An EEG investigation of visual spatial working memory and schizophrenia

Background. Patients with schizophrenia exhibit widespread and severe cognitive deficits, including abnormal feature integration and working memory dysfunction. The disconnectivity theory of schizophrenia states that abnormal functional relationships between neural networks may underlie these and other symptom-related deficits. Methods . This theory was addressed by the use of electroencephalographic (EEG) recordings during a delayed match-to-sample paradigm. Recording EEG activity during a visual delayed match-to-sample task enabled the investigation of both visual stimulus encoding and large scale network activity during the delay interval, when a stimulus must be maintained in memory. Event-related potentials (ERPs) measured stimulus encoding to both the first and second stimulus, while power and coherence analyses on EEG activity during the delay interval measured the strength of neural activation at different frequencies and its coordination with other areas of the brain, respectively. Results. Behaviorally, schizophrenia patients had a higher visual threshold on the task. ERP amplitudes (P1, N1 and P2) to the first and second stimulus did not differ between groups, nor were there group differences in coherence during the delay interval. Power analyses of averaged waveforms revealed lower activation at theta, alpha and low beta frequencies by schizophrenia patients whereas single trial analyses revealed a trend toward lower theta power at frontal sites towards the end of the delay interval only. Finally, negative slow wave amplitude during the delay interval revealed greater slow wave negativity by the end of the delay interval by control subjects compared to schizophrenia patients. Conclusions. These findings indicate that negative slow wave and theta activity may be sensitive measures of group differences between non-psychiatric controls and those with schizophrenia. Differences between single-trial and averaged power analyses may indicate subtle increases in the temporal variability of the neural signal for schizophrenia patients, even when controlling for task difficulty. Finally, these results are somewhat surprising in that several measures of neural processing (sensory ERPs, power and coherence) suggest similar functional mechanisms between schizophrenia patients and controls once task difficulty is held constant across subjects.