Visual target modulation of functional connectivity networks revealed by self-organizing group ICA

We applied a data-driven analysis based on self-organizing group independent component analysis (sogICA) to fMRI data from a three-stimulus visual oddball task. SogICA is particularly suited to the investigation of the underlying functional connectivity and does not rely on a predefined model of the experiment, which overcomes some of the limitations of hypothesis-driven analysis. Unlike most previous applications of ICA in functional imaging, our approach allows the analysis of the data at the group level, which is of particular interest in high order cognitive studies. SogICA is based on the hierarchical clustering of spatially similar independent components, derived from single subject decompositions. We identified four main clusters of components, centered on the posterior cingulate, bilateral insula, bilateral prefrontal cortex, and right posterior parietal and prefrontal cortex, consistently across all participants. Post hoc comparison of time courses revealed that insula, prefrontal cortex and right fronto-parietal components showed higher activity for targets than for distractors. Activation for distractors was higher in the posterior cingulate cortex, where deactivation was observed for targets. While our results conform to previous neuroimaging studies, they also complement conventional results by showing functional connectivity networks with unique contributions to the task that were consistent across subjects. SogICA can thus be used to probe functional networks of active cognitive tasks at the group-level and can provide additional insights to generate new hypotheses for further study.     

Aberrant functional network connectivity in psychopathy from a large (N = 985) forensic sample

Psychopathy is a personality disorder characterized by antisocial behavior, lack of remorse and empathy, and impaired decision making. The disproportionate amount of crime committed by psychopaths has severe emotional and economic impacts on society. Here we examine the neural correlates associated with psychopathy to improve early assessment and perhaps inform treatments for this condition. Previous resting‐state functional magnetic resonance imaging (fMRI) studies in psychopathy have primarily focused on regions of interest. This study examines whole‐brain functional connectivity and its association to psychopathic traits. Psychopathy was hypothesized to be characterized by aberrant functional network connectivity (FNC) in several limbic/paralimbic networks. Group‐independent component and regression analyses were applied to a data set of resting‐state fMRI from 985 incarcerated adult males. We identified resting‐state networks (RSNs), estimated FNC between RSNs, and tested their association to psychopathy factors and total summary scores (Factor 1, interpersonal/affective; Factor 2, lifestyle/antisocial). Factor 1 scores showed both increased and reduced functional connectivity between RSNs from seven brain domains (sensorimotor, cerebellar, visual, salience, default mode, executive control, and attentional). Consistent with hypotheses, RSNs from the paralimbic system—insula, anterior and posterior cingulate cortex, amygdala, orbital frontal cortex, and superior temporal gyrus—were related to Factor 1 scores. No significant FNC associations were found with Factor 2 and total PCL‐R scores. In summary, results suggest that the affective and interpersonal symptoms of psychopathy (Factor 1) are associated with aberrant connectivity in multiple brain networks, including paralimbic regions.     

Altered functional organization within and between resting-state networks in chronic subcortical infarction

This study aimed to investigate the changes in functional connectivity (FC) within each resting-state network (RSN) and between RSNs in subcortical stroke patients who were well recovered in global motor function. Eleven meaningful RSNs were identified via functional magnetic resonance imaging data from 25 subcortical stroke patients and 22 normal controls using independent component analysis. Compared with normal controls, stroke patients exhibited increased intranetwork FC in the sensorimotor (SMN), visual (VN), auditory (AN), dorsal attention (DAN), and default mode (DMN) networks; they also exhibited decreased intranetwork FC in the frontoparietal network (FPN) and anterior DMN. Stroke patients displayed a shift from no FC in controls to negative internetwork FC between the VN and AN as well as between the VN and SMN. Stroke patients also exhibited weakened positive (anterior and posterior DMN; posterior DMN and right FPN) or negative (AN and right FPN; posterior DMN and dorsal SMN) internetwork FC when compared with normal controls. We suggest that subcortical stroke may induce connectivity changes in multiple functional networks, affecting not only the intranetwork FC within RSNs but also the internetwork FC between these RSNs.     

Functional network connectivity during rest and task conditions: A comparative study

Functional connectivity (FC) examines temporal statistical dependencies among distant brain regions by means of seed‐based analysis or independent component analysis (ICA). Spatial ICA also makes it possible to investigate FC at the network level, termed functional network connectivity (FNC). The dynamics of each network (ICA component), which may consist of several remote regions is described by the ICA time‐course of that network; hence, FNC studies statistical dependencies among ICA time‐courses. In this article, we compare comprehensively FNC in the resting state and during performance of an auditory oddball (AOD) task in 28 healthy subjects on relevant (nonartifactual) brain networks. The results show global FNC decrease during the performance of the task. In addition, we show that specific networks enlarge and/or demonstrate higher activity during the performance of the task. The results suggest that performing an active task like AOD may be facilitated by recruiting more neurons and higher activation of related networks rather than collaboration among different brain networks. We also evaluated the impact of temporal filtering on FNC analyses. Results showed that the results are not significantly affected by filtering. Hum Brain Mapp 34:2959–2971, 2013. © 2012 Wiley Periodicals, Inc.     

Functional and structural changes in the memory network associated with left temporal lobe epilepsy

Understanding functional plasticity in memory networks associated with temporal lobe epilepsy (TLE) is central to predicting memory decline following surgery. However, the extent of functional reorganization within memory networks remains unclear. In this preliminary study, we used novel analysis methods assessing network‐level changes across the brain during memory task performance in patients with TLE to test the hypothesis that hippocampal functions may not readily shift between hemispheres, but instead may show altered intra‐hemispheric organization with unilateral damage. In addition, we wished to relate functional differences to structural changes along specific fibre pathways associated with memory function. Nine pre‐operative patients with intractable left TLE and 10 healthy controls underwent functional MRI during complex scene encoding. Diffusion tensor imaging was additionally performed in the same patients. In our study, we found no evidence of inter‐hemispheric shifts in memory‐related activity in TLE using standard general linear model analysis. However, tensor independent component analysis revealed significant reductions in functional connectivity between bilateral MTL, occipital and left orbitofrontal regions among others in left TLE. This altered orbitofrontal activity was directly related to measures of fornix tract coherence in patients (P < 0.05). Our results suggest that specific fibre pathways, potentially affected by MTL neurodegeneration, may play a central role in functional plasticity in TLE and highlight the importance of network‐based analysis approaches. Relative to standard model‐based methods, novel objective functional connectivity analyses may offer improved sensitivity to subtle changes in the distribution of memory functions relevant for surgical planning in TLE. Hum Brain Mapp, 2009. © 2009 Wiley‐Liss, Inc.     

Dissociable relations between amygdala subregional networks and psychopathy trait dimensions in conduct‐disordered juvenile offenders

Psychopathy is a serious psychiatric phenomenon characterized by a pathological constellation of affective (e.g., callous, unemotional), interpersonal (e.g., manipulative, egocentric), and behavioral (e.g., impulsive, irresponsible) personality traits. Though amygdala subregional defects are suggested in psychopathy, the functionality and connectivity of different amygdala subnuclei is typically disregarded in neurocircuit‐level analyses of psychopathic personality. Hence, little is known of how amygdala subregional networks may contribute to psychopathy and its underlying trait assemblies in severely antisocial people. We addressed this important issue by uniquely examining the intrinsic functional connectivity of basolateral (BLA) and centromedial (CMA) amygdala networks in relation to affective, interpersonal, and behavioral traits of psychopathy, in conduct‐disordered juveniles with a history of serious delinquency (N = 50, mean age = 16.83 ± 1.32). As predicted, amygdalar connectivity profiles exhibited dissociable relations with different traits of psychopathy. Interpersonal psychopathic traits not only related to increased connectivity of BLA and CMA with a corticostriatal network formation accommodating reward processing, but also predicted stronger CMA connectivity with a network of cortical midline structures supporting sociocognitive processes. In contrast, affective psychopathic traits related to diminished CMA connectivity with a frontolimbic network serving salience processing and affective responding. Finally, behavioral psychopathic traits related to heightened BLA connectivity with a frontoparietal cluster implicated in regulatory executive functioning. We suggest that these trait‐specific shifts in amygdalar connectivity could be particularly relevant to the psychopathic phenotype, as they may fuel a self‐centered, emotionally cold, and behaviorally disinhibited profile. Hum Brain Mapp 37:4017–4033, 2016. © 2016 The Authors Human Brain Mapping Published by Wiley Periodicals, Inc.     

Resting-state functional connectivity in major depression: abnormally increased contributions from subgenual cingulate cortex and thalamus.

BACKGROUND: Positron emission tomography (PET) studies of major depression have revealed resting-state abnormalities in the prefrontal and cingulate cortices. Recently, fMRI has been adapted to examine connectivity within a specific resting-state neural network--the default-mode network--that includes medial prefrontal and anterior cingulate cortices. The goal of this study was to examine resting-state, default-mode network functional connectivity in subjects with major depression and in healthy controls. METHODS: Twenty-eight subjects with major depression and 20 healthy controls underwent 5-min fMRI scans while resting quietly. Independent component analysis was used to isolate the default-mode network in each subject. Group maps of the default-mode network were compared. A within-group analysis was performed in the depressed group to explore effects of depression refractoriness on functional connectivity. RESULTS: Resting-state subgenual cingulate and thalamic functional connectivity with the default-mode network were significantly greater in the depressed subjects. Within the depressed group, the length of the current depressive episode correlated positively with functional connectivity in the subgenual cingulate. CONCLUSIONS: This is the first study to explore default-mode functional connectivity in major depression. The findings provide cross-modality confirmation of PET studies demonstrating increased thalamic and subgenual cingulate activity in major depression. Further, the within-subject connectivity analysis employed here brings these previously isolated regions of hypermetabolism into the context of a disordered neural network. The correlation between refractoriness and subgenual cingulate functional connectivity within the network suggests that a quantitative, resting-state fMRI measure could be used to guide therapy in individual subjects.     

Neural intrinsic connectivity networks associated with risk aversion in old age

Risk aversion is associated with several important real world outcomes. Although the neurobiological correlates of risk aversion have been studied in young persons, little is known of the neurobiological correlates of risk aversion among older persons. Resting-state functional MRI data were collected on 134 non-demented participants of the Rush Memory and Aging Project, a community-based cohort study of aging. Risk aversion was measured using a series of standard questions in which participants were asked to choose between a certain monetary payment ($15) versus a gamble in which they could gain more than $15 or gain nothing, with potential gains varied across questions. Participants determined to be “high” (n = 27) and “low” (n = 27) in risk aversion were grouped accordingly. Using a spherical seed region of interest in the anterior cingulate cortex, voxel-wise functional connectivity network similarities were observed in bilateral frontal, anterior and posterior cingulate, insula, basal ganglia, temporal, parietal, and thalamic regions. Differences in functional connectivity were observed such that those low in risk aversion had greater connectivity to clusters in the superior, middle, and medial frontal regions, as well as cerebellar, parietal, occipital, and inferior temporal regions. Those high in risk aversion had greater connectivity to clusters in the inferior and orbital frontal, parahippocampal, and insula regions, as well as thalamic, parietal, precentral gyrus, postcentral gyrus, and middle temporal regions. Similarities and differences in functional connectivity patterns may reflect the historical recruitment of specific brain regions as a network in the active processing of risk in older adults.     

Hallucination‐ and speech‐specific hypercoupling in frontotemporal auditory and language networks in schizophrenia using combined task‐based fMRI data: An fBIRN study

Hypercoupling of activity in speech‐perception‐specific brain networks has been proposed to play a role in the generation of auditory‐verbal hallucinations (AVHs) in schizophrenia; however, it is unclear whether this hypercoupling extends to nonverbal auditory perception. We investigated this by comparing schizophrenia patients with and without AVHs, and healthy controls, on task‐based functional magnetic resonance imaging (fMRI) data combining verbal speech perception (SP), inner verbal thought generation (VTG), and nonverbal auditory oddball detection (AO). Data from two previously published fMRI studies were simultaneously analyzed using group constrained principal component analysis for fMRI (group fMRI‐CPCA), which allowed for comparison of task‐related functional brain networks across groups and tasks while holding the brain networks under study constant, leading to determination of the degree to which networks are common to verbal and nonverbal perception conditions, and which show coordinated hyperactivity in hallucinations. Three functional brain networks emerged: (a) auditory‐motor, (b) language processing, and (c) default‐mode (DMN) networks. Combining the AO and sentence tasks allowed the auditory‐motor and language networks to separately emerge, whereas they were aggregated when individual tasks were analyzed. AVH patients showed greater coordinated activity (deactivity for DMN regions) than non‐AVH patients during SP in all networks, but this did not extend to VTG or AO. This suggests that the hypercoupling in AVH patients in speech‐perception‐related brain networks is specific to perceived speech, and does not extend to perceived nonspeech or inner verbal thought generation.     

Method for multimodal analysis of independent source differences in schizophrenia: combining gray matter structural and auditory oddball functional data

The acquisition of both structural MRI (sMRI) and functional MRI (fMRI) data for a given study is a very common practice. However, these data are typically examined in separate analyses, rather than in a combined model. We propose a novel methodology to perform independent component analysis across image modalities, specifically, gray matter images and fMRI activation images as well as a joint histogram visualization technique. Joint independent component analysis (jICA) is used to decompose a matrix with a given row consisting of an fMRI activation image resulting from auditory oddball target stimuli and an sMRI gray matter segmentation image, collected from the same individual. We analyzed data collected on a group of schizophrenia patients and healthy controls using the jICA approach. Spatially independent joint-components are estimated and resulting components were further analyzed only if they showed a significant difference between patients and controls. The main finding was that group differences in bilateral parietal and frontal as well as posterior temporal regions in gray matter were associated with bilateral temporal regions activated by the auditory oddball target stimuli. A finding of less patient gray matter and less hemodynamic activity for target detection in these bilateral anterior temporal lobe regions was consistent with previous work. An unexpected corollary to this finding was that, in the regions showing the largest group differences, gray matter concentrations were larger in patients vs. controls, suggesting that more gray matter may be related to less functional connectivity in the auditory oddball fMRI task.     

Reduced frontotemporal functional connectivity in schizophrenia associated with auditory hallucinations.

BACKGROUND: We used functional magnetic resonance imaging (fMRI) to investigate the frontotemporal disconnection hypothesis of schizophrenia. METHODS: Eight DSM-IV schizophrenia patients and 10 control subjects were studied with fMRI while they thought of the missing last word in 128 visually presented sentences. The fMRI data were analyzed comparing the effect of sentence completion (vs. rest) using a random effects analysis. RESULTS: There were no significant group differences in regional brain responses. Correlation coefficients between left temporal cortex (x = -54, y = -42, z = 3) and left dorsolateral prefrontal cortex (x = -39, y = 12, z = 24) were significantly lower in the schizophrenic group and were negatively correlated with the severity of auditory hallucinations. CONCLUSIONS: Previous demonstrations of hypofrontality in schizophrenia may reflect particular task requirements. Frontotemporal functional connectivity is reduced in schizophrenia and may be associated with auditory hallucinations.     

Altered motor network activation and functional connectivity in adult Tourette's syndrome

Tourette's syndrome (TS) is a developmental neuropsychiatric disorder characterized by motor and vocal tics as well as psychiatric comorbidities. Disturbances of the fronto-striatal-thalamic pathways responsible for motor control and impulse inhibition have been previously described in other studies. Although differences in motor performance are well recognized, imaging data elucidating the neuronal correlates are scarce. Here, we examined 19 adult TS patients (13 men, aged 22-52 years, mean = 34.3 years) and 18 age- and sex-matched controls (13 men, aged 24-57 years, mean = 37.6 years) in a functional magnetic resonance imaging study at 1.5 T. We corrected for possible confounds introduced by tics, motion, and brain-structural differences as well as age, sex, comorbidities, and medication. Patients and controls were asked to perform a sequential finger-tapping task using their right, left, and both hands, respectively. Task performance was monitored by simultaneous MR-compatible video recording. Although behavioral data obtained during scanning did not show significant differences across groups, we observed differential neuronal activation patterns depending on both handedness (dominant vs. nondominant) and tapping frequency in frontal, parietal, and subcortical areas. When controlling for open motor performance, a failure of deactivation in easier task conditions was found in the subgenual cingulate cortex in the TS patients. In addition, performance-related functional connectivity of lower- and higher-order motor networks differed between patients and controls. In summary, although open performance was comparable, patients showed different neuronal networks and connectivity patterns when performing increasingly demanding tasks, further illustrating the impact of the disease on the motor system.     

Attribution of emotions to body postures: An independent component analysis study of functional connectivity in autism

The ability to interpret others' body language is a vital skill that helps us infer their thoughts and emotions. However, individuals with autism spectrum disorder (ASD) have been found to have difficulty in understanding the meaning of people's body language, perhaps leading to an overarching deficit in processing emotions. The current fMRI study investigates the functional connectivity underlying emotion and action judgment in the context of processing body language in high‐functioning adolescents and young adults with autism, using an independent components analysis (ICA) of the fMRI time series. While there were no reliable group differences in brain activity, the ICA revealed significant involvement of occipital and parietal regions in processing body actions; and inferior frontal gyrus, superior medial prefrontal cortex, and occipital cortex in body expressions of emotions. In a between‐group analysis, participants with autism, relative to typical controls, demonstrated significantly reduced temporal coherence in left ventral premotor cortex and right superior parietal lobule while processing emotions. Participants with ASD, on the other hand, showed increased temporal coherence in left fusiform gyrus while inferring emotions from body postures. Finally, a positive predictive relationship was found between empathizing ability and the brain areas underlying emotion processing in ASD participants. These results underscore the differential role of frontal and parietal brain regions in processing emotional body language in autism. Hum Brain Mapp 35:5204–5218, 2014. © 2014 Wiley Periodicals, Inc .     

Structural network topology in limbic encephalitis is associated with amygdala enlargement,memory performance and serostatus

Limbic encephalitis (LE) forms a spectrum of autoimmune diseases involving temporal lobe epilepsy and memory impairment. Imaging features of LE are known to depend on the associated antibody and to occur on the brain network level. However, first studies investigating brain networks in LE have either focused on one distinct antibody subgroup or on distinct anatomical regions. In this study, brain graphs of 17 LE patients with autoantibodies against glutamic acid decarboxylase 65 (GAD-LE), four LE patients with autoantibodies against leucine-rich glioma-inactivated 1, five LE patients with autoantibodies against contactin-associated protein-like 2, 26 age- and gender-matched healthy control subjects, and 20 epilepsy control patients with hippocampal sclerosis were constructed based on T1-weighted structural magnetic resonance imaging scans and diffusion tensor imaging. GAD-LE showed significantly altered global network topology in terms of integration and segregation as compared to healthy controls and patients with hippocampal sclerosis (P < .01, analysis of variance with Tukey-Kramer post hoc tests). Linear regression linked global network measures with amygdala volume and verbal memory performance (P < .05). Alterations of local network topology show serotype dependence in hippocampus, amygdala, insula, and various cortical regions. Our findings reveal serotype-dependent patterns of structural connectivity and prove the relevance of in silico network measures on clinical grounds.     

Dysregulation of working memory and default‐mode networks in schizophrenia using independent component analysis,an fBIRN and MCIC study

Deficits in working memory (WM) are a consistent neurocognitive marker for schizophrenia. Previous studies have suggested that WM is the product of coordinated activity in distributed functionally connected brain regions. Independent component analysis (ICA) is a data‐driven approach that can identify temporally coherent networks that underlie fMRI activity. We applied ICA to an fMRI dataset for 115 patients with chronic schizophrenia and 130 healthy controls by performing the Sternberg Item Recognition Paradigm. Here, we describe the first results using ICA to identify differences in the function of WM networks in schizophrenia compared to controls. ICA revealed six networks that showed significant differences between patients with schizophrenia and healthy controls. Four of these networks were negatively task‐correlated and showed deactivation across the posterior cingulate, precuneus, medial prefrontal cortex, anterior cingulate, inferior parietal lobules, and parahippocampus. These networks comprise brain regions known as the default‐mode network (DMN), a well‐characterized set of regions shown to be active during internal modes of cognition and implicated in schizophrenia. Two networks were positively task‐correlated, with one network engaging WM regions such as bilateral DLPFC and inferior parietal lobules while the other network engaged primarily the cerebellum. Our results suggest that DLPFC dysfunction in schizophrenia might be lateralized to the left and intrinsically tied to other regions such as the inferior parietal lobule and cingulate gyrus. Furthermore, we found that DMN dysfunction in schizophrenia exists across multiple subnetworks of the DMN and that these subnetworks are individually relevant to the pathophysiology of schizophrenia. In summary, this large multsite study identified multiple temporally coherent networks, which are aberrant in schizophrenia versus healthy controls and suggests that both task‐correlated and task‐anticorrelated networks may serve as potential biomarkers. Hum Brain Mapp, 2009. © 2009 Wiley‐Liss, Inc.     

Dynamic functional network connectivity in Huntington's disease and its associations with motor and cognitive measures

Dynamic functional network connectivity (dFNC) is an expansion of traditional, static FNC that measures connectivity variation among brain networks throughout scan duration. We used a large resting‐state fMRI (rs‐fMRI) sample from the PREDICT‐HD study (N = 183 Huntington disease gene mutation carriers [HDgmc] and N = 78 healthy control [HC] participants) to examine whole‐brain dFNC and its associations with CAG repeat length as well as the product of scaled CAG length and age, a variable representing disease burden. We also tested for relationships between functional connectivity and motor and cognitive measurements. Group independent component analysis was applied to rs‐fMRI data to obtain whole‐brain resting state networks. FNC was defined as the correlation between RSN time‐courses. Dynamic FNC behavior was captured using a sliding time window approach, and FNC results from each window were assigned to four clusters representing FNC states, using a k‐means clustering algorithm. HDgmc individuals spent significantly more time in State‐1 (the state with the weakest FNC pattern) compared to HC. However, overall HC individuals showed more FNC dynamism than HDgmc. Significant associations between FNC states and genetic and clinical variables were also identified. In FNC State‐4 (the one that most resembled static FNC), HDgmc exhibited significantly decreased connectivity between the putamen and medial prefrontal cortex compared to HC, and this was significantly associated with cognitive performance. In FNC State‐1, disease burden in HDgmc participants was significantly associated with connectivity between the postcentral gyrus and posterior cingulate cortex, as well as between the inferior occipital gyrus and posterior parietal cortex.     

Altered intra‐ and inter‐network functional coupling of resting‐state networks associated with motor dysfunction in stroke

Motor functions are supported through functional integration across the extended motor system network. Individuals following stroke often show deficits on motor performance requiring coordination of multiple brain networks; however, the assessment of connectivity patterns after stroke was still unclear. This study aimed to investigate the changes in intra‐ and inter‐network functional connectivity (FC) of multiple networks following stroke and further correlate FC with motor performance. Thirty‐three left subcortical chronic stroke patients and 34 healthy controls underwent resting‐state functional magnetic resonance imaging. Eleven resting‐state networks were identified via independent component analysis (ICA). Compared with healthy controls, the stroke group showed abnormal FC within the motor network (MN), visual network (VN), dorsal attention network (DAN), and executive control network (ECN). Additionally, the FC values of the ipsilesional inferior parietal lobule (IPL) within the ECN were negatively correlated with the Fugl‐Meyer Assessment (FMA) scores (hand + wrist). With respect to inter‐network interactions, the ipsilesional frontoparietal network (FPN) decreased FC with the MN and DAN; the contralesional FPN decreased FC with the ECN, but it increased FC with the default mode network (DMN); and the posterior DMN decreased FC with the VN. In sum, this study demonstrated the coexistence of intra‐ and inter‐network alterations associated with motor‐visual attention and high‐order cognitive control function in chronic stroke, which might provide insights into brain network plasticity following stroke.     

Decreased efficiency of task-positive and task-negative networks during working memory in schizophrenia

Working memory (WM) is one of the most impaired cognitive processes in schizophrenia. Functional magnetic resonance imaging (fMRI) studies in this area have typically found a reduction in information processing efficiency but have focused on the dorsolateral prefrontal cortex. In the current study using the Sternberg Item Recognition Test, we consider networks of regions supporting WM and measure the activation of functionally connected neural networks over different WM load conditions. We used constrained principal component analysis with a finite impulse response basis set to compare the estimated hemodynamic response associated with different WM load condition for 15 healthy control subjects and 15 schizophrenia patients. Three components emerged, reflecting activated (task-positive) and deactivated (task-negative or default-mode) neural networks. Two of the components (with both task-positive and task-negative aspects) were load dependent, were involved in encoding and delay phases (one exclusively encoding and the other both encoding and delay), and both showed evidence for decreased efficiency in patients. The results suggest that WM capacity is reached sooner for schizophrenia patients as the overt levels of WM load increase, to the point that further increases in overt memory load do not increase fMRI activation, and lead to performance impairments. These results are consistent with an account holding that patients show reduced efficiency in task-positive and task-negative networks during WM and also partially support the shifted inverted-U-shaped curve theory of the relationship between WM load and fMRI activation in schizophrenia.