Declarative memory (DM) impairments are reported in schizophrenia and in unaffected biological relatives of patients. However, the neural correlates of successful and unsuccessful encoding, mediated by the medial temporal lobe (MTL) memory system, and the influence of disease-related genetic liability remain under explored. This study employed an event-related functional MRI paradigm to compare activations for successfully and unsuccessfully encoded associative face-name stimuli between 26 schizophrenia patients (mean age: 33, 19 m/7f), 30 controls (mean age: 29, 24 m/6f), and 14 unaffected relatives of patients (mean age: 40, 5 m/9f). Compared to controls or unaffected relatives, patients showed hyper-activations in ventral visual stream and temporo-parietal cortical association areas when contrasting successfully encoded events to fixation. Follow-up hippocampal regions-of-interest analysis revealed schizophrenia-related hyper-activations in the right anterior hippocampus during successful encoding; contrasting successful versus unsuccessful events produced schizophrenia-related hypo-activations in the left anterior hippocampus. Similar hippocampal hypo-activations were observed in unaffected relatives during successful versus unsuccessful encoding. Post hoc analyses of hippocampal volume showed reductions in patients, but not in unaffected relatives compared to controls. Findings suggest that DM encoding deficits are attributable to both disease-specific and genetic liability factors that impact different components of the MTL memory system. Hyper-activations in temporo-occipital and parietal regions observed only in patients suggest the influence of disease-related factors. Regional hyper- and hypo-activations attributable to successful encoding occurring in both patients and unaffected relatives suggest the influence of schizophrenia-related genetic liability factors.
Schizophrenia is characterized by a generalized cognitive impairment with pronounced deficits in memory and executive function (Reichenberg and Harvey, 2007 and Ranganath et al., 2008). Specifically, patients with schizophrenia experience impairments in declarative memory (DM) (Aleman et al., 1999, Weiss and Heckers, 2001 and Ranganath et al., 2008), which includes everyday memories of events (episodic memory) and facts (semantic memory) (Eichenbaum and Cohen, 2001). DM impairments are also reported in unaffected relatives of patients and increase with degree of biological relatedness, suggesting the involvement of schizophrenia genetic liability factors (Faraone et al., 2000 and Whyte et al., 2005).
The hippocampus and medial temporal lobe (MTL) are essential for DM (Eichenbaum and Cohen, 2001). Prefrontal and posterior association regions also act to mediate memory processing (Sperling et al., 2010 and Wang et al., 2010). Functional imaging studies of DM tasks in healthy subjects confirm MTL involvement and illustrate that regional activation is influenced by task characteristics, how information is learned, and whether encoding is successful (Buckner and Koutstaal, 1998 and Preston et al., 2005).
DM relies on the successful encoding, storage, and retrieval of information. DM deficits in patients with schizophrenia and non-symptomatic relatives appear particularly attributable to encoding difficulties (Cirillo and Seidman, 2003). Since different network components contribute to the type and stage of DM processing (Brewer and Moghekar, 2002), encoding deficits may relate to dysfunctions confined to specific MTL regions and/or to disturbances in connected cortical regions. Although more frequently focused on attempted encoding, several DM studies have demonstrated altered neural activity in hippocampal, parahippocampal, and connected prefrontal regions in schizophrenia (Heckers, 2001, Achim and Lepage, 2005 and Ragland et al., 2009). Fewer fMRI studies have examined DM in unaffected relatives (MacDonald et al., 2009), and none have dissociated disturbances in regional activity by examining encoding success for associative stimuli exclusively.
To identify the subcomponents of the MTL memory system affected by schizophrenia and disease-related genetic liability, we employed a validated event-related fMRI design (Sperling et al., 2003) to compare blood-oxygen-level-dependent (BOLD) responses for successful DM encoding in schizophrenia patients, first-degree unaffected biological relatives of patients, and community controls. The DM task, including novel associative face-name stimuli, is shown to elicit MTL and regionally specific hippocampal activations during successful encoding in controls. We hypothesized that patients would show differences in the magnitude of task-related brain activity in the MTL and associated cortical regions. Further, we predicted that relatives of patients, sharing approximately half of their genes with schizophrenia probands, would show intermediate abnormalities. Since successful encoding elicits greater neural activity in the anterior hippocampus (Sperling et al., 2003), hippocampal regions-of-interest (ROI) analyses were also conducted. Finally, post hoc analysis of structural imaging data examined differences in hippocampal volumes across groups.
Table 1 includes demographic, clinical, and memory performance information by group. Patients and controls were similar in age, F(1, 54) = 2.86, p = .10, and gender, χ2 (1, 49) = .01, p = .94; all groups were of similar socioeconomic status, F(2, 61) = 2.59, p = .08. However, relatives were older than controls (F(1, 42) = 10.33, p < .01, although not older than patients, p > .07. Relatives also differed in gender compared to patients and controls (χ2(1, 43) = 8.33, p < .01 and χ2(1, 39) = 6.54, p = .01). Although performing above chance, patients showed poorer memory performance than controls, F(1, 48) = 7.94, p < .01. Relatives did not differ in performance from patients p > .7 or controls p > .09 ( Table 1). Age, gender, and performance from the post-scan memory test were included as covariates in all fMRI analysis.
