The Presupplementary Area Within the Language Network: A Resting State Functional Magnetic Resonance Imaging Functional Connectivity Analysis

Abstract

The presupplementary motor area (pre-SMA) is involved in volitional selection. Despite the lateralization of the language network and different functions for both pre-SMA, few studies have reported the lateralization of pre-SMA activity and very little is known about the possible lateralization of pre-SMA connectivity. Via functional connectivity analysis, we sought to understand how the language network may be connected to other intrinsic connectivity networks (ICNs) through the pre-SMA. We performed a spatial independent component analysis of resting state functional magnetic resonance imaging in 30 volunteers to identify the language network. Subsequently, we applied seed-to-voxel functional connectivity analyses centered on peaks detected in the pre-SMA. Three signal peaks were detected in the pre-SMA. The left rostral pre-SMA intrinsic connectivity network (LR ICN) was left lateralized in contrast to bilateral ICNs associated to right pre-SMA peaks. The LR ICN was anticorrelated with the dorsal attention network and the right caudal pre-SMA ICN (RC ICN) anticorrelated with the default mode network. These two ICNs overlapped minimally. In contrast, the right rostral ICN overlapped the LR ICN. Both right ICNs overlapped in the ventral attention network (vATT). The bilateral connectivity of the right rostral pre-SMA may allow right hemispheric recruitment to process semantic ambiguities. Overlap between the right pre-SMA ICNs in vATT may contribute to internal thought to external environment reorientation. Distinct ICNs connected to areas involved in lexico-syntactic selection and phonology converge in the pre-SMA, which may constitute the resolution space of competing condition-action associations for speech production.

Introduction

The supplementary motor complex (SMC) is a medial frontal structure. Within the SMC there are major differences in connectivity between the presupplementary motor area (pre-SMA), which is mainly connected to prefrontal regions, and the supplementary area proper (SMA proper), which projects mainly to primary motor regions. The limit between these two areas corresponds roughly to the vertical plane intersecting the anterior commissure (Nachev et al., 2008; Picard and Strick, 1996, 2001). The pre-SMA projects to the inferior frontal cortex (Swann et al., 2012) and plays a role in high-level planning (Picard and Strick, 2001) including speech initiation and execution (Alario et al., 2006; Price, 2010; Segaert, 2012).

The pre-SMA has been discussed widely in the literature as concerns of the neurophysiology of language, but very little as concerns of functional connectivity. This may be due to the nondomain-specific involvement of the pre-SMA in volitional selection (Tremblay and Gracco, 2009). However the pre-SMA has been associated with various, overt or covert, linguistic tasks (Abrahams et al., 2003; Alario et al., 2006; Bohland and Guenter, 2006; Crosson et al., 2001; Price, 2010; Segaert et al., 2012; Tremblay and Gracco, 2006) as well as the production of nonspeech-related volitional movement (Botvinick et al., 1999; Fan et al., 2007; Mars et al., 2005, 2007; Nachev et al., 2005; Rushworth et al., 2002; Ullsperger and von Cramon, 2001; Van Oostende et al., 1997; Xue et al., 2008).

Thus, it may be of interest to focus on the intrinsic connectivity networks (ICNs) associated with this area to better understand language physiology and highlight the coordination of semantic, lexico-syntactic, and phonological aspects of language.

ICNs may be defined as sets of brain areas that are synchronous in the low frequency range (<0.1 Hz) (Biswal et al., 1995). The role of these spontaneous oscillations is still poorly understood (Duyn, 2011) but their correlation structure (functional connectivity) is believed to reflect the neuroanatomical substrate of task-induced activity (Fox et al., 2005; Smith et al., 2009).

Thus, anatomical connectivity provides indications on possible functional connectivity patterns.

The precise anatomical connectivity of the pre-SMA was recently described by diffusion-tensor-imaging-based tractography; a white matter frontal aslant tract (FAT) connects the pre-SMA to the pars opercularis and pars triangularis of the inferior frontal gyrus (IFG_po and IFG_tri, respectively). FAT is left lateralized (Catani et al., 2012) and underlies verbal fluency in primary progressive aphasia (Catani et al., 2013), thus confirming the major role of the left pre-SMA in lexical selection.

Further, there appears to be a functional parcellation in the right and left pre-SMAs that participates in different aspects of motor and lexical selection control involved in speech production (Alario et al., 2006; Bohland and Guenter, 2006; Ghosh et al., 2008; Segaert et al., 2012; Swann et al., 2012; Xue et al., 2008).

Specifically, lexical selection is processed by a circuitry involving the left rostral pre-SMA (Alario et al., 2006), whereas motor and speech inhibition are processed by the right pre-SMA (Xue et al., 2008).

From a methodological point of view, the location of the pre-SMA adjacent to the medial sagittal plane is a challenging obstacle to investigation (Zhang et al., 2012) and only a few studies have addressed the laterality of pre-SMA activity during linguistic tasks.

Further, despite its left lateralization, the literature shows that the lateralization of the language network is dependent on the syntactic or semantic nature of the task being executed as well as the difficulty and amount of semantic integration required (Ben-Artzi et al., 2009; Brown et al., 2006; Diaz et al., 2011; Dronkers et al., 2004; Faust and Lavidor; 2003; Faust et al., 2008; Ferstl et al., 2008; Gandour et al., 2003; Kuperberg et al., 2008; Mashal et al., 2008; Marslen-Wilson and Tyler, 2007; Papoutsi et al., 2011; Rapp et al., 2004, 2010; Ross, 1981; Rota et al., 2009; Schmidt and Seger, 2009; Schmidt et al., 2007; Snijders et al., 2010; Tyler et al., 2010; Wilson et al., 2012). It is however unclear how the left lateralized language network coordinates with its right counterpart.

Due to its major function in selection, we expect that the pre-SMA plays a role in this coordination. As the left pre-SMA is involved in lexical selection in the left hemisphere, we suspect that the right pre-SMA plays a role in the control of this extended bilateral system supporting semantic (Binder et al., 2009) and speech comprehension (Bozic et al., 2010).

It is important to note that the right pre-SMA is not only involved in speech and motor control but it is also a key area of the ventral attention network (vATT) (Fox et al., 2006), which also comprises the IFG_po. These two areas are positioned along the pathway of the right FAT. Thus, the right pre-SMA may support some attentional aspects of linguistic processes involving vATT. In line with this interpretation, it has been suggested that during linguistic social interaction, memory captures attention in a bottom-up process involving vATT (Cabeza, 2008).

Apart from considerations on connectivity, pathology provides indications on the possible functions of the left and right pre-SMAs. Lesions of the SMC result in apparently contradictory syndromes, such as an urge to move and akinesia, although both involve the impairment of volitional selection (Nachev et al., 2008). However, dominant hemisphere SMC surgical resection most frequently results in an akinetic syndrome characterized by transient profound postoperative motor and speech deficit. Classically these deficits resolve within weeks or months (Laplane et al., 1977; Krainik et al., 2001, 2003; Rostomily, 1991; Russel and Kelly, 2003) in most but not all patients (Kim et al., 2013; Rosenberg et al., 2010). The precise mechanism allowing for such a recovery remains uncertain. An intuitive hypothesis is that the right pre-SMA steps in to compensate for the impairment in verbal fluency (Krainik et al., 2004). However, a recent study reported that recovery does not involve contralateral pre-SMA recruitment but residual ipsilateral SMC activation (Rosenberg et al., 2010). This observation strongly supports the argument that the left and right pre-SMAs have different functions and most likely different connectivity patterns.

Thus, the aim of this study was to systematically assess differences and overlaps in ICNs associated with peak activities of the pre-SMA isolated within the language network using functional magnetic resonance imaging (fMRI).

To insure that pre-SMA coordinates were not confounded by the coactivation of networks other than language during the execution of a specific task, we performed resting state fMRI. Indeed, the language network has been isolated by functional connectivity analysis in the absence of any specific task (Cordes et al., 2000; Hampson et al., 2002; Lee et al., 2012) and this method has been validated against the gold standard of language network identification (Mitchell et al., 2013).

In a first step, we performed a spatial independent component (IC) analysis to isolate the pre-SMA within the language network at rest. In a second step, we used these coordinates to perform a functional connectivity analysis. These steps allowed us to avoid the circularity of some successive analyses on the same data set (Suresh et al., 2011) and thus to understand how the language network is functionally connected to other ICNs through the pre-SMA.

Materials and Methods

Subjects

Thirty volunteers, all neurologically intact, right-handed, healthy, full-time physicians at our institution [mean age=34 years (25–57 years), 10 women] were included in this study after providing their informed consent. The study was approved by the local ethics committee.

fMRI acquisition

All datasets were acquired on a 1.5-T MR scanner (Magnetom Avanto; Siemens Medical Systems, Erlangen, Germany) equipped with a circular, polarized head coil.

Participants were placed supinely with the head immobilized by foam pads and straps. During the resting state session they were instructed to stare at a red fixation cross projected on a black screen above them.

An echo planar imaging sequence was used to acquire functional series for each participant (repetition time=3000 msec, echo time=50 msec, flip angle=90°, 32 axial slices interleaved, 5.0-mm-thick, 0-mm-gap, in plane matrix=64×64, yielding a voxel size of 3×3×5 mm³, field of view=240 mm), covering the whole brain including the cerebellum. We acquired 207 functional volumes per session. A T1-weighted anatomical three-dimensional dataset covering the whole brain was also obtained (176 contiguous axial slices, in-plane matrix 256×256, yielding a voxel size of 1×1×1 mm³) to coregister and normalize echo planar images.

Image preprocessing

fMRI data were analyzed using SPM5 software (Wellcome Department of Imaging Neuroscience, University College, London, United Kingdom; www.fil.ion.ucl.ac.uk/spm) running under Matlab (The MathWorks, Natick, MA). The first three acquisition volumes of each functional series were discarded to allow for stabilization of longitudinal magnetization.

Preprocessing was conducted as follows. First, each volunteer's native space images were corrected for time delays between different slices. Then, they were realigned to the first volume and unwrapped to correct for head movement and susceptibility distortions. The vectors of the estimated movements provided by SPM were inspected to ensure that the amplitude does not exceed 3 mm along each of the three axes. The anatomical dataset was segmented using VBM 5.0 toolbox for SPM. Coregistration of images from different modalities was achieved using gray matter segmentation as a reference image and a mean echo planar of the slice-timed and motion-corrected unwrapped image as the source image. Normalization of echo planar images to the standard template provided by the Montreal Neurological Institute (MNI-template) was then based on the coregistered gray matter segments. Functional images were written out to a final resolution of 3×3×3 mm³ before being smoothed by an isotropic 6-mm Gaussian kernel in full-width at half-maximum.

Spatial IC analysis

Importantly, as the literature attributes a variety of functions to different subareas within the pre-SMA, we purposely did not draw any prior hypothesis as to the coordinates of the pre-SMs during analysis. Thus, we took advantage of spatial independent component analysis (sICA) of networks to isolate the coordinates of component peaks within the pre-SMA in the language network.

Identification of the language network

sICA for the identification of large-scale networks was performed using NetBrainWork (www.imed.jussieu.fr/en/equipes/e1/netbrainwork/netbrainwork.php) running under Matlab. Briefly, after computing spatial ICs on each individual run (40 ICs per run), a hierarchical clustering of the 1200 ICs (40 ICs×30 subjects) was performed. From this similarity tree, representative classes of the population were determined according to the maximization of two indices (representativeness and unicity). Then, a fixed-effect group map of t-scores was computed for each class from all ICs belonging to these classes. These t-maps were thresholded at p<0.05 false discovery rate (FDR), corrected for multiple comparisons and with a minimal cluster extent of 5 voxels.

Identification of the language network was performed using the following criteria: left lateralized network, presenting maximal activity within all subdivisions of the IFG [pars opercularis (Marslen-Wilson and Tyler, 2007), orbitalis and triangularis (Bozic et al., 2010)] and also the angular gyrus (ANG) (Vigneau et al., 2006); middle temporal gyrus (MTG) with a peak of activity in its mid-posterior part (Devlin et al., 2004; Dronkers et al., 2004); temporal poles (Binder et al., 2011); bilateral caudate (Crosson et al., 2003); right cerebellum (Jansen et al., 2005); and dorsomedial prefrontal cortex including rostral pre-SMA (Alario et al., 2006).

Defining regions of interest in the language network

Identification of local peak maxima on the FDR-corrected t-maps of the language network was processed using the ANATOMY toolbox v1.5 for SPM.

Regions of interest (ROIs) masks were generated using the WFU PickAtlas toolbox for SPM. A 3-mm-radius ROI was drawn centered on the coordinates of the detected peaks within the pre-SMA between the coronal plane passing through the anterior commissure and the posterior limit of the superior medial frontal gyrus. The ROIs were then coregistered with the smoothed normalized functional images, resulting in a volume of ∼100 mm³.

Seed-based functional connectivity analysis

Functional connectivity analysis was performed using Conn v13.i software (www.nitrc.org/project/conn) running under Matlab.

Preprocessing of functional series was further performed by regressing the following covariables as sources of confounding in such a way as to correct for signal inhomogeneities due to biological movements and to remove spurious variance.

(1) The five principal components of the brain oxygen level dependent signal of the white matter and cerebro-spinal fluid obtained from individual normalized masks and their first-order derivatives.

(2) The six principal components of each of the six estimated individual motion parameters obtained from the realignment step of preprocessing and their first-order derivatives.

Signals of functional data were temporally detrended and then bandpass filtered from 0.008 to 0.09 Hz.

First-level analysis included extraction of the signal of each language network ROI within the individual gray matter mask. Then we computed correlation maps between each seed ROI and all other voxels within the template brain mask provided by the Conn v13.i software.

After Fisher Z′ transformation, a second-level random-effect group analysis was performed, employing one sample t-tests to obtain seed-to-voxel thresholded connectivity maps. Differences between the connectivity maps of pre-SMAs were obtained by paired t-test. All results were thresholded at p<0.05 FDR corrected at the voxel level, extent threshold 5 voxels.

Overlap between ICNs

To understand how information could be exchanged between ICNs associated with the pre-SMA, we analyzed their overlap by determining the conjunction masks of pairs of pre-SMA ICN maps. To reduce the false-positive risk, we only considered coordinates of local peaks of correlation in pre-SMA ICNs as determined by the ANATOMY toolbox. Peaks of correlation falling in the masks or within a distance <3 mm from the masks were selected. Pairs of peaks of correlation within the same anatomic area were considered as potential bridges of connectivity between ICNs if their absolute distance along the three axes was <18 mm.

Results

The main ICNs identified by sICA are shown in Supplementary Figure S1 (Supplementary Data are available online at www.liebertpub.com/brain). The language network is shown in Figure 1. Corresponding coordinates and t values of the 53 local maxima of the language network appear in Table 1.

FIG. 1.

Language network as identified by spatial independent component analysis. No activity was detected in the supplementary area proper. Pre-SMA, presupplementary motor area; PCC, posterior cingulate cortex; CAU, caudate nucleus. Group analysis. Significant clusters of interest (COIs) were projected onto template brain slices in the MNI stereotactic space. The color bar indicates t values. COIs are shown at voxel-level p _FDRcorr<0.05, cluster-size threshold 5 voxels. FDR, false discovery rate; MNI, Montreal Neurological Institute.

Table 1.

Local Peak Maxima of the Language Network Identified by Independent Components Analysis

Region	Cluster size	T²⁹	x	y	z
[Left inferior frontal gyrus (p. orbitalis)]	1402	12.6	−49	23	−10
(Left middle frontal gyrus)		10.89	−41	11	55
(Left superior medial gyrus)		10.53	−4	38	35
(Left superior medial gyrus)		10.39	−4	38	50
(Left rostral pre-SMA)		10.21	−8	23	60
(Left superior medial gyrus)		10.09	−4	45	40
(Left superior medial gyrus)		9.94	−8	30	60
[Left inferior frontal gyrus (p. triangularis)]		9.47	−53	19	5
(Left middle temporal gyrus)		9.38	−56	−38	−5
(Right superior medial gyrus)		9.08	8	53	30
(Left superior medial gyrus)		9.01	−11	60	30
(Left superior medial gyrus)		8.56	−8	56	35
(Right rostral pre-SMA)		7.79	8	23	60
(Right superior medial gyrus)		7.68	8	49	40
(Left middle frontal gyrus)		7.1	−38	23	45
(Left middle temporal gyrus)		7.05	−64	−23	−15
(Left middle frontal gyrus)		6.83	−19	49	30
[Left inferior frontal gyrus (p. orbitalis)]		6.52	−26	19	−25
(Left superior frontal gyrus)		5.85	−30	53	0
(Left middle temporal gyrus)		5.39	−60	−11	−20
(Left inferior temporal gyrus)		5.16	−53	−4	−30
(Left inferior temporal gyrus)		4.98	−49	4	−40
(Left inferior temporal gyrus)		4.84	−45	8	−35
(Right caudal pre-SMA)		4.81	8	8	70
[Left inferior frontal gyrus (p. triangularis)]		4.78	−49	15	30
(Left medial temporal pole)		4.69	−49	15	−30
(Right middle frontal gyrus)		4.21	26	60	20
(Right middle frontal gyrus)		3.92	23	56	25
(Left angular gyrus)	323	13.68	−53	−60	30
[Right cerebellum (Crus 1)]	223	10.07	23	−86	−30
[Right cerebellum (Crus 2)]		7.3	23	−86	−40
[Right cerebellum (Crus 1)]		6.06	38	−64	−30
[Right cerebellum (VIII)]		4.67	38	−60	−50
[Right cerebellum (VIII)]		4.04	30	−64	−50
[Right cerebellum (Crus 2)]		3.55	38	−71	−45
[Right inferior frontal gyrus (p. orbitalis)]	101	6.91	49	23	−10
[Right inferior frontal gyrus (p. triangularis)]		6.48	56	23	5
[Right inferior frontal gyrus (p. orbitalis)]		4.37	34	23	−15
(Right temporal pole)		4.36	49	19	−20
(Right middle temporal gyrus)	58	5.98	68	−41	0
(Right middle temporal gyrus)		5.83	64	−34	−10
(Right middle temporal gyrus)		4.33	53	−26	−15
(Right middle temporal gyrus)		4.05	68	−15	−15
(Right angular gyrus)	57	6.2	53	−53	30
(Right inferior parietal lobule)		4.76	56	−60	40
[Left cerebellum (Crus 2)]	45	5.85	−23	−83	−40
(Left caudate nucleus)	38	5.96	−15	4	15
(Left posterior cingulate cortex)	37	6.6	−4	−53	30
(Right caudate nucleus)	32	6.91	15	15	10
(Left middle cingulate cortex)	29	6.66	−4	−15	40
(Right middle frontal gyrus)	10	3.96	45	23	45
[Right inferior frontal gyrus (p. opercularis)]		3.85	53	23	35
(Right medial temporal pole)	8	4.93	53	11	−30
[Right cerebellum (IX)]	7	4.44	8	−56	−50

pre-SMA peaks appear in bold.

FDR, false discovery rate; MNI, Montreal Neurological Institute; pre-SMA, presupplementary motor area.

Within the language network, no peak of activity was detected in the SMA proper. Pre-SMA activity was detected bilaterally with only three local maxima. Maximal activity was present in the left rostral pre-SMA at coordinates (−8, 23, 60). Two peaks were detected on the right side, one within the rostral pre-SMA at coordinates symmetric to its left homologue at (8, 23, 60) and the other in the caudal pre-SMA at (8, 8, 70) (Supplementary Fig. S2).

Connectivity patterns and differences between pre-SMA ICNs

ICNs associated with the three pre-SMA ROIs are shown in Figure 2 and described in Supplementary Tables S1–S3.

FIG. 2.

Intrinsic connectivity network (ICN) maps associated with the three pre-SMA regions of interest isolated from the language network. (A) Left rostral pre-SMA ICN. (B) Right rostral pre-SMA ICN. (C) Right caudal pre-SMA ICN. ICNs are shown in hot and projected on coronal and sagittal slices of a template brain in the MNI stereotactic space. The color bar indicates t values. ICNs shown at voxel-level p _FDRcorr<0.05, cluster-size threshold 5 voxels.

Positive correlations of the rostral pre-SMA

The left rostral pre-SMA ICN (LR ICN) and right rostral pre-SMA ICN (RR ICN) included the posterior cingulate cortex (PCC), orbitomedial prefrontal cortex (FMO), both ANG, and both MTG extending to the temporal poles and also the anterior cingulate cortex (ACC), middle cingulate cortex (MCC), superior frontal gyrus (SFG), middle frontal gyrus (MFG), and IFG extending to the anterior insula/frontal operculum (AIFO).

Negative correlations of the rostral pre-SMA

The LR presented a negative correlation with a set of areas, including the bilateral supramarginal gyrus, superior intraparietal sulcus (sIPS), superior parietal lobule, and left middle temporo-occipital junction (MT+). These areas are known to be major nodes of the dorsal attention network (dATT) (Supplementary Fig. 3A). The RR presented the same negative correlation pattern, remarkably limited to the left hemisphere however (Supplementary Fig. 3B).

Differences in connectivity of the rostral pre-SMA

The LR presented no connectivity with the right IFG or right temporo-parietal junction (TPJ) and only limited connectivity with the right hemisphere. This pattern contrasted with the notably symmetric connectivity of the RR in both hemispheres.

The RR ICN overlapped the LR ICN and differed from it by greater connectivity with a set of areas, including the right TPJ extending to the inferior parietal lobule and the lower bank of the sIPS, the right dorsal ACC (dACC), and the right AIFO and IFG. All these areas are nodes of the vATT (Corbetta and Shulman, 2002; Fox et al., 2006) (Table 2 and Fig. 3). Conversely, a statistical comparison showed that the LR presented greater connectivity than the RR ICN, with few sparse small clusters (Supplementary Table S4). Further differences of left pre-SMA ICN consisted in a continuous pattern of connectivity from the pre-SMA to the dorsolateral prefrontal cortex (DLPFC) of the MFG through the rostral pre-motor dorsal area (pre-PMd) shown in Figure 4 (Abe and Takashi, 2009; Picard and Strick, 2001).

FIG. 3.

Paired t-test of left rostral (LR) and right rostral (RR) pre-SMA ICN maps. The greater connectivity of the RR ICN compared with the LR ICN is shown in hot and projected on a template brain in the MNI stereotactic space. IPS, intraparietal sulcus; TPJ, temporo-parietal junction; AIFO, anterior insula/frontal operculum; IFGtri, inferior frontal gyrus pars triangularis; MTGpost, posterior middle temporal gyrus; LR ICN, left rostral pre-SMA ICN; RR ICN, right rostral pre-SMA ICN. The color bar indicates t values. Map shown at voxel-level p _FDRcorr<0.05, cluster-size threshold 5 voxels.

FIG. 4.

Left rostral presupplementary area connectivity pattern (shown in hot) through convexity projected on a template brain in the MNI stereotactic space. pre-PMd, rostral premotor dorsal cortex; DLPFC, dorsolateral prefrontal cortex. The color bar indicates t values. Map is shown at voxel-level p _FDRcorr<0.05, cluster-size threshold 5 voxels.

Table 2.

Paired t-Test: Greater Connectivity of the Right Rostral Compared with the Left Rostral Presupplementary Motor Area Intrinsic Connectivity Network

Region	Cluster size	T²⁹	x	y	z
[Right supramarginal gyrus (TPJ)]	308	5.96	60	−42	24
[Right superior temporal gyrus (TPJ)]		5.96	63	−45	21
(Right middle temporal gyrus)		5.7	51	−30	−9
(Right superior intraparietal sulcus)		4.49	42	−48	57
[Right inferior frontal gyrus (p. opercularis)]	207	6.65	45	12	27
(Right middle frontal gyrus)		4.58	39	0	63
(Right precentral gyrus)		4.49	48	9	45
(Right temporal pole)	182	4.66	48	15	−12
(Right anterior insula frontal operculum)		4.58	33	15	−9
(Right rolandic operculum)		4.15	57	6	6
(Right insula lobe)		4.15	36	15	0
[Right inferior frontal gyrus (p. opercularis)]		3.97	57	18	15
(Right rostral pre-SMA)	169
(Right pre-SMA)		5.01	6	9	54
(Left middle cingulate cortex)		5.01	0	15	39
(Right dorsal anterior cingulate cortex)		3.8	3	15	27
(Right superior frontal gyrus)	115	5.44	24	60	30
(Right middle frontal gyrus)		5.1	27	60	24
[Right inferior frontal gyrus (p. triangularis)]	50	4.92	48	36	0
(Right middle frontal gyrus)		4.06	42	42	9
(Right middle cingulate cortex)	38	5.61	6	−24	42
(Right parahippocampal gyrus)	32	4.49	24	15	−27
[Right inferior frontal gyrus (p. orbitalis)]		4.23	18	27	−12
(Right rectal gyrus)		3.89	27	21	−24
[Right inferior frontal gyrus (p. orbitalis)]		3.71	15	27	−18
(Left insula lobe)	21	4.49	−36	6	3
[Left cerebellum (Crus 1)]	13	4.75	−33	−69	−27
[Left cerebellum (Crus 1)]		3.97	−33	−78	−24
(Right middle frontal gyrus)	10	3.89	48	27	45
[Left cerebellum (Crus 2)]	8	4.23	−18	−75	−36
[Left cerebellum (Crus 2)]	7	4.49	−42	−63	−48
[Left cerebellum (Crus 1)]	6	3.97	−48	−57	−30
(Left supramarginal gyrus)	5	4.32	−66	−33	21
(Left superior temporal gyrus)		4.32	−63	−33	18

The significant local peak maxima were obtained under a threshold of p _FDRcorr<0.05 at the voxel level, cluster-size threshold 5 voxels; T²⁹ =T value for 29 degrees of freedom; anatomical labels were derived from the anatomy toolbox for SPM and MRIcron; TPJ, temporo-parietal junction; anatomic areas belonging to the attention networks appear in bold; in cases where multiple peaks appear in the same anatomic area of a cluster, only the maximal peak is reported; x,y,z=the original SPM coordinates of the MNI space in millimeters.

Positive correlations of right caudal pre-SMA

The right caudal pre-SMA ICN (RC ICN) was slightly right lateralized and presented a positive connectivity with posterior dACC, MCC, both AIFO and TPJ, and also DLPFC with a moderate right lateralization for right AIFO and TPJ. This network appeared to share similarities with the vATT. However, a main difference with the two other ICNs and also the vATT was the large pattern of connectivity with the primary motor and the premotor dorsal (PMd) cortex below the notch, the rolandic operculum, and also the bilateral supra-marginal gyrus and anterior MFG (Fig. 5).

FIG. 5.

Right caudal presupplementary area connectivity pattern (shown in hot) projected on a template brain in the MNI stereotactic space. (A) Connectivity with primary motor and premotor areas; white dashed arrows indicate central sulcus. (B) Connectivity with verbal working memory areas. preCG, precentral gyrus; RolOp, rolandic operculum; aMFG, anterior middle frontal gyrus; SMG, supra-marginal gyrus. The color bar indicates t values. Map is shown at voxel-level p _FDRcorr<0.05, cluster-size threshold 5 voxels.

Negative correlations of right caudal pre-SMA

The RC ICN was negatively correlated with the FMO, PCC, precuneus, bilateral ANG, and DLPFC, all components of the default mode network (DMN) (Supplementary Fig. S4 and Supplementary Table S5).

Overlap between ICNs

As previously indicated, the LR and RR largely overlapped within a set of areas, including nearly all clusters of the LR ICN. The LR and RC ICNs overlapped marginally.

The RR and RC overlapped within the dACC, bilateral AIFO, TPJ, and rostral boundaries of the Brodmann areas 9 and 46. This overlap included major nodes of the vATT (Fig. 6).

FIG. 6.

Mask of the overlap between the right rostral and right caudal presupplementary area ICNs. Legends in the figure correspond to circled areas. dACC, dorsal anterior cingulate cortex; SFS, superior frontal sulcus. Mask is projected on a template brain in the MNI stereotactic space.

The analysis of peak correlation pairs within the intersection of the three pre-SMA ICNs yielded only one direct link (left putamen) between the LR and RC ICNs. Analysis of pairs of peak correlation within the intersection of the RR and RC ICNs revealed a set of distributed potential connections with a particular distribution along the posterior and anterior banks of the right TPJ and also in the dorsal and ventral banks of the rostral superior frontal sulci (Supplementary Fig. S5 and Supplementary Table S6).

Finally, as expected due to their large overlap, extensive connections were found between the LR and RR ICNs.

Figure 7 summarizes the main results.

FIG. 7.

Main results of the pre-SMA functional connectivity study. White stars indicate the ICNs of the presupplementary motor area (pre-SMA ICN) associated to local maxima isolated in the pre-SMA within the language network. White crosses indicate overlap between pre-SMA ICNs linked by white lines. Yellow arrow indicates that compared LR ICN, RR ICN presents greater connectivity with vATT. Blue arrows indicate pre-SMA ICN anticorrelated networks. Blue dashed arrow indicates that RR ICN presents only a partial spatial anticorrelation with dATT. RC ICN, right caudal pre-SMA ICN; DMN, default mode network; dATT, dorsal attention network; vATT, ventral attention network.

Discussion

To our knowledge this study is the first to report the connectivity pattern of segregated areas within the pre-SMA inside the language network.

Our main finding is that there is no unique pre-SMA peak of activity within the language network but rather three peaks, each presenting a distinct connectivity pattern. Particularly, two of these peaks are located on the right hemisphere, thus confirming our hypothesis that the right pre-SMA is involved in language processing.

Not surprisingly, the LR ICN is left lateralized, shares many similarities with the language network, and encompasses major nodes, including the IFG_tri involved in lexical selection (Fiebach et al., 2002) and syntactic processing (Snijders et al., 2009).

In contrast, the RR ICN appears bilaterally and symmetrically and overlaps the LR ICN. Another characteristic of the RR ICN is its similarities with the vATT, particularly evident when comparing RR and LR connectivity patterns.

Finally, the RC ICN also presents similarities with the vATT, which is not surprising since the coordinates for the RC pre-SMA (8, 8, 70) are close to those of the vATT medial prefrontal activity (8, 0, 67) (Fox et al., 2006). However, the RC ICN differed from the two other ICNs by its unique extensive connectivity pattern with the PMd and the primary motor cortex of the mouth and larynx and also with the bilateral anterior MFG and SMG, which are the main areas involved in verbal working memory (Deschamps et al., 2014; Mottaghy et al., 2003). This suggests that the RC ICN is more specifically involved in phonological working memory.

These results are in accordance with our hypothesis that within the language network, the right pre-SMA may be more connected to the attentional network. Further, there appears to be a functional architecture with two main characteristics: an extensive overlap of left and right rostral ICNs and, in the right hemisphere, the interposition of the vATT between two networks seemingly supporting different functions. This latter characteristic is similar to what has been demonstrated for anticorrelated networks where nodes of a third network play a gate-keeping role for task achievement (Higo et al., 2011; Menon and Uddin, 2010; Spreng et al., 2010, 2013; Sridharan et al., 2008).

Earlier pre-SMA functional connectivity studies

Two previous articles addressed SMC functional connectivity using resting state fMRI. The first focused on validating a new functional parcellation method. The authors confirmed feasibility by showing a functional delineation between SMA and pre-SMA (Kim et al., 2010). The second, a more fine-grained study on an important cohort, systematically assessed the functional connectivity of 24 ROIs along the SMC. The RR ICN and RC ICN that we found correspond to the connectivity maps associated with that study ROIs H10 and H6, respectively (Zhang et al., 2012); the only difference is that our RR pre-SMA was anticorrelated to only a part of the dATT.

Zhang et al. provided a detailed description of the gradual shift of connectivity along the y axis and identified some slightly significant differences in connectivity between hemispheres. However, those authors positioned their ROIs in the midline and thus, by study design, they could not detect the major lateralized patterns of connectivity that we did.

Extensive overlap between the LR and RR ICNs

In our study, the LR ICN was largely confined to the left hemisphere. This finding is coherent with the lexical selection function attributed to the left rostral pre-SMA (Alario et al., 2006) and with the importance of the left hemisphere in the core lexico-syntactic process of language, as indicated by lesion studies (Papoutsi et al., 2011; Wilson et al., 2012). The main clusters of the LR ICN are distributed along the dorsal and ventral pathways of language (Saur et al., 2008) and the left FAT. These tracts are known to support phonological, semantic, and lexical selection processes of language, respectively.

The spatial extension of the RR ICN into the right counterparts of the LR ICN clusters is reminiscent of the recruitment of these areas by semantic ambiguities, complexity and prosody (Bozic et al., 2010; Brown et al., 2006; Ferstl et al., 2008; Kuperberg et al., 2008; Marslen-Wilson and Tyler, 2007; Rapp et al., 2004, 2010; Rota et al., 2009) as the left hemisphere would preferentially code for the dominant meanings of words (Faust and Lavidor, 2003).

This recruitment of the right hemisphere for coarse-coded information and less-salient stimuli has been formulated in terms of fine/coarse coding and graded salience theory (Jung-Beeman, 2005; Mashal et al., 2008).

This interpretation is supported by the involvement of the right IFG_tri in the RR ICN. It has been shown that within the right hemisphere, IFG_tri inhibits the selection process in case of uncertainty (Levy and Wagner, 2011).

In the semantic domain of word processing, such a particular function for the right IFG_tri fits with the right hemisphere's capacity to maintain a wide range of meanings without selection for later generation of critical ambiguous words (Faust et al., 2008; Jung-Beeman, 2005).

The precise mechanism allowing such cooperation between hemispheres may be the language-specific left-to-right transcallosal inhibition of the IFG (Thiel et al., 2006). The modulation of this inhibition should permit right hemisphere recruitment when ambiguities necessitate retaining multiple meanings in working memory.

Anticorrelation pattern of the pre-SMA

The gradient of connectivity with the vATT that we found between the left and right pre-SMAs is also present with the dATT, the right caudal ICN showing the lesser anticorrelation with it. Thus, it appears that, within the pre-SMAs, there is a left-to-right and rostro-caudal gradient of increasing connectivity with attentional networks.

A question arises as to why the LR ICN is more spatially anticorrelated to key areas of the dATT, such as the sIPS, than its right counterpart.

Orienting attention to semantic categories (Cristescu et al., 2006) activates the dATT, which is also involved in evaluating semantic distances (Mahon and Caramazza, 2010; Pinel et al., 2001). It has been shown that in linguistic episodic memory tasks, “know” (weak) memory activates the dATT more than “remember” (strong) memory does (Kim, 2010).

Thus, the fact that the correlation structure of the right rostral pre-SMA is preserved with the right sIPS may reflect the involvement of the right hemisphere in the evaluation of semantic distances when processing semantic ambiguities.

Overlap between RR and RC ICN includes main areas of the vATT

In the right hemisphere, we identified a systematized set of peaks of correlation within the vATT representing potential links between the RR and RC ICNs.

The vATT may function as a system to switch between internally and externally directed activity (Corbetta et al., 2008). Such a coupling between internally and externally directed thoughts would appear to be particularly important in linguistic social interaction. In our data, this particular function of the vATT seems to be reflected by the fact that it links ICNs presenting a particular anticorrelation pattern, that is, an LR IC anticorrelated to the dATT, a network known to be involved in the focus of attention on external activity, and a RC ICN anticorrelated to the DMN known to be involved in internally directed thoughts (Buckner et al., 2008).

Corbetta et al. hypothesized that the link between a reorientation task and theory of mind cognition, both involving the right TPJ, is seen in how subjects shift continually between another's viewpoint and processing perceptual evidence from their own viewpoint. Consequently, the attention signal in the TPJ may reflect the shift between one's self-perspective and another's viewpoint (Corbetta et al., 2008).

We illustrated that the right TPJ is a node within language control systems and this finding seems to support Corbetta et al.'s intriguing hypothesis. vATT interposition between RR ICN and RC ICN, in this context, suggests that vATT may play a role in switching between speech and the search for semantics.

Nondomain-specific pre-SMA function: volitional selection

We previously mentioned that the SMC is involved in a nondomain function represented by volitional selection and the existence of a gradient of connectivity, with the rostral SMC connected to prefrontal areas involved in cognitive functions and the caudal SMC connected to primary motor areas.

A recent effective connectivity study showed that the pre-SMA is involved in generating and sequencing speech-related motor planning (Hartwigsen et al., 2013). The authors of that study found a driving input to pre-SMA and an increase in facilitatory effective connectivity to the left premotor dorsal (PMd) when sequencing complex pseudo-words. The coordinates for the pre-SMA in that study are close to those of our RC ICN. Further, the coordinates given for their PMd and PMv correspond, respectively, to those we found for the PMd/primary motor cortex below the notch and the rolandic operculum, presenting thus a pattern of connectivity with the phonological motor area similar to our RC ICN.

Due to their different functions it is important to distinguish the pre-PMd from the PMd and the primary motor areas.

The precise role of pre-PMd has been highlighted by an elegant study by Abe et al. (2007) consisting in a “delayed-encoding recognition task” involving “chunks” in long-term memory.

Abe et al. show that in the left hemisphere the IFG processes re-encoding chunks from long-term memory, as the DLPFC processes segmenting in elementary structures, which are recombined and bound by the pre-PMd. This study and others [for a review see Abe and Takashi (2009)] show that, similar to the SMC, the PMd complex is involved in a nondomain-specific process along the same rostro-caudal gradient. Nachev et al. (2008) proposed that SMC may constitute a contingent space for competing condition-action associations resulting in volitional choice.

It is noteworthy that our LR pre-SMA presents a major connectivity pattern with the IFG, and a remarkable continuum of connectivity to the pre-PMd extending to the DLPFC. Thus, the LR ICN along with the FAT and associated U-shaped tracts connecting the SFG (pre-PMd) to the MFG (DLPFC) and the MFG to the IFG (Catani et al., 2012) may contribute to the multiple processes described by Abe et al. and ultimately to the construction of new bounded structures in the lexico-syntactic domain within the pre-PMd. In contrast, the RC ICN appears to include areas that are most likely involved in phonological working memory. According to Nachev, we speculate that the convergence of the lexico-syntactic items and phonological codes within the pre-SMA space may permit the resolution of competing conditions (encoded in pre-PMd) and actions (encoded in the motor cortex) for speech production.

Limitations of our study

Our study does have some limits. First, the functions attributed to the networks we isolated in our connectivity analysis are not based on a formal demonstration, for example, by a specific activation paradigm. Thus, our interpretation is indirect and largely founded on the identification of associations between specific areas, each known to be involved in a particular function. However, we did deliberately perform a resting state functional analysis. As previously mentioned, the language network was isolated by functional connectivity analysis in the absence of any specific task. This method has been validated against the gold standard method, which comprises direct cortical stimulation in awake patients during neurosurgical procedures (Mitchell and et al., 2013). Our choice of methodology was based on the well-known fact that IC analysis captures intranetwork connectivity. However, seed-to-voxel analysis and generalized linear model (GLM) analysis of task-induced activity capture intra- and internetwork connectivity (Suresh et al., 2011). Our methodology allowed us to avoid the inherent circularity of successive GLM/seed-to-voxel connectivity analyses and thus to understand how the language network is connected to the ICNs most likely involved in lexico-syntactic, semantic, and phonological aspects of language.

Second, we found several anticorrelations between the ICNs we identified. It has been argued that methods such as regressing out a global signal in order to eliminate biases due to non-neural sources may lead to spurious negative correlations (Murphy et al., 2009). However, the “conn” toolbox we used did not regress global signal but instead applied a component-based noise correction method (Behzadi et al., 2007). Accordingly, we likely prevented spurious anticorrelation (Chai et al., 2012).

Conclusions

We conclude that functional connectivity analysis of the three pre-SMA subregions found in the language network reveals a particular pattern, with a left ICN anticorrelated to the dATT, a right ICN anticorrelated to the DMN, and a third right network overlapping the two others. The bilateral connectivity of the RR pre-SMA, along with this overlapping network in the left hemisphere, may allow right hemispheric recruitment to process semantic ambiguities. It may also provide a functional architecture for fine-coarse coding and the actuation of functions corresponding to those encompassing the graded salience theory. The particular overlap between the right rostral and caudal pre-SMA ICNs in the vATT may contribute to the fluid reorientation of attention from internal thoughts to the external environment, and also to the switching between speech and search for semantics when one is engaged in linguistic social interaction. Finally, the functional architecture of distinct ICNs connected with areas involved in lexico-syntactic selection on one hand, and the verbal working memory and phonology on the other hand, both converging in pre-SMA, supports the view that the SMC may constitute the space of resolution of competing condition-action associations.

Footnotes

Author Disclosure Statement

The authors report no competing financial interests exist.

References

Abe

, Takashi

. 2009. Functional coupling underlying motor and cognitive functions of the dorsal premotor cortex. Behav Brain Res, 198:13–23.

Abe

, Hanakawa

, Takayama

, Kuroki

, Ogawa

, Fukuyama

. 2007. Functional coupling of human prefrontal and premotor areas during cognitive manipulation. J Neurosci, 27:3429–3438.

Abrahams

, Goldstein

, Simmons

, Brammers

, Williams

SCR

, Giampetro

, et al. 2003. Functional magnetic resonance imaging of verbal fluency and confrontation naming using compressed image acquisition to permit overt responses. Hum Brain Mapp, 20:29–40.

Alario

, Chainay

, Lehericy

, Cohen

. 2006. The role of the supplementary motor area (SMA) in word production. Brain Res, 1076:129–143.

Behzadi

, Restom

, Liau

, Liu

. 2007. A component based noise correction method (CompCor) for BOLD and perfusion based fMRI. Neuroimage, 37:90–101.

Ben-Artzi

, Faust

, Moeller

. 2009. Hemispheric asymmetries in discourse processing: evidences from false memory for lists and texts. Neuropsychologia, 47:430–438.

Binder

, Desai

, Graves

, Conant

. 2009. Where is the semantic system? A critical review and meta-analysis of 120 functional neuroimaging studies. Cereb Cortex, 19:2767–2796.

Binder

, Gross

, Allendorfer

, Bonilha

, Chapin

, Edwards

, Grabowski

, Langfitt

, Loring

, Lowe

, Koenig

, Morgan

, Ojemann

, Rorden

, Szaflarski

, Tivarus

, Weaver

. 2011. Mapping anterior temporal lobe language areas with fMRI: a multicenter normative study. Neuroimage, 54:1465–1475.

Biswal

, Yetkin

, Haughton

, Hyde

. 1995. Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magn Reson Med, 34:537–541.

10.

Bohland

, Guenter

. 2006. An fMRI investigation of syllable sequence production. Neuroimage, 32:821–841.

11.

Botvinick

, Nystrom

, Fissel

, Carter

, Cohen

. 1999. Conflict monitoring versus selection-for-action in anterior cingulate cortex. Nature, 402:179–181.

12.

Bozic

, Tyler

, Ives

, Randall

, Marslen-Wilson

. 2010. Bihemispheric foundations for human speech comprehension. Proc Natl Acad Sci U S A, 107:17439–17444.

13.

Brown

, Martinez

, Parsons

. 2006. Music and language side by side in the brain: a PET study of the generation of melodies and sentences. Eur J Neurosci, 23:2791–2803.

14.

Buckner

, Andrews-Hanna

, Schacter

. 2008. The brain's default network: anatomy, function, and relevance to disease. Ann N Y Acad Sci, 1124:1–38.

15.

Cabeza

. 2008. Role of parietal regions inepisodic memory retrieval: The dual attentional processes hypothesis. Neuropsychologia, 46:1813–1827.

16.

Catani

, Dell'acqua

, Vergani

, Malik

, Hodge

, Roy

, et al. 2012. Short frontal lobe connections of the human brain. Cortex, 48:273–291.

17.

Catani

, Mesulam

, Jakobsen

, Malik

, Martersteck

, Wieneke

, et al. 2013. A novel frontal pathway underly verbal fluency in primary progressive aphasia. Brain, 136:2619–2628.

18.

Chai

, Castanon

, Ongur

, Whitfield-Gabrieli

. 2012. Anticorrelations in resting state networks without global signal regression. Neuroimage, 59:1420–1428.

19.

Corbetta

, Patel

, Shulman

. 2008. The reorienting system of the human brain: from environment to theory of mind. Neuron, 58:306–324.

20.

Corbetta

, Shulman

. 2002. Control of goal-directed and stimulus-driven attention in the brain. Nat Rev Neurosci, 3:201–215.

21.

Cordes

, Haughton

, Arfanakis

, Wendt

, Turski

, Moritz

, et al. 2000. Mapping functionally related regions of brain with functional connectivity MR imaging. AJNR, 21:1636–1644.

22.

Cristescu

, Devlin

, Nobre

. 2006. Orienting attention to semantic categories. Neuroimage, 33:1178–1187.

23.

Crosson

, Benefield

, Cato

, Sadek

, Moore

, Wierenga

, Gopinath

, Soltysik

, Bauer

, Auerbach

, Gokcay

, Leonard

, Briggs

. 2003. Left and right basal ganglia and frontal activity during language generation: contributions to lexical, semantic, and phonological processes. J Int Neuropsychol Soc, 9:1061–1077.

24.

Crosson

, Sadek

, Maron

, Gökçay

, Mohr

, Auerbach

, et al. 2001. Relative shift in activity from medial to lateral frontal cortex during internally versus externally guided word generation. J Cogn Neurosci, 13:272–283

25.

Deschamps

, Baum

, Graco

. 2014. On the role of supramarginal gyrus in phonological processing and verbal working memory: evidence from rTMS studies. Neuropsychologia, 53:39–46.

26.

Devlin

, Jamison

, Matthews

, Gonnerman

. 2004. Morphology and the internal structure of words. Proc Natl Acad Sci U S A, 101:14984–14988.

27.

Diaz

, Barett

, Hogstrom

. 2011. The influence of sentence novelty and figurativeness on brain activity. Neuropsychologia, 49:320–330.

28.

Dronkers

, Wilkins

, Van

Jr. , Redfern

, Jaeger

. 2004. Lesion analysis of the brain areas involved in language comprehension. Cognition, 92:145–177.

29.

Duyn

. 2011. Spontaneous fMRI activity during resting wakefulness and sleep. Prog Brain Res, 193:295–305.

30.

Fan

, Kolster

, Ghajar

, Suh

, Knight

, Sarkar

, et al. 2007. Response anticipation and response conflict: an event-related potential and functional magnetic resonance imaging study. J Neurosci, 27:2272–2282.

31.

Faust

, Ben-Hartzi

, Harel

. 2008. Hemispheric asymmetries in semantic processing: evidence from false memories for ambiguous words. Brain Lang, 105:220–228.

32.

Faust

, Lavidor

. 2003. Semantically convergent and semantically divergent priming in the cerebral hemispheres: lexical decision and semantic judgment. Brain Res Cogn Brain Res, 17:585–597.

33.

Ferstl

, Neumann

, Bogler

, von Cramon

. 2008. The extended language network: a meta-analysis of neuroimaging studies on text comprehension. Hum Brain Mapp, 29:581–593.

34.

Fiebach

, Friederici

, Müller

, Von Cramon

. 2002. fMRI evidences of a dual route to the mental lexicon in visual word recognition. J Cogn Neurosci, 14:11–23.

35.

Fox

, Corbetta

, Snyder

, Vincent

, Raichle

. 2006. Spontaneous neuronal activity distinguishes human dorsal and ventral attention systems. Proc Natl Acad Sci U S A, 103:10046–10051.

36.

Fox

, Snyder

, Vincent

, Corbetta

, Van

, Raichle

. 2005. The human brain is intrinsically organized into dynamic, anticorrelated functional networks. Proc Natl Acad Sci U S A, 102:9673–9678.

37.

Gandour

, Wong

, Dzemidzic

, Lowe

, Tong

, Li

. 2003. A cross-linguistic fMRI study of perception of intonation and emotion in Chinese. Hum Brain Mapp, 18:149–157.

38.

Ghosh

, Tourville

, Guenther

. 2008. A neuroimaging study of premotor lateralization and cerebellar involvement in the production of phonemes and syllables. J Speech Lang Hear Res, 51:1183–1202.

39.

Hampson

, Peterson

, Skudlarski

, Gatenby

, Gore

. 2002. Detection of functional connectivity using temporal correlations in MR images. Hum Brain Mapp, 15:247–262.

40.

Hartwigsen

, Saur

, Price

, Baumgaertner

, Ulmer

, Siebner

. 2013. Increased facilitatory connectivity from the pre-SMA to the left dorsal premotor cortex during pseudoword repetition. J Cogn Neurosci, 25:580–594.

41.

Higo

, Mars

, Boorman

, Buch

, Rushworth

. 2011. Distributed and causal influence of frontal operculum in task control. Proc Natl Acad Sci U S A, 108:4230–4235.

42.

Jansen

, Floel

, Van

, Konrad

, Rotte

, Forster

, Deppe

, Knecht

. 2005. Crossed cerebro-cerebellar language dominance. Hum Brain Mapp, 24:165–172.

43.

Jung-Beeman

. 2005. Bilateral brain processes for comprehending natural language. Trends Cogn Sci, 9:512–518.

44.

Kim

. 2010. Dissociating the roles of the default-mode, dorsal, and ventral networks in episodic memory retrieval. Neuroimage, 50:1648–1657.

45.

Kim

, Lee

, Jo

, Kim

, Lee

, Kim

, et al. 2010. Defining functional SMA and pre-SMA subregions in human MFC using resting-state fMRI: functional connectivity based parcellation method. Neuroimage, 49:2375–2386.

46.

Kim

, Kim

, Lee

, Han

, Kim

, et al. 2013. Risk factor analysis of the development of new neurological deficits following supplementary motor area resection. J Neurosurg, 119:7–14.

47.

Krainik

, Duffau

, Capelle

, Cornu

, Boch

, Mangin

, et al. 2004. Role of the healthy hemisphere in recovery after resection of the supplementary motor area. Neurology, 62:1323–1332.

48.

Krainik

, Lehéricy

, Duffau

, Capelle

, Chainay

, Cornu

, et al. 2003. Postoperative speech disorder after medial frontal surgery: role of the supplementary motor area. Neurology, 60:587–594.

49.

Krainik

, Lehéricy

, Duffau

, Vlaicu

, Poupon

, Capelle

, et al. 2001. Role of the supplementary motor area in motor deficit following medial frontal lobe surgery. Neurology, 57:871–888.

50.

Kuperberg

, Sitnikova

, Lakshmanan

. 2008. Neuroanatomical distinctions within the semantic system during sentence comprehension: evidence from functional magnetic resonance imaging. Neuroimage, 40:367–388.

51.

Laplane

, Talairach

, Meininger

, Bancaud

, Orqoqozo

. 1977. Clinical consequences of corticectomies involving the supplementary motor area in man. J Neurol Sci, 34:301–314.

52.

Lee

, Hacker

, Snyder

, Corbetta

, Zhang

, Leuthardt

, Shimony

. 2012. Clustering of resting state networks. PLoS One, 7:e40370.

53.

Levy

, Wagner

. 2011. Cognitive control and right ventrolateral prefrontal cortex: reflexive reorienting, motor inhibition, and action updating. Ann N Y Acad Sci, 1224:40–62.

54.

Mahon

, Caramazza

. 2010. Judging semantic similarities: an event-related fMRI study with auditory word stimuli. Neuroscience, 69:279–286.

55.

Mars

, Coles

, Grol

, Holroyd

, Nieuwenhuis

, Hulstijn

, Toni

. 2005. Neural dynamic of error processing in medial prefrontal cortex. Neuroimage, 28:1007–1013.

56.

Mars

, Piekema

, Coles

, Hulstijn

, Toni

. 2007. On the programming and deprogramming of actions. Cereb Cortex, 17:2972–2979.

57.

Marslen-Wilson

, Tyler

. 2007. Morphology, language and the brain: the decompositional substrate for language comprehension. Philos Trans R Soc Lond B Biol Sci, 362:823–836.

58.

Mashal

, Faust

, Hendler

, Jung-Beeman

. 2008. Hemispheric differences in processing the literal interpretation of idioms: converging evidence from behavioral and fMRI studies. Cortex, 44:848–860.

59.

Menon

, Uddin

. 2010. Saliency, switching, attention and control: a network model of insula function. Brain Struct Funct, 214:655–667.

60.

Mitchell

, Hacker

, Breshears

, Szrama

, Sharma

, Bundy

, et al. 2013. A novel data-driven approach to preoperative mapping of functional cortex using resting-state functional magnetic resonance imaging. Neurosurgery, 73:969–982.

61.

Mottaghy

, Pascual-Leone

, Kemma

, Töpper

, Herzog

, Müller-Gartner

, et al. 2003. Modulation of a brain-behavior relationship in verbal working memory by rTMS. Cogn Brain Res, 15:241–249.

62.

Murphy

, Birn

, Handwerker

, Jones

, Bandettini

. 2009. The impact of global signal regression on resting state correlations: are anti-correlated networks introduced?. Neuroimage, 44:893–905.

63.

Nachev

, Kennard

, Husain

. 2008. Functional role of the supplementary and pre-supplementary motor areas. Nature, 9:856–869.

64.

Nachev

, Rees

, Parton

, Kennard

, Husain

. 2005. Volition and conflictin human medial frontal cortex. Curr Biol, 15:122–128.

65.

Papoutsi

, Stamatakis

, Griffiths

, Marslen-Wilson

, Tyler

. 2011. Is left fronto-temporal connectivity essential for syntax? Effective connectivity, tractography and performance in left-hemisphere damaged patients. Neuroimage, 58:656–664.

66.

Picard

, Strick

. 1996. Motor areasof the medial wall: a review of their location and functional activation. Cereb Cortex, 6:342–353.

67.

Picard

, Strick

. 2001. Imaging the premotor areas. Curr Opin Neurobiol, 11:663–672.

68.

Pinel

, Dehaene

, Rivière

, LeBihan

. 2001. Modulation parietal activation by semantic distance in a number comparison task. Neuroimage, 14:1013–1026.

69.

Price

. 2010. The anatomy of language: a review of 100 fMRI studies published in 2009. Ann N Y Acad Sci, 1191:62–88.

70.

Rapp

, Leube

, Erb

, Grodd

, Kircher

. 2004. Neural correlates of metaphor processing. Brain Res Cogn Brain Res, 20:395–402.

71.

Rapp

, Mutschler

, Wild

, Erb

, Lengsfeld

, Saur

, Grodd

. 2010. Neural correlates of irony comprehension: the role of schizotypal personality traits. Brain Lang, 113:1–12.

72.

Rosenberg

, Nossek

, Liebling

, Fried

, Schapira-Lister

, Hendler

, et al. 2010. Prediction of neurological deficits and recovery after surgery in the supplementary motor area: a prospective study in 26 patients. J Neurosurg, 113:1152–1163.

73.

Ross

. 1981. The aprosodias. Funcional-anatomic organization of the affective components of language in the right hemisphere. Arch Neurol, 38:561–569.

74.

Rostomily

, Berger

, Ojeman

, Lettich

. 1991. Postoperative deficit and functional recovery following removal of tumors involving the dominant hemisphere supplementary motor area. J Neurosurg, 75:62–68.

75.

Rota

, Sitaram

, Veit

, Erb

, Weiskopf

, Dogil

, Birbaumer

. 2009. Self-regulation of regional cortical activity using real-time fMRI: the right inferior frontal gyrus and linguistic processing. Hum Brain Mapp, 30:1605–1614.

76.

Rushworth

, Hadland

, Paus

, Sipila

. 2002. Role of the human medial frontal cortex in task switching: a combined fMRI and TMS study. J Neurophysiol, 87:2577–2592.

77.

Russel

, Kelly

. 2003. Incidence and clinical evolution of postoperative deficits after volumetric stereotactic resection of glial neoplasms involving the supplementary motor area. Neurosurgery, 52:506–516.

78.

Saur

, Kreher

, Schnell

, Kümerer

, Kellmeyer

, Vry

, et al. 2008. Ventral and dorsal pathway for language. Proc Natl Acad Sci U S A, 105:18035–18040.

79.

Schmidt

, DeBuse

, Seger

. 2007. Right hemisphere metaphor processing? Characterizing the lateralization of semantic process. Brain Lang, 100:127–141

80.

Schmidt

, Seger

. 2009. Neural correlate of metaphor processing: the role of figurativeness, familiarity and difficulty. Brain Cogn, 71:375–386.

81.

Segaert

, Menenti

, Weber

, Petersson

, Hagoort

. 2012. Shared syntax in language production and language comprehension—an FMRI study. Cereb Cortex, 22:1662–1670.

82.

Smith

, Fox

, Miller

, Glahn

, Fox

, Mackay

, Filippini

, Watkins

, Toro

, Laird

, Beckmann

. 2009. Correspondence of the brain's functional architecture during activation and rest. Proc Natl Acad Sci U S A, 106:13040–13045.

83.

Snijders

, Petersson

, Hagoort

. 2010. Effective connectivity of cortical and subcortical regions during unification of sentences structure. Neuroimage, 52:1633–1644.

84.

Snijders

, Vosse

, Kempen

, Van Berjum

, Petersson

, Hagoorth

. 2009. Retrieval and unification of syntactic structure in sentence comprehension: an fMRI study using word-category ambiguities. Cereb Cortex, 19:1493–1503.

85.

Spreng

, Sepulcre

, Turner

, Stevens

, Schacter

. 2013. Intrinsic architecture underlying the relations among the default, dorsal attention and frontoparietal control networks of the human brain. J Cogn Neurosci, 25:74–86.

86.

Spreng

, Stevens

, Chamberlain

, Gilmore

, Schacter

. 2010. Default network activity, coupled with the frontoparietal control network, supports goal-directed cognition. Neuroimage, 53:303–317.

87.

Sridharan

, Levitin

, Menon

. 2008. A critical role for the right fronto-insular cortex in switching between central-executive and default-mode networks. Proc Natl Acad Sci U S A, 105:12569–12574.

88.

Suresh

, Caffo

, van Zijl

PCM

, Pekar

. 2011. On the relationship between seed-based and ICA-based measures of functional connectivity. Magn Reson Med, 66:644–657.

89.

Swann

, Cai

, Conner

, Pieters

, Claffey

, George

, Aron

, Tandon

. 2012. Roles for the pre-supplementary motor area and the right inferior frontal gyrus in stopping action: electrophysiological responses and functional and structural connectivity. Neuroimage, 59:2860–2870.

90.

Thiel

, Schumacher

, Wienhard

, Gairing

, Kracht

, Wagner

, Haupt

, Heiss

. 2006. Direct demonstration of transcallosal disinhibition in language networks. J Cereb Blood Flow Metab, 26:1122–1127.

91.

Tremblay

, Gracco

. 2006. Contribution of the frontal lobe to externally and internally specified verbal responses: fMRI evidence. Neuroimage, 33:947–957.

92.

Tremblay

, Gracco

. 2009. Contribution of the pre-SMA to the production of words and non-speech oral motor gestures, as revealed by repetitive transcranial magnetic stimulation (rTMS). Brain Res, 1268:112–124.

93.

Tyler

, Wright

, Randall

, Marslen-Wilson

, Stamatakis

. 2010. Reorganization of syntactic processing following left-hemisphere brain damage: does right-hemisphere activity preserve function?. Brain, 133:3396–3408.

94.

Ullsperger

, von Cramon

. 2001. Subprocesses of performance monitoring: a dissociation of error processing and response competition revealed by event-related fMRI and ERPs. Neuroimage, 6:181–190.

95.

Van Oostende

, Van Hecke

, Sunaert

, Nuttin

, Marchal

. 1997. FMRI sudies of the supplementary motor area and the premotor cortex. Neuroimage, 6:181–190.

96.

Vigneau

, Beaucousin

, Herve

, Duffau

, Crivello

, Houde

, et al. 2006. Meta-analyzing left hemisphere language areas: phonology, semantics, and sentence processing. Neuroimage, 30:1414–1432.

97.

Wilson

, Galantucci

, Tartaglia

, Gorno-Tempini

. 2012. The neural basis of syntactic deficits in primary progressive aphasia. Brain Lang, 122:190–198.

98.

Xue

, Aron

, Poldrack

. 2008. Common neural substrates for inhibition of spoken and manual responses. Cereb Cortex, 18:1923–1932.

99.

Zhang

, Ide

, Li

. 2012. Resting state functional connectivity of the medial superior frontal cortex. Cereb Cortex, 22:99–111.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.17 MB

0.11 MB

0.16 MB

0.08 MB

0.07 MB

0.04 MB

0.02 MB

0.03 MB

0.02 MB