Cortical thickness and subcortical gray matter reductions in neuropsychiatric systemic lupus erythematosus

Abstract

Within systemic lupus erythematosus (SLE) patients can be divided into groups with and without central nervous system involvement, the latter being subcategorized as neuropsychiatric systemic lupus erythematosus (NPSLE). While a number of research groups have investigated NPSLE, there remains a lack of consistent application of this diagnostic criteria within neuroimaging studies. Previous neuroimaging research suggests that SLE patients have reduced subcortical and regional gray matter volumes when compared to controls, and that these group differences may be driven by SLE patients with neuropsychiatric symptoms. The current study sought to compare measures of cortical thickness and subcortical structure volume between NPSLE, SLE, and healthy controls. We hypothesized that patients with NPSLE (N = 21) would have thinner cortex and reduced subcortical gray matter volumes when compared to SLE (N = 16) and control subjects (N = 21). All subjects underwent MRI examinations on a 1.5 Tesla Siemens Sonata scanner. Anatomical reconstruction and segmentation were performed using the FreeSurfer image analysis suite. Cortical and subcortical volumes were extracted from FreeSurfer and analyzed for group differences, controlling for age. The NPSLE group exhibited decreased cortical thickness in clusters of the left frontal and parietal lobes as well as in the right parietal and occipital lobes compared to control subjects. Compared to the SLE group, the NPSLE group exhibited comparable thinning in clusters of the frontal and temporal lobes. Controlling for age, we found that between group effects for subcortical gray matter structures were significant for the thalamus (F = 3.06, p = .04), caudate nucleus (F = 3.19, p = .03), and putamen (F = 4.82, p = .005). These results clarify previous imaging work identifying cortical atrophy in a mixed SLE and NPSLE group, and suggest that neuroanatomical abnormalities are specific to SLE patients diagnosed with neuropsychiatric symptoms. Future work should help elucidate the underlying mechanisms underlying the emerging neurobiological profile seen in NPSLE, as well as clarify the apparent lack of overlap between cortical thinning and functional activation results and other findings pointing to increased functional activation during cognitive tasks.

Figure 1. Group comparison of cortical thickness relationships to age for the entire group of NPSLE, SLE, and control subjects (N = 58).

Images show clusters of lower (blue clusters) cortical thickness values related to age. Clusters are displayed in the range of p≤.01 to p≤.0001 (color scale shows −log (10) p-value). Top left  =  left lateral hemisphere; Bottom left  =  left medial hemisphere; Top right  =  right lateral hemisphere; Bottom right  =  right medial hemisphere. Light blue regions indicate regions where age and cortical thickness were significantly related (p<.05) corrected for multiple comparisons (FDR).

Figure 2. Scatterplot showing individual cortical thickness values (in millimeters) obtained from the left rostrofrontal cortex for the sample.

Red squares  =  NPLSE patients; Blue circles  =  SLE patients (blue circles); Black diamonds  =  control subjects, stratified by age. All three groups show decreasing trends for cortical thickness in all significant regions (light blue in Figure 1).

Figure 3. Group comparison of cortical thickness differences between NPSLE patients (N = 21) and control subjects (N = 21).

Images show clusters of lower (blue clusters) cortical thickness values controlling for age. Clusters are displayed in the range of p≤.01 to p≤.0001 (color scale shows −log (10) p-value). Clusters which survived FDR correction for multiple correction (p≤.05 are encircled). Top left  =  left lateral hemisphere; Bottom left  =  left medial hemisphere; Top right  =  right lateral hemisphere; Bottom right  =  right medial hemisphere.

Figure 4. Group comparison of cortical thickness differences between SLE patients (N = 16) and control subjects (N = 21).

Images show clusters of higher (red clusters) and lower (blue clusters) cortical thickness values controlling for age. Clusters are displayed in the range of p≤.01 to p≤.0001 (color scale shows −log (10) p-value). No clusters survived FDR correction for multiple correction (p≤.05). Top left  =  left lateral hemisphere; Bottom left  =  left medial hemisphere; Top right  =  right lateral hemisphere; Bottom right  =  right medial hemisphere.

Figure 5. Group comparison of cortical thickness differences between SLE patients (N = 16) and NPSLE patients (N = 21).

Images show clusters of lower (blue clusters) cortical thickness values controlling for age. Clusters are displayed in the range of p≤.01 to p≤.0001 (color scale shows −log (10) p-value). Clusters which survived FDR correction for multiple correction (p≤.05 are encircled). Top left  =  left lateral hemisphere; Bottom left  =  left medial hemisphere; Top right  =  right lateral hemisphere; Bottom right  =  right medial hemisphere.