Dissertation committee:
Παπακωνσταντίνου Ολυμπία, Καθηγήτρια, Ιατρική Σχολή, ΕΚΠΑ
Ευσταθόπουλος Ευστάθιος, Καθηγητής, Ιατρική Σχολή, ΕΚΠΑ
Φερεντίνος Παναγιώτης, Αναπληρωτής Καθηγητής, Ιατρική Σχολή, ΕΚΠΑ
Κελέκης Νικόλαος, Καθηγητής, Ιατρική Σχολή, ΕΚΠΑ
Σμυρνής Νικόλαος, Καθηγητής, Ιατρική Σχολή, ΕΚΠΑ
Σεϊμένης Ιωάννης, Καθηγητής, Ιατρική Σχολή, ΕΚΠΑ
Καραβασίλης Ευστράτιος, Επίκουρος Καθηγητής, Τμήμα Ιατρικής, ΔΠΘ
Summary:
Bipolar disorder (BD) belongs to the broader category of affective disorders, is characterised
by a dysfunction in the ability to regulate emotion and affects a significant proportion of the
population. Accurate and timely recognition of symptoms and diagnosis of BD is at the heart
of research efforts in BD, and the contribution of magnetic resonance imaging may be
important in this direction. Magnetic resonance neuroimaging has highlighted brain changes
at both the gray matter and white matter levels in regions involved in the etiopathogenesis of
BD. Among these regions, the cerebellum has a critical role in mood regulation. However,
although cerebellar gray matter alterations have been previously reported in BD, cerebrocerebellar
white matter connections remain poorly studied. On the other hand, magnetic
resonance spectroscopy, which offers the possibility to measure in vivo the chemical profile
of the brain and to understand the biochemical pathophysiology of BD, does not yet provide
consistently repeatable findings in areas involved in the pathophysiology of the disease,
possibly due to different methodological approaches and heterogeneity in the clinical
characteristics of patients. One of these characteristics relates to predominant polarity (PP)
and onset polarity (OP), which have important clinical and therapeutic implications in BD.
PP is defined as polarity that occurs during at least two-thirds of lifetime mood episodes and
distinguishes patients with BD into patients who experience predominantly depressive
episodes (PP-D), patients who experience predominantly manic or hypomanic episodes (PPM),
and patients who do not meet any of the aforementioned criteria (PP-U). On the other
hand, EP is defined as the polarity of the first episode in BD (depressive or manic). The
neurobiological bases of PP and OP from a neuroimaging perspective remain largely
unknown.
Purpose:
The main aim of the present study is to investigate the neuroanatomical profile of polarity
subphenotypes (PP and OP) in euthymic BD patients using a whole-brain multiparametric
neuroimaging protocol to evaluate brain gray matter (volumetry and cortical thickness), white
matter integrity of the major projection, commissural and association pathways of the brain
and cerebro-cerebellar connectivity (tractography) , as well as major metabolites in the
cingulate gyrus and hippocampus associated with the etiopathogenesis of BD (spectroscopy).
Materials and Methods:
42 euthymic patients with BD (BD-I=30, BD-II=12) were enrolled in the study, who were
classified into sub-groups according to their PP (PP-D, PP-M, PP-U) and OP (OP-D, OP-M)
(OP-Μ Ν=17, OP-D Ν=25, PP-Μ Ν=12, PP-D Ν=14, PP-U Ν=16). A total of 42 healthy
controls with similar demographic characteristics were also included. A whole-brain
multiparametric neuroimaging protocol was applied with sequences specific for further
investigation of gray matter volume and cortical thickness (anatomical high-resolution 3D T1
anatomical sequence), white matter integrity (30-directional diffusion tensor imaging
sequence) and the metabolite profile of the cingulate gyrus and hippocampus (single-voxel
1H spectral imaging sequence). For all imaging data, quality control procedures were
followed before further processing with specific software and extraction of individual
quantitative indices, separately for the right and left cerebral hemispheres. The final data
together with the demographic data of the participants and the available clinical data of the
patients were entered into a database for further statistical processing using univariate and
multivariate statistical models and correction for multiple comparisons.
Results:
Study 1: BD patients exhibited higher fractional anisotropy (FA) in fronto-ponto-cerebellar
tracts bilaterally compared to HC. Subphenotype-specific FA profiles were identified within
the BD cohort. Regarding PP subgroups, we found FA changes in a) left contralateral frontoponto-
cerebellar tract (PP-D > HC) and b) contralateral/ipsilateral fronto-ponto-cerebellar
tracts bilaterally (PP-M > HC). Regarding OP subgroups, we observed FA changes in a)
left/right contralateral fronto-ponto-cerebellar tracts (OP-D > HC) and b) all fronto-pontocerebellar,
most parieto-ponto-cerebellar and right contralateral occipito-ponto-cerebellar
tracts (OP-M > HC). In general, greater and more widespread cerebro-cerebellar changes
were observed in OP-M patients than in OP-D patients compared to HC. OP-M showed
higher FA compared to OP-D patients in several afferent WM tracts. No GM differences
were identified between BD and HC and across BD subgroups.
Study 2: Phenotype-associated cortical thickness abnormalities and volumetric alterations
were identified, but no WM changes ascertained. Specifically, we found a main effect of OP
on GM volume of left middle frontal gyrus and of OP and PP (either or both) on cortical
thickness of various regions previously implicated in BD, i.e. inferior frontal gyrus-pars
opercularis (left) and pars orbitalis (bilateral), left lateral orbitofrontal gyrus, bilateral medial
segment of the superior frontal gyrus, left planum polare, right anterior cingulate gyrus, left
anterior and posterior insula, bilateral frontal operculum (both OP and PP); left anterior and
posterior orbitofrontal gyrus, left transverse temporal gyrus, right posterior insula (only OP);
and right medial frontal cortex (only PP). Based on the magnitude of differences on pairwise
comparisons, we found a large effect of OP on cortical thickness in a single region (left
anterior orbitofrontal gyrus) (OP-M > OP-D), while PP subgroups showed large or medium
effect size differences in cortical thickness (PP-M > PP-D) in a wider array of regions (right
medial frontal cortex, left frontal operculum, left inferior frontal gyrus-pars opercularis,
bilateral medial segment of the superior frontal gyrus). For most regions, PP-D patients
showed the greatest decreases in cortical thickness compared to HC while PP-M showed the
smallest, with PP-U showing an “unspecified” pattern mostly lying in-between PP-D and PPM
Study 3 (metabolite profiles of the cingulate gyrus and hippocampus). A significant effect of
PP on the metabolite profile was observed, with significant differences in the tNAA/tCho
ratio in the anterior cingulate gyrus (PP-M < PP-U). In addition, a large magnitude of
difference was observed between healthy and patients with PP-M (HC > PP-M) and between
patients with PP-D and PP-U (PP-D < PP-U) and a medium magnitude of difference between
healthy and patients with PP-D (HC > PP-D). A significant effect of OP on the metabolite
profile was observed, with significant differences in the anterior cingulate gyrus (Ins/tNAA,
Ins/tCr, Ins/tCho), as well as in the right hippocampus (tCho/tCr). Individual pairwise
comparisons revealed significant differences in ratios between patients with OP-M and OPD:
Ins/tNAA (OP-M > OP-D), Ins/tCr (OP-M > OP-D), Ins/tCho (OP-M > OP-D) and also
found a significant difference in Ins/tCho of the anterior cingulate gyrus between OP-D and
HC (OP-D < HC). Finally, a significant difference was observed in the tCho/tCr ratio of the
right hippocampus (OP-M > HC). All comparisons with a statistically significant difference
after correction for multiple comparisons were followed by a large magnitude of difference
(OP-M > OP-D: Ins/tNAA, Ins/tCr, Ins/tCho and HC > OP-D: Ins/tCho).
Conclusions:
This is the first study to date to investigate differences in gray matter volume and whole-brain cortical thickness, integrity of cerebro-cerebellar and cortical white matter connections, and
metabolite profiles, taking into account subphenotypes of polarity (PP and OP) in both type I
and type II BD patients. We identified extensive changes in cortical thickness and faint
volumetric changes associated with differences between one or more polarity subgroups and
the healthy group, whereas we did not detect changes in major projection, commissural and
association pathways within the cerebral hemispheres. Through a more thorough study,
however, of the cerebellum to the cerebral hemispheres white matter connection (cerebrocerebellar
pathways), we highlighted a distinct profile of change in cerebro-cerebellar
integrity comparing polarity subphenotypes and the healthy group. Finally, we demonstrated
significant differences in the metabolic profile of mainly the anterior cingulate gyrus and
right hippocampus, especially in direct comparisons between polarity subphenotypes. The
findings of the present thesis on the one hand provide strong evidence that both PP and OP
can be considered as neurobiological determinants in AD, and on the other hand highlight the
crucial contribution of quantitative multiparametric neuroimaging markers in the study of AD
with important not only research but also clinical implications.
