Neurobiological mapping of abnormal gray matter developmental trajectories in psychosis
Poster No:
488
Submission Type:
Abstract Submission
Authors:
Natalia García-San-Martín1, Richard Bethlehem2, Agoston Mihalik2, Jakob Seidlitz3, Isaac Sebenius2, Claudio Alemán-Morillo1, Lena Dorfschmidt3, Golia Shafiei4, Víctor Ortiz-García de la Foz5, Kate Merritt6, Anthony David6, Sarah Morgan2, Miguel Ruiz-Veguilla7, Rosa Ayesa-Arriola5, Javier Vázquez-Bourgon5, Aaron Alexander-Bloch4, Bratislav Misic8, Edward Bullmore2, John Suckling2, Benedicto Crespo-Facorro5,7,1, Rafael Romero-García1,2
Institutions:
1University of Seville, Seville, Spain, 2University of Cambridge, Cambridge, United Kingdom, 3The Children’s Hospital of Philadelphia, Philadelphia, PA, 4University of Pennsylvania, Philadelphia, PA, 5Marqués de Valdecilla University Hospital, Cantabria, Spain, 6University College London, London, United Kingdom, 7Virgen del Rocío University Hospital, Seville, Spain, 8McGill University, Montreal, Canada
First Author:
Co-Author(s):
Benedicto Crespo-Facorro
Marqués de Valdecilla University Hospital|Virgen del Rocío University Hospital|University of Seville
Cantabria, Spain|Seville, Spain|Seville, Spain
Marqués de Valdecilla University Hospital|Virgen del Rocío University Hospital|University of Seville
Cantabria, Spain|Seville, Spain|Seville, Spain
Rafael Romero-García
University of Seville|University of Cambridge
Seville, Spain|Cambridge, United Kingdom
University of Seville|University of Cambridge
Seville, Spain|Cambridge, United Kingdom
Introduction:
The psychosis spectrum encompasses a heterogeneous range of clinical conditions associated with abnormal brain development. The molecular and micro-architectural attributes that account for structural deviations from typical neurodevelopment are still unknown.
Methods:
MRI data (T1-w) from first-degree relatives of schizophrenia (SCZ) and schizoaffective disorder (SAD) patients (n=160; age=42.3±15.7), chronic SCZ-SAD patients (n=587; age=37.2±12), and healthy controls (HC; n =38,232; age=51.9±23.7) were obtained from ABCD, ASRB, BSNIP, UCLA CNP LA5c, MCIC, and UKB datasets. Individuals who had Psychotic Experiences (PE) rated as 'suspected', 'definite', or 'clinical' if they had additional signs of social impairment or help-seeking, formed the PE group (ALSPAC cohort; n=157; age=21.5±1.4). HC (n=269; age=22.3±1.5) were rated as not having had PEs. Individuals who had experienced a First Episode of Psychosis (n=352; age=31.4±8.8), along with their HC (n=195; age=30.6±7.7), were obtained from PAFIP cohort.
We used GAMLSS (Stasinopoulos & Rigby, 2007) to benchmark regional cortical volumes of psychosis-related groups against normative trajectories from ~100,000 participants (Bethlehem et al., 2022). Thus, age, sex, and site-normalized measures of cortical atypicalities across the lifespan ranked brain volumes within a range of 0 to 1 (centiles), which were compared with HC using the Wilcoxon test (FDR-corrected).
Following Hansen et al., 2022 methodology, we explored the associations between centiles and the spatial maps of 46 neurobiological features classified under 6 types: neurotransmitter, cell type, layer thickness, microstructure, cortical expansion, and metabolism.
The Principal Component Analysis-Canonical Correlation Analysis (PCA-CCA) model captured associations (weights) between the neurobiological maps and the regional centiles that the resulting linear combination of the maps constituted the predicted centiles (Mihalik et al., 2022). A spatial autocorrelation-preserving permutation test (spin test) was used to assess the statistical significance of the models and weights. Finally, a set of loadings was computed to ascertain the extent to which each neurobiological feature contributed to the predicted centiles.
We used GAMLSS (Stasinopoulos & Rigby, 2007) to benchmark regional cortical volumes of psychosis-related groups against normative trajectories from ~100,000 participants (Bethlehem et al., 2022). Thus, age, sex, and site-normalized measures of cortical atypicalities across the lifespan ranked brain volumes within a range of 0 to 1 (centiles), which were compared with HC using the Wilcoxon test (FDR-corrected).
Following Hansen et al., 2022 methodology, we explored the associations between centiles and the spatial maps of 46 neurobiological features classified under 6 types: neurotransmitter, cell type, layer thickness, microstructure, cortical expansion, and metabolism.
The Principal Component Analysis-Canonical Correlation Analysis (PCA-CCA) model captured associations (weights) between the neurobiological maps and the regional centiles that the resulting linear combination of the maps constituted the predicted centiles (Mihalik et al., 2022). A spatial autocorrelation-preserving permutation test (spin test) was used to assess the statistical significance of the models and weights. Finally, a set of loadings was computed to ascertain the extent to which each neurobiological feature contributed to the predicted centiles.
Results:
We found a generalized decrease in centiles across all groups compared to HC (Fig. 1a top). The relatives showed significant decreases in a great number of regions, while FEP and chronic in the majority of them (Fig. 1a middle). The PCA-CCA predicted centiles closely resembled empirical centiles for significant models (Fig. 1a bottom; all groups except PE; Pspin<0.05), where the groups with the lowest centiles, FEP and chronic, revealed the strongest correlation between them (Fig. 1b).
All models exhibited significant loadings (Fig. 2a; Pspin<0.05). Groups with the lowest centiles showed a greater number of significant negative loadings. Synapse density and 5-HT2A stood out in all groups, neurotransmitters and cortical expansion in relatives and the chronic group, metabolism and microstructure in FEP, and layer thickness in the chronic group.
The greatest negative overlapping loadings across all groups indicated a high presence of neurotransmitters, excitatory neurons, synapse density, and metabolism in regions where centiles are consistently low (Fig. 2b). The greatest positive loadings, consistent across all groups except chronic, indicating a high presence of neurotransmitters and cell types in regions closer to neurotypical centiles.
All models exhibited significant loadings (Fig. 2a; Pspin<0.05). Groups with the lowest centiles showed a greater number of significant negative loadings. Synapse density and 5-HT2A stood out in all groups, neurotransmitters and cortical expansion in relatives and the chronic group, metabolism and microstructure in FEP, and layer thickness in the chronic group.
The greatest negative overlapping loadings across all groups indicated a high presence of neurotransmitters, excitatory neurons, synapse density, and metabolism in regions where centiles are consistently low (Fig. 2b). The greatest positive loadings, consistent across all groups except chronic, indicating a high presence of neurotransmitters and cell types in regions closer to neurotypical centiles.
Conclusions:
We identified group-specific volume deviations below the expected trajectory for different psychosis-related groups based on a normative centile method. We revealed an overlapping spatial distribution of the neurobiological features, which are highly co-localized with the abnormal developmental trajectories. These findings help understand the vulnerability factors that may underlie atypical brain maturation in different conditions and stages of psychosis.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 1
Modeling and Analysis Methods:
Multivariate Approaches 2
Keywords:
Schizophrenia
STRUCTURAL MRI
Other - Normative modelling
1|2Indicates the priority used for review
Provide references using author date format
Hansen, J. Y., Shafiei, G., Markello, R. D., Smart, K., Cox, S. M. L., Nørgaard, M., Beliveau, V., Wu, Y., Gallezot, J. D., Aumont, É., Servaes, S., Scala, S. G., DuBois, J. M., Wainstein, G., Bezgin, G., Funck, T., Schmitz, T. W., Spreng, R. N., Galovic, M., … Misic, B. (2022). Mapping neurotransmitter systems to the structural and functional organization of the human neocortex. Nature Neuroscience, 25(11), 1569–1581. https://doi.org/10.1038/s41593-022-01186-3
Mihalik, A., Chapman, J., Adams, R. A., Winter, N. R., Ferreira, F. S., Shawe-Taylor, J., & Mourão-Miranda, J. (2022). Canonical Correlation Analysis and Partial Least Squares for Identifying Brain–Behavior Associations: A Tutorial and a Comparative Study. In Biological Psychiatry: Cognitive Neuroscience and Neuroimaging (Vol. 7, Issue 11, pp. 1055–1067). Elsevier Inc. https://doi.org/10.1016/j.bpsc.2022.07.012
Stasinopoulos, D. M., & Rigby, R. A. (2007). Generalized Additive Models for Location Scale and Shape (GAMLSS) in R. Journal of Statistical Software, 23(7). https://doi.org/10.18637/jss.v023.i07