Considered one of the core cognitive impairments in the illness, working memory disturbances have been intensely studied in schizophrenia. While functional neuroimaging studies have provided important anatomic localization of associated neural deficits (primarily to the dorsolateral prefrontal cortex), characterizing the precise nature of the physiologic disturbances remains an active line of research. Disturbances in the ability of neural networks to coordinate activity through synchronized oscillations, which have been shown to underlie a number of perceptual and cognitive processes in the illness, are also thought to provide the neurophysiologic basis for working memory impairments in the illness (Basar-Eroglu et al., 2007; Cho et al., 2006). This recent paper by Haenschel and colleagues makes a valuable contribution to this literature, with a systematic examination of the component processes of working memory function in early-onset schizophrenia patients, indexing synchrony across a broad spectrum of frequencies.

Using a modified Sternberg paradigm with abstract visual objects, Haenschel et al. examined evoked and induced synchronous activity in two regions of interest (using anterior versus posterior electrodes) across theta, alpha, beta, and gamma bands during the encoding, maintenance and retrieval phases of the task. The behavioral data showed poorer performance with increasing working memory load in patients compared with controls. In the EEG data, the most prominent findings were during the encoding phase, with evoked theta, alpha and beta activity being reduced in patients; in controls, but not patients, activity during encoding in these bands predicted task performance. The maintenance phase was divided into an earlier phase, which showed comparable induced anterior electrode alpha and posterior gamma increases in both groups, and a later phase, which showed a shift in the peak of induced gamma to lower working memory load conditions for patients. During retrieval, evoked theta, and induced theta and gamma, were reduced for patients. Modulations in synchrony tended to have a bias towards either anterior or posterior electrodes, but the between-group differences largely did not interact with these topographic biases.

A notable strength of this study is its systematic evaluation of the different phases of working memory performance, in contrast with previous studies of working memory in schizophrenia that used synchrony measures. The latter limited examination to delay-related activity or could not unambiguously associate activity with distinct working memory processes due to task characteristics (e.g., in the N-back task).

Parsing out the component subprocesses allowed Haenschel et al. to highlight the importance of evoked activity disturbances during the encoding period. It was of interest that these disturbances were limited to frequency bands lower than the gamma band, which was affected during the maintenance and retrieval phases in this task. This band has been found to be affected in previous studies of working memory in schizophrenia.

The unique pattern of spectral disturbances in evoked and induced activity for each of the different working memory phases provides a rich characterization of the EEG measures of oscillatory synchrony disturbances in schizophrenia. The challenge now is to work towards providing a parsimonious account of the findings, as it would seem unlikely that each finding of disturbance would map uniquely to an independent pathophysiologic process. For instance, examining cross-frequency interactions (e.g., Canolty et al., 2006) could contribute towards establishing such a coherent account. Similarly, as each phase of working memory performance depends on the integrity of prior phases, there would likely be dependencies between activations across the different phases of task performance. Elucidating the relative interdependencies versus independence of the findings could have critical therapeutic import, both in guiding development of cognitive remediation strategies as well as providing a framework for evaluating any novel interventions. For instance, if disturbances in gamma oscillations during the maintenance and retrieval phases are found to be causally related to the disturbances in lower frequencies during encoding, therapeutic efforts may then be most fruitfully directed at remediating the encoding-related disturbances.

Alternatively, it may be that the observed disturbances in synchrony across the three phases all have a common cause due to generic disturbances in mechanisms that support synchronous oscillations (perhaps, with encoding phase oscillations being the most sensitive index of such disturbances). These might include alterations in parvalbumin- and somatostatin-containing GABA interneurons that are conserved across multiple cortical areas (Hashimoto et al. , 2008) and that may give rise to disturbances in gamma (Bartos et al., 2007) versus theta (Beierlein et al., 2000) oscillations, respectively. Presumably, remediation of the respective neurobiologic disturbances could result in improvements that are frequency band-specific but span multiple working memory subprocesses.

In summary, this study represents a valuable contribution to our understanding of oscillatory disturbances associated with working memory impairments in schizophrenia. It also provides a nice point of departure for examining other aspects that merit further study, such as the generalizability of findings to later- onset schizophrenia, unmedicated patients, relatives of patients and to tasks that tap verbal working memory processes. Also, as none of the activations indexed for patients were predictive of their performance, further work is necessary to establish the neural basis for how patients actually do perform such tasks, which can help to inform novel cognitive retraining paradigms or pharmacologic interventions for remediation of working memory impairments in schizophrenia.

References:

Bartos M, Vida I, Jonas P. Synaptic mechanisms of synchronized gamma oscillations in inhibitory interneuron networks. Nat Rev Neurosci. 2007;8:45-56. Abstract

Basar-Eroglu C, Brand A, Hildebrandt H, Karolina Kedzior K, Mathes B, Schmiedt C. Working memory related gamma oscillations in schizophrenia patients. Int J Psychophysiol. 2007;64:39-45. Abstract

Beierlein M, Gibson JR, Connors BW. A network of electrically coupled interneurons drives synchronized inhibition in neocortex. Nat Neurosci. 2000; 3:904-910. Abstract

Canolty RT, Edwards E, Dalal SS, Soltani M, Nagarajan SS, Kirsch HE, Berger MS, Barbaro NM, Knight RT. High gamma power is phase-locked to theta oscillations in human neocortex. Science. 2006;313:1626-1628. Abstract

Cho RY, Konecky RO, Carter CS. Impairments in frontal cortical gamma synchrony and cognitive control in schizophrenia. Proc Natl Acad Sci. 2006;103:19878-19883. Abstract

View all comments by Raymond Cho

  I recommend this paper

I thought it was interesting that oscillatory activity in the early-onset schizophrenia patients was actually fairly intact, compared to the control subjects. There was no difference in alpha power between controls and patients during the delay period, and a gamma power deficit in the patients only appeared in the late part of the delay period. Furthermore, this late gamma deficit was supported statistically by post-hoc tests following only a trend level interaction with the group at p = .079. This isn't a very robust effect.

If the authors had elected to use the standard statistical criterion for post-hoc testing of alpha = .05, this deficit would not have been detected and the paper's conclusion would have been that delay-period oscillatory activity is not affected in early-onset schizophrenia patients. It would seem that the most robust oscillation deficits were in the retrieval period in the theta and gamma bands, which of course are very interesting. (The evoked theta deficit in the encoding period was apparently due to the P1 ERP, as the authors reported in their paper in Archives of General Psychiatry, Haenschel et al., 2007.)

It would be interesting to know whether delay-period oscillations show more pronounced deficits in older and/or more chronic schizophrenia patients. Given the robustness of delay-period alpha in working memory tasks, and its possible role in suppressing task-irrelevant processing (e.g., work by Ole Jensen and colleagues: Jokisch and Jensen, 2007), one might predict that alpha deficits would be associated with working memory errors in schizophrenia patients.

View all comments by Kevin Spencer