|Home | About | Journals | Submit | Contact Us | Français|
Schizophrenia patients and their first-degree relatives exhibit impaired working memory (WM) performance, possibly reflecting the disorder’s genetic diathesis (Glahn et al., 2003). Attempts to determine the neural bases of these WM deficits suggest that reduced efficiency of certain critical neurocognitive mechanisms may be involved (Callicott et al., 2003; Manoach, 2003; Karlsgodt et al., 2007). Consistent with an inefficiency hypothesis, we demonstrated that compared to controls, patients and their non-schizophrenic co-twins exhibit relatively larger changes in the electrophysiological signatures of the stimulus-encoding and memory-consolidation stages of WM performance, per unit increase in memory demands (Bachman et al., submitted). Along with encoding and consolidation, successful maintenance of information over a delay requires action of WM control functions, including suppression of potential interference when target stimuli are no longer displayed. Inhibiting cognitive processing of potential distracters is critical for preventing competition between task-relevant and task-irrelevant representations (Postle, 2005), and has been shown behaviorally to be differentially impaired in schizophrenia (Oltmanns & Neale, 1975).
Event-related desynchronization (ERD) and synchronization (ERS) of upper, or “fast,” EEG alpha frequency band activity, a measure of power decrease (ERD) or increase (ERS) as a percentage of pre-stimulus power (Neuper & Pfurtscheller, 2001), have been associated consistently with individual differences in neurocognitive efficiency measured during performance of challenging cognitive tasks, including paradigms that place heavy demands on WM (Neubauer et al., 2006). Research into the mechanisms underlying these fluctuations in power has demonstrated that alpha ERD reflects increased excitability of active cerebral cortex, whereas, alpha ERS appears to reflect decreased excitability, or large-scale inhibition, of cortex (Neuper & Pfurtscheller, 2001).
Consistent with this inhibitory function, alpha ERS is evident during the delay period of WM tasks, correlates in magnitude with memory load, and peaks over cortical areas responsible for processing task-modality specific sensory information (Sauseng et al., 2005; Jensen et al., 2002). Collectively, these findings suggest that alpha ERS reflects ‘top-down’ gating of sensory areas to prevent encoding of goal-irrelevant stimuli while task-relevant information is held actively in mind (reviewed, Klimesch et al., 2007).
Fast alpha ERS, therefore, should be greater in amplitude during the delay period of a delayed-match-to-sample WM task among individuals less efficient at gating task-irrelevant stimuli. It was predicted that, across groups, delay-period ERS would increase with memory load; moreover, patients and their co-twins were expected to display stronger load effects than controls, reflecting a hypothesized decrease in neurocognitive efficiency. Alternatively, inefficient inhibition of posterior cortical areas not directly related to WM would be associated with increased posterior ERS among patients and their co-twins, insensitive to WM load manipulation.
Participants were recruited from the cohort of same-sex twin pairs born in Finland between 1940 and 1957 and in which one cotwin received a DSM-III-R diagnosis of schizophrenia and the other did not meet criteria for any psychotic disorder. Additional subject identification and recruitment details are provided elsewhere (Cannon et al., 1998). Twenty-nine of these discordant pairs participated, although factors including hardware problems and noise, movement, and apparent electromagnetic inference – likely related to construction nearby the laboratory – rendered data from 16 twin sets unusable. Since only intact twin pairs were included, patients and their non-schizophrenic co-twins were eliminated from the study sample at equivalent rates. Another 4 discordant pairs were excluded because their behavioral performance fell below chance, leaving 9 discordant pairs. After screening for any psychotic disorder, Cluster A diagnosis, or history of psychosis-related treatment or work disability in any first-degree relatives, 9 demographically-matched, control twin pairs were recruited from the same database. The data contamination discussed above did not affect the discordant and control pairs differentially (16 discordant sets and 14 control sets eliminated).
Pairs did not differ in age (discordant: M=47.97 years; control: M=48.67), F[1,35]=0.066, p=0.936) or sex (4 female pairs per group), and all were right-handed and white. Substance abuse histories were also matched. Zygosity type was collapsed due to the size of the overall sample.
After the study was described, written informed consent was obtained. The study protocol was approved by the ethics committees or Institutional Review Boards of the Universities of Helsinki, Pennsylvania, and California, Los Angeles, and the Uusimaa Hospital District in Helsinki.
The match-to-sample task (Glahn et al., 2003) consists of 2000 ms exposure to a memory set of 1, 3, 5, or 7 locations arrayed around a fixation point, followed by a 3000 ms delay during which only the fixation point is visible, and then a single-probe dot prompting the participant to decide whether it matches the location of any stimuli in the memory set. 50% of trials were true positive and 50% were true negative; trial sequence was ordered randomly. Participants received feedback during initial practice trials.
EEG was recorded in a magnetically-shielded room using a whole-head EEG/MEG instrument (Elekta Neuromag, Ltd.). Signals were obtained from 66 equally-spaced Ag/AgCl electrodes referenced to the nose (Virtanen et al., 1996), and digitized at 500 Hz, with a 0.01–160 Hz band-pass filter. Vertical and horizontal electro-oculograms (EOG) were recorded with a bandpass of 0.5–30 Hz.
Continuous data were subjected to independent components analysis (ICA; Jung et al., 2000) for identification and removal of artifacts. On average, 16.1% (S.E.M.=1.3%), 14.0% (S.E.M.=1.9%), and 14.3% (S.E.M.=1.9%) of components were removed from the control, cotwin, and patient data, respectively, with no significant difference between the groups, F(2,35)=0.55, p=0.58. Conditioning of the EEG data included band-pass filtering (zero-phase shift, 0.01–70 Hz). Epochs were sorted for accuracy and only correct trials were retained.
As recommended by Klimesch (1999), peak individual alpha frequency (IAF) was ascertained for each subject. Each participant’s fast alpha band ranged from IAF at the lower bound to (IAF+2 Hz) at the upper bound. ERD/ERS was calculated as a percentage of power decrease (ERD) or increase (ERS), relative to the 200 ms pre-stimulus baseline (Neuper & Pfurtscheller, 2001). Consistent with published reports (Neubauer et al., 2006), electrodes were collapsed into linearly-derived channels (created from clusters of 7 contiguous EEG channels) representing grossly-defined brain regions (see Figure caption).
Data analysis employed a general linear mixed model ANOVA (SAS Institute, Cary, NC), with set size entered as a repeated measure and twin-pair membership was entered as a random variable (Satterthwaite option used to correct for non-independence of observations). Group differences and Group × Set Size interactions were tested for accuracy (% correct), reaction time (RT), and fast alpha ERD/ERS magnitude. Hypotheses were modeled as contrast statements within the ANOVA.
There were significant effects of group, F(2,27)=3.75, p=0.036, and memory set size, F(3,105)=50.90, p<0.001, on accuracy. Consistent with studies using the same task (e.g., Glahn al., 2003), Group × Set Size was not significant, F(6,105)=1.32, p=0.255. RT results paralleled the accuracy data with significant effects of group, F(2,28)=9.09, p=0.009; and memory set size, F(3,106)=49.65, p<0.001; but not Group × Set Size, F(6,105)=0.33, p=0.920.
During the delay, there were significant effects of group, F(2,33) 4.92, p=0.014, memory load, F(3,511)=11.72, p<0.001, and scalp region, F(3,511)=3.19, p=0.024 (anterior vs. posterior, not included) on fast alpha ERS. Critically, the Group × Memory Load interaction was significant, F(6,511)=4.16, p=0.004, as was the contrast predicting that patients and their co-twins would show greater ERS than controls, most markedly at higher load levels, F(1,511)=14.11, p<0.001. Despite this increased load-sensitivity among patients and their co-twins, topographic distribution of ERS activity did not differ as evidenced by the absence of Group × Scalp Region, F(6,511)=0.93, p=0.473, Memory Load × Scalp Region, F(9,511)=0.60, p=0.797, or Group × Load × Scalp Region, F(18,511)=0.48, p=0.966, effects.1 There were no significant associations between electrophysiological measures and symptom severity or medication exposure.
Schizophrenia patients and their co-twins were found to display a larger increase in ERS magnitude with increasing memory loads, relative to controls. Given that delay-period alpha ERS likely reflects the activity of a ‘top-down’ (Klimesch et al., 2007) interference control mechanism serving to buffer task-relevant information against competition from potential distracters (Klimesch et al., 2007; Jensen et al., 2002), these results suggest an increased physiological cost of screening out potentially-interfering information among patients and their co-twins. Therefore, rather than failing to inhibit encoding of task-irrelevant stimuli, interference control appears to function in affected groups, but at greatly reduced efficiency.
Although these conclusions rely upon a relatively small sample, several safeguards were implemented in order to preserve our inferential power. These include use of ICA to remove non-cerebral artifacts while leaving EEG signals unaltered (Jung et al., 2000), reliance on a minimum number of accurate, artifact-free epochs per condition, and implementation of statistical tests that do not assume equal variance between groups. Finally, exclusion of subjects performing below chance at two or more memory loads and inclusion of only correct trials removed any uncertainty inherent in consideration of cognitive processes leading to incorrect responses.
The necessity of combining monozygotic and dizygotic cotwins into a single category precludes stating conclusively that genetic factors account for patients’ and their co-twins’ apparently homologous decrements in neurocognitive efficiency. The presence of these abnormalities among patients’ co-twins, however, does suggest that they occur independently of diagnosis-related factors, such as antipsychotic medication exposure, significant symptom expression, and history of institutionalization.
Thus, present results provide an electrophysiological perspective on WM capacity limitations associated with schizophrenia and its inherited diathesis, demonstrating that patients and their non-schizophrenic co-twins suffer from a reduction in neurocognitive efficiency during the process of active information maintenance, in addition to decreased efficiency evident during the late perceptual encoding period (Bachman et al., submitted).
The authors wish to thank Ulla Mustonen, Pirjo Käki, and Eila Voipio for their contributions to subject recruitment and evaluation, Antti Tanskanen for his contributions to the register searches, Kauko Heikkilä for his contributions to data management of the Finnish Twin Cohort Study, Theo van Erp for statistical consultation, and the Finnish twins for participation in the study.
Previous Presentation Portions of this work were presented at the meeting of the Society for Psychophysiology Research in Vancouver, BC, October 25 – 29, 2006.
1On an exploratory basis, we did compare group averaged data (correct trials only) including subjects performing at or below chance with group data excluding those same subjects, and found virtually no difference in the pattern of ERP amplitude results.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.