T165. ULTRA-HIGH FIELD MORPHOMETRY IN DRUG-NAïVE FIRST EPISODE PSYCHOSIS

Abstract Background Structural neuroimaging studies report distributed grey matter volume (GMV) deficits in drug-naïve first episode psychosis (FEP), though their relevance to symptom burden and cognitive deficits is currently unclear. When compared to studies in medicated patients and/or patients with established later-stage of psychosis, the GMV deficits reported have been limited in both spatial distribution and effect size, indicating the possibility of stage-specific progression during the clinical course of psychosis. TOPSY (Tracking Outcomes of Psychosis) is one of the first studies intending to track the neurobiological trajectory using ultra-high field (7T) imaging starting from a drug-naïve first episode state. Here, we report the initial findings from the voxel-based morphometry (VBM) of GMV. To our knowledge, this is the first VBM report from drug-naïve FEP subjects obtained using a 7T MRI acquisition. Methods We used ultra-high field (7 Tesla) MRI in 28 patients with FEP (satisfying criterion A of DSM-5 schizophrenia) and 18 controls, to evaluate differences in the grey matter. Volume in a voxelwise manner. FEP and controls were matched for age, sex and parental socioeconomic status. Patients were recruited at an early intervention unit (PEPP, London Ontario) and had active psychotic symptoms at the time of scanning. We also obtained abbreviated PANSS (8 items) scores to index the severity of psychosis. Analysis was done using SPM12, after DARTEL based registration and segmentation but without spatial smoothing. 2-tailed voxelwise T-test with FDR correction (p=0.05, 5% rate for false positives) was used. We used multiple regression analysis to predict the scores from processing speed measure (modified Symbol Substitution Test) and the severity of Delusions and Unusual Thought Content (P1 and G9), the 2 symptoms for which most subjects sought treatment in the first place. Results Patients had a significant reduction in GMV in left fusiform gyrus (Hedge’s g = 1.98, T= 6.7), and increased GMV in the right precuneus (Hedge’s g = 1.63, T= 5.5) and lingual cortex (Hedge’s g = 1.19, T= 4.0). We did not find any other areas of significant GMV change. Of these 3 circumscribed GMV changes, reduced fusiform GMV was found among FEP patients with lower processing speed (ß=0.45, p=0.04), higher severity of delusions (ß=-0.43, p=0.049) and unusual thought content (ß=-0.59, p=0.01). Increased precuneus GMV was found among FEP patients with higher severity of delusions (ß=0.62, p=0.008) and unusual thought content (ß=0.50, p=0.03). Right lingual changes were not related to the severity of delusions or processing speed scores. Discussion Our findings suggest that (1) GMV deficits are minimal in drug-naïve FEP subjects, with large effect-size changes concentrated around face processing (fusiform) region (2) GMV increases co-occur with GMV reduction especially in those with most severe delusions and cognitive deficits indicating a role for compensatory plasticity. Subtle early brain structural changes appear to predict symptom burden and cognitive deficits at the time of first clinical presentation with psychosis. Focusing on treatments that manipulate the structure of fusiform cortex could potentially reduce the severity of some of the early symptoms in FEP.

early neurodevelopmental pathologies. It is unclear, however, whether such structural changes might be evident across the schizophrenia spectrum, involving at-risk subjects as well as even healthy subjects with subclinical or attenuated psychotic(-like) symptoms Methods: We analysed high-resolution MRI scans (3 Tesla, T1-weighted MPRAGE, 1x1x1mm resolution) from n=177 healthy subjects with no current or previous psychiatric condition recruited from the local community. Subjects completed the SCL90R, a general symptom checklist (i.e. self-rating of symptoms), which includes subscales for psychoticism (with subclinical psychotic/-like symptoms) and paranoid ideation. We used the CAT12 toolbox to analyse both gyrification using the absolute mean curvature approach (Luders et al., NeuroImage 2006, as well as cortical thickness and voxel-based morphometry (VBM). Results: Correcting for effects of age and gender, we found a significant negative correlation between SCL90R psychoticism scores and gyrification in a left prefrontal / frontopolar cluster, but no similar finding for wither cortical thickness analysis nor analyses of the paranoid ideation subscale. Discussion: Our results suggest that prefrontal gyrification might be a marker for psychotic phenotypes spanning a spectrum from subclinical symptom expression to frank psychosis. This association seems linked to gyrification (rather than other markers of brain structure), which would suggest a relative specificity. Hence, this would be consistent with the assumption that gyrification is related to early neurodevelopmental effects, which lead to liabity to experiencing psychotic symptoms later in life, and might thus serve as an imaging phenotype for early risk detection and intervention in high-risk groups. Background: Structural neuroimaging studies report disrupted morphological relationship in the grey matter volume (structural covariance) in patients with schizophrenia, indicating an impairment in functional and/ or developmental plasticity. To our knowledge, no studies have examined the alterations in structural covariance across the entire brain in drug-naïve first episode psychosis. TOPSY (Tracking Outcomes of Psychosis) is one of the first studies intending to track the neurobiological trajectory using ultra-high field (7T) imaging starting from a drug-naïve first episode state. Here, we report the initial findings from the structural covariance of grey matter volume. To our knowledge, this is the first structural covariance analysis being reported using a 7T anatomical MRI acquisition. Methods: We used ultra-high field (7 Tesla) MRI in 28 patients with FEP (satisfying criterion A of DSM-5 schizophrenia) and 18 controls, to estimate grey matter volume in a voxelwise manner. FEP and controls were matched for age, sex and parental socioeconomic status. Patients were recruited at an early intervention unit (PEPP, London Ontario) and had active psychotic symptoms at the time of scanning. Morphometric analysis was done using SPM12, after DARTEL based registration and segmentation but without spatial smoothing on 160 brain regions (6mm spheres) obtained using Dosenbach's atlas. Correlation matrix for each group was constructed from 160*160 pearson correlation coefficients, followed by estimation of a bias matrix for each subject using jack-knife bias estimation. Bias values for each pair of nodes in an individual subject quantified the contribution of that subject to the overall within-group covariance. Higher positive values meant greater covariance between the two given nodes in that subject, relative to the rest of the group. These bias matrices can be considered equivalent to demeaned and normalised matrices of structural covariance. Structural covariance across all possible regional pairwise connections was tested using 2-tailed voxelwise T-test with FDR correction (p=0.05, 5% rate for false positives). Results: Patients had a significant reduction in structural covariance affecting between right posterior insula and right precentral gyrus (within sensorimotor network, t=3.86, Hedge's g = 1.15); between right posterior insula and left ventral prefrontal cortex (between sensorimotor and salience network, t=3.71, Hedge's g = 1.10); and between right anterior cingulate cortex and right dorsal prefrontal cortex (between sensorimotor and default-mode network, t=3.10, Hedge's g = 0.92). There were no pairwise connections with increased structural covariance among FEP subjects compared to healthy controls. Discussion: Our findings suggest that (1) structural covariance is disrupted even by the time of first-episode of psychosis; thus, the disruptions in morphological relationships reported in schizophrenia are not explicable by antipsychotic usage or illness duration (2) sensorimotor network regions show a predominant disruption in structural covariance, affecting morphological relationships with both salience and default mode regions. The functional and developmental plasticity of sensorimotor networks may be crucial for the early trajectory of psychosis.

Robarts Research Institute
Background: Structural neuroimaging studies report distributed grey matter volume (GMV) deficits in drug-naïve first episode psychosis (FEP), though their relevance to symptom burden and cognitive deficits is currently unclear. When compared to studies in medicated patients and/or patients with established later-stage of psychosis, the GMV deficits reported have been limited in both spatial distribution and effect size, indicating the possibility of stage-specific progression during the clinical course of psychosis. TOPSY (Tracking Outcomes of Psychosis) is one of the first studies intending to track the neurobiological trajectory using ultra-high field (7T) imaging starting from a drug-naïve first episode state. Here, we report the initial findings from the voxelbased morphometry (VBM) of GMV. To our knowledge, this is the first VBM report from drug-naïve FEP subjects obtained using a 7T MRI acquisition. Methods: We used ultra-high field (7 Tesla) MRI in 28 patients with FEP (satisfying criterion A of DSM-5 schizophrenia) and 18 controls, to evaluate differences in the grey matter. Volume in a voxelwise manner. FEP and controls were matched for age, sex and parental socioeconomic status. Patients were recruited at an early intervention unit (PEPP, London Ontario) and had active psychotic symptoms at the time of scanning. We also obtained abbreviated PANSS (8 items) scores to index the severity of psychosis. Analysis was done using SPM12, after DARTEL based registration and segmentation but without spatial smoothing. 2-tailed voxelwise T-test with FDR correction (p=0.05, 5% rate for false positives) was used. We used multiple regression analysis to predict the scores from processing speed measure (modified Symbol Substitution Test) and the severity of Delusions and Unusual Thought Content (P1 and G9), the 2 symptoms for which most subjects sought treatment in the first place.
Results: Patients had a significant reduction in GMV in left fusiform gyrus (Hedge's g = 1.98, T= 6.7), and increased GMV in the right precuneus (Hedge's g = 1.63, T= 5.5) and lingual cortex (Hedge's g = 1.19, T= 4.0). We did not find any other areas of significant GMV change. Of these 3 circumscribed GMV changes, reduced fusiform GMV was found among FEP patients with lower processing speed (ß=0.45, p=0.04), higher severity of delusions (ß=-0.43, p=0.049) and unusual thought content (ß=-0.59, p=0.01). Increased precuneus GMV was found among FEP patients with higher severity of delusions (ß=0.62, p=0.008) and unusual thought content (ß=0.50, p=0.03). Right lingual changes were not related to the severity of delusions or processing speed scores. Discussion: Our findings suggest that (1) GMV deficits are minimal in drugnaïve FEP subjects, with large effect-size changes concentrated around face processing (fusiform) region (2) GMV increases co-occur with GMV reduction especially in those with most severe delusions and cognitive deficits indicating a role for compensatory plasticity. Subtle early brain structural changes appear to predict symptom burden and cognitive deficits at the time of first clinical presentation with psychosis. Focusing on treatments that manipulate the structure of fusiform cortex could potentially reduce the severity of some of the early symptoms in FEP.

T166. SPATIAL INCOHERENCE OF LARGE-SCALE CORTICAL NETWORKS RELATES TO FORMAL THOUGHT DISORDER IN SCHIZOPHRENIA: A 7T MRI-BASED THICKNESS STUDY
Lena Palaniyappan* ,1 , Ali Al-Radaideh 2 , Penny Gowland 3 , Peter Liddle 3 1 University of Western Ontario; 2 The Hashemite University; 3 University of Nottingham, Background: The thickness of cerebral cortex varies across individuals as well as across different regions within an individual. Shared trophic or plastic influences such as repeated task-related recruitment of extant brain regions results in morphological covariance within large-scale brain networks. Pathological processes disrupting functional co-activation can result in higher than expected degree of variability within large-scale networks in an individual level, resulting in spatial incoherence. We studied spatial incoherence of cortical thickness in 17 cortical networks identified on the basis of well-known patterns of intrinsic connectivity, to identify the spatially incoherent networks and relate them to differences in severity of thought disorder among patients with schizophrenia. Methods: Ultra-high field 7T anatomical MRI scans (MPRAGE) were obtained from 20 subjects in a clinically stable, medicated early stage of schizophrenia, and 19 sex, parental socioeconomic-status and age matched healthy controls. Cortical thickness was estimated using Freesurfer v5.0, across 17 networks based on the parcellation scheme of Yeo et al. We computed within-network coefficient of variation in thickness (CVT) across vertices that constitute each network. Higher CVT of a network in a subject indicates higher spatial incoherence within the network for that individual. Independent 2-tailed t-tests were used to compare CVT of 17 networks between the 2 groups with FDR-corrected p=0.05 considered as statistically significant. We related CVT of affected networks to the scores of positive and negative Formal Thought Disorder measured using Thought and Language Index in patients. Results: Salience Network (aka Ventral Attention Network as per Yeo atlas), Default Mode Network and Central Executive Network (aka dorsal Attention Network in Yeo atlas) showed most significant reduction in MRI-derived cortical thickness (networks #8, #12, #15 as well as #16 of Yeo atlas). Only the Salience and Executive Networks (network #8 and #12) showed higher coefficient of variation in patients compared to controls, indicating either a failure of coordinated maturation or co-ordinated function. Higher spatial incoherence of Salience Network related to reduced mean thickness of Central Executive Network in patients with schizophrenia; this relationship was not seen in healthy controls (Fisher's z test, p=0.02). Both higher coefficient of variation in Salience Network and lower mean thickness in Central Executive Network predicted the severity of positive but not negative thought disorder scores.

Discussion:
Our results indicate that (1) large-scale cortical networks involved in information processing (Salience and Executive Networks) show spatial incoherence in schizophrenia (2) the degree of spatial incoherence relates to the severity of disorganisation of thoughts and language in patients. Spatial incoherence may be the result of a dysmaturational or functional dysplastic effect reflecting inefficient cortical recruitment in schizophrenia. Within-subject morphological variability carries useful information that can potentially explain the elusive neural basis of complex symptoms such as formal thought disorder.

University of Nottingham
Background: The structural integrity of the anterior cingulum has been repeatedly observed to be abnormal in patients with schizophrenia. Reduced glutathione levels, indicating oxidative stress, is associated with reduced structural integrity of cingulum bundle in patients with schizophrenia. Variations in neuregulin-1, a well-established candidate marker for schizophrenia, results in oligodendrocyte dysfunction and defective myelination, and is shown to affect the structural integrity of the anterior cingulum in patients with schizophrenia. While the evidence to date has been obtained using diffusion tensor imaging, abnormal tract-specific changes in myelin content can be more directly inferred by combining multiple modalities of WM imaging such as diffusion tensor (DTI) and magnetization transfer imaging (MTI) in parallel. Methods: We used ultra-high resolution (7 Tesla) MTI in 17 patients with schizophrenia and 20 controls, to evaluate the macromolecular (predominantly myelin) content of the brain. Immediately after the 7T scanning, we also obtained a 3T diffusion tensor image (DTI) and undertook probabilistic tractography using FSL software (AutoPtx, ProbTrackX) to delineate anterior cingulum bilaterally. Unpaired t tests were used for group comparisons along with estimates of Cohen's d or Hedge's g for effect sizes. Results: Patients had a significant reduction in magnetization transfer ratio (MTR) in right (Cohen's d=0.91, p=0.007) but not left (d=0.03, p=0.92) cingulum bundle. There was also a trend level reduction in fractional anisotropy of right (d=0.60, p=0.07) but not left (d=0.47, p=0.17) cingulum bundle. We did not find any significant relationship between the 3 major symptom dimensions of schizophrenia (Reality Distortion, Disorganization, Psychomotor Poverty) and Cingulum MTR. Patients with Schneiderian delusions (n=5) showed a significantly reduced MTR of left cingulum compared to patients (n=12) with no Schneiderian delusions (Hedges' g=1.36, p=0.02). Discussion: Our findings suggest that MTR changes in anterior cingulum, resulting from either dysmyelination or neuroinflammation, is present in clinically stable patients with schizophrenia despite their medicated status. We lacked sufficient power to detect association between MTR changes of cingulum and symptom dimensions. Nevertheless, our results suggest that MTR changes are of higher magnitude than changes in fractional anisotropy, indicating the sensitivity of measuring myelination as a biological marker of white matter aberrations in schizophrenia.