Youths with BP
ADHD had distinct differences in subcortical structures compared to youths with ADHD. For example, youths with ADHD differed significantly from youths with BP
ADHD and youths with BP in the basal ganglia (caudate and putamen) and in the amygdala. There were no differences in subcortical structures between the youths with BP alone and those with BP
ADHD. These data indicate that morphometric subcortical volumes in youths with BP
ADHD are more similar to those with BP and do not share neuroanatomic correlates with the ADHD group.
Compared to HC, youths with BP
ADHD had smaller TCV and a moderately larger nucleus accumbens volume. In addition, youths with BP were also noted to have a trend toward larger accumbens volumes compared to HC. These findings have been reported by this group previously (Ahn et al. 2007
; Frazier et al. 2008
). It is of interest that the study by Ahn and colleagues noted a trend toward a larger right nucleus accumbens in a combined sample of youths with BP (n
46, 76% with co-morbid ADHD). In another study, our group compared subcortical (limbic and basal ganglia) structures between four groups of youths—those with BP with and without psychosis, those with schizophrenia (SCZ) and HC. We found that both BP groups (with and without psychosis) had enlarged nucleus accumbens, and that youths with BP
psychosis did not share any neuroanatomic findings with the SCZ group (Frazier 2008
). Overall, our findings suggest that abnormally large nucleus accumbens is specific to BP and may be associated with increased illness severity (symptoms of ADHD and psychosis).
Inconsistent with the literature, this study did not find differences in regions of the basal ganglia in ADHD as compared to HC after correction for multiple comparisons (Castellanos et al. 1994
; Aylward et al. 1996
; Filipek et al. 1997
; Overmeyer et al. 2001
; Castellanos et al. 2002
; Wellington et al. 2006
; McAlonan et al. 2007
; Wang et al. 2007
). However, the volumetric raw means and the large effect sizes for regions of the basal ganglia including the caudate (d
0.7) and putamen (d
0.7) indicate that these regions are smaller in youths with ADHD compared to HC. Our lack of finding is likely due to the small sample size of the groups. We did find that the caudate was significantly smaller in the ADHD group compared to both BP groups. For the amygdala and putamen, the ADHD group had significantly smaller volumes as compared to both BP groups and the BP
ADHD group, respectively. Youths with BP and BP
ADHD were not found to have any differences in striatal regions in comparison to HC. In summary, our findings suggest regions of the basal ganglia are more likely to be involved in the pathophysiology of ADHD rather than BP.
Our study indicates that there is not an overlap in subcortical abnormalities in youths with BP
ADHD as compared to ADHD youths, which suggests that youths with BP
+ADHD may not have a “true” ADHD co-morbidity. However, there is emerging evidence that adults with BP
ADHD may share similar neuroanatomical correlates with those with ADHD in cortical structures. In fact, in a similar study in adults with BP
ADHD, BP alone and ADHD, distinct differences between BP
ADHD and BP were found in cortical but not subcortical structures (Biederman et al. 2007
). Furthermore, adults with BP
ADHD had neuroanatomic similarities to adults with both ADHD and BP alone (Biederman et al. 2007
). In addition, two functional (f)MRI studies comparing youths with BP and BP
ADHD have also found distinct differences in activation between the groups in cortical but not subcortical (including limbic) structures (Adler et al. 2005
; Leibenluft et al. 2007
). Therefore, it is possible that youths with BP
ADHD have shared cortical neuroanatomic correlates of both BP and ADHD. Unfortunately, the fMRI studies did not include an ADHD group, so it is difficult to determine if the differences in cortical activation are due to ADHD co-morbidity or to a BP subtype difference.
Youths with ADHD were found to have a reversal of the normal symmetry of the thalamus as compared to HC as well as both BP groups. Thalamic volume abnormalities have not previously been reported in ADHD. However, thalamic injury in youths with closed head trauma increases the risk of secondary ADHD (Gerring et al. 2000
). The pathophysiologic underpinnings of ADHD are thought to involve fronto-striatial circuits. The role of the thalamus as an interconnecting relay station between frontal and subcortical structures makes it an interesting target for future investigation.
Our findings in this study are consistent with our hypothesis of a reduction in hippocampal volumes and an increase in nucleus accumbens volumes in youths with BP. However, unlike other studies (Blumberg et al. 2003
; DelBello et al. 2004
; Chang et al. 2005
; Chen et al. 2004
; Blumberg et al. 2005
; Dickstein et al. 2005
), we did not find reduced amygdala volumes in either BP group. However, we did find an inverse relationship between amygdala volume and MRS scores, suggesting that children with BP who have more significant symptoms have smaller amygdala volumes, thus implicating this structure in the BP presentation. The lack of amygdala abnormalities in BP has been reported by this group elsewhere (Frazier et al. 2005b
), and our group speculated that this could be due to methodological differences in amygdala measurement, the younger age of our sample (mean age range of prior studies was 13.4–16.3 as compared to our study with mean 10.7), and the variable inclusion rates of co-morbid ADHD (ranged from 10 to 80%) in studies. Interestingly, we did find reduced volumes for the amygdala in youths with ADHD. Therefore, abnormal amygdala findings in the BP literature could be due to the inclusion of youths with co-morbid ADHD. In support of this hypothesis, youths with ADHD have been found to have a smaller basolateral complex of the amygdala compared to HC (Plessen et al. 2006
This study included both male and female subjects, and sex–structure interactions were assessed. Sex was noted to be an important factor for TCV and putamen and a moderate sex-by-diagnosis interaction was found for the hippocampus, which has been reported by this group elsewhere (Frazier et al. 2005b
; Frazier et al. 2008
). The impact of sex on subcortical structures has been reported by others (for review, see Durston et al. 2001
). Normatively, hippocampal and caudate volumes are larger in females and cerebral volumes and amygdala volumes are larger in males (Durston et al. 2001
; Goldstein et al. 2001
). There have been very few neuroimaging studies that have investigated sex differences in cortical and subcortical structures in neuropsychiatric illness (Frazier et al. 2008
); therefore, further research is warranted in this area of investigation.
Nucleus accumbens volume in the BP
ADHD group was the only region found to correlate negatively with number of medications. The effects of psychotropic medications, such as antidepressants, mood stabilizers, and antipsychotics, on brain structures, particularly subcortical brain structures, remain unknown. However, there have been several recent studies, which have attempted to explore the impact of medications on gray matter (GM) and white matter (WM) and on specific regions of the brain. For instance, Castellanos and colleagues found greater WM deficits in unmedicated ADHD youths than medicated youths (Castellanos et al. 2002
). Furthermore, an increase in GM volumes has been reported in individuals taking lithium (Sassi et al. 2002
; Monkul et al. 2007
). An exploratory analysis of youths with BP found that individuals with past lithium or valproate exposure tended to have greater amygdala GM volumes than subjects with BP without exposure (Chang et al. 2005
). Basal ganglia enlargement has been noted with typical but not atypical antipsychotics (Chakos et al. 1995
; Frazier et al. 1996
; Corson et al. 1999
). Overall, the impact of at least some medications on brain structures may be significant and, in part, account for the discrepant findings of neuroanatomic structures in the neuropsychiatric literature. Further investigations of the impact of medications on GM and WM and on specific brain regions in neuropsychiatric disorders are needed.
The findings in this study should be interpreted with caution given its limitations, which include the cross-sectional nature, the use of multiple diagnostic comparisons, and the relatively small sample sizes. Furthermore, many subjects were taking psychotropic medication at the time of the study. Last, we included youths in different mood states at the time of scan, which may be a potential confound.