Symposia at the 2017 ESCAP Congress (Geneva) on 22q11 Deletion Syndrome, organized by the Swiss National Centre of Competence in Research – Synapsy.
Neuroimaging markers of psychosis in 22q11DS
Chair Marie Schaer - Co-chair Maria Carmela Padula.
Presentations by Maria Padula, Zora Kikinis, Daniela Zoller and Corrado Sandini.
See below for all abstracts
View the other symposia of the Synapsy 22q11 deletion syndrome:
Overview of the behavioral phenotype and available clinical interventions
Original Synapsys Symposium abstract by Maria C. Padula (Developmental Imaging and Psychopathology Laboratory, Department of Psychiatry, University of Geneva School of Medicine, Geneva, Switzerland) et al. on Altered structural network architecture is predictive of the presence of psychotic symptoms in patients with 22q11.2 deletion syndrome (ESCAP 2017 Congress in Geneva, Switzerland).
Maria C. Padula1*, Elisa Scariati1, Marie Schaer1, Corrado Sandini1, Maude Schneider1,2, Dimitri Van De Ville3,4, Stephan Eliez1,5
1. Developmental Imaging and Psychopathology Laboratory, Department of Psychiatry, University of Geneva School of medicine, Geneva, Switzerland.
2. Center for Contextual Psychiatry, Department of Neurosciences, KU Leuven, Belgium.
3. Medical Image Processing Lab, Institute of Bioengineering, Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland.
4. Department of radiology and medical informatics, University of Geneva, Geneva, Switzerland.
5. Department of Genetic Medicine and Development, University of Geneva School of medicine, Geneva, Switzerland.
22q11.2 deletion syndrome (22q11DS) represents a homogeneous model of schizophrenia particularly suitable for the search of neural biomarkers of psychosis. Impairments in structural connectivity related to the presence of psychotic symptoms have been reported in patients with 22q11DS. However, studies investigating connectivity differences in patients with 22q11DS with different symptomatic profiles are still scarce and need further investigation. In this study we investigated differences in structural connectivity in patients with and without mild to attenuated positive symptoms of psychosis using univariate and multivariate approaches.
Twenty-seven patients with 22q11DS with attenuated positive psychotic symptoms (psy+) were selected as having a score higher than 3 in at least one of the positive subscales of the Structured Interview for Prodromal Syndromes. As comparison group, 27 non-symptomatic patients (psy-) were selected and individually matched for age and gender. Different structural connectivity measures were compared between the two groups, including the number of connections between pairs of brain regions, graph theory measures and diffusion measures. Group comparisons were performed using Mann Whitney U-test and a multivariate classification. In particular, a naïve Bayes classifier was trained and tested using leave-one-subject-out cross validation (LOOCV).
The univariate comparison of connectivity measures between psy + and psy- patients did not give significant results. However, the multivariate analysis revealed that altered structural network architecture significantly discriminate patients with and without attenuated positive psychotic symptoms with a max accuracy of 70%. Among the regions contributing to the significant classification there were the superior fontal and the anterior cingulate cortices.
Our results showed that differences in patients with 22q11DS with and without mild to severe positive psychotic symptoms are subtle and emerge only when using a multivariate approach. In particular, measures of impaired network architecture allowed to significantly discriminate the two group of patients, and pointed to dysconnecivity of frontal and limbic regions in patients with higher positive symptoms severity. Therefore, alterations in structural network architecture may represent a potential biomarker for an increased risk of psychosis in patients with 22q11DS.
Original Synapsys Symposium abstract by Daniela Zöller (Medical Image Processing Laboratory, Institute of Bioengineering, École Polytechnique Fédérale de Lausanne (EPFL), Department of Radiology and Medical Informatics, University of Geneva, Developmental Imaging and Psychopathology Laboratory, Office Médico-Pédagogique, Department of Psychiatry, University of Geneva) et al. on Multivariate BOLD signal variability alterations in psychosis in 22q11.2 deletion syndrome (ESCAP 2017 Congress in Geneva, Switzerland).
Daniela Zöller (a,b,c), Marie Schaer (c), Maria Carmela Padula (c), Elisa Scariati (c), Dimitri Van De Ville (a,b), Stephan Eliez (c).
a. Medical Image Processing Laboratory, Institute of Bioengineering, ´Ecole Polytechnique F´ed´erale de Lausanne (EPFL).
b. Department of Radiology and Medical Informatics, University of Geneva.
c. Developmental Imaging and Psychopathology Laboratory, Office Médico-Pédagogique, Department of Psychiatry, University of Geneva.
Although rarely considered in fMRI studies, blood oxygenation level dependent (BOLD) signal variance might present a potential marker for brain dysfunction. Chromosome 22q11.2 deletion syndrome (22q11DS) is a neurodevelopmental disorder coming with a high prevalence of schizophrenia of 30% to 40%. In this study, we investigated multivariate alterations in BOLD signal variance and its agerelationship related to psychotic symptoms.
We included 19 patients with 22q11DS who presented mild to severe positive psychotic symptoms (PS+, mean age = 17.16±3.92), defined by a score of 3 or more in one of the positive subscales of the SIPS. We compared them to 19 patients without positive psychotic symptoms (PS-, mean age = 17.21±4.05), who were individually matched for age and motion. After conventional pre-processing of the resting-state fMRI scans, BOLD signal variance was determined by calculating the temporal standard deviation of voxel-wise time series. We then applied Partial Least Squares correlation (PLSC) to reveal alterations in BOLD signal variance and in its relationship with age. PLSC maximizes the correlation between two data sets in a multivariate way and is, therefore, very well suited to analyze the relationship between high-dimensional brain data and other subject-specific design variables, such as diagnosis and age.
PLSC analysis of PS+ and PS- patients resulted in two significant components. The first component showed brain areas with altered age-relationship of BOLD signal variance in PS+. It revealed a negative correlation between BOLD signal variance and age in the PS+ group in prefrontal regions. This relationship was not evident in the PS- group. Prefrontal areas are known to be structurally a?ected in schizophrenia and 22q11DS. The second component detected brain areas with altered BOLD signal variance in the PS+ group. BOLD signal variance in the PS+ group was lower in the superior frontal cortex and higher in the orbitofrontal and anterior cingulare cortices (ACC). Dysconnectivity of the ACC has been reported in 22q11DS and linked to psychotic symptoms.
This is the first study investigating BOLD signal variance in psychotic patients within a cohort of 22q11DS. Thanks to the multivariate characteristic of our approach we were able to reveal patterns of alterations and age-relationship specific to psychotic patients, which suggest BOLD variability as a potential marker for psychosis.
Original Synapsys Symposium abstract by Zora Kikinis (Harvard University, Office of Technology Development, Cambridge, MA, USA) on Investigation of Heterogeneity in Cortical Microstructure in Individuals with 22q11 Deletion Syndrome: a Diffusion MRI study (ESCAP 2017 Congress in Geneva, Switzerland).
The gray matter of the brain has traditionally been investigated using structural Magnetic Resonance Imaging (sMRI). sMRI provides information about measures such as volume, cortical thickness and cortical areas, but not about the microstructural organization of the cortex. Diffusion MRI (dMRI) allows the detection of changes in the microstructure of tissues and could be used for the investigation of the effects of abnormal neurodevelopment in the cortical gray matter. In this study, we use dMRI to explore changes in microstructure in gray matter in individuals with 22q11Deletion Syndrome (22q11DS). We apply a new measure, the Heterogeneity in Fractional Anisotropy (HFAt), to explore the variability of microstructure within given areas of interest.
We acquired dMRI and sMRI scans from 56 subjects with 22q11DS and 30 healthy controls, mean age 21 years. sMRIs were used to extract and to parcellate gray matter into discrete anatomical regions (FreeSurfer). We then grouped those regions into three functional areas, namely the primary, paralimbic and associative. dMRI data were corrected for cerebrospinal fluid contamination using Free Water Correction first and then we calculated the HFAt in gray matter for each of the three functional brain areas. To investigate the relationship between microstructure and function we used the Stroop Interference test. This subtest is a measure of selective attention and response inhibition mediated by the paralimbic area.
The HFAt was significantly increased in the associative area in the 22q11DS group, t(84)=-3.7, p<0.001, CI. 95 -0.01-0.03, d=0.85, as well as in the paralimbic area, t(84)=-5.5, p<0.001, CI. 95 -0.01-0.05, d=1.2, suggesting increased heterogeneity of gray matter microstructure in these regions. There were no changes in HFAt in the primary area. There was a negative correlation between the performance on the Stroop Interference test and HFAt in the paralimbic area in the 22q11DS group, r=-0.37, n=56, p=0.005. This suggests that greater variability in gray matter is associated with lower performance on the neuropsychological test.
This is the first report of microstructural changes in gray matter in individuals with 22q11DS using dMRI. dMRI is thus a useful tool, and HFAt is a useful measure, for assessing microstructural changes in gray matter of patients with 22q11DS. HFAt is also correlated with the cognitive measure mediated by paralimbic areas of the brain.
Original Synapsys Symposium abstract by Corrado Sandini (Developmental Imaging and Psychopathology Laboratory, Department of Psychiatry, University of Geneva School of Medicine, Geneva, Switzerland) et al. on Alterations of structural covariance networks in 22q11DS in relation to psychotic symptoms, a cross-sectional and longitudinal investigation (ESCAP 2017 Congress in Geneva, Switzerland).
Corrado Sandini1, Daniela Zoler1 , Elisa Scariati1, Maria Carmela Padula1, Maude Schneider1,6, Marie Schaer1,3, Dimitri Van De Ville 4,5, Stephan Eliez1,2
1 Developmental Imaging and Psychopathology Laboratory, University of Geneva School of medicine, Geneva, Switzerland.
2 Department of Genetic Medicine and Development, University of Geneva School of medicine, Geneva, Switzerland.
3 Stanford Cognitive and Systems Neuroscience Laboratory, Stanford University School of medicine, California, USA.
4 Departement of Radiology and Medical Informatics,University of Geneva, Switzerland.
5 Institute of Bioengineering, Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland.
6 Center for Contextual Psychiatry, Research Group Psychiatry, Department of Neuroscience, KU Leuven, Leuven, Belgium.
22q11.2 Deletion Syndrome is a genetic disorder that is considered as a model to study the pathogenesis of psychosis. Structural covariance is a method of exploring the architecture of human brain networks. Alterations of brain network architecture have been previously linked to psychosis.
In the present study we explore alterations of structural covariance networks (SCNs) specifically linked to psychotic symptoms. We also implement a novel longitudinal approach to describe a deviant maturation of SCNs in 22q11DS.
We acquired structural T1 weighted images in the largest neuroimaging cohort of 22q11DS to date: 124 patients (M/F=65/56) each with 1.8 scans (220 scans) and in 121(M/F=59/65) healthy controls (HC) each with 2.1 scans (253 scans). For cross-sectional analysis patients were divided according to the presence of at least moderate psychotic symptoms. SCNs were constructed with a standard protocol using Freesurfer. We employed graph-theory to investigate alterations of network architecture. For longitudinal investigation we implemented a novel sliding-window approach that allowed precise description of developmental trajectories.
Psychotic patients (PPs) selectively presented alterations of SCNs, which were more segregated and less segregated, both compared to HC and non-psychotic patients (NPPs). PPs additionally displayed aberrant connectivity in the anterior cingulate portion of the salience network. Our longitudinal analysis on the entire cohort revealed that the aberrant increase in segregation and reduction of integration coincided with the period of maximal vulnerability to psychosis, during mid to late adolescence. Interestingly this aberrant maturation was preceded by a lack of development during late-childhood, that is the critical period for the development of SCNs in controls.
Our results confirm that alterations of brain network architecture and aberrant connectivity of the salience network are important in the pathogenesis of psychosis.
A novel longitudinal approach allowed to demonstrate deviant development of SCNs in 22q11DS, with an aberrant maturation in during adolescence, preceded by a lack of maturation during late-childhood. It could be hypothesized that the reduced maturation during late-childhood could predispose to the subsequent aberrant maturation during adolescence. If this were confirmed, it would highlight late-childhood as a critical period for early neuro-protective intervention.