Longitudinal Inference of Multimodal Cortical and Hippocampal Connectivity in Psychosis Subtypes
Poster No:
654
Submission Type:
Abstract Submission
Authors:
Jana Totzek1,2, Mallar Chakravarty1,2, Ridha Joober1,2, Ashok Malla1,2, Jai Shah1,2, Alexandra Young3, Martin Lepage1,2, Katie Lavigne1,2
Institutions:
1McGill University, Montreal, Quebec, 2Douglas Research Centre, Montreal, Canada, 3University College London, London, London
First Author:
Co-Author(s):
Introduction:
The hippocampus is the neural correlate which shows the largest reduction in volume in psychosis (van Erp et al., 2016), while reductions in morphometric hippocampal-cortical connectivity predict negative symptoms as mediated by verbal memory (Makowski et al., 2020). Machine-learning approaches suggest that there are two distinct subtypes of neural atrophy in psychosis, of which one begins in the hippocampus (Jiang et al., 2023). To date it remains open when multimodal hippocampal connectivity deviates in relation to other cortical modules across subtypes of psychosis.
Methods:
We measured morphometric and resting-state functional connectivity. To derive morphometric connectivity, we sampled data from 175 patients with first episode (FEP) and enduring psychosis, and 117 non-clinical controls. 18 hippocampal and adjacent white matter volumes were derived through MAGeT (Pipitone et al., 2014), while cortical thickness was extracted through CIVET 2.1.1 (Ad-Dab'bagh et al., 2006), and parcellated into 62 DKT regions (Klein & Tourville, 2012). Structural covariance matrices were derived for patients and controls separately, and subject-specific structural covariance matrices through the jackknife bias estimation procedure (Ajnakina et al., 2021). The hippocampal regions were grouped into a hippocampal module while the cortical regions were grouped into the 7 Yeo modules (Makowski et al., 2020; Yeo et al., 2011), and we estimated the graph-theoretical participation coefficient as a measure of intermodular connectivity (Rubinov & Sporns, 2010). To derive resting-state functional connectivity, we sampled data from 66 FEP patients, and 51 controls. We used the CONN toolbox to extract functional correlation matrices between the 18 hippocampal MAGeT-regions and 46 cortical regions of the Harvard-Oxford atlas. The hippocampal regions were grouped into a hippocampal module and the 46 cortical regions were grouped into the Yeo modules to derive participation coefficients. We used the average participation coefficients of all 8 morphometric and 8 functional modules as input for two separate Subtype and Stage Inference (SuStaIn) analyses (Young, 2018). SuStaIn is a machine-learning algorithm which merges disease progression modeling and clustering, allowing us to derive connectivity progression patterns across psychosis subtypes.
Results:
Following 10-fold cross-validation, SuStaIn resulted in two models with three subtypes each. In the morphometric and functional analyses, Subtype 0 included individuals with normal-range connectivity on all markers. In the morphometric analysis, Subtype 1 progressed from decreased somatomotor network connectivity to decreased dorsal attention (DAN), frontoparietal (FPN), visual, salience, and hippocampal network connectivity, followed by increased limbic and default mode network (DMN) connectivity, while Subtype 2 progressed from increased DMN and limbic network connectivity to decreased hippocampal, salience, DAN, FPN, visual, and somatomotor network connectivity. Different patterns emerged in the functional SuStaIn analysis. Subtype 1 progressed from decreased FPN, limbic, visual, DMN, and hippocampal network connectivity toward increased DAN, somatomotor, and salience network connectivity, while Subtype 2 progressed from increased somatomotor, salience, and DAN connectivity toward decreased hippocampal, DMN, visual, limbic, and FPN connectivity.
·SuStaIn Results
Conclusions:
We found that hippocampal connectivity was the first to decrease after an increase in cortical connectivity across modalities. Our results are consistent with Jiang et al. (2023) and extend these findings into the field of multimodal connectivity, while underlining the role of bidirectional modeling in psychosis. Future work should address the relationship between the identified subtypes and clinical features of psychosis to evaluate the clinical utility of SuStaIn in using these neural progression patterns as a potential predictor of clinical outcomes.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 1
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural)
fMRI Connectivity and Network Modeling
Other Methods
Novel Imaging Acquisition Methods:
Multi-Modal Imaging 2
Keywords:
ADULTS
Computational Neuroscience
FUNCTIONAL MRI
Machine Learning
Psychiatric Disorders
Schizophrenia
STRUCTURAL MRI
1|2Indicates the priority used for review
Provide references using author date format
Ajnakina, O. (2021). Structural covariance of cortical gyrification at illness onset in treatment resistance: a longitudinal study of first-episode psychoses. Schizophrenia bulletin, 47(6), 1729-1739.
Jiang, Y. (2023). Neuroimaging biomarkers define neurophysiological subtypes with distinct trajectories in schizophrenia. Nature Mental Health, 1(3), 186-199.
Klein, A. (2012). 101 labeled brain images and a consistent human cortical labeling protocol. Frontiers in neuroscience, 6, 171.
Makowski, C. (2020). Altered hippocampal centrality and dynamic anatomical covariance of intracortical microstructure in first episode psychosis. Hippocampus, 30(10), 1058-1072.
Pipitone, J. (2014). Multi-atlas segmentation of the whole hippocampus and subfields using multiple automatically generated templates. Neuroimage, 101, 494-512.
Rubinov, M. (2010). Complex network measures of brain connectivity: uses and interpretations. Neuroimage, 52(3), 1059-1069.
van Erp, T.G. (2016). Subcortical brain volume abnormalities in 2028 individuals with schizophrenia and 2540 healthy controls via the ENIGMA consortium. Molecular psychiatry, 21(4), 547-553.
Yeo, B.T. (2011). The organization of the human cerebral cortex estimated by intrinsic functional connectivity. Journal of neurophysiology.
Young, A.L. (2018). Uncovering the heterogeneity and temporal complexity of neurodegenerative diseases with Subtype and Stage Inference. Nature communication