Schizophrenia is a debilitating illness, characterized by significant impairment in functioning across the social, emotional, and cognitive domains. Although there is strong evidence to suggest that life stressors precede symptom exacerbation in schizophrenia, important gaps in the research literature still remain. First, the biological mechanisms through which stress may contribute to the onset and expression of clinical symptoms have not been well defined. Second, there is considerable variability across schizophrenia patients in terms of their susceptibility to an elevation in symptoms following life stress.

According to the neural-diathesis stress model of schizophrenia, cortisol interacts with a biological vulnerability for heightened dopamine neurotransmission to influence illness expression. Utilizing this framework, the present investigation sought to clarify whether the specific stressor and patient characteristics that lead to heightened cortisol release also place patients at greater risk for stress-induced relapse. Furthermore, this study was designed so as to elucidate whether cortisol serves as one of the biological mechanisms through which life stress gets "under the skin" to influence symptom expression in schizophrenia.

Drawing upon data obtained from 125 schizophrenia patients and 95 healthy individuals, results indicated that clinical symptoms were most closely linked to acute and chronic stressors characterized by social evaluative content. Specific individual difference factors, including low trait positive affect, a reduced use of adaptive coping styles, early life adversity, and the Met allele of the catechol-O-methyltransferase Val158Met (COMT) polymorphism, placed patients with schizophrenia at greater risk for symptom exacerbation following the experience of social evaluative stress. In healthy individuals, low trait positive affect and maladaptive coping styles moderated the relationship between life stress and depressive symptoms. Results also indicated that the specific individual characteristics moderating the relationships between social evaluative episodic stress and clinical symptoms also influenced the magnitude of the cortisol response to a laboratory psychosocial stressor, the Trier Social Stress Test. These findings have implications for not only furthering our understanding of the psychobiological determinants of stress reactivity in schizophrenia, but can contribute significantly to the development of specific treatment and management strategies aimed at resolving stress in the lives of individuals with this debilitating condition.

The present dissertation evaluated the P50 component of the event-related brain potential (ERP), an index of sensory gating measured during a paired-stimulus paradigm, as an intermediate phenotype between clinical symptoms and neuronal mechanisms that might also contribute to higher-order cognitive dysfunction in schizophrenia. Despite considerable evidence that schizophrenia patients exhibit impaired suppression of the P50 to the second of two identical stimuli, the clinical significance and neural mechanisms associated with the P50 gating deficit remain poorly understood. A series of three related studies sought to more fully define the clinical and neurobiological correlates of P50 suppression deficits in schizophrenia. In Study 1, comparison of 52 schizophrenia patients and 41 healthy participants confirmed a P50 suppression deficit in patients, replicating prior research. Investigation of the clinical correlates of the P50 deficit in patients revealed associations between impaired suppression and clinician ratings of attentional difficulties as well as poor working memory performance. In examining the neural substrates of P50 gating with diffusion tensor imaging, Study 2 found associations between P50 deficits and compromised structural integrity of white matter tracts connecting brain regions previously implicated in the generation of P50 suppression in schizophrenia. Study 3 examined whether P50 impairments are amenable to cognitive training and predictive of treatment outcome. Compared to a control intervention without cognitive targets, patients who completed cognitive remediation demonstrated significant improvements in P50 suppression. Baseline P50 ratios were also predictive of improvement as evidenced by post-treatment clinician ratings and performance on an attention task, thereby indicating that patients with stronger P50 suppression are likely to receive the most benefit from cognitive remediation. All relationships observed in the present set of studies were specific to P50 and did not extend to the N100 component of the ERP. By characterizing the relationship of P50 to clinical symptoms and cognitive dysfunction in schizophrenia, isolating biological mechanisms that might be involved in P50 suppression, and evaluating its amenability to cognitive training, the present dissertation provided support for P50 as a viable biomarker for guiding the development of interventions that target cognitive impairments in this chronic and debilitating mental illness.

The period following a first episode of schizophrenia (SZ) is considered a critical window for treating cognitive impairment, and cognitive training (CT) interventions have shown promising efficacy. Cognitive impairment in SZ is associated with widespread alterations in functional neural network connectivity. However, research has not identified consistent associations between specific connectivity measures and individual domains of cognitive impairment. Furthermore, few studies have examined whether CT works by promoting normalization of neural networks (i.e., reduction of pre-existing disturbances) and/or by fostering compensatory network processes. In the present studies, we used graph-theory connectivity analysis to evaluate properties of intrinsic functional network organization and their relationship to cognitive function in first-episode SZ before and after CT. Using resting-state fMRI data from 45 young adults with SZ and 32 healthy comparison participants (HC), we tested for group differences in graph properties of the ‘task-positive’ cingulo-opercular (CON) and frontoparietal (FPN) networks and the whole-brain network, as well as their relationship with global and domain-specific measures of cognitive function prior to CT. SZ and HC did not differ substantially on measures of network organization, and across groups, generalized cognitive performance was positively associated with centrality of a node in right dorsal anterior cingulate cortex (dACC). Cognitive function in SZ was more positively associated with whole-brain functional segregation than in HC. In SZ, cognitive performance appeared to be decoupled from bilateral dACC node centrality, though effects were non-significant after correcting for multiple comparisons. Next, using longitudinal data from 17 SZ who completed a six-month course of CT and 20 HC who received no intervention, we tested for network normalization and compensation following CT. Although effects were non-significant after correcting for multiple comparisons, we found preliminary evidence of compensation in SZ following CT. Centrality of a node in right dACC decreased after CT, and this decrease corresponded with gains in processing speed. Reduced centrality of two other nodes in right dACC following CT was associated with gains in working memory and generalized cognitive performance. Results illuminate a potential role for compensation in cognitive recovery from SZ and may contribute to increased precision of future CT interventions.

Individuals with depression and anxiety have difficulty inhibiting their response to emotional information. Depression has been related to more attentional bias to negative information as well as less attentional bias to positive information. Anxiety has been associated with more attentional bias to high arousal, negative (or threat-related) stimuli. Independent of depression and anxiety symptoms, stress across the lifespan has also been linked to alterations in emotion inhibition along with increased risk for the development of internalizing psychopathology in adulthood. Three studies were conducted in order to identify behavioral and neural correlates of a novel emotion inhibition task, and evaluate a potential mechanistic pathway linking stress exposure to psychopathology via alterations to emotion inhibition. To better characterize the role that distracting emotional contexts plays in disruption of inhibitory control, the Image-Color-Word Stroop (IMCWS) task was developed. Participants (N=133) completed the IMCWS as well as questionnaires assessing life stress, resilience, depression, and anxiety. EEG data were collected on a subset of participants (N = 72) during completion of the IMCWS task. The inclusion of emotionally valenced, high-arousal scenes resulted in longer reaction times during high-conflict conditions of the superimposed color-word Stroop, indicating that the IMCWS task successfully enhanced the automatic recruitment of emotion processing to be inhibited. Further, the IMCWS was found to alter event-related potential (ERP) components commonly associated with allocation of attention and emotion processing (e.g., N250, P300) and inhibitory control processes (N450), suggesting that emotion inhibition during the IMCWS involved increased early attentional allocation to emotional contexts followed by increased recruitment of inhibitory control mechanisms. Current life stress (CLS) was found to be predictive of anhedonic depression, and both were associated with alterations in emotion inhibition. Whereas CLS was associated with more biased attention and, therefore, reduced emotion inhibition during the presentation of negative-valence context images, anhedonic depression was found to be related to less biased attention (i.e., increased emotion inhibition) during the presentation of positive-valence context images. Emotion inhibition was not found to play a mechanistic role in the pathway from CLS to anhedonic depression. Rather, results suggest distinct, yet complementary impacts of CLS and anhedonic depression on emotion inhibition function. Given the comorbidity of CLS and depression, CLS should be considered in future studies of emotion regulation and cognitive control disruptions in depression.

Early-onset psychosis (EOP), or onset of overt psychosis prior to age 18, is associated with particularly severe neurocognitive and social impairment, and poor prognosis (Frangou, 2010; Vyas & Gogtay, 2012). While significant impairments have been documented in specific cognitive domains (Fioravanti, Carlone, Vitale, Cinti, & Clare, 2005; Heinrichs & Zakzanis, 1998), working memory (WM) in particular is an area that has been theorized to represent a reliable cognitive endophenotype of schizophrenia. However, the majority of literature to date has focused on adult-onset schizophrenia, in which relevant developmental processes have already unfolded. Adolescence provides a privileged opportunity to understand how psychosis-related abnormalities in the structure and function of neural networks affect the relationships between social cognition, neurocognition and functioning, as well as an opportunity to define biomarkers for development of treatments to improve functional outcomes.

As such, the first study examined the relationship between individual WM capacity and task-based neural activation and functional connectivity. Results indicated that, relative to typically developing controls, patients with EOP have poorer WM performance, lower overall WM capacity, reduced neural activity in WM-associated brain regions, and reduced coupling of WM-associated regions. Additionally, EOP patients evidenced greater neural activation and connectivity with increasing age, suggesting an atypical developmental trajectory along with general inefficiency of WM circuitry.

Additional insight into the neurodevelopmental processes relevant to psychosis can come from examining genetically defined high-risk cohorts, such as 22q11.2 deletion syndrome (22q11DS). The second study evaluated the phenotypic overlap between EOP and 22q11DS by investigating the profile of neurocognitive and social cognitive impairment in individuals with EOP relative to 22q11DS patients and healthy controls. Despite greater overall cognitive impairment in the 22q11DS group (with the exception of verbal fluency, for which EOP patients evidenced greater impairment), patients with EOP and 22q11DS had comparable deficits in processing speed. 22q11DS and EOP patients also evidenced similar patterns of relationships between cognitive measures and psychotic symptoms. However, neurocognition and social cognition largely did not predict future functioning among the groups.

The third and final study examined plasticity-based cognitive training (CT) as a potential intervention for such cognitive impairments in EOP. Results from the feasibility study indicated that slightly over half of participants were able to complete at least 10 hours of CT; dropout was primarily due to finding training boring. However, patients completing training showed higher post- versus pre-CT social functioning, and pre- to post-CT reductions in general and anxiety symptoms. This investigation into neurocognitive and social cognitive dysfunction in EOP, which incorporates perspectives from an fMRI paradigm and a comparison to a genetic high-risk cohort, has the potential to generate knowledge on a clinically significant and understudied area, and potentially produce new treatment targets for EOP.

Schizophrenia is a biologically complex disorder usually characterized by a decline in cognition and functioning. Current medications do not improve cognition and functioning, thus there is an effort to create a more targeted, objective approach for understanding the relationship between neurobiological and psychological changes in order to facilitate treatment development. I reviewed these methods, termed “quantitative traits”, and discussed their usefulness in neurobiological and genetic research. I then tested whether some measurements of brain activity, measured using electroencephalography, in response to auditory and visual stimuli were associated with performance on cognitive tests and general functioning in typically developing individuals and in adolescents with schizophrenia. I found that improved processing of auditory stimuli predicted improvements in auditory memory and functioning in healthy individuals. I also found that adolescents with schizophrenia showed impairments in processing auditory and visual stimuli relative to typically developing individuals, and these impairments were related to symptom severity. This research will help us to understand how quantitative measures of brain activity are related to cognition and functioning in schizophrenia, which may lead to a clearer understanding of the biological mechanisms involved in each, and eventual improvements in treatment and outcomes.

Individuals with schizophrenia (SZ) commonly demonstrate attentional deficits, which in turn may contribute to higher-level cognitive dysfunction and clinical symptoms associated with the disorder. In EEG studies, these deficits are reliably observed during an auditory paired-stimulus paradigm, in which two identical stimuli are presented 500 ms apart. Relative to healthy comparison participants (HC), SZ typically exhibit a smaller N100 component of the event-related brain potential (ERP) following the first stimulus, indicative of impaired allocation of attention. HC generally demonstrate a decrease in N100 amplitude in response to the second stimulus relative to the first (i.e., N100 suppression), whereas SZ reliably demonstrate less suppression, indicative of inhibitory filtering dysfunction. A separate body of research using fMRI has linked salience and inhibitory processing deficits in SZ to insufficient suppression of the default mode network (DMN) of the brain, which supports internally-based processes and self-referential processing. The salience network (SAL), a higher-order system responsible for monitoring stimulus salience, regulates DMN as needed for tasks requiring attention to external stimuli. Given the shared constructs associated with N100 responses and the connectivity between DMN and SAL, it was hypothesized that these N100 variables and neural network measures capture disorder-related dysfunction in shared mechanisms of salience and inhibitory processing and would thus be related. To evaluate these relationships in SZ (N = 52) and HC (N = 25), a multimodal approach was employed using resting-state fMRI, resting-state EEG, and EEG recorded during the paired-stimulus paradigm. Among SZ, resting state hyperconnectivity within DMN was observed using both oscillatory EEG and fMRI data, with DMN hyperconnectivity observed specifically in the EEG beta frequency band. SZ and HC generally demonstrated opposing relationships between resting state DMN connectivity and N100 variables, such that greater intra-DMN connectivity was associated with attenuated N100 amplitude and higher N100 ratio scores in SZ. SZ also demonstrated greater resting state DMN-to-SAL effective connectivity than SAL-to-DMN connectivity, whereas HC demonstrated no such asymmetry. In oscillatory EEG collected during the paired-stimulus paradigm, SZ demonstrated less evoked theta-band connectivity within DMN following presentation of the auditory stimulus. No relationships between task-based network connectivity measures and N100 variables were identified. Present findings demonstrate specific profiles of DMN dysfunction in SZ that differentially relate to attentional processes indexed by measures obtained from the N100 ERP component during the paired-stimulus paradigm. Multimodal data fusion demonstrated that EEG and fMRI assessments of DMN were related yet also captured unique aspects of network function associated with N100 phenomena in SZ. Future work can build upon this research to further clarify how alterations in these neural mechanisms contribute to disorder-related impairments and inform promising targets for individualized prevention and treatment.

Cardiac vagal tone (CVT) reflects the contribution of the parasympathetic nervous system to cardiac regulation and is hypothesized to contribute to the regulation of inflammation and expression of psychopathology. CVT may play an important role in sleep disturbance and insomnia treatment. Through a series of four studies, this dissertation interrogated the associations between CVT, inflammation, and clinical symptoms in sleep disturbed samples and evaluated possible roles for CVT in insomnia intervention. Studies 1 and 2 proposed a specific model: chronic sleep disturbance reduces CVT with downstream effects on inflammation and clinical symptoms. Studies 1 and 2 then evaluated cross-sectional associations between these variables and tested exploratory mediation models in two samples with poor sleep quality: 106 older adults with insomnia and 39 patients with first episode schizophrenia. Overall, results did not support the mediation models in either sample. However, findings revealed novel associations between CVT and cellular inflammatory processes in older adults and extended a negative association between CVT and systemic inflammation to schizophrenia, consistent with the cholinergic anti-inflammatory pathway. Studies 3 and 4 then evaluated possible roles for CVT in a randomized controlled trial which compared active behavioral treatments (Cognitive Behavioral Therapy for Insomnia and Tai Chi Chih) to sleep education in older adults with insomnia. Study 3 revealed that active behavioral interventions led to improvements in insomnia and clinical symptoms without up-regulating CVT, indicating that CVT is unlikely to be a mechanism supporting insomnia recovery. Study 4 then determined whether individual differences in CVT at pre-treatment moderated the effects of active behavioral interventions on key trial outcomes. Results showed that CVT did not moderate treatment effects. Null findings from Studies 3 and 4 underscore the need for further identification and evaluation of candidate mechanisms and moderators to maximize existing behavioral treatments for insomnia. Taken together, present results add to the mixed literature on CVT. Findings simultaneously challenge the assumption that CVT is a reliable correlate of clinical state and extend prior research on the role of the anti-inflammatory pathway in two unique samples with poor sleep quality.