston
KJ
Unified segmentation
Neuroimage
 , 
2005
, vol. 
26
 (pg. 
839
-
851
)
Bandettini
PA
Wong
EC
Jesmanowicz
A
Hinks
RS
Hyde
JS
Spin-echo and gradient-echo EPI of human brain activation using BOLD contrast: a comparative study at 1.5 T
NMR Biomed
 , 
1994
, vol. 
7
 (pg. 
12
-
20
)
Benson
DF
Whitaker
H
Whitaker
HA
Neurologic correlates of anomia
Studies in neurolinguistic
 , 
1979
New York
Academic Press
Binder
JR
Desai
RH
Graves
WW
Conant
LL
Where is the semantic system? A critical review and meta-analysis of 120 functional neuroimaging studies
Cereb Cortex
 , 
2009
, vol. 
19
 (pg. 
2767
-
2796
)
Binney
RJ
Embleton
KV
Jefferies
E
Parker
GJ
Ralph
MA
The ventral and inferolateral aspects of the anterior temporal lobe are crucial in semantic memory: evidence from a novel direct comparison of distortion-corrected fMRI, rTMS, and semantic dementia
Cereb Cortex
 , 
2010
, vol. 
20
 (pg. 
2728
-
2738
)
Brainard
DH
The Psychophysics Toolbox
Spat Vis
 , 
1997
, vol. 
10
 (pg. 
433
-
436
)
Brambati
SM
Benoit
S
Monetta
L
Belleville
S
Joubert
S
The role of the left anterior temporal lobe in the semantic processing of famous faces
Neuroimage
 , 
2010
, vol. 
53
 (pg. 
674
-
681
)
Bredart
S
Valentine
T
Calder
A
Gassi
L
An interactive activation model of face naming
Q J Exp Psychol A
 , 
1995
, vol. 
48
 (pg. 
466
-
486
)
Brown
AS
A review of the tip-of-the-tongue experience
Psychol Bull
 , 
1991
, vol. 
109
 (pg. 
204
-
223
)
Bruce
V
Young
A
Understanding face recognition
Br J Psychol
 , 
1986
, vol. 
77
 
Pt 3
(pg. 
305
-
327
)
Burton
AM
Bruce
V
Johnston
RA
Understanding face recognition with an interactive activation model
Br J Psychol
 , 
1990
, vol. 
81
 
Pt 3
(pg. 
361
-
380
)
Butler
CR
Brambati
SM
Miller
BL
Gorno-Tempini
ML
The neural correlates of verbal and nonverbal semantic processing deficits in neurodegenerative disease
Cogn Behav Neurol
 , 
2009
, vol. 
22
 (pg. 
73
-
80
)
Butterworth
B
Howard
D
McLoughlin
P
The semantic deficit in aphasia: the relationship between semantic errors in auditory comprehension and picture naming
Neuropsychologia
 , 
1984
, vol. 
22
 (pg. 
409
-
426
)
Damasio
H
Grabowski
TJ
Tranel
D
Hichwa
RD
Damasio
AR
A neural basis for lexical retrieval
Nature
 , 
1996
, vol. 
380
 (pg. 
499
-
505
)
Damasio
H
Tranel
D
Grabowski
T
Adolphs
R
Damasio
A
Neural systems behind word and concept retrieval
Cognition
 , 
2004
, vol. 
92
 (pg. 
179
-
229
)
De Panfilis
C
Schwarzbauer
C
Positive or negative blips? The effect of phase encoding scheme on susceptibility-induced signal losses in EPI
Neuroimage
 , 
2005
, vol. 
25
 (pg. 
112
-
121
)
Deichmann
R
Gottfried
JA
Hutton
C
Turner
R
Optimized EPI for fMRI studies of the orbitofrontal cortex
Neuroimage
 , 
2003
, vol. 
19
 (pg. 
430
-
441
)
Devlin
JT
Russell
RP
Davis
MH
Price
CJ
Wilson
J
Moss
HE
Matthews
PM
Tyler
LK
Susceptibility-induced loss of signal: comparing PET and fMRI on a semantic task
Neuroimage
 , 
2000
, vol. 
11
 (pg. 
589
-
600
)
Duvernoy
HM
The human brain: surface, three-dimensional sectional anatomy with MRI, and blood supply
 , 
1999
Wien (Austria)
Springer
Evans
JJ
Heggs
AJ
Antoun
N
Hodges
JR
Progressive prosopagnosia associated with selective right temporal-lobe atrophy—a new syndrome
Brain
 , 
1995
, vol. 
118
 (pg. 
1
-
13
)
Friston
KJ
Ashburner
J
Kiebel
S
Nichols
T
Penny
W
Statistical parametric mapping
 , 
2007
London
Academic Press
Gainotti
G
Different patterns of famous people recognition disorders in patients with right and left anterior temporal lesions: a systematic review
Neuropsychologia
 , 
2007
, vol. 
45
 (pg. 
1591
-
1607
)
Gainotti
G
Barbier
A
Marra
C
Slowly progressive defect in recognition of familiar people in a patient with right anterior temporal atrophy
Brain
 , 
2003
, vol. 
126
 (pg. 
792
-
803
)
Gainotti
G
Miceli
G
Caltagirone
C
Silveri
MC
Masullo
C
The relationship between type of naming error and semantic-lexical discrimination in aphasic patients
Cortex
 , 
1981
, vol. 
17
 (pg. 
401
-
410
)
Gainotti
G
Silveri
MC
Villa
G
Miceli
G
Anomia with and without lexical comprehension disorders
Brain Lang
 , 
1986
, vol. 
29
 (pg. 
18
-
33
)
Gati
JS
Menon
RS
Ugurbil
K
Rutt
BK
Experimental determination of the BOLD field strength dependence in vessels and tissue
Magn Reson Med
 , 
1997
, vol. 
38
 (pg. 
296
-
302
)
Geschwind
N
The varieties of naming error
Cortex
 , 
1967
, vol. 
3
 (pg. 
96
-
112
)
Gorno-Tempini
ML
Brambati
SM
Ginex
V
Ogar
J
Dronkers
NF
Marcone
A
Perani
D
Garibotto
V
Cappa
SF
Miller
BL
The logopenic/phonological variant of primary progressive aphasia
Neurology
 , 
2008
, vol. 
71
 pg. 
1227
 
Gorno-Tempini
ML
Cipolotti
L
Price
CJ
Category differences in brain activation studies: where do they come from?
Proc Biol Sci
 , 
2000
, vol. 
267
 (pg. 
1253
-
1258
)
Gorno-Tempini
ML
Dronkers
NF
Rankin
KP
Ogar
JM
Phengrasamy
L
Rosen
HJ
Johnson
JK
Weiner
MW
Miller
BL
Cognition and anatomy in three variants of primary progressive aphasia
Ann Neurol
 , 
2004
, vol. 
55
 pg. 
335
 
Gorno-Tempini
ML
Hillis
AE
Weintraub
S
Kertesz
A
Mendez
M
Cappa
SF
Ogar
JM
Rohrer
JD
Black
S
Boeve
BF
, et al.  . 
Classification of primary progressive aphasia and its variants
Neurology
 , 
2011
, vol. 
76
 (pg. 
1006
-
1014
)
Gorno-Tempini
ML
Hutton
C
Josephs
O
Deichmann
R
Price
C
Turner
R
Echo time dependence of BOLD contrast and susceptibility artifacts
Neuroimage
 , 
2002
, vol. 
15
 (pg. 
136
-
142
)
Gorno-Tempini
ML
Price
CJ
Identification of famous faces and buildings: a functional neuroimaging study of semantically unique items
Brain
 , 
2001
, vol. 
124
 (pg. 
2087
-
2097
)
Gorno-Tempini
ML
Price
CJ
Josephs
O
Vandenberghe
R
Cappa
SF
Kapur
N
Frackowiak
RS
The neural systems sustaining face and proper-name processing
Brain
 , 
1998
, vol. 
121
 
Pt 11
(pg. 
2103
-
2118
)
Grossman
M
Mickanin
J
Onishi
K
Hughes
E
D'Esposito
M
Ding
XS
Alavi
A
Reivich
M
Progressive nonfluent aphasia: language, cognitive, and PET measures contrasted with probable Alzheimer's disease
J Cogn Neurosci
 , 
1996
, vol. 
8
 (pg. 
135
-
154
)
Henry
ML
Gorno-Tempini
ML
The logopenic variant of primary progressive aphasia
Curr Opin Neurol
 , 
2010
, vol. 
23
 (pg. 
633
-
637
)
Hillis
AE
Rapp
B
Romani
C
Caramazza
A
Selective impairment of semantics in lexical processing
Cogn Neuropsychol
 , 
1990
, vol. 
7
 (pg. 
191
-
243
)
Hodges
JR
Patterson
K
Oxbury
S
Funnell
E
Semantic dementia. Progressive fluent aphasia with temporal lobe atrophy
Brain
 , 
1992
, vol. 
115
 
Pt 6
(pg. 
1783
-
1806
)
Howard
D
Lexical anemia—or the case of the missing lexical entries
Q J Exp Psychol A
 , 
1995
, vol. 
48
 (pg. 
999
-
1023
)
Howard
D
Gatehouse
C
Distinguishing semantic and lexical word retrieval deficits in people with aphasia
Aphasiology
 , 
2006
, vol. 
20
 (pg. 
921
-
950
)
Indefrey
P
Levelt
WJ
The spatial and temporal signatures of word production components
Cognition
 , 
2004
, vol. 
92
 (pg. 
101
-
144
)
Insausti
R
Juottonen
K
Soininen
H
Insausti
AM
Partanen
K
Vainio
P
Laakso
MP
Pitkanen
A
MR volumetric analysis of the human entorhinal, perirhinal, and temporopolar cortices
AJNR Am J Neuroradiol
 , 
1998
, vol. 
19
 (pg. 
659
-
671
)
Kanwisher
N
McDermott
J
Chun
MM
The fusiform face area: a module in human extrastriate cortex specialized for face perception
J Neurosci
 , 
1997
, vol. 
17
 (pg. 
4302
-
4311
)
Kruger
G
Kastrup
A
Glover
GH
Neuroimaging at 1.5 T and 3.0 T: comparison of oxygenation-sensitive magnetic resonance imaging
Magn Reson Med
 , 
2001
, vol. 
45
 (pg. 
595
-
604
)
Lambon Ralph
MA
Patterson
K
Generalization and differentiation in semantic memory: insights from semantic dementia
Ann N Y Acad Sci
 , 
2008
, vol. 
1124
 (pg. 
61
-
76
)
Lambon Ralph
MA
Sage
K
Roberts
J
Classical anomia: a neuropsychological perspective on speech production
Neuropsychologia
 , 
2000
, vol. 
38
 (pg. 
186
-
202
)
McCarthy
G
Puce
A
Gore
JC
Allison
T
Face-specific processing in the human fusiform gyrus
J Cogn Neurosci
 , 
1997
, vol. 
9
 (pg. 
605
-
610
)
Mesulam
MM
Slowly progressive aphasia without generalized dementia
Ann Neurol
 , 
1982
, vol. 
11
 (pg. 
592
-
598
)
Mesulam
MM
Primary progressive aphasia—a 25-year retrospective
Alzheimer Dis Assoc Disord
 , 
2007
, vol. 
21
 (pg. 
S8
-
S11
)
Mesulam
MM
Weintraub
S
Spectrum of primary progressive aphasia
Baillieres Clin Neurol
 , 
1992
, vol. 
1
 (pg. 
583
-
609
)
Mummery
CJ
Patterson
K
Hodges
JR
Price
CJ
Functional neuroanatomy of the semantic system: divisible by what?
J Cogn Neurosci
 , 
1998
, vol. 
10
 (pg. 
766
-
777
)
Nickels
L
Howard
D
A frequent occurrence—factors affecting the production of semantic errors in aphasic naming
Cogn Neuropsychol
 , 
1994
, vol. 
11
 (pg. 
289
-
320
)
Ojemann
JG
Akbudak
E
Snyder
AZ
McKinstry
RC
Raichle
ME
Conturo
TE
Anatomic localization and quantitative analysis of gradient refocused echo-planar fMRI susceptibility artifacts
Neuroimage
 , 
1997
, vol. 
6
 (pg. 
156
-
167
)
Papinutto
N
Jovicich
J
Optimization of brain tissue contrast in structural images at 4T: a computer simulation and validation study
 , 
2008
Valencia (Spain)
European Society for Magnetic Resonance in Medicine and Biology
Patterson
K
Nestor
PJ
Rogers
TT
Where do you know what you know? The representation of semantic knowledge in the human brain
Nat Rev Neurosci
 , 
2007
, vol. 
8
 (pg. 
976
-
987
)
Perani
D
Schnur
T
Tettamanti
M
Gorno-Tempini
M
Cappa
SF
Fazio
F
Word and picture matching: a PET study of semantic category effects
Neuropsychologia
 , 
1999
, vol. 
37
 (pg. 
293
-
306
)
Price
CJ
The anatomy of language: a review of 100 fMRI studies published in 2009
Ann N Y Acad Sci
 , 
2010
, vol. 
1191
 (pg. 
62
-
88
)
Price
CJ
Mummery
CJ
Moore
CJ
Frakowiak
RS
Friston
KJ
Delineating necessary and sufficient neural systems with functional imaging studies of neuropsychological patients
J Cogn Neurosci
 , 
1999
, vol. 
11
 (pg. 
371
-
382
)
Rankin
KP
Gorno-Tempini
ML
Allison
SC
Stanley
CM
Glenn
S
Weiner
MW
Miller
BL
Structural anatomy of empathy in neurodegenerative disease
Brain
 , 
2006
, vol. 
129
 (pg. 
2945
-
2956
)
Robinson
S
Windischberger
C
Rauscher
A
Moser
E
Optimized 3 T EPI of the amygdalae
Neuroimage
 , 
2004
, vol. 
22
 (pg. 
203
-
210
)
Schwartz
MF
Kimberg
DY
Walker
GM
Faseyitan
O
Brecher
A
Dell
GS
Coslett
HB
Anterior temporal involvement in semantic word retrieval: voxel-based lesion-symptom mapping evidence from aphasia
Brain
 , 
2009
, vol. 
132
 (pg. 
3411
-
3427
)
Schwartz
MF
Marin
OS
Saffran
EM
Dissociations of language function in dementia: a case study
Brain Lang
 , 
1979
, vol. 
7
 (pg. 
277
-
306
)
Schwarzbach
J
A simple framework (ASF) for behavioral and neuroimaging experiments based on the psychophysics toolbox for MATLAB
Behav Res Methods
 , 
2011
Seghier
ML
Fagan
E
Price
CJ
Functional subdivisions in the left angular gyrus where the semantic system meets and diverges from the default network
J Neurosci
 , 
2010
, vol. 
30
 (pg. 
16809
-
16817
)
Semenza
C
Retrieval pathways for common and proper names
Cortex
 , 
2006
, vol. 
42
 (pg. 
884
-
891
)
Simmons
WK
Martin
A
The anterior temporal lobes and the functional architecture of semantic memory
J Int Neuropsychol Soc
 , 
2009
, vol. 
15
 (pg. 
645
-
649
)
Simmons
WK
Reddish
M
Bellgowan
PS
Martin
A
The selectivity and functional connectivity of the anterior temporal lobes
Cereb Cortex
 , 
2010
, vol. 
20
 (pg. 
813
-
825
)
Snowden
JS
Thompson
JC
Neary
D
Knowledge of famous faces and names in semantic dementia
Brain
 , 
2004
, vol. 
127
 (pg. 
860
-
872
)
Tranel
D
Impaired naming of unique landmarks is associated with left temporal polar damage
Neuropsychology
 , 
2006
, vol. 
20
 (pg. 
1
-
10
)
Tranel
D
Damasio
H
Damasio
AR
A neural basis for the retrieval of conceptual knowledge
Neuropsychologia
 , 
1997
, vol. 
35
 (pg. 
1319
-
1327
)
Triantafyllou
C
Hoge
RD
Krueger
G
Wiggins
CJ
Potthast
A
Wiggins
GC
Wald
LL
Comparison of physiological noise at 1.5 T, 3 T and 7 T and optimization of fMRI acquisition parameters
Neuroimage
 , 
2005
, vol. 
26
 (pg. 
243
-
250
)
Tsapkini
K
Frangakis
CE
Hillis
AE
The function of the left anterior temporal pole: evidence from acute stroke and infarct volume
Brain
 , 
2011
 
doi: 10.3758/s13428-011-0106-8
Tsukiura
T
Suzuki
C
Shigemune
Y
Mochizuki-Kawai
H
Differential contributions of the anterior temporal and medial temporal lobe to the retrieval of memory for person identity information
Hum Brain Mapp
 , 
2008
, vol. 
29
 (pg. 
1343
-
1354
)
Valentine
T
Brennen
T
Brèdart
S
The cognitive psychology of proper names: on the importance of being ernest
 , 
1996
London
Routledge
Vandenberghe
R
Price
C
Wise
R
Josephs
O
Frackowiak
RS
Functional anatomy of a common semantic system for words and pictures
Nature
 , 
1996
, vol. 
383
 (pg. 
254
-
256
)
Visser
M
Jefferies
E
Lambon Ralph
MA
Semantic processing in the anterior temporal lobes: a meta-analysis of the functional neuroimaging literature
J Cogn Neurosci
 , 
2010
, vol. 
22
 (pg. 
1083
-
1094
)
Walker
GM
Schwartz
MF
Kimberg
DY
Faseyitan
O
Brecher
A
Dell
GS
Coslett
HB
Support for anterior temporal involvement in semantic error production in aphasia: new evidence from VLSM
Brain Lang
 , 
2011
, vol. 
117
 (pg. 
110
-
122
)
Warrington
EK
The selective impairment of semantic memory
Q J Exp Psychol
 , 
1975
, vol. 
27
 (pg. 
635
-
657
)
Weiskopf
N
Hutton
C
Josephs
O
Deichmann
R
Optimal EPI parameters for reduction of susceptibility-induced BOLD sensitivity losses: a whole-brain analysis at 3 T and 1.5 T
Neuroimage
 , 
2006
, vol. 
33
 (pg. 
493
-
504
)
Zaitsev
M
Hennig
J
Speck
O
Point spread function mapping with parallel imaging techniques and high acceleration factors: fast, robust, and flexible method for echo-planar imaging distortion correction
Magn Reson Med
 , 
2004
, vol. 
52
 (pg. 
1156
-
1166
)
Zeng
H
Constable
RT
Image distortion correction in EPI: comparison of field mapping with point spread function mapping
Magn Reson Med
 , 
2002
, vol. 
48
 (pg. 
137
-
146
)