Edited by: Feng Liu, Tianjin Medical University General Hospital, China
Reviewed by: Wi Hoon Jung, Daegu University, South Korea; Yanmei Tie, Harvard Medical School, United States
*Correspondence: Xiang Wang,
This article was submitted to Neuroimaging and Stimulation, a section of the journal Frontiers in Psychiatry
†These authors have contributed equally to this work
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Recent evidence from nonhuman primates studies (
Earlier animal studies have suggested that projections from the cerebral cortex to striatum have a topographic organization, such as rostral areas of cerebral cortex projected to the rostral striatum and caudal areas projected to the caudal striatum (
Gradually, functional subregions had been introduced according to the distribution of corticostriatal inputs. For example, Alexander and colleagues proposed a striatum model composed of five segregated and parallel functional loops. In their model, each definable striatal area receives input from a particular cortical area and sends efferent to specific basal ganglia nuclei that, ultimately, project back to the same part of the cortex by way of the thalamus (
In the last two decades, investigators began to examine functional and structure striatal subregions and corticostriatal circuitry of the human brain
Striatal dysfunction has long been thought to be a fundamental element in schizophrenia in the different hypothesis of the etiology of schizophrenia, no matter its neurochemical dopamine hypothesis (
Meta-analytic connectivity modeling (MACM) is a large-scale, unbiased, task-dependent, and data-driven approach to generate a precise, comprehensive functional connectivity map (
Therefore, this study aims to combine the task-dependent MACM and task-independent RSFC approaches to seven anatomical connectivity-based striatum subregions described by Tziortzi et al. (
Striatal ROIs were defined based on the Oxford-GSK-Imanova connectivity striatal atlases (
Oxford-GSK-Imanova striatal connectivity atlas (adapted from Tziortzi et al. (
MACM identifies brain regions that are coactivated above chance with a particular ROI across a large number of functional neuroimaging experiments (
Forty-nine participants with the FES were recruited from the Department of Psychiatry of the Second Xiangya Hospital of Central South University, Changsha, China. All participants were diagnosed with schizophrenia using the Structural Clinical Interview for
Thirty-one healthy adult volunteers were recruited aged between 16 and 40 years, right-handed, free from clinically significant illness, current or previous history of neurological or psychiatric diagnosis, and alcohol or drug addiction. Individuals with a family history of psychiatric illness among their first-degree relatives were also excluded from the healthy group. In the end, 27 healthy controls were adopted in analysis, and four were eliminated for data quality or head movements.
Demographic and clinical characteristics of participants in the two groups were compared using χ2 tests for categorical variables and independent-samples
Informed consent was obtained from all subjects, and the study obtained the approval of the Institutional Ethical Board of the Second Hospital of Xiangya, Central South University.
Data were acquired on a 3.0-T Intera Achieva X (Phillips, Holland) whole-body MRI system equipped with a 20-channel Head Matrix Coil. To help stabilize head position, each subject was fitted with a thermoplastic mask fastened to holders on the head coil. During functional scans, subjects viewed a black background and were instructed to relax, stay still, stay awake, and keep their eyes open.
Three-dimensional structural MRI images (T1-weighted) were acquired from the sagittal plane using spoiled gradient echo pulse sequence, with scanning parameters: repetition time (TR) = 8.5 ms, time to echo (TE) = 3.743 ms, flip angle = 8°, field of view (FOV) = 256 × 256 mm, matrix = 256 × 256, voxels = 1 × 1 × 1 mm, slices number = 180, slice thickness =1 mm, gaps = 0 mm.
Resting-state functional images were obtained using a blood oxygenation level–dependent (BOLD) contrast-sensitive gradient echo echo-planar sequence, and we acquired 206 images in total, with scanning parameters: TR = 2,000 ms, TE = 30 ms, flip angle = 90°, FOV = 240 × 240 mm, matrix = 64 × 64, voxels = 3.75 × 3.75 × 4 mm, slices number = 36, slice thickness = 4 mm, gap = 0 mm.
Data were processed in SPM8 (University College; London, UK;
For each subject, functional analysis was performed between each striatal ROI and the rest of the brain in a voxel-wise manner. To improve normality, the correlation coefficients in each voxel were transformed to
RSFC analysis of HC. One-sample
Conjunction analysis of MACM and RSFC. To delineate areas showing task-dependent and task-independent connectivity with striatum subregions, we performed MACM-RSFC conjunction analysis with strict minimum statistics (
RSFC analysis between FES and HC. For each striatal subregion, two-sample
A total of 1,023 studies published no later than May 31, 2018, were included in our meta-analysis. These studies corresponded to 1,366 experiments and 2,976 experimental conditions, with a total of 16,369 subjects and 22,026 activation locations. Detailed descriptions of the striatal ROIs can be found in
The number of articles, experiments, conditions, subjects, and locations identified by each striatal subregion in the Brainmap database.
Striatal subregions | Papers | Experiments | Conditions | Subjects | Locations |
---|---|---|---|---|---|
Str_limbic | 250 | 358 | 728 | 4,277 | 4,915 |
Str_executive | 484 | 676 | 1,450 | 7,888 | 11,107 |
Str_rostral-motor | 80 | 93 | 223 | 1,242 | 1,850 |
Str_caudal-motor | 115 | 135 | 311 | 1,543 | 2,335 |
Str_parietal | 74 | 84 | 202 | 1,113 | 1,447 |
Str_occipital | 16 | 16 | 54 | 228 | 236 |
Str_temporal | 4 | 4 | 8 | 78 | 136 |
For simplicity, the abbreviations Str_limbic, Str_executive, Str_rostral-motor, Str_caudal-motor, Str_parietal, Str_occipital, and Str_temporal were adopted wherein Str means striatal.
The MACM results are shown in
Task-dependent meta-analytic connectivity modeling (MACM) results for the seven striatum subregions. The result of temporal subregion was not demonstrated because of no significant connectivity at the current threshold. The first column (left): six striatum subregions (ROIs); the last column (right): connectivity modeling results of six striatum subregions (FDR correction with a statistical threshold of
The Str_rostral-motor subregion (
Activity of the Str_parietal subregion (
Histograms of the BDs and its top 15 subcategories for each striatal ROI in our MACM are presented in
Behavior domains (BDs) and their subcategories associated with each striatal subregion in MACM analysis.
Paradigm classes (PCs) associated with each striatal subregion in MACM analysis.
BDs that were overrepresented among experiments showing regional coactivation with both the Str_rostral-motor and Str_caudal-motor subregions were action (especially for the execution subcategory) and cognition (especially for the language and attention subcategories). PCs related to Str_rostral-motor coactivation were primarily finger tapping, tone monitoring, and flexion/extension; PCs related to Str_caudal-motor coactivation were finger tapping, reward, and pain monitoring.
BDs that were overrepresented among experiments showing regional coactivation with the Str_parietal subregion were cognition and action. PCs that had significant associations with Str_parietal were reward, flexion/extension, finger tapping, pain monitoring, and go/no-go. BDs that were overrepresented among experiments showing regional coactivation with Str_occipital were cognition, emotion, and action. PCs related to Str_occipital were primarily face monitoring and reward.
Results of RSFC of healthy controls were detailed as described in Supplementary Material (Results: 2. Task-independent RSFC analysis). As shown in
Overlapping results of task-dependent MACM and task-independent RSFC for the seven striatal subregions. Due to few/no significant connectivity results of occipital/temporal striatum subregion in MACM analysis at the current threshold, only functional connectivity modeling of five striatal subregions is presented. RSFC results are depicted in yellow in the top rows (FDR correction with a statistical threshold of
The demographic and clinical data from the 45 patients and 27 healthy controls are shown in
Comparing to HC, FES group showed weaker RSFC between Str_limbic and bilateral anterior cingulate gyrus (ACC, BA24/32), mPFC (BA32), left insula, and also right thalamus and left putamen (FDR corrected at the whole brain voxel level, with a significance threshold set at
Demographic and clinical variables.
Schizophrenia (n = 45) | Healthy controls (n = 27) | Statistical test | ||
---|---|---|---|---|
Age (years) | 21.31 ± 5.50 | 22.56 ± 3.25 | 1.207 | 0.231 |
Sex (male/female) | 25/20 | 15/12 | 1.000 | 0.596 |
Duration of illness (month) | 10.98 ± 8.09 | |||
PANSS sum-score | 81.70 ± 11.87 | |||
Positive sum-score | 20.95 ± 6.39 | |||
Negative sum-score | 19.89 ± 9.05 | |||
General psychiatric | 40.86 ± 7.91 |
Between-group functional connectivity analysis in HC and FES.
Cluster | Anatomical Region | BA | X | Y | Z | Volume (mm3) | |
---|---|---|---|---|---|---|---|
Str_limbic | |||||||
Control > schizophrenia1 | |||||||
1 | L medial frontal gyrus | −6 | −3 | 51 | 69 | 3.89 | |
R medial frontal gyrus | 32 | 6 | 3 | 48 | 3.67 | ||
L cingulate gyrus | 24 | −6 | 6 | 39 | 2.78 | ||
2 | R insula | 33 | 18 | 6 | 40 | 3.67 | |
3 | R thalamus | 15 | −12 | 6 | 63 | 3.65 | |
4 | L putamen | −30 | 3 | 0 | 77 | 3.42 | |
L insula | −36 | 12 | 3 | 3.21 | |||
5 | R cingulate gyrus | 32 | 3 | 21 | 42 | 1 | 2.93 |
6 | R anterior cingulate | 3 | 6 | −6 | 1 | 2.82 | |
Str_executive | |||||||
Control > schizophrenia2 | |||||||
1 | L precentral | −45 | −6 | 48 | 16 | 4.08 | |
2 | L insula | −30 | 18 | 9 | 20 | 3.97 | |
3 | L cingulum_mid | −9 | 6 | 33 | 5 | 3.95 | |
4 | R supp_motor_area | 6 | 3 | 46 | 4 | 3.52 | |
5 | R thalamus (ventral lateral nucleus) | 15 | −18 | 12 | 8 | 3.49 | |
6 | R insula | 36 | 15 | 6 | 3 | 3.45 |
1Results were reported at a height threshold of p < 0.001(uncorrected) and an extent threshold of p < 0.05 (FDR corrected).
2Results were reported at a height threshold of p < 0.001(uncorrected).
Between-group functional connectivity difference of limbic and executive striatal subregions in HC and FES. Comparing with HC, patients with FES showed reduced functional connectivity between limbic striatum subregion (top region) and thalamus, mPFC, ACC, IFG, and insula (yellow areas with FDR corrected and an extent threshold of
The present study examines the dysfunction connectivity of striatum in FES based on the integrated functional model of the striatum that arises from the functional connections of the seven structural connectivity-based striatum subregions. Our functional connectivity model results for the corticostriatal subcircuits of these seven subregions were highly consistent with the structural connectivity evidence that was used to subdivide the striatum originally, especially for subregions connected with the PFC, such as Str_limbic, Str_executive, Str_rostal-motor, and Str_caudal-motor. We further observed considerable overlap between task-dependent MACM results and task-independent RSFC analysis results. More importantly, abnormal functional connectivity of limbic and executive of striatum subregions was identified in schizophrenia.
The striatal subregions adopted in this study were derived from Tziortzi and colleagues’ (
In our MACM analysis results, limbic subregion showed significant functional connectivity with mPFC/anterior cingulate (BA24/32) and ventrolateral PFC/IFG (BA47). The corresponding BD and PC analysis showed those coactivations were related mainly to the emotion and cognition domain, and the performance of reward task paradigms. And the Str_executive subregion was found to have widespread functional connectivity with the cortex, especially with the frontoparietal cortex. Again, the BD and PC analysis showed paradigms that are dependent upon explicit memory and language. Both motor subregions, Str_rostral-motor and Str_caudal-motor, shared similar functional connectivity. However, Str_rostral-motor coactivated mainly with the precentral gyrus, whereas Str_caudal showed more coactivation with the postcentral gyrus and dorsal cingulate gyrus, which implies that these two subregions’ circuits are responsible for related but distinct functions. More detailed discussions of the functional connectivity pattern for each striatal subregion were presented in the discussion part of
It should be noticed that insula (BA13) was strongly coactivated with nearly all of the striatum subregions except for Str_occipital and Str_temporal. In Tziortzi and colleagues’ connectivity striatal atlases, probabilistic connections of striatum voxels were related with the frontal, parietal, occipital, and temporal lobes, but not with the insular lobe. However, a meta-analysis of caudate and putamen functional connectivity showed patterns that are consistent with these structural projections, particularly between the putamen and dorsal posterior insula and between the caudate and anterior ventral insula (
Another brain region identified by this MACM study was the thalamus, which was coactivated with almost all striatum subregions. Specifically, the Str_limbic and Str_executive subregions showed functional connectivity with the dorsomedial thalamic nucleus, whereas the Str_rostral-motor, Str_caudal-motor, and Str_parietal subregions showed functional connectivity with the ventral posterior lateral thalamic nuclei. Previous studies demonstrated that the dorsomedial nucleus was highly interconnected with the PFC, which suggests that it may be involved in modulating cognitive functions, emotional reaction, and regulation of alertness, whereas the ventrolateral nuclei have been strongly associated with voluntary movement (
Except for consistent structure connectivity and task-related functional connectivity, results of the task-independent RSFC analysis in healthy subjects also overlapped with those of the task-dependent MACM, providing further evidentiary support for the corticostriatal functional connectivity map. However, the RSFC analysis results were more widely distributed than those of MACM, which may arise from the fundamental differences of the two analytical approaches and the small sample size of healthy controls comparing with bigger Brainmap datasets. In summary, our task-dependent and task-independent functional analysis results using ROIs extracted from corticostriatal structural connectivity atlases showed consistent functional and structural connectivity of corticostriatal circuit, providing a good template that can be extended to the clinical population to examine dysfunctional in corticostriatal circuit.
Our study identified abnormal functional connectivity in limbic and executive striatum subregions in FES. Specifically, compared with HC, significantly reduced functional connectivity between the limbic subregion and thalamus, mPFC, ACC, IFG, and insula was identified in the FES group. And FES also showed decreased functional connectivity between executive subregions and thalamus, insula, and SMA.
Limbic striatum subregion, on the one hand, presented reduced functional connectivity with some areas in the default mode network (DMN) in schizophrenia, like mPFC and ACC, which are consistent with previous studies. Previous researches suggested DMN is associated with internal cognition and self-related processing and to be deactivated in goal-oriented tasks, and a failure to suppress DMN always leads to impaired performance in tasks (
On the other hand, limbic striatum subregion also showed decreased functional connectivity with the salient network (SN), for example, medial ACC and insula, which are responsible for receiving sensory information from subcortex area and switching from resting state to goal-oriented task state (
Dysfunctional connectivity of executive striatum subregion was mainly focused on thalamus (ventral lateral nucleus) and SMA. Executive striatum subregion constituted of brain areas that have the highest structure connectivity with BA9, BA9/46, and BA10 of the dorsolateral prefrontal cortex related to central-executive network (CEN). Contrary to DMN, CEN was triggered by external goal-oriented task but silence during rest. The identified RSFC between Str_executive and related motor brain areas, such as ventral lateral nucleus in thalamus involving in the motor pathway, as well as SMA associated with the execution of motor (
There are several potential methodological limitations in this study. It is important to note that the BOLD imaging method used in all adopted MRI articles was not sensitive enough to detect all the connections between striatum and cortical areas as rigid representations of anatomic connectivity. However, as a meta-analysis, our study did report some steady and reliable connections by combing thousands of articles using functional neuroimaging experiments, which give a supplement for the anatomic connectivity. Another major limitation of this MACM study was the database employed. Even though Brainmap is one of the largest databases of published functional and structural neuroimaging experiments with coordinate-based results, it is not exhaustive. Finally, the relatively small sample size compared with bigger datasets from Brainmap, the statistical method, the analysis threshold, and also the correction method used in present study constituted a limitation of picking up sensitive connections in RSFC analysis.
Our MACM and RSFC results obtained from task-dependent and task-independent functional connectivity confirm and extend previous findings indicating that the organization of cortical inputs to the striatum is highly ordered into subdivisions of the striatum, especially that of frontal and parietal lobe afferents. More importantly, using this connectivity pattern, we further confirm the reduced functional connectivity between limbic striatum subregions and thalamus and some brain areas in DMN (e.g., mPFC, ACC) and SN (e.g., insula) and also between executive striatum subregions and thalamus, SMA, and insula in FES, which supports the important role of the corticostriatal-thalamic loop in the pathophysiology of schizophrenia.
The datasets generated for this study are available on request to the corresponding author.
The studies involving human participants were reviewed and approved by the Second Xiangya Hospital of Central South University. The patients/participants provided their written informed consent to participate in this study.
XW, BZ, and PL conceived the presented idea and wrote the manuscript, in consultation with DÖ and XSW. BZ analyzed the data with support from PL and XJ, and PL also contributed to the figure presented in the manuscript. WS and SY helped with revision of the article.
This work was supported by the National Natural Science Foundation of China (XW, grant 31671144) (PL, grant 61473221), Shanghai Municipal Science and Technology Major Project (grant 2018SHZDZX01), and Hunan Provincial Natural Science Foundation of China (XSW, grant 2019JJ40362).
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
MACM, meta-analytic connectivity modeling; DTI, diffusion tensor imaging; ROIs, regions of interests; RSFC, resting-state functional connectivity; FES, first-episode schizophrenia; DLPFC, dorsolateral prefrontal cortex; PET, positron emission tomography; mPFC, medial prefrontal cortex; BDs, behavior domains; PCs, paradigm classes; ALE, activation likelihood estimation; FDR, false discovery rate; FWE, family-wise error rate; EPI, echo-planar imaging; MNI, Montreal Neurological Institute; BOLD, blood oxygenation level dependent; BA, Brodmann area; MFG, middle frontal gyrus; IFG, inferior frontal gyrus; STG, superior temporal gyrus; SMA, supplementary motor area; ACC, anterior cingulate cortex; Str_limbic, Str_executive, Str_rostral-motor, Str_caudal-motor, Str_parietal, Str_occipital and Str_temporal, striatum subregions of limbic, executive, rostral-motor, caudal-motor, parietal, occipital and temporal; DMN, default mode network; SN, salient network; CEN, central-executive network.
The Supplementary Material for this article can be found online at: