Multi-model order spatially constrained ICA reveals highly replicable group differences and consistent predictive results from resting data: A large N fMRI schizophrenia study.

Brain functional networks identified from resting functional magnetic resonance imaging (fMRI) data have the potential to reveal biomarkers for brain disorders, but studies of complex mental illnesses such as schizophrenia (SZ) often yield mixed results across replication studies. This is likely due in part to the complexity of the disorder, the short data acquisition time, and the limited ability of the approaches for brain imaging data mining. Therefore, the use of analytic approaches which can both capture individual variability while offering comparability across analyses is highly preferred. Fully blind data-driven approaches such as independent component analysis (ICA) are hard to compare across studies, and approaches that use fixed atlas-based regions can have limited sensitivity to individual sensitivity. By contrast, spatially constrained ICA (scICA) provides a hybrid, fully automated solution that can incorporate spatial network priors while also adapting to new subjects. However, scICA has thus far only been used with a single spatial scale (ICA dimensionality, i.e., ICA model order). In this work, we present an approach using multi-objective optimization scICA with reference algorithm (MOO-ICAR) to extract subject-specific intrinsic connectivity networks (ICNs) from fMRI data at multiple spatial scales, which also enables us to study interactions across spatial scales. We evaluate this approach using a large N (N > 1,600) study of schizophrenia divided into separate validation and replication sets. A multi-scale ICN template was estimated and labeled, then used as input into scICA which was computed on an individual subject level. We then performed a subsequent analysis of multiscale functional network connectivity (msFNC) to evaluate the patient data, including group differences and classification. Results showed highly consistent group differences in msFNC in regions including cerebellum, thalamus, and motor/auditory networks. Importantly, multiple msFNC pairs linking different spatial scales were implicated. The classification model built on the msFNC features obtained up to 85% F1 score, 83% precision, and 88% recall, indicating the strength of the proposed framework in detecting group differences between schizophrenia and the control group. Finally, we evaluated the relationship of the identified patterns to positive symptoms and found consistent results across datasets. The results verified the robustness of our framework in evaluating brain functional connectivity of schizophrenia at multiple spatial scales, implicated consistent and replicable brain networks, and highlighted a promising approach for leveraging resting fMRI data for brain biomarker development.

Multi-spatial-scale dynamic interactions between functional sources reveal sex-specific changes in schizophrenia.

We introduce an extension of independent component analysis (ICA), called multiscale ICA, and design an approach to capture dynamic functional source interactions within and between multiple spatial scales. Multiscale ICA estimates functional sources at multiple spatial scales without imposing direct constraints on the size of functional sources, overcomes the limitation of using fixed anatomical locations, and eliminates the need for model-order selection in ICA analysis. We leveraged this approach to study sex-specific and sex-common connectivity patterns in schizophrenia. Results show dynamic reconfiguration and interaction within and between multi-spatial scales. Sex-specific differences occur (a) within the subcortical domain, (b) between the somatomotor and cerebellum domains, and (c) between the temporal domain and several others, including the subcortical, visual, and default mode domains. Most of the sex-specific differences belong to between-spatial-scale functional interactions and are associated with a dynamic state with strong functional interactions between the visual, somatomotor, and temporal domains and their anticorrelation patterns with the rest of the brain. We observed significant correlations between multi-spatial-scale functional interactions and symptom scores, highlighting the importance of multiscale analyses to identify potential biomarkers for schizophrenia. As such, we recommend such analyses as an important option for future functional connectivity studies.

Weighted average of shared trajectory: A new estimator for dynamic functional connectivity efficiently estimates both rapid and slow changes over time.

BACKGROUND: Dynamic functional network connectivity (dFNC) of the brain has attracted considerable attention recently. Many approaches have been suggested to study dFNC with sliding window Pearson correlation (SWPC) being the most well-known. SWPC needs a relatively large sample size to reach a robust estimation but using large window sizes prevents us to detect rapid changes in dFNC. NEW METHOD: Here we first calculate the gradients of each time series pair and use the magnitude of these gradients to calculate weighted average of shared trajectory (WAST) as a new estimator for dFNC. RESULTS: Using WAST to compare healthy control and schizophrenia patients using a large dataset, we show disconnectivity between different regions associated with schizophrenia. In addition, WAST results reveals patients with schizophrenia stay longer in a connectivity state with negative connectivity between motor and sensory regions than do healthy controls. COMPARISON WITH EXISTING METHODS: We compare WAST with SWPC and multiplication of temporal derivatives (MTD) using different simulation scenarios. We show that WAST enables us to detect very rapid changes in dFNC (undetected by SWPC) while MTD performance is generally lower. CONCLUSIONS: As large window sizes are unable to detect short states, using shorter window size is desirable if the estimator is robust enough. We provide evidence that WAST requires fewer samples (compared to SWPC) to reach a robust estimation. As a result, we were able to identify rapidly varying dFNC patterns undetected by SWPC while still being able to robustly estimate slower dFNC patterns.

Reduced Amplitude of Low-Frequency Brain Oscillations in the Psychosis Risk Syndrome and Early Illness Schizophrenia.

Low-frequency oscillations (LFOs) of the blood oxygen level-dependent (BOLD) signal are gaining interest as potential biomarkers sensitive to neuropsychiatric pathology. Schizophrenia has been associated with alterations in intrinsic LFOs that covary with cognitive deficits and symptoms. However, the extent to which LFO dysfunction is present before schizophrenia illness onset remains unknown. Resting-state FMRI data were collected from clinical high-risk (CHR; n=45) youth, early illness schizophrenia (ESZ; n=74) patients, and healthy controls (HCs; n=85) aged 12-35 years. Age-adjusted voxelwise fractional amplitude of low-frequency fluctuations (fALFF; 0.01-0.08 Hz) of the BOLD signal was compared among the three groups. Main effects of Group (p<0.005 height threshold, familywise error cluster-level corrected p<0.05) were followed up via Tukey-corrected pairwise comparisons. Significant main effects of Group (p<0.05) revealed decreased fALFF in ESZ and CHR groups relative to HCs, with values in the CHR group falling between those of ESZ and HC groups. These differences were identified primarily in posterior cortex, including temporoparietal regions, extending into occipital and cerebellar lobes. Less LFO activity was related to greater symptom severity in both CHR and ESZ groups in several of these posterior cortical regions. These data support an intermediate phenotype of reduced posterior cortical LFO amplitude in CHR individuals, with resting fALFF values smaller than in HCs but higher than in ESZ patients. Findings indicate that LFO magnitude alterations relate to clinical symptoms and predate psychosis onset but are more pronounced in the early stages of schizophrenia.

A novel method for quantifying scanner instability in fMRI.

A method was developed to quantify the effect of scanner instability on functional MRI data by comparing the instability noise to endogenous noise present when scanning a human. The instability noise was computed from agar phantom data collected with two flip angles, allowing for a separation of the instability from the background noise. This method was used on human data collected at four 3 T scanners, allowing the physiological noise level to be extracted from the data. In a "well-operating" scanner, the instability noise is generally less than 10% of physiological noise in white matter and only about 2% of physiological noise in cortex. This indicates that instability in a well-operating scanner adds very little noise to functional MRI results. This new method allows researchers to make informed decisions about the maximum instability level a scanner can have before it is taken off line for maintenance or rejected from a multisite consortium. This method also provides information about the background noise, which is generally larger in magnitude than the instability noise.

Brain-performance correlates of working memory retrieval in schizophrenia: a cognitive modeling approach.

Correlations of cognitive functioning with brain activation during a sternberg item recognition paradigm (SIRP) were investigated in patients with schizophrenia and in healthy controls studied at 8 sites. To measure memory scanning times, 4 response time models were fit to SIRP data. The best fitting model assumed exhaustive serial memory scanning followed by self-terminating memory search and involved one intercept parameter to represent SIRP processes not contributing directly to memory scanning. Patients displayed significantly longer response times with increasing memory load and differed on the memory scanning, memory search, and intercept parameters of the best fitting probability model. Groups differed in the correlation between the memory scanning parameter and linear brain response to increasing memory load within left inferior and left middle frontal gyrus, bilateral caudate, and right precuneus. The pattern of findings in these regions indicated that high scanning capacity was associated with high neural capacity among healthy subjects but that scanning speed was uncoupled from brain response to increasing memory load among schizophrenia patients. Group differences in correlation of the best fitting model's scanning parameter with a quadratic trend in brain response to increasing memory load suggested inefficient or disordered patterns of neural inhibition among individuals with schizophrenia, especially in the left perirhinal and entorhinal cortices. The results show at both cognitive and neural levels that disordered memory scanning contributes to deficient SIRP performance among schizophrenia patients.