Neurocognitive And Neuroimaging Characteristics Of Chronic Deficit Schizophrenia:Evidences From Resting-state Whole Brain FMRI

BackgroundSchizophrenia is a highly heterogeneous psychiatric disorder. A major barrier to identify the neurobiological underpinnings of schizophrenia is its current nosology that reflects a clinical syndrome rather than a single disease entity. Deficit schizophrenia (DS), introduced by Carpenter and Kirkpatrick, represents a clinically homogeneous subgroup of patients characterized by the presence of primary and enduring negative symptoms which presents as a trait-like feature from the first episode of schizophrenia and lasts during periods of clinical stability. Patients with DS differ from those with non-deficit schizophrenia (NDS) in a variety of clinical aspects such as the higher rate of family history, male and summer birth, poorer treatment response and long-term clinical outcome, which supports the deficit syndrome as a distinct subgroup in schizophrenia. Notably, the diagnosis of deficit and non-deficit schizophrenia subtypes are characterized by a high degree of longitudinal stability, which is suitable for the cross-sectional and retrospective study. To date, numerous studies focusing on the different cognitive components impairment of DS patients compared with NDS and healthy controls (HC) have yielded a broad range of findings. The majority of studies consistently reported that DS and NDS patients exhibited more severe impairments than HC in attention and executive functions, and the DS patients exhibited more severe impairments than NDS patients in general cognitive abilities, olfaction, executive functions, social cognition and language functions. However, previous reports on the other cognitive domains of DS patients varied considerably. Though early neurocognitive studies have reported a significantly poorer performance in DS patients on tests sensitive to frontal and parietal lobe dysfunction, recent studies and meta-analysis of neurocognitive findings suggested the neuropsychological impairment of deficit patients does not reflect either frontoparietal dysfunction or any other anatomically defined pattern of impairment. The inconsistencies of these various reports do not allow a congruent understanding of the characteristic neurocognitive profile of DS, which might be attributed to the various clinical and methodological factors including variation of psychiatric symptoms severity, small sample size and different neuropsychological assessment tools. More importantly than all of that, up to now, the relationship between neurocognitive impairments and clinical variables in DS and NDS patients has been rarely investigated.Numerous structural and functional magnetic resonance imaging (MRI) studies have been made to investigate the pathological and therapeutic mechanism on schizophrenia. Up to now, the MRI studies do not support the hypothesis that DS represents the more severe end of a severity continuum in schizophrenia, but fail to identify clear morphological correlates of the deficit syndrome. The discrepancies of existing neuroimaging studies on DS may be attributed to the differences in imaging techniques among studies, focus of local brain region or single neuronal circuits rather than the whole-brain level, small sample sizes and heterogeneous treatment conditions (patients were drug-free in some studies, drug-treated in others). It is worth noting that emerging evidences demonstrate the schizophrenia as a "miswiring" or dysconnectivity disorder in multiple neuronal circuits rather than a focal pathology in a single region. Recently, the schizophrenia network dysfunction hypothesis further indicate that the aberrant network structure and activity may contribute to the symptoms and cognitive impairment through its interaction with schizophrenia pathogenesis. Therefore, it would be of value to explore the unique characteristics of brain networks in DS patients based on the whole brain level and investigate the relationships between these alterative metrics on neuronal circuits and the clinical and cognitive variables. Resting-State functional MRI (R-fMRI) can capture the dynamics of information communication among brain regions, which has been identified as the ideal tool for investigating brain functional network. R-fMRI has many advantages over other imaging methods, like the non-invasiveness, no requirement about the injection of contrast agent, relatively high spatial and temporal resolution and repeatability without ill effects. Using R-fMRI, previous neuroimaging studies have reported the widespread abnormal functional connectivity or amplitude of low-frequency fluctuations (ALFF) across the brain in schizophrenia patients. Also, recent studies using complex network analyses based upon graph theory have demonstrated abnormal topological organization of whole-brain networks in schizophrenia, such as a loss of small-worldness and network modularity, and a redistribution of hubs. Notably, only one study until now reported that patients with DS exhibited disruption of topological organization in structural brain networks in comparison with either NDS patients or HC by measuring inter-regional cortical thickness correlation derived from structural MRI data. Besides, the neuroimaging studies on DS is still stagnant in the state of brain gray/white matter volume or cortical thickness, or the level of local brain region functional activation. No research has been made to explore the abnormalities of whole-brain functional networks in DS patients based on R-fMRI. Therefore, R-fMRI-based studies which focus on the topological organization and the ALFF (and/or functional connectivity) alterations in whole-brain functional networks in DS patients would provide crucial insights into the understanding of pathophysiological mechanisms of DS and may provide new imaging biomarkers for the therapy of the negative symptoms in schizophrenia. In the present study, we recruited the chronic schizophrenia patients with high degree of homogeneity and well matched controls with age, handedness and gender. The diagnoses of deficit and non-deficit schizophrenia were made according to the Chinese version of the Schedule for the Deficit Syndrome (SDS, the gold standard) and the variances between DS and NDS were strictly restricted due to the confounders including gender, type and dosage of antinsvchotic drug. fluctuations of psvchiatric symptoms and social environment. For all participants, we assessed their cognitive function using multi-domain neurocognitive tests and constructed their whole-brain functional networks using the R-fMRI data. The present study yielded three main research directions. (1) We sought to investigate the severity and characteristic patterns of neurocognitive impairments in DS and NDS patients and their relationships with clinical variables. (2) We sought to determine whether there were commonalities and distinctions in the topological abnormalities of whole-brain functional networks between the DS and NDS groups as compared to the controls by using R-fMRI and graph theoretical approaches. (3) We sought to explore the aberrant regional cerebral function by measuring the fALFF (a normalized ALFF measure) among DS, NDS and HC. Then, areas with fALFF alterations between DS and NDS were used as seeds in whole-brain functional connectivity analysis to investigate whether there were abnormalities in the salience network (SN) between DS and NDS. Furthermore, we explored the relationships between the network parameter alterations and clinical symptoms and performances on neuropsychological tests in the two patient subgroups, respectively. Finally, the study is expected to advance our current understanding of the characteristic patterns of neurocognitive impairments and neurobiological mechanism underlying the functional network dysfunction in DS, and shed lights on exploring network-based biomarkers for the therapy of the negative symptoms in schizophrenia in the further.Part 1. Neurocognitive Impairments in Deficit Schizophrenia and TheRelationships with Symptom Dimensions and Other Clinical Variables Objective:We sought to investigate the severity and characteristic patterns of neurocognitive impairments in DS and NDS patients and their relationships with clinical variables. Methods:Attention, ideation fluency, cognitive flexibility and visuospatial memory function were assessed in 40 DS male patients,57 NDS male patients, and 52 male HC by a comprehensive neuropsychological battery. The core cognitive domain and its mediation patterns with clinical variables were investigated in DS and NDS patients respectively with a series of statistical processes including general linear model (GLM) analysis, profile analysis, Pearson correlation and the multiple regression analysis.Results:Both schizophrenia subgroups had overall more severe cognitive impairments than controls while DS performed worse on every neuropsychological measure except the Stroop interference than the NDS patients with age and education as the covariates. Profile analysis found significantly different patterns of cognitive profiles between two patients group mainly due to their differences in attention and cognitive flexibility functions. Age, education, illness duration and negative symptoms were found to have the correlations with cognitive impairments in the NDS group, while only age and the negative symptoms were correlated with the cognitive impairments in the DS group. Multiple regression analyses revealed that sustained attention and cognitive flexibility were the core impaired cognitive domains mediating other cognitive functions in DS and NDS patients respectively.Conclusion:DS patients exemplified worse in almost all cognitive domains than NDS patients. Sustained attention and cognitive flexibility might be the key impaired cognitive domains for DS and NDS patients respectively. The present study suggested the DS as a specific subgroup of schizophrenia.Part 2. Convergence and Divergence of Brain Network Dysfunctions in Deficit and Non-deficit SchizophreniaObjective:We sought to determine whether there were commonalities and distinctions in the topological abnormalities of whole-brain functional networks between the DS and NDS groups as compared to the controls by using R-fMRI and graph theoretical approaches. Furthermore, we explored the relationships between the network parameter alterations and clinical symptoms and performances on neuropsychological tests in the two patient subgroups, respectively.Methods:R-MRI and graph theory approaches were employed to investigate the topological organization of the functional brain networks of 114 male participants including 33 DS,41 NDS and 40 HC. The clinical variables were analyzed by the univariate one-way analysis of covariance (ANCOVA). The routine imaging preprocessing was carried out using the SPM8 and GRETNA software. To construct the brain functional network, the images of each brain was parcellated into 90 regions of interest (ROIs,45 for each hemisphere) using the automated anatomically labeling (AAL) atlas. To measure interregional resting-state functional connectivity, Pearson correlation coefficients between each pair of ROIs were calculated, thus generating a 90 x 90 correlation matrix for each subject. Then, we calculated the topological organization parameters (both global and regional network measures) of whole-brain functional networks among the three groups based upon the graph theory approaches. For statistical analysis, first, ANCOVAs were used to detect the among-group difference of topological metrics with the age, education, illness duration and chlorpromazine (CPZ)-equivalent daily dose as the covariates. Then, multiple linear regression analyses were performed to examine the relationships between the clinical and cognitive variables and the topological properties in each patient group.Results:The ANCOVA analysis showed significant differences in education (F=6.685, p=0.002) but not age (F=1.464, p=0.236) among the three groups. Least-significant difference (LSD) post hoc comparisons revealed shorter education periods for DS (p=0.001) and NDS (p=0.006) patients relative to HC subjects, while the two patient subgroups did not differ significantly (p=0.473). The two patient subgroups had no significant differences in the mean age of onset, smoking and antipsychotic medicine types and dosage (chlorpromazine equivalents). The DS patients showed more severe psychopathological total symptom and negative symptom (all ps<0.001) than NDS but not in either positive, affect or disorganized syndrome (all ps>0.172). All of the three groups exhibited typical small-world network architecture. However, at the whole-brain level, the NDS group exhibited significantly lower local network efficiency than the HC group, suggesting a relatively subtle random configuration. In contrast, the DS group exhibited more randomization of brain network organization, as characterized by significantly smaller local clustering, shortest path length, local efficiency and greater global efficiency. At the nodal level, both the NDS and DS groups showed commonly higher regional nodal connectivity in the limbic system [left inferior frontal gyrus (orbital part), olfactory cortex and right hippocampus], and lower regional connectivity mainly in the right inferior occipital gyrus, putamen and pallidum than the controls. Importantly, the DS group exhibited higher regional nodal connectivity in the left middle temporal gyrus (temporal pole) and right inferior temporal gyrus (ITG), and lower in the right inferior frontal gyrus (triangular part) as compared to the NDS group. The degree in right putamen and pallidum was negatively correlated with SANS total score in the NDS group. The nodal efficiency in the right ITG, smaller shortest path length and greater global efficiency exhibited correlations with the BPRS total score in the DS group.Conclusion:We demonstrated the convergence and divergence of the topological disorganization of whole-brain functional networks in patients with DS and NDS, which provides crucial insights into the understanding of differential pathophysiological mechanisms of the two schizophrenic subtypes.Part 3. The comparison of ALFF and functional connectivity in Deficit and Non-deficit SchizophreniaObjective:This study investigated the characteristics of ALFF in DS and NDS. Furthermore, we explored the alterations of whole-brain functional connectivity based on the areas with ALFF alterations between DS and NDS, and the relationships between the alterations of resting-state fMRI signals and clinical symptoms and performances on neuropsychological tests in the two patient subgroups, respectively.Methods:The sample was the same as Part2. Regional cerebral function was evaluated by measuring the voxelwise fractional ALFF (fALFF), a normalized ALFF measure, among DS, NDS and HC groups. Areas with fALFF alterations between DS and NDS were used as seeds in whole-brain functional connectivity analysis to investigate whether there were abnormalities in the SN between DS and NDS. GLMs were used to detect the among-group differences of resting-state fMRI signals with the age, education, illness duration and chlorpromazine (CPZ)-equivalent daily dose as the covariates. A threshold of p=0.05 after AlphaSim correction was used, an approach which utilizes Monte Carlo simulations to correct for multiple comparisons.Results:Both the patient subgroups exhibited regional hypoactivity in the sensorimotor area, visual cortex and the frontoparietal circuit, as well as the hyperactivity in the precuneus, middle part of cingulum and limbic system when compared with the HC. Moreover, the NDS group demonstrated the higher fALFF than the HC in the left thalamus, caudate and hippocampus. The DS patients had lower fALFF in the left insula and extended into frontotemporal cortex compared with the NDS and HC. When compared with the DS and HC, the NDS patients the lower fALFF in bilateral visual cortex and higher fALFF in bilateral insula and anterior cingulum cortex (ACC). Increased functional connectivity were found in DS group compared with the NDS group in the salience network (insula-visual cortex and temporooccipital circuits) and frontooccipital, frontotemporal circuits when the functional networks were constructed based on the seeds with fALFF alterations between DS and NDS. Furthermore, the functional connectivity between left insula and visual cortex was positively correlated with the cognitive flexibility in DS group, with trends toward positive correlations with the BPRS total score and the visuospatial memory. The functional connectivity between ACC and visual cortex showed positive correlation with the visuospatial memory in the NDS group.Conclusion:There were different alterant patterns of fALFF and functional connectivity in the salience network between DS and NDS. The DS group demonstrated the specific hypoactivity in left insula while the NDS group showed hyperactivity in bilateral insula, which suggested the DS as a specific subgroup of schizophrenia.