Abstract

The thalamus is a small brain structure that relays neuronal signals between subcortical and cortical regions. Abnormal thalamocortical connectivity in schizophrenia has been reported in previous studies using blood-oxygenation-level-dependent (BOLD) functional MRI (fMRI) performed at 3T. However, anatomically the thalamus is not a single entity, but is subdivided into multiple distinct nuclei with different connections to various cortical regions. We sought to determine the potential benefit of using the enhanced sensitivity of BOLD fMRI at ultra-high magnetic field (7T) in exploring thalamo-cortical connectivity in schizophrenia based on subregions in the thalamus. Seeds placed in thalamic subregions of 14 patients and 14 matched controls were used to calculate whole-brain functional connectivity. Our results demonstrate impaired thalamic connectivity to the prefrontal cortex and the cerebellum, but enhanced thalamic connectivity to the motor/sensory cortex in schizophrenia. This altered functional connectivity significantly correlated with disease duration in the patients. Remarkably, comparable effect sizes observed in previous 3T studies were detected in the current 7T study with a heterogeneous and much smaller cohort, providing evidence that ultra-high field fMRI may be a powerful tool for measuring functional connectivity abnormalities in schizophrenia. Further investigation with a larger cohort is merited to validate the current findings.

1. Introduction

The thalamus is an important deep gray matter structure that relays information between subcortical and cortical regions. Altered thalamocortical connectivity has been reported in schizophrenia (SCZ) (Anticevic et al., 2014;Anticevic et al., 2015;Cetin et al., 2014;Klingner et al., 2014). The thalamus can be subdivided into multiple distinct nuclei with different anatomical connections to various cortical regions (Behrens et al., 2003a;Behrens et al., 2003b). To better assess the relationship between schizophrenia and thalamocortical connectivity, it is therefore critical to investigate the functional connectivity between sub-regions in the thalamus and the cortex, rather than averaging connectivity across the entire thalamus. Indeed, findings using 3.0 Tesla (3T) MRI support a differential connectivity between specific subregions and the cortex (Anticevic et al., 2014;Anticevic et al., 2015). As the sensitivity of blood-oxygenation-level dependent (BOLD) functional MRI (fMRI) is expected to show a supra-linear enhancement with field strength (Uludag et al., 2009), ultra-high field (7.0 Tesla, or 7T, and beyond) BOLD fMRI provides a powerful tool for the investigation of the functional connectivity of these small sub-thalamic regions, potentially using much smaller sample sizes than are needed at 3T. Here, we test this hypothesis by using resting state BOLD fMRI at 7T in a group of schizophrenia patients and matched controls to measure functional connectivity between sub-regions in the thalamus defined by the Oxford thalamic connectivity atlas (Behrens et al., 2003a;Behrens et al., 2003b) and the whole brain on a voxel-by-voxel basis.

2. Methods

3. Results

Demographic information and clinical measures are shown inTable 1. Age and sex were matched between schizophrenia patients and control subjects (P > 0.1). The average numbers of cigarettes smoked per day were comparable between the two groups. Schizophrenia patients had significantly higher BPRS scores (P < 0.001), and slightly lower total MoCA scores (P < 0.05) compared to control subjects. No significant differences were found in motion parameters derived from the SPM realignment routine between the two groups (data not shown).

Tables 2 and3 summarize the main findings in the group comparisons. Hyper-connectivity (Table 2) and hypo-connectivity (Table 3) between thalamic sub-regions and various cortical areas were detected in schizophrenia patients compared to controls, which include both primary targets and other regions (non-primary, shaded inTables 2 and3) of each seed. Hyper-connectivity was observed between all thalamic sub-regions and the sensory/motor cortex and the temporal cortex. Hyper-connectivity to some regions in the parietal cortex was also detected from several thalamic sub-regions. Hypo-connectivity in schizophrenia patients was found between thalamic sub-regions and the frontal cortex, the cingulate cortex, the caudate and putamen, and the cerebellum. Most of these changes were detected in both hemispheres in corresponding regions, although the cluster sizes varied between the left and right hemispheres in some regions.Figure 2 displays the regions with significantly decreased or increased thalamic functional connectivity in schizophrenia patients compared to controls on MNI normalized anatomical images (significant regions from each sub-thalamic seed combined on the same figure).

Significant positive correlation was found between disease duration in schizophrenia patients and functional connectivity between the thalamus and the sensory/motor cortex (Figure 3). Significant negative correlation was observed between disease duration and functional connectivity between the thalamus and the prefrontal cortex (Figure 3). Age, smoking status and medication dosage were included as covariates in the correlation analysis. Note that here all sub-thalamic regions that showed significant difference in functional connectivity with the sensory/motor or the prefrontal cortex in the group comparisons were combined (union) in the correlation analysis, as results from individual seed region did not survive the significance tests, possibly due to insufficient sensitivity. The disease durations of the fourteen patients included in this analysis can be split into two sub-groups with moderate (n = 8) or long (n = 6) disease duration. However, the sample size of each sub-group was too small to detect significant differences between them. Functional connectivity did not significantly correlate with BPRS or MoCA scores (including subscales for both BPRS and MoCA), or antipsychotic medication dosage in schizophrenia patients.

Functional connectivity between the thalamus and the sensory/motor cortex, and functional connectivity between the thalamus and the prefrontal cortex also showed significant negative correlation (Figure 4). Here, all sub-thalamic regions that showed significant difference in functional connectivity with the sensory/motor or the prefrontal cortex in the group comparisons were also combined asFigure 3.

The key findings including group differences and correlations remained significant when all analysis was repeated without global signal regression (GSR). The results without GSR are reported inSupplemental Tables S2 andS3.

4. Discussion

In this study, resting state BOLD fMRI was performed at 7T to examine functional connectivity between sub-thalamic regions and the other brain areas in schizophrenia patients. Our results showed impaired thalamic connectivity to the prefrontal cortex and the cerebellum, but enhanced thalamic connectivity to the motor/sensory cortex in schizophrenia. The negative correlation between the hyper-connectivity in the thalamus-sensory/motor circuit and the hypo-connectivity in the thalamus-prefrontal circuit may imply a compensatory mechanism in the brain, where additional cortical regions are recruited to subserve functions that have been impaired due to the pathology of the disorder (Anticevic et al., 2014;Anticevic et al., 2015). We also found significant correlations between altered functional connectivity and disease duration in the patients. Our main findings (altered thalamic connectivity to the cortex) in the current study are consistent with previous large scale studies conducted at 3T and lower fields(Anticevic et al., 2014;Anticevic et al., 2015;Cetin et al., 2014;Klingner et al., 2014). In the present study, we were able to detect comparable effect sizes with a much smaller cohort (n = 14, as compared to n = 90 in (Anticevic et al., 2014) and n = 222 in (Anticevic et al., 2015)). We attribute this primarily to the enhanced sensitivity of BOLD fMRI at higher magnetic field.

We used the Oxford thalamic connectivity atlas(Behrens et al., 2003a;Behrens et al., 2003b) to define the seed regions within the thalamus for the functional connectivity analysis in this study. Each sub-region (seed) in this atlas has a primary anatomical connection based on a probability threshold of 25% calculated from diffusion MRI measures. Therefore, many voxels within one sub-region can still have anatomical connections to brain regions other than their assigned primary target (Behrens et al., 2003a;Johansen-Berg et al., 2005). This seems to be the case for functional connectivity as well. In our data, each thalamic sub-region also had functional connections to multiple brain areas other than its primary target. Similar results have been reported in previous 3T studies (Anticevic et al., 2014;Anticevic et al., 2015). Alternatively, data driven approaches can be carried out to segment the thalamus based on the functional connectivity to various brain regions (Anticevic et al., 2014;Anticevic et al., 2015). As such procedures typically require much larger sample size, they will be incorporated in our subsequent studies.

The global signal regression (GSR) step in functional connectivity analysis may potentially confound group comparison results(Murphy et al., 2009), as the magnitudes of the global brain signal may vary under different physiological and pathological conditions(Yang et al., 2014). We therefore performed the same analysis with and without the GSR step in preprocessing. Similar to reports in the literature (Murphy et al., 2009), the results without GSR seemed to show somewhat less anti-correlations compared to results with GSR. However, our main findings including the group level differences in functional connectivity and correlations were consistent between the two methods.

Regional brain atrophy, which has been documented in schizophrenia (Kreczmanski et al., 2005;Loeber et al., 1999;Meda et al., 2008;Ota et al., 2014;Pearlson et al., 1997;Walther et al., 2011), may be another potential confounding effect for the functional connectivity measures. To rule out this confounding factor, we included gray matter volume derived from high resolution anatomical images as a covariate in group comparisons and correlation analyses in this study.

Tobacco use is always more prevalent in schizophrenia patients than the general population, and chronic smoking may significantly alter brain functions. Therefore, most control subjects recruited for this study were regular smokers, in order to group match the schizophrenia patients. The current smoking status of each study participant was recorded in cigarettes smoked per day, and the average numbers were comparable between controls and patients. Moreover, smoking status was included as a covariate for all statistical analyses here. Full exploration of the effects of smoking on functional connectivity will require a more detailed investigation, including more refined measures of cigarette exposure and a sample size sufficient to search for a dose-response relationship between smoking and measures of connectivity. Note that the functional connectivity data in controls in the current study may also be different from normal subjects in other studies due to tobacco use.

Finally, an important caveat is that the patients, not controls, in this study were all receiving antipsychotic medicines. Several studies have shown that antipsychotic drugs and their therapeutic effect can restore and increase functional connectivity in the ventral tegmental area (midbrain) (Hadley et al., 2014), striatum (Sarpal et al., 2015), dorsal attention network (Kraguljac et al., 2016a), and hippocampus (Kraguljac et al., 2016b) in schizophrenia patients. A further complication is that different pharmacological classes of antipsychotics may have different effects on brain function (Goozee et al., 2014). In our data, no significant correlation was found between functional connectivity and medication dosage in patients; and medication dosage was included as a covariate when assessing correlations between functional connectivity and disease duration, BPRS and MOCA scores (including subscales). Besides, decreased functional connectivity was found between the thalamus and the prefrontal cortex, and the cerebellum, whereas the studies mentioned above showed increased functional connectivity after antipsychotic drug administration. Future study is warranted to determine if abnormalities in functional connectivity are also presenting in unmedicated patients, and if antipsychotics differentially affect connectivity of specific brain regions.

The small sample size is a fundamental limitation of the current study. The heterogeneity of the study population (the patients can be split into two sub-groups with moderate (n = 8) or long (n = 6) disease duration) and differential effects of acute and chronic substance abuse other than tobacco may also confound the results of group comparisons and correlation analysis. Future studies with larger samples will be necessary to validate our findings, analyze additional clinical variables, and explore the relationship of thalamic anatomy and connectivity in further detail.

5. Conclusion

The enhanced sensitivity of BOLD fMRI at ultra-high field (7T) allowed the detection of altered functional connectivity between sub-regions in the thalamus and cortex in schizophrenia patients from a cohort much smaller compared to previous 3T studies. Our results therefore support the further development of ultra-high field functional connectivity measures as biomarkers for therapeutic trials in schizophrenia, which could greatly benefit from the capacity to detect functional changes with fewer subjects. In addition, our data support further investigation into the potential value of using functional connectivity, as determined by ultra-high field methods, to track disease progression and evaluate therapeutic interventions at the level of individual patients.

Supplementary Material

Acknowledgments

The authors thank Mr. Joseph S. Gillen, Mrs. Terri Lee Brawner, Ms. Kathleen A. Kahl, and Ms. Ivana Kusevic for experimental assistance. This project was supported by a generous donation from Mr. Jose Brito, the ABCD Charitable Trust, and by the National Center for Research Resources and the National Institute of Biomedical Imaging and Bioengineering of the National Institutes of Health through resource grant P41 EB015909, and by the National Institute of Mental Health of the National Institutes of Health through grant R21 MH107016.

Role of funding source

This work was supported by NIH grants R21 MH107016 and P41 EB015909, the ABCD Charitable Trust, and a generous donation from Mr. Jose Brito.

Study funding: This study was supported by the resource grant NCRR NIBIB P41 EB015909, the grant NIMH R21 MH107016, the ABCD Charitable Trust, and a generous donation from Mr. Jose Brito.

Abbreviations:

3T

3.0 Tesla

7T

7.0 Tesla

SCZ

Schizophrenia

BOLD

Blood-oxygenation-level dependent

fMRI

Functional MRI

BPRS

Brief Psychiatric Rating Scale

MoCA

Montreal Cognitive Assessment

GRE

Gradient-echo

EPI

Echo-planar imaging

SPM

Statistical Parametric Mapping

GSR

Global signal regression

Footnotes

References

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.