Establishing task- and modality-dependent dissociations between the semantic and default mode networks

Gina F. Humphreys; Paul Hoffman; Maya Visser; Richard J. Binney; Matthew A. Lambon Ralph

doi:10.1073/pnas.1422760112

Edited by Steven E. Petersen, Washington University School of Medicine, St. Louis, MO, and accepted by the Editorial Board May 11, 2015 (received for review December 2, 2014)

Significance

Functional neuroimaging has established that most cognitive functions are supported by distributed neural networks. Hundreds of studies have investigated the semantic network (SN) and the default mode network (DMN) (neural deactivation when undertaking a variety of tasks). These stable networks are increasingly used as biomarkers in neurological and psychiatric investigations. Despite implicating overlapping neural regions and shared cognitive mechanisms, the relationship between the two networks has received minimal attention. Analyses of a large multitask distortion-corrected functional MRI (fMRI) dataset established that both networks fractionate, depending on the semantic nature of the task, stimulus type, modality, and task difficulty. The implications for the SN, variability in the DMN and its cognitive coherence, and interpretation of resting-state fMRI data are discussed.

Abstract

The default mode network (DMN) and semantic network (SN) are two of the most extensively studied systems, and both are increasingly used as clinical biomarkers in neurological studies. There are strong theoretical reasons to assume a relationship between the networks, as well as anatomical evidence that they might rely on overlapping cortical regions, such as the anterior temporal lobe (ATL) or angular gyrus (AG). Despite these strong motivations, the relationship between the two systems has received minimal attention. We directly compared the SN and DMN using a large (n = 69) distortion-corrected functional MRI (fMRI) dataset, spanning a range of semantic and nonsemantic tasks that varied input modality. The results showed that both networks fractionate depending on the semantic nature of the task, stimulus type, modality, and task difficulty. Furthermore, despite recent claims that both AG and ATL are semantic hubs, the two areas responded very differently, with results supporting the role of ATL, but not AG, in semantic representation. Specifically, the left ATL was positively activated for all semantic tasks, but deactivated during nonsemantic task performance. In contrast, the left AG was deactivated for all tasks, with the level of deactivation related to task difficulty. Thus, ATL and AG do not share a common interest in semantic tasks, but, rather, a common “disinterest” in nonsemantic tasks. The implications for the variability in the DMN, its cognitive coherence, and interpretation of resting-state fMRI data are discussed.

Two substantial bodies of research literature, spanning cognitive and clinical neuroscience fields, have been dedicated to exploring the function and components of the semantic network (SN) and the default mode network (DMN). The DMN is an anatomically defined network that shows task-related deactivation during many goal-directed tasks (i.e., rest > task) (1) and can be reliably delineated using techniques such as independent components analysis (ICA) of resting-state functional MRI (fMRI) (2). The SN is a fronto-temporo-parietal network that is sensitive to semantic content in comparisons of semantic tasks > rest/nonsemantic control tasks (3). Although investigations of the DMN and SN have been primarily independent of each other, there are good reasons to compare the two networks directly. First, the networks might share common cognitive functions. One prominent theory suggests that during “rest,” the brain is engaged in the activation of rich conceptual representations, and thus default-mode processing places strong demands on the semantic system (4). Secondly, the DMN and SN engage some common anatomical areas. The DMN consistently includes medial prefrontal cortex, parietal areas [angular gyrus (AG), precuneus, posterior cingulate cortex (PCC)], and, somewhat more variably, the lateral anterior temporal lobe (ATL) and hippocampus (1, 5). Some of these areas are considered central to semantic processing. For instance, both the ATL and AG have been proposed to be “semantic hubs” that help to represent multimodal semantic representations (6 ⇓–8). However, despite these strong motivations, only a handful of studies have directly compared the two networks; even fewer have (i) used methods to maximize the likelihood of detecting ATL activations (9), and (ii) none have compared results across a range of semantic and nonsemantic tasks to establish the functional generality of each network. Accordingly, we investigated the similarities and differences in the SN and DMN using a large (n = 69) distortion-corrected fMRI dataset, spanning a range of semantic and nonsemantic tasks that varied the input modality. By comparing the patterns of task-related activation and deactivation, it proved possible to determine when the networks converge and deviate and to reveal task- and modality-dependent responses in both networks.

Comparison between the DMN and SN is challenging because of apparent inconsistencies in both bodies of literatures. The function of the DMN is a hotly debated issue, with proposed functions including mind wandering, monitoring the external environment, internally directed thought, goal-directed thought, thinking about the past or future, or considering alternative perspectives (1, 5, 10 ⇓–12). Furthermore, some subcomponents of the DMN—in particular the ATL—are inconsistently reported across studies, leading to suggestions that the DMN might be made up of multiple subsystems, each serving distinct functions (semantic memory, episodic memory, decision making, and affective and sensory processing) (5, 13 ⇓–15). The inconsistent involvement of ATL might also relate to a series of methodological challenges associated with imaging this region (9).

There is also a lack of clarity within the SN literature. For instance, semantic processing typically engages a fronto-temporo-parietal network (3); however, the role of certain regions within the network is currently under debate. A wealth of converging evidence from neuropsychology, transcranial magnetic stimulation (TMS), PET, and distortion-corrected fMRI suggests that regions within the ATL are crucial in transmodal semantic representation (16 ⇓⇓⇓–20). However, it is currently unclear as to whether the AG serves a similar function. A meta-analysis of semantic neuroimaging studies found that the AG consistently exhibited sensitivity to semantic manipulations (3), yet the overall AG activation for semantic tasks is often negative with respect to rest (21, 22). Other evidence is more consistent with a role for dorsal AG/intraparietal sulcus (IPS) in executive aspects of semantic processing, rather than semantic representations per se (23). Secondly, the AG has been implicated in numerous cognitive domains outside of the field of semantic cognition, suggesting a more domain-general cognitive function (attention, episodic memory, numerical processing, and syntax; ref. 24). Indeed, one possibility is that the AG forms part of a domain-general processing network that is involved in “automatic” or “stimulus-driven” task-processes and is anticorrelated with the “executive” dorsal parietal cortex (IPS) (24 ⇓–26). Together, these inconsistencies in both the current SN and DMN literatures make establishing the relationship between the two networks difficult.

A clearer picture might emerge through direct, within-study comparisons. The few existing single-task investigations have found that certain DMN components can show semantic sensitivity (21, 22, 27, 28). For instance, the AG sometimes shows less deactivation for semantic compared with nonsemantic tasks (22, 27). However, the results across studies have been inconsistent; some studies show widespread overlap across multiple DMN and SN areas (AG, ATL, medial prefrontal cortex, cingulate) (21), whereas others find limited overlap (22, 27, 28). Indeed, across studies, no region has been consistently reported in both networks. In addition to the various factors noted above, the lack of clarity might relate to a failure of existing studies to take into account task-dependent variations, because the networks’ neural responses are likely to vary based on factors such as stimulus type, modality, or task difficulty (19, 29 ⇓–31)—which might be clarified by directly comparing DMN and SN across multiple tasks and modalities. One final limitation of the existing work is the tendency to focus solely on areas of DMN–SN overlap and to ignore any large divergences between the networks. This limitation is important, because if large portions of the SN are not involved in DMN, it questions the core theoretical assumption that the DMN’s core function is semantic.

To clarify the relationship between the SN and DMN, we conducted a large-scale investigation of the similarities and differences in the SN and DMN by comparing results from multiple semantic and nonsemantic tasks that varied in stimulus type (words, pictures, environmental sounds, numbers, and pattern matching), task difficulty, and input modality (visual and auditory). Critically, data were acquired by using a distortion-corrected fMRI protocol, promoting detection of signal from all parts of the ATL. By comparing the pattern of task-related activation and deactivation, it was possible to determine where SN and DMN converge and segregate, as well as to reveal task- and modality-dependent responses in the networks. In addition, given their potential pivotal role in semantic cognition, we explored the roles of the ATL and AG in more detail.

Results

The DMN was identified by determining areas with greater activation during rest compared with task periods (rest > task). To determine semantic-dependent variations in the network, this contrast was performed separately for the semantic and nonsemantic tasks, and we looked for commonalities across all tasks. These analyses revealed semantic-variant and -invariant effects (Fig. 1 and Table S1). All results reported below (unless otherwise specified) were thresholded by using a voxel height threshold P < 0.001 and cluster-corrected by using family-wise error (FWE) P < 0.05. Significant deactivation relative to rest during semantic task performance (rest > semantics) was found in a fronto-temporo-parietal network, which included the bilateral inferior parietal lobe (IPL) [AG and supramarginal gyrus (SMG)], medial structures (PCC and medial frontal cortex), the right ATL, the bilateral auditory cortex [Heschl's gyrus and superior temporal gyrus (STG)], the bilateral hippocampus, and frontal areas (left superior orbital gyrus, bilateral ACC, right middle frontal gyrus, and left insula). The pattern of results for the nonsemantic tasks was similar, but with the notable addition of the ATL and IFG bilaterally (this difference was confirmed to be significant, as reported below). Conjunction analyses confirmed a common pattern of deactivation for semantic and nonsemantic tasks of bilateral IPL (AG, SMG, precuneus, and PCC), right ATL, bilateral auditory cortex (Heschl's gyrus and STG), right hippocampus, and frontal areas (bilateral middle frontal gyrus, left superior orbital gyrus, and left ACC). Thus, certain DMN “components” (ATL and IFG) appear to vary depending on the semantic content of the task, whereas other components are recruited for both task types.

Fig. 1.

Positive (red) and negative (blue) activation for semantic tasks vs. rest; the nonsemantic tasks vs. rest; their conjunction; and semantic > nonsemantic tasks (uncorrected, P < 0.001).

To determine the SN, we examined task-related activations (task > rest) for the semantic tasks and compared these activations to the pattern from the nonsemantic tasks (Fig. 1). Overall, semantic tasks were found to activate a fronto-temporo-parietal network and visual cortex (all semantics > rest). This network included the anterior and posterior temporal cortex (left fusiform gyrus, left temporal pole, and left middle temporal gyrus), frontal areas (bilateral IFG, left precentral gyrus, and right middle orbital gyrus), lateral superior parietal cortex (bilateral IPS/SPL), and left putamen. Note that parts of this network, the left ATL (especially fusiform gyrus) and IFG, were deactivated for the nonsemantic tasks; thus, certain parts of the DMN are sensitive to semantic content. The SN showed some notable differences in activation, as well as some commonalities compared with the nonsemantic tasks. First, similar positive activation for the nonsemantic tasks (nonsemantic > rest) was found in parietal (bilateral IPS and right SMG) and occipital areas; however, little frontal and ATL activation was found compared with the semantic tasks (although there was some restricted recruitment of the left temporal pole). Indeed, these differences between the semantic and nonsemantic tasks were confirmed by conducting direct comparisons (semantic > nonsemantic), which revealed significantly stronger recruitment of semantic tasks within bilateral IFG, bilateral ATL (temporal pole and fusiform gyrus), bilateral pMTG, and right middle orbital gyrus. A very small cluster within the left AG also showed a stronger response to the semantic compared with nonsemantic tasks; however, this difference was only at a reduced statistical threshold (P < 0.001, uncorrected) and, unlike the left ATL, reflected differential deactivation (i.e., greater deactivation for the nonsemantic vs. semantic tasks).

These analyses provide several key findings. First, they show that certain components of the DMN are common to both semantic and nonsemantic tasks, including IPL and medial structures (PCC and medial frontal cortex), as well as right ATL and auditory cortex (although see below). In contrast, other areas of the DMN show task-dependent responses. In particular, the left ATL and bilateral IFG, which are positively activated in the semantic task, form part of the DMN during the performance of nonsemantic tasks (i.e., they are sensitive to both semantics > rest and rest > nonsemantic tasks). Finally, and importantly, the AG and ATL showed dissociable responses. In particular, the left ATL was positively activated for semantic tasks and deactivated for nonsemantic tasks, whereas the AG was deactivated by both semantic and nonsemantic tasks (albeit moderately more strongly for the nonsemantic tasks). Thus, this result provides convincing evidence that the ATL and AG serve distinct cognitive functions.

In the next analysis, we investigated whether the SN and DMN vary depending on the type of semantic task. To perform this analysis, we compared tasks involving written words against pictures (i.e., tasks that share the same modality but differ in verbal vs. nonverbal content; Fig. 2 and Table S1). When including only reading-based semantic tasks, the DMN (rest > reading) and the SN (reading > rest) were similar to that revealed by the general semantic analysis described above: The reading-based tasks activated left ATL relative to rest (fusiform gyrus and temporal pole) but deactivated right ATL. However, the pattern was different for the picture-based semantic tasks, which showed bilateral positive ATL activation (temporal pole and fusiform gyrus) and comparatively little ATL deactivation in either hemisphere. Direct contrasts between reading- and picture-based semantic tasks confirmed that picture tasks showed stronger bilateral ATL engagement (fusiform and inferior temporal gyrus), as well as occipital areas (Fig. S1). In contrast, reading-based tasks engaged left IFG, left pMTG, bilateral superior frontal gyrus, and right hippocampus more strongly compared with picture tasks. Note that the AG showed significant deactivation for both task types, with no significant differences between the two. The medial structures (PCC and medial frontal cortex) were also equivalently deactivated for both tasks. Therefore, these results, combined with those from the more general analysis above, clearly demonstrate that, unlike “core DMN regions” (PCC, medial frontal, and AG), the ATL is involved in both SN and DMN, but its recruitment varies depending on stimulus type.

Fig. 2.

The effects of stimulus-type and stimulus-modality. Positive (red) and negative (blue) activation for reading tasks vs. rest; picture tasks vs. rest; visual tasks vs. rest; and auditory tasks vs. rest (uncorrected, P < 0.001).

Fig. S1.

The direct contrast of picture tasks and reading tasks showing greater right ATL activation (fusiform and inferior temporal gyrus) for the picture tasks [red, pictures > reading; blue, reading > pictures (uncorrected, P < 0.001)].

The influence of input modality was also investigated to examine modality-dependent and -independent responses. In the first overall analysis, auditory cortex was found to form part of the DMN (and this finding was common regardless of the semantic nature of the task). Previous evidence has shown that sensory areas are deactivated when they are not central to task performance (29), and thus we expected that this effect would be modality-dependent, driven mainly by tasks from the visual modality (the majority of tasks). To test this prediction, we separately examined the results for the visual and auditory semantic tasks (Fig. 2). Our predictions were confirmed: The tendency for the auditory cortex to form part of the DMN was driven by the visual tasks (rest > visual tasks). When including only the auditory tasks (rest > auditory tasks), the sensory cortices included in the DMN shifted to include parts of visual rather than auditory cortex. This difference was significant in a direct comparison between the two networks. We also examined areas that were invariant to input modality by conducting a conjunction analysis across visual and auditory semantic tasks. This analysis showed that for the DMN, bilateral ventral parietal cortex (AG and SMG), medial structures (PCC and medial frontal cortex), right ATL, and left superior orbital gyrus were common to visual and auditory tasks (rest > visual and rest > auditory). Conversely, for the SN, the left ATL, pMTG, and IPS were common (visual > rest and auditory > rest).

The analyses reported above confirm that the ATL and AG show a different pattern of activation (and deactivation) to semantic and nonsemantic tasks. To examine the relationship between these areas further, the percent signal change from each region was correlated across tasks. We also examined the relationship between these regions and the IPS, an area that may form an anticorrelated network with the AG (26). Consistent with their profile on the task-based results, the correlation analysis showed no significant relationship between the responses of AG and ATL (r = −0.34, P = 0.18), providing further evidence that these regions respond dissimilarly. In contrast, a strong negative correlation between IPS and AG was found (r = −0.64, P = 0.006), consistent with the proposal that the AG and IPS are anticorrelated networks. Additionally, there was a positive trend in the correlation between ATL and IPS activation (r = 0.45, P = 0.07). Finally, we examined the extent to which activation in each area varied depending on task difficulty by correlating percent signal change with the average RT for the same task. This analysis showed that the AG was negatively correlated with task difficulty (r = −0.612, P = 0.04), whereas the IPS showed a trend toward a moderate positive correlation (r = 0.55, P = 0.08), and the ATL showed a positive, but nonsignificant, relationship (r = 0.51, P = 0.11).

Discussion

The aim of this large-scale (69-participant, multitask) investigation was to clarify the relationship between the DMN and SN. The results indicate that both networks are highly task- and modality-dependent (see summary in Table 1). Certain DMN areas are sensitive to the semantic nature of the task. Specifically, the involvement of ATL regions in the DMN (and left IFG to some extent) was found to vary depending on the level of semantic involvement (semantic vs. nonsemantic) and semantic stimulus type (pictures vs. written words). In particular, these areas were positively activated during semantic tasks, but were deactivated during nonsemantic task performance, and hence form a part of the DMN only for nonsemantic tasks. Activation in other areas was independent of task, but was instead influenced by input modality or task difficulty. Specifically, primary sensory cortices were deactivated for tasks presented in their nonspecialized modality. In contrast, the AG was insensitive to both modality and task, but was more strongly deactivated for more difficult tasks.

View this table:

Table 1.

Results summary for each region

Given that both ATL and AG have been proposed as potentially critical regions for the DMN or SN, the second aim of this study was to compare responses in the ATL and AG across tasks. The results showed that the AG and ATL responded very differently to each task, thus implying distinct cognitive functions. In particular, the polarity of activation in the ATL depended on semantic content; it was positively activated for semantic tasks and deactivated for nonsemantic tasks (with laterality varying depending on stimulus type). In contrast, the AG was deactivated by all tasks, with the degree of deactivation relating to task difficulty rather than semantic content per se. Indeed, direct correlations between the AG and ATL activity found no evidence of a significant relationship between the two areas. The AG, unlike the ATL, also showed an inverse relationship with the extent of IPS activation, an area considered central to the “multiple-demand” executive processing system (25, 32).

The present data might have strong implications for interpretation of networks identified in resting-state functional connectivity studies (as revealed by interregional correlations, independent component analysis, etc.). It is understandably the case that the occurrence of positive correlations in functional connectivity between regions is interpreted as evidence of a common underlying cognitive function. For instance, functional connectivity between the ATL, AG, and frontal cortices has been interpreted as evidence of a semantic processing network (33). However, the present data generate a second alternative hypothesis, which can be explored in future studies (e.g., by careful deconstruction of the network components observed in task-dependent ICA; cf. ref. 34). Specifically, the ATL and AG did not show common activation for semantic tasks but, rather, common deactivation during nonsemantic task performance. Accordingly, positive interregional time-course correlations might not reflect a common interest in semantic tasks, but, rather, a common “disinterest” in the tasks that deactivate both regions. If correct, one must be cautious in interpreting the results from connectivity studies alone. Indeed, the present study highlights how cognitive interpretation and fractionation of a distributed network can be aided by combining data from experimentally driven task-based investigations.

The present results suggest that when a neural region is not critical to task function it is deactivated. This finding was true across a variety of structures and processes. For instance, auditory areas were deactivated during visual processing (and vice versa), and semantic areas were deactivated during the performance of nonsemantic tasks. These findings are consistent with two proposals about neural activation. The first is the “limited-capacity” model of neural processing, in which neural resources are finite, and thus alternative cognitive processes are competitive in nature (35). Under these circumstances, the most efficient strategy is to down-regulate a particular neural system if it is unnecessary/disruptive to task performance. Therefore, according to this hypothesis, the changing task dependencies within DMN areas can be explained by variations in the neural computations required for a particular task. The second potentially related proposal is that there are online plasticity mechanisms to balance metabolic energy consumption against task performance (36). Thus, if a region’s neurocomputational function is not required for the current task, its activity is down-regulated to save metabolic energy.

By combining data across a variety of different semantic and nonsemantic tasks, we were able to clarify the relationship between the DMN and SN. The overlap was particularly clear for regions within the ATL. These data are consistent with a convergence of results from neuropsychology, TMS, functional neuroimaging, and intracranial recordings, which points to these ATL regions underpinning a transmodal semantic representational hub (7, 8, 16, 20, 37). This outcome is highly consistent with the “semantic hypothesis” for the DMN. Although originally proposed for the AG rather than the ATL, this hypothesis suggests that during rest, the brain is engaged in detailed conceptual-language processes that draw on the SN (4). The response characteristics of the ATL fit perfectly with this hypothesis. It does not seem to hold, however, for other parts of the DMN (including the AG)—many of which are insensitive to the semantic demands of the task and are, instead, influenced by modality or task difficulty.

The present study strongly suggests that, rather than serving one single cognitive function, the DMN is best viewed as a dynamic patchwork reflecting variable deactivation of several subsystems, each serving distinct computations. Indeed, this view is consistent with recent claims that the DMN consists of multiple dissociable, but interacting, components that serve a variety of cognitive function (semantic memory, episodic memory, decision making, and affective and sensory processing) (5, 13 ⇓–15). In short, although the DMN network is a relatively consistent neuroimaging phenomenon (reproducible across a range of imaging data and types of analysis; cf. ref. 7), it would appear that it is not a coherent, homogeneous cognitive entity.

In addition to the DMN, this study has implications for semantic models. Some theories propose that the AG is a semantic hub, with a function akin to the ATL (3, 6). The results from this study, however, show that the AG and ATL exhibit very different responses across tasks. Although the brain has multiple tertiary association cortices [indeed, both the ATL and AG have been shown to rank among the highest functionally interconnected areas (38)], the present data provide convincing evidence that the ATL and AG serve distinct cognitive functions. These data, combined with the wealth of converging evidence from neuropsychology, fMRI, and TMS, provide compelling support for the core role of the ATL, rather than the AG, in semantic representation. Although the present data do not exclude the possibility that the AG is involved in semantic processing in some way, at the very least these results show that the AG does not perform a similar role to the ATL in semantic cognition.

The present results showed that the left ATL was positively engaged during all semantic tasks, whereas the right ATL was sensitive to the stimulus type, with pictures generating positive activation, whereas written words led to deactivation. This finding adds to a growing body of literature comparing processing in the left vs. right ATL. A seminal magnetoencephalography study of semantic processing across modalities, as well as a recent large-scale meta-analysis of functional imaging studies (37, 39), found that ATL activation for written words or speech production is strongly left-lateralized, whereas other forms of semantic tasks (pictures, auditory words, and auditory sounds) show bilateral ATL engagement (see also ref. 19). These and parallel neuropsychological data on left vs. right ATL differences have been formally considered in a number of implemented computational models of semantic processing (40, 41). The key ideas from these models are that semantic representation may be supported by regions within the ATLs, bilaterally, with differential patterns of activation or impairments in unilateral ATL patients arising from the effects of asymmetric connectivity with input and output areas. Thus, the relatively greater importance of the left ATL for spoken tasks and for written-word comprehension would follow from differentially higher connectivity to left-hemisphere–biased speech production systems (cf. ref. 42) and the left posterior ventral occipitotemporal cortex, which exhibits greater involvement in the visual processes that underpin written-word recognition (43). An alternative hypothesis is that each ATL supports discrete semantic functions (44). Irrespective of the exact cognitive interpretation, this finding has a strong methodological implication for the current semantic neuroimaging literature. By far, the most commonly used form of stimulus in semantic fMRI studies is the written word. Although written words have obvious practical and logistic advantages (visual presentation of stimuli is much easier than auditory in the scanner, and written words allow the full range of concrete, abstract, emotion, etc., concepts to be probed), it is clear that written words—unlike pictures, spoken words, and sounds—generate a strongly left-lateralized pattern of activation. This result could encourage the apparent conclusion that semantic processing is predominately left-lateralized, when, in fact, for all modalities other than written words, it appears to be much more bilateral in form.

The medial structures (PCC and medial frontal cortex) and AG were found to be deactivated by all task types. This finding is consistent with observations that these regions form the most reliable and highly connected components of the DMN (5, 38, 45) and are considered to be core parts of the DMN (45). The medial structures are particular active in tasks involving autobiographical memory, theory of mind, and episodic memory retrieval, leading to suggestions that they may be involved in self-projection (projecting oneself to a different context) or internal mentation (self-directed thought) (10, 12).

Finally, the question remains as to the core computation of the AG. The present data showed that: (i) the AG is deactivated by semantic and nonsemantic tasks, and the magnitude of deactivation relates to task difficulty; and (ii) AG deactivation is anticorrelated with activation of the dorsal parietal cortex (IPS), suggesting that the two networks serve opposing task functions. According to models of attention, dorsal and ventral parietal cortices are involved in top-down vs. bottom-up attentional processes, respectively (46). Thus, it is plausible that dorsal and ventral areas are implicated in (relatively domain-general) executive vs. automatic processing, respectively. With regard to AG’s core function, a recent large-scale multidomain meta-analysis of 386 neuroimaging studies also showed that the AG (i) deactivates for a wide variety of domains (including semantic tasks, executively demanding decisions, etc.) but (ii) is positively engaged by a variety of different domains (episodic retrieval, numerical tasks, sentence-level tasks, etc.) (24). Thus, it appears that the AG serves a more domain-general function and is not specialized for semantic processing. One possibility is that the parietal cortex acts as a multimodal online buffer of incoming internal or external information (24). Within this system, the dorsal and ventral parietal cortex serve counterpointed roles; the ventral system automatically buffers input, whereas the dorsal system is involved in top-down executive processing of buffered information. Indeed, this hypothesis is consistent with evidence for the role of the parietal cortex in working and short-term memory (47). According to this theory, the continual automatic buffering of additional information by the ventral parietal cortex can be disruptive during the performance of some, but not all, goal-directed executively demanding tasks. Hence, during difficult task performance, activation of this region is suppressed.

Methods

Tasks.

Data were collected from seven semantic tasks plus modality- and RT-matched nonsemantic tasks from across four fMRI studies (n = 69). Each study included at least one semantic condition and one nonsemantic control condition from the same modality. The tasks are described in detail elsewhere (17 ⇓–19, 48); however, crucially for the present study, the paired semantic/nonsemantic tasks varied in stimulus type and modality: picture tasks (×2), written word tasks (×3), auditory word tasks (×1), and environmental sounds tasks (×1) (Table 2).

View this table:

Table 2.

Details of each fMRI task

Scanning.

Images were acquired on a 3T Philips Achieva scanner by using an eight-element SENSE head coil with a sense factor of 2.5. The data from each study were collected by using the same distortion-corrected fMRI technique (see Table S2 for individual study parameters). Following the standard method for distortion-corrected spin-echo fMRI (49), the images were acquired with a single-direction k space traversal and a left–right phase encoding direction. A “prescan” was acquired before each run, consisting of 10 volumes of dual-direction k space traversal spin-echo echo-planar imaging scans. This acquisition provided 10 pairs of images matching the functional time series, but with distortions in both phase-encoding directions (10 left–right and 10 right–left). These scans were used in the distortion-correction procedure. The correction was computed by using the method reported in ref. 49, in which each image from functional time series is registered to the mean of the prescan images by using a six-parameter rigid-body transformation. Subsequently, a spatial transformation matrix was calculated from the prescan images, consisting of the spatial remapping necessary to correct the distortion. This transformation was then applied to each of the coregistered functional images.

Analyses.

By using SPM5, data were motion-corrected and coregistered to the anatomical T1. Images were then spatially normalized to Montreal Neurological Institute standard space, resampled to 3 × 3 × 3-mm dimensions, and smoothed with a Gaussian filter of FHWM = 8 mm. First- and second-level analyses were carried out by using SPM8. At the first level, a general linear model analysis was performed by modeling each condition as a separate regressor using a boxcar function convolved with the canonical hemodynamic response function. Contrasts were calculated for each condition vs. rest (task > rest). In the second-level analysis, all data were entered into a single ANOVA model with each “study” included as a separate level. This method thereby controls for any cross-study confounds. In addition, “subject” was added as a covariate to the model to control for any subject effects. T contrasts were computed to examine overall effects of semantics (semantics > rest, nonsemantic > rest, and semantics > nonsemantics), stimulus type (picture tasks > rest, word tasks > rest, pictures > words, and words > pictures), and modality (visual > rest, auditory > rest, visual > auditory, and auditory > visual). These contrasts-of-interest were computed from the same omnibus ANOVA that modeled both study and subject. Accordingly, the overall model accounts for the variance associated with any study- or subject-specific variations and then reveals the activation differences, which are attributable uniquely to the contrast-of-interest. Unless otherwise stated, a standard voxel height threshold P < 0.001, cluster corrected by using FWE P < 0.05, was used.

Correlation analyses were performed to determine the relationship between AG, ATL (anterior fusiform gyrus), and IPS activation and task difficulty. Task difficulty was determined based on the average reaction time for each task, across participants. These mean reaction times were then correlated with the participant-average percent signal change from an AG, ATL, and IPS region of interest (ROI). The ROIs were defined based on the voxels showing significant activation for the contrast of semantics > rest (ATL), all tasks > rest (IPS), or rest > all tasks (AG) from the higher-level analysis.

Acknowledgments

This work was supported by Medical Research Council Grant MR/J004146/1.

Footnotes

↵¹To whom correspondence may be addressed. Email: matt.lambon-ralph{at}manchester.ac.uk or gina.humphreys{at}manchester.ac.uk.

Author contributions: G.F.H. and M.A.L.R. designed research; G.F.H., P.H., M.V., and R.J.B. performed research; G.F.H., P.H., M.V., R.J.B., and M.A.L.R. analyzed data; and G.F.H., P.H., M.V., R.J.B., and M.A.L.R. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission. S.E.P. is a guest editor invited by the Editorial Board.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1422760112/-/DCSupplemental.

References

↵
1. Raichle ME, et al.
(2001) A default mode of brain function. Proc Natl Acad Sci USA 98(2):676–682
.
OpenUrl Abstract/FREE Full Text
↵
1. Yeo BTT,
2. Krienen FM,
3. Chee MWL,
4. Buckner RL
(2013) Estimates of segregation and overlap of functional connectivity networks in the human cerebral cortex. Neuroimage 88C:212–227
.
OpenUrl CrossRef PubMed
↵
1. Binder JR,
2. Desai RH,
3. Graves WW,
4. Conant LL
(2009) Where is the semantic system? A critical review and meta-analysis of 120 functional neuroimaging studies. Cereb Cortex 19(12):2767–2796
.
OpenUrl Abstract/FREE Full Text
↵
1. Binder JR, et al.
(1999) Conceptual processing during the conscious resting state. A functional MRI study. J Cogn Neurosci 11(1):80–95
.
OpenUrl CrossRef PubMed
↵
1. Buckner RL,
2. Andrews-Hanna JR,
3. Schacter DL
(2008) The brain’s default network: Anatomy, function, and relevance to disease. Ann N Y Acad Sci 1124:1–38
.
OpenUrl CrossRef PubMed
↵
1. Binder JR,
2. Desai RH
(2011) The neurobiology of semantic memory. Trends Cogn Sci 15(11):527–536
.
OpenUrl CrossRef PubMed
↵
1. Patterson K,
2. Nestor PJ,
3. Rogers TT
(2007) Where do you know what you know? The representation of semantic knowledge in the human brain. Nat Rev Neurosci 8(12):976–987
.
OpenUrl CrossRef PubMed
↵
1. Lambon Ralph MA
(2013) Neurocognitive insights on conceptual knowledge and its breakdown. Philos Trans R Soc B 369(1634):20120392
.
OpenUrl CrossRef PubMed
↵
1. Visser M,
2. Embleton KV,
3. Jefferies E,
4. Parker GJ,
5. Lambon Ralph MA
(2010) The inferior, anterior temporal lobes and semantic memory clarified: Novel evidence from distortion-corrected fMRI. Neuropsychologia 48(6):1689–1696
.
OpenUrl CrossRef PubMed
↵
1. Buckner RL,
2. Carroll DC
(2007) Self-projection and the brain. Trends Cogn Sci 11(2):49–57
.
OpenUrl CrossRef PubMed
↵
1. Mason MF, et al.
(2007) Wandering minds: The default network and stimulus-independent thought. Science 315(5810):393–395
.
OpenUrl Abstract/FREE Full Text
↵
1. Andrews-Hanna JR
(2012) The brain’s default network and its adaptive role in internal mentation. Neuroscientist 18(3):251–270
.
OpenUrl Abstract/FREE Full Text
↵
1. Laird AR, et al.
(2009) Investigating the functional heterogeneity of the default mode network using coordinate-based meta-analytic modeling. J Neurosci 29(46):14496–14505
.
OpenUrl Abstract/FREE Full Text
↵
1. Andrews-Hanna JR,
2. Reidler JS,
3. Sepulcre J,
4. Poulin R,
5. Buckner RL
(2010) Functional-anatomic fractionation of the brain’s default network. Neuron 65(4):550–562
.
OpenUrl CrossRef PubMed
↵
1. Smith SM, et al.
(2009) Correspondence of the brain’s functional architecture during activation and rest. Proc Natl Acad Sci USA 106(31):13040–13045
.
OpenUrl Abstract/FREE Full Text
↵
1. Pobric G,
2. Jefferies E,
3. Lambon Ralph MA
(2007) Anterior temporal lobes mediate semantic representation: Mimicking semantic dementia by using rTMS in normal participants. Proc Natl Acad Sci USA 104(50):20137–20141
.
OpenUrl Abstract/FREE Full Text
↵
1. Binney RJ,
2. Embleton KV,
3. Jefferies E,
4. Parker GJ,
5. Lambon Ralph MA
(2010) 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 20(11):2728–2738
.
OpenUrl Abstract/FREE Full Text
↵
1. Visser M,
2. Jefferies E,
3. Embleton KV,
4. Lambon Ralph MA
(2012) Both the middle temporal gyrus and the ventral anterior temporal area are crucial for multimodal semantic processing: Distortion-corrected fMRI evidence for a double gradient of information convergence in the temporal lobes. J Cogn Neurosci 24(8):1766–1778
.
OpenUrl CrossRef PubMed
↵
1. Visser M,
2. Lambon Ralph MA
(2011) Differential contributions of bilateral ventral anterior temporal lobe and left anterior superior temporal gyrus to semantic processes. J Cogn Neurosci 23(10):3121–3131
.
OpenUrl CrossRef PubMed
↵
1. Vandenberghe R,
2. Price C,
3. Wise R,
4. Josephs O,
5. Frackowiak RS
(1996) Functional anatomy of a common semantic system for words and pictures. Nature 383(6597):254–256
.
OpenUrl CrossRef PubMed
↵
1. Wirth M, et al.
(2011) Semantic memory involvement in the default mode network: A functional neuroimaging study using independent component analysis. Neuroimage 54(4):3057–3066
.
OpenUrl CrossRef PubMed
↵
1. Seghier ML,
2. Fagan E,
3. Price CJ
(2010) Functional subdivisions in the left angular gyrus where the semantic system meets and diverges from the default network. J Neurosci 30(50):16809–16817
.
OpenUrl Abstract/FREE Full Text
↵
1. Noonan KA,
2. Jefferies E,
3. Visser M,
4. Lambon Ralph MA
(2013) Going beyond inferior prefrontal involvement in semantic control: Evidence for the additional contribution of dorsal angular gyrus and posterior middle temporal cortex. J Cogn Neurosci 25(11):1824–1850
.
OpenUrl CrossRef PubMed
↵
1. Humphreys GF,
2. Lambon Ralph MA
(September 9, 2014) Fusion and fission of cognitive functions in the human parietal cortex. Cereb Cortex doi:10.1093/cercor/bhu198
.
OpenUrl Abstract/FREE Full Text
↵
1. Fedorenko E,
2. Duncan J,
3. Kanwisher N
(2013) Broad domain generality in focal regions of frontal and parietal cortex. Proc Natl Acad Sci USA 110(41):16616–16621
.
OpenUrl Abstract/FREE Full Text
↵
1. Fox MD, et al.
(2005) The human brain is intrinsically organized into dynamic, anticorrelated functional networks. Proc Natl Acad Sci USA 102(27):9673–9678
.
OpenUrl Abstract/FREE Full Text
↵
1. Seghier ML,
2. Price CJ
(2012) Functional heterogeneity within the default network during semantic processing and speech production. Front Psychol 3:281
.
OpenUrl CrossRef PubMed
↵
1. Shapira-Lichter I,
2. Oren N,
3. Jacob Y,
4. Gruberger M,
5. Hendler T
(2013) Portraying the unique contribution of the default mode network to internally driven mnemonic processes. Proc Natl Acad Sci USA 110(13):4950–4955
.
OpenUrl Abstract/FREE Full Text
↵
1. Laurienti PJ, et al.
(2002) Deactivation of sensory-specific cortex by cross-modal stimuli. J Cogn Neurosci 14(3):420–429
.
OpenUrl CrossRef PubMed
↵
1. Fransson P
(2006) How default is the default mode of brain function? Further evidence from intrinsic BOLD signal fluctuations. Neuropsychologia 44(14):2836–2845
.
OpenUrl CrossRef PubMed
↵
1. Mayer JS,
2. Roebroeck A,
3. Maurer K,
4. Linden DE
(2010) Specialization in the default mode: Task-induced brain deactivations dissociate between visual working memory and attention. Hum Brain Mapp 31(1):126–139
.
OpenUrl PubMed
↵
1. Duncan J
(2010) The multiple-demand (MD) system of the primate brain: Mental programs for intelligent behaviour. Trends Cogn Sci 14(4):172–179
.
OpenUrl CrossRef PubMed
↵
1. Pascual B, et al.
(2015) Large-scale brain networks of the human left temporal pole: A functional connectivity MRI study. Cereb Cortex 25(3):680–702
.
OpenUrl Abstract/FREE Full Text
↵
1. Geranmayeh F,
2. Wise RJS,
3. Mehta A,
4. Leech R
(2014) Overlapping networks engaged during spoken language production and its cognitive control. J Neurosci 34(26):8728–8740
.
OpenUrl Abstract/FREE Full Text
↵
1. Handy TC
(2000) Capacity theory as a model of cortical behavior. J Cogn Neurosci 12(6):1066–1069
.
OpenUrl CrossRef PubMed
↵
1. Attwell D,
2. Laughlin SB
(2001) An energy budget for signaling in the grey matter of the brain. J Cereb Blood Flow Metab 21(10):1133–1145
.
OpenUrl CrossRef PubMed
↵
1. Marinkovic K, et al.
(2003) Spatiotemporal dynamics of modality-specific and supramodal word processing. Neuron 38(3):487–497
.
OpenUrl CrossRef PubMed
↵
1. Buckner RL, et al.
(2009) Cortical hubs revealed by intrinsic functional connectivity: Mapping, assessment of stability, and relation to Alzheimer’s disease. J Neurosci 29(6):1860–1873
.
OpenUrl Abstract/FREE Full Text
↵
1. Rice GE,
2. Lambon Ralph MA,
3. Hoffman P
(March 13, 2015) The roles of left vs. right anterior temporal lobes in conceptual knowledge: An ALE meta-analysis of 97 functional neuroimaging studies. Cereb Cortex doi:10.1093/cercor/bhv024
.
OpenUrl Abstract/FREE Full Text
↵
1. Schapiro AC,
2. McClelland JL,
3. Welbourne SR,
4. Rogers TT,
5. Lambon Ralph MA
(2013) Why bilateral damage is worse than unilateral damage to the brain. J Cogn Neurosci 25(12):2107–2123
.
OpenUrl CrossRef PubMed
↵
1. Lambon Ralph MA,
2. McClelland JL,
3. Patterson K,
4. Galton CJ,
5. Hodges JR
(2001) No right to speak? The relationship between object naming and semantic impairment: neuropsychological evidence and a computational model. J Cogn Neurosci 13(3):341–356
.
OpenUrl CrossRef PubMed
↵
1. Blank SC,
2. Scott SK,
3. Murphy K,
4. Warburton E,
5. Wise RJS
(2002) Speech production: Wernicke, Broca and beyond. Brain 125(Pt 8):1829–1838
.
OpenUrl Abstract/FREE Full Text
↵
1. Cohen L, et al.
(2002) Language-specific tuning of visual cortex? Functional properties of the visual word form area. Brain 125(Pt 5):1054–1069
.
OpenUrl Abstract/FREE Full Text
↵
1. Drane DL, et al.
(2013) Famous face identification in temporal lobe epilepsy: Support for a multimodal integration model of semantic memory. Cortex 49(6):1648–1667
.
OpenUrl CrossRef PubMed
↵
1. Andrews-Hanna JR,
2. Smallwood J,
3. Spreng RN
(2014) The default network and self-generated thought: Component processes, dynamic control, and clinical relevance. Ann N Y Acad Sci 1316:29–52
.
OpenUrl CrossRef PubMed
↵
1. Corbetta M,
2. Shulman GL
(2002) Control of goal-directed and stimulus-driven attention in the brain. Nat Rev Neurosci 3(3):201–215
.
OpenUrl CrossRef PubMed
↵
1. Postle BR
(2015) The cognitive neuroscience of visual short-term memory. Current Opinion in Behavioral Sciences 1(0):40–46
.
OpenUrl CrossRef
↵
1. Hoffman P,
2. Binney RJ,
3. Lambon Ralph MA
(2015) Differing contributions of inferior prefrontal and anterior temporal cortex to concrete and abstract conceptual knowledge. Cortex 63:250–266
.
OpenUrl CrossRef PubMed
↵
1. Embleton KV,
2. Haroon HA,
3. Morris DM,
4. Lambon Ralph MA,
5. Parker GJM
(2010) Distortion correction for diffusion-weighted MRI tractography and fMRI in the temporal lobes. Hum Brain Mapp 31(10):1570–1587
.
OpenUrl CrossRef PubMed

Article Alerts

Email Article

Citation Tools

Request Permissions

Article Classifications

Biological Sciences
Psychological and Cognitive Sciences

Proceedings of the National Academy of Sciences: 112 (25)

Table of Contents

Submit

[1] ↵
Raichle ME, et al.
(2001) A default mode of brain function. Proc Natl Acad Sci USA 98(2):676–682
.
OpenUrl Abstract/FREE Full Text

[2] Raichle ME, et al.

[3] ↵
Yeo BTT,
Krienen FM,
Chee MWL,
Buckner RL
(2013) Estimates of segregation and overlap of functional connectivity networks in the human cerebral cortex. Neuroimage 88C:212–227
.
OpenUrl CrossRef PubMed

[4] Yeo BTT,

[5] Krienen FM,

[6] Chee MWL,

[7] Buckner RL

[8] ↵
Binder JR,
Desai RH,
Graves WW,
Conant LL
(2009) Where is the semantic system? A critical review and meta-analysis of 120 functional neuroimaging studies. Cereb Cortex 19(12):2767–2796
.
OpenUrl Abstract/FREE Full Text

[9] Binder JR,

[10] Desai RH,

[11] Graves WW,

[12] Conant LL

[13] ↵
Binder JR, et al.
(1999) Conceptual processing during the conscious resting state. A functional MRI study. J Cogn Neurosci 11(1):80–95
.
OpenUrl CrossRef PubMed

[14] Binder JR, et al.

[15] ↵
Buckner RL,
Andrews-Hanna JR,
Schacter DL
(2008) The brain’s default network: Anatomy, function, and relevance to disease. Ann N Y Acad Sci 1124:1–38
.
OpenUrl CrossRef PubMed

[16] Buckner RL,

[17] Andrews-Hanna JR,

[18] Schacter DL

[19] ↵
Binder JR,
Desai RH
(2011) The neurobiology of semantic memory. Trends Cogn Sci 15(11):527–536
.
OpenUrl CrossRef PubMed

[20] Binder JR,

[21] Desai RH

[22] ↵
Patterson K,
Nestor PJ,
Rogers TT
(2007) Where do you know what you know? The representation of semantic knowledge in the human brain. Nat Rev Neurosci 8(12):976–987
.
OpenUrl CrossRef PubMed

[23] Patterson K,

[24] Nestor PJ,

[25] Rogers TT

[26] ↵
Lambon Ralph MA
(2013) Neurocognitive insights on conceptual knowledge and its breakdown. Philos Trans R Soc B 369(1634):20120392
.
OpenUrl CrossRef PubMed

[27] Lambon Ralph MA

[28] ↵
Visser M,
Embleton KV,
Jefferies E,
Parker GJ,
Lambon Ralph MA
(2010) The inferior, anterior temporal lobes and semantic memory clarified: Novel evidence from distortion-corrected fMRI. Neuropsychologia 48(6):1689–1696
.
OpenUrl CrossRef PubMed

[29] Visser M,

[30] Embleton KV,

[31] Jefferies E,

[32] Parker GJ,

[33] Lambon Ralph MA

[34] ↵
Buckner RL,
Carroll DC
(2007) Self-projection and the brain. Trends Cogn Sci 11(2):49–57
.
OpenUrl CrossRef PubMed

[35] Buckner RL,

[36] Carroll DC

[37] ↵
Mason MF, et al.
(2007) Wandering minds: The default network and stimulus-independent thought. Science 315(5810):393–395
.
OpenUrl Abstract/FREE Full Text

[38] Mason MF, et al.

[39] ↵
Andrews-Hanna JR
(2012) The brain’s default network and its adaptive role in internal mentation. Neuroscientist 18(3):251–270
.
OpenUrl Abstract/FREE Full Text

[40] Andrews-Hanna JR

[41] ↵
Laird AR, et al.
(2009) Investigating the functional heterogeneity of the default mode network using coordinate-based meta-analytic modeling. J Neurosci 29(46):14496–14505
.
OpenUrl Abstract/FREE Full Text

[42] Laird AR, et al.

[43] ↵
Andrews-Hanna JR,
Reidler JS,
Sepulcre J,
Poulin R,
Buckner RL
(2010) Functional-anatomic fractionation of the brain’s default network. Neuron 65(4):550–562
.
OpenUrl CrossRef PubMed

[44] Andrews-Hanna JR,

[45] Reidler JS,

[46] Sepulcre J,

[47] Poulin R,

[48] Buckner RL

[49] ↵
Smith SM, et al.
(2009) Correspondence of the brain’s functional architecture during activation and rest. Proc Natl Acad Sci USA 106(31):13040–13045
.
OpenUrl Abstract/FREE Full Text

[50] Smith SM, et al.

[51] ↵
Pobric G,
Jefferies E,
Lambon Ralph MA
(2007) Anterior temporal lobes mediate semantic representation: Mimicking semantic dementia by using rTMS in normal participants. Proc Natl Acad Sci USA 104(50):20137–20141
.
OpenUrl Abstract/FREE Full Text

[52] Pobric G,

[53] Jefferies E,

[54] Lambon Ralph MA

[55] ↵
Binney RJ,
Embleton KV,
Jefferies E,
Parker GJ,
Lambon Ralph MA
(2010) 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 20(11):2728–2738
.
OpenUrl Abstract/FREE Full Text

[56] Binney RJ,

[57] Embleton KV,

[58] Jefferies E,

[59] Parker GJ,

[60] Lambon Ralph MA

[61] ↵
Visser M,
Jefferies E,
Embleton KV,
Lambon Ralph MA
(2012) Both the middle temporal gyrus and the ventral anterior temporal area are crucial for multimodal semantic processing: Distortion-corrected fMRI evidence for a double gradient of information convergence in the temporal lobes. J Cogn Neurosci 24(8):1766–1778
.
OpenUrl CrossRef PubMed

[62] Visser M,

[63] Jefferies E,

[64] Embleton KV,

[65] Lambon Ralph MA

[66] ↵
Visser M,
Lambon Ralph MA
(2011) Differential contributions of bilateral ventral anterior temporal lobe and left anterior superior temporal gyrus to semantic processes. J Cogn Neurosci 23(10):3121–3131
.
OpenUrl CrossRef PubMed

[67] Visser M,

[68] Lambon Ralph MA

[69] ↵
Vandenberghe R,
Price C,
Wise R,
Josephs O,
Frackowiak RS
(1996) Functional anatomy of a common semantic system for words and pictures. Nature 383(6597):254–256
.
OpenUrl CrossRef PubMed

[70] Vandenberghe R,

[71] Price C,

[72] Wise R,

[73] Josephs O,

[74] Frackowiak RS

[75] ↵
Wirth M, et al.
(2011) Semantic memory involvement in the default mode network: A functional neuroimaging study using independent component analysis. Neuroimage 54(4):3057–3066
.
OpenUrl CrossRef PubMed

[76] Wirth M, et al.

[77] ↵
Seghier ML,
Fagan E,
Price CJ
(2010) Functional subdivisions in the left angular gyrus where the semantic system meets and diverges from the default network. J Neurosci 30(50):16809–16817
.
OpenUrl Abstract/FREE Full Text

[78] Seghier ML,

[79] Fagan E,

[80] Price CJ

[81] ↵
Noonan KA,
Jefferies E,
Visser M,
Lambon Ralph MA
(2013) Going beyond inferior prefrontal involvement in semantic control: Evidence for the additional contribution of dorsal angular gyrus and posterior middle temporal cortex. J Cogn Neurosci 25(11):1824–1850
.
OpenUrl CrossRef PubMed

[82] Noonan KA,

[83] Jefferies E,

[84] Visser M,

[85] Lambon Ralph MA

[86] ↵
Humphreys GF,
Lambon Ralph MA
(September 9, 2014) Fusion and fission of cognitive functions in the human parietal cortex. Cereb Cortex doi:10.1093/cercor/bhu198
.
OpenUrl Abstract/FREE Full Text

[87] Humphreys GF,

[88] Lambon Ralph MA

[89] ↵
Fedorenko E,
Duncan J,
Kanwisher N
(2013) Broad domain generality in focal regions of frontal and parietal cortex. Proc Natl Acad Sci USA 110(41):16616–16621
.
OpenUrl Abstract/FREE Full Text

[90] Fedorenko E,

[91] Duncan J,

[92] Kanwisher N

[93] ↵
Fox MD, et al.
(2005) The human brain is intrinsically organized into dynamic, anticorrelated functional networks. Proc Natl Acad Sci USA 102(27):9673–9678
.
OpenUrl Abstract/FREE Full Text

[94] Fox MD, et al.

[95] ↵
Seghier ML,
Price CJ
(2012) Functional heterogeneity within the default network during semantic processing and speech production. Front Psychol 3:281
.
OpenUrl CrossRef PubMed

[96] Seghier ML,

[97] Price CJ

[98] ↵
Shapira-Lichter I,
Oren N,
Jacob Y,
Gruberger M,
Hendler T
(2013) Portraying the unique contribution of the default mode network to internally driven mnemonic processes. Proc Natl Acad Sci USA 110(13):4950–4955
.
OpenUrl Abstract/FREE Full Text

[99] Shapira-Lichter I,

[100] Oren N,

[101] Jacob Y,

[102] Gruberger M,

[103] Hendler T

[104] ↵
Laurienti PJ, et al.
(2002) Deactivation of sensory-specific cortex by cross-modal stimuli. J Cogn Neurosci 14(3):420–429
.
OpenUrl CrossRef PubMed

[105] Laurienti PJ, et al.

[106] ↵
Fransson P
(2006) How default is the default mode of brain function? Further evidence from intrinsic BOLD signal fluctuations. Neuropsychologia 44(14):2836–2845
.
OpenUrl CrossRef PubMed

[107] Fransson P

[108] ↵
Mayer JS,
Roebroeck A,
Maurer K,
Linden DE
(2010) Specialization in the default mode: Task-induced brain deactivations dissociate between visual working memory and attention. Hum Brain Mapp 31(1):126–139
.
OpenUrl PubMed

[109] Mayer JS,

[110] Roebroeck A,

[111] Maurer K,

[112] Linden DE

[113] ↵
Duncan J
(2010) The multiple-demand (MD) system of the primate brain: Mental programs for intelligent behaviour. Trends Cogn Sci 14(4):172–179
.
OpenUrl CrossRef PubMed

[114] Duncan J

[115] ↵
Pascual B, et al.
(2015) Large-scale brain networks of the human left temporal pole: A functional connectivity MRI study. Cereb Cortex 25(3):680–702
.
OpenUrl Abstract/FREE Full Text

[116] Pascual B, et al.

[117] ↵
Geranmayeh F,
Wise RJS,
Mehta A,
Leech R
(2014) Overlapping networks engaged during spoken language production and its cognitive control. J Neurosci 34(26):8728–8740
.
OpenUrl Abstract/FREE Full Text

[118] Geranmayeh F,

[119] Wise RJS,

[120] Mehta A,

[121] Leech R

[122] ↵
Handy TC
(2000) Capacity theory as a model of cortical behavior. J Cogn Neurosci 12(6):1066–1069
.
OpenUrl CrossRef PubMed

[123] Handy TC

[124] ↵
Attwell D,
Laughlin SB
(2001) An energy budget for signaling in the grey matter of the brain. J Cereb Blood Flow Metab 21(10):1133–1145
.
OpenUrl CrossRef PubMed

[125] Attwell D,

[126] Laughlin SB

[127] ↵
Marinkovic K, et al.
(2003) Spatiotemporal dynamics of modality-specific and supramodal word processing. Neuron 38(3):487–497
.
OpenUrl CrossRef PubMed

[128] Marinkovic K, et al.

[129] ↵
Buckner RL, et al.
(2009) Cortical hubs revealed by intrinsic functional connectivity: Mapping, assessment of stability, and relation to Alzheimer’s disease. J Neurosci 29(6):1860–1873
.
OpenUrl Abstract/FREE Full Text

[130] Buckner RL, et al.

[131] ↵
Rice GE,
Lambon Ralph MA,
Hoffman P
(March 13, 2015) The roles of left vs. right anterior temporal lobes in conceptual knowledge: An ALE meta-analysis of 97 functional neuroimaging studies. Cereb Cortex doi:10.1093/cercor/bhv024
.
OpenUrl Abstract/FREE Full Text

[132] Rice GE,

[133] Lambon Ralph MA,

[134] Hoffman P

[135] ↵
Schapiro AC,
McClelland JL,
Welbourne SR,
Rogers TT,
Lambon Ralph MA
(2013) Why bilateral damage is worse than unilateral damage to the brain. J Cogn Neurosci 25(12):2107–2123
.
OpenUrl CrossRef PubMed

[136] Schapiro AC,

[137] McClelland JL,

[138] Welbourne SR,

[139] Rogers TT,

[140] Lambon Ralph MA

[141] ↵
Lambon Ralph MA,
McClelland JL,
Patterson K,
Galton CJ,
Hodges JR
(2001) No right to speak? The relationship between object naming and semantic impairment: neuropsychological evidence and a computational model. J Cogn Neurosci 13(3):341–356
.
OpenUrl CrossRef PubMed

[142] Lambon Ralph MA,

[143] McClelland JL,

[144] Patterson K,

[145] Galton CJ,

[146] Hodges JR

[147] ↵
Blank SC,
Scott SK,
Murphy K,
Warburton E,
Wise RJS
(2002) Speech production: Wernicke, Broca and beyond. Brain 125(Pt 8):1829–1838
.
OpenUrl Abstract/FREE Full Text

[148] Blank SC,

[149] Scott SK,

[150] Murphy K,

[151] Warburton E,

[152] Wise RJS

[153] ↵
Cohen L, et al.
(2002) Language-specific tuning of visual cortex? Functional properties of the visual word form area. Brain 125(Pt 5):1054–1069
.
OpenUrl Abstract/FREE Full Text

[154] Cohen L, et al.

[155] ↵
Drane DL, et al.
(2013) Famous face identification in temporal lobe epilepsy: Support for a multimodal integration model of semantic memory. Cortex 49(6):1648–1667
.
OpenUrl CrossRef PubMed

[156] Drane DL, et al.

[157] ↵
Andrews-Hanna JR,
Smallwood J,
Spreng RN
(2014) The default network and self-generated thought: Component processes, dynamic control, and clinical relevance. Ann N Y Acad Sci 1316:29–52
.
OpenUrl CrossRef PubMed

[158] Andrews-Hanna JR,

[159] Smallwood J,

[160] Spreng RN

[161] ↵
Corbetta M,
Shulman GL
(2002) Control of goal-directed and stimulus-driven attention in the brain. Nat Rev Neurosci 3(3):201–215
.
OpenUrl CrossRef PubMed

[162] Corbetta M,

[163] Shulman GL

[164] ↵
Postle BR
(2015) The cognitive neuroscience of visual short-term memory. Current Opinion in Behavioral Sciences 1(0):40–46
.
OpenUrl CrossRef

[165] Postle BR

[166] ↵
Hoffman P,
Binney RJ,
Lambon Ralph MA
(2015) Differing contributions of inferior prefrontal and anterior temporal cortex to concrete and abstract conceptual knowledge. Cortex 63:250–266
.
OpenUrl CrossRef PubMed

[167] Hoffman P,

[168] Binney RJ,

[169] Lambon Ralph MA

[170] ↵
Embleton KV,
Haroon HA,
Morris DM,
Lambon Ralph MA,
Parker GJM
(2010) Distortion correction for diffusion-weighted MRI tractography and fMRI in the temporal lobes. Hum Brain Mapp 31(10):1570–1587
.
OpenUrl CrossRef PubMed

[171] Embleton KV,

[172] Haroon HA,

[173] Morris DM,

[174] Lambon Ralph MA,

[175] Parker GJM

Main menu

User menu

Search

Establishing task- and modality-dependent dissociations between the semantic and default mode networks

Significance

Abstract

Results

Discussion

Methods

Tasks.

Scanning.

Analyses.

Acknowledgments

Footnotes

References

Citation Manager Formats

Article Classifications

You May Also be Interested in

Similar Articles

Articles

PNAS Portals

Information

Main menu

User menu

Search

Significance

Abstract

Results

Discussion

Methods

Tasks.

Scanning.

Analyses.

Acknowledgments

Footnotes

References

Citation Manager Formats

Article Classifications

Sign up for Article Alerts

Jump to section

You May Also be Interested in

Similar Articles

Articles

PNAS Portals

Information