Synchronous neural oscillations and cognitive processes Lawrence M. Ward Department of Psychology, University of British Columbia, 2136 West Mall, Vancouver, B.C. V6T 1Z4, Canada The central problem for cognitive neuroscience is to describe how cognitive processes arise from brain pro- cesses. This review summarizes the recent evidence that synchronous neural oscillations reveal much about the origin and nature of cognitive processes such as memory, attention and consciousness. Memory pro- cesses are most closely related to theta and gamma rhythms, whereas attention seems closely associated with alpha and gamma rhythms. Conscious awareness may arise from synchronous neural oscillations occur- ring globally throughout the brain rather than from the locally synchronous oscillations that occur when a sen- sory area encodes a stimulus. These associations between the dynamics of the brain and cognitive pro- cesses indicate progress towards a unified theory of brain and cognition. A popular stance in modern cognitive science is that cognitive processes arise from functionally organized brain processes [1], hence the discipline of cognitive neuroscience, which seeks to understand how this comes about. How can we identify the brain processes closely associated with cognitive processes and understand how the former give rise to the latter? Clearly, a first stage is to describe the functional anatomy of the brain: which areas show increased activity when a particular cognitive task is being performed. Informed by modern functional imaging techniques such as PET and fMRI, we have made an impressive beginning on this task. But cognitive processes are not static; they are dynamic. Even the simplest percept, memory or decision is a process that unfolds in time [2]. A popular way to think about the relationship between brain dynamics and cognitive dynamics is to describe the sequence of brain areas that ‘light up’ during the various stages in the performance of a cognitive task, like the sequence of bumpers hit by a pinball shot from its spring [3–5]. This approach is limited, however, because it cannot describe in any detail what is going on in those lit- up areas. Moreover, it doesn’t seem to be able to cope fully with the emerging view of brain processes as reverberations of reentrant activity in a complex neural network [6,7].A complementary approach that begins with the latter view is to try to describe how the oscillatory activity of the brain, as revealed by the electroencephalogram (EEG) and the magnetoencephalogram (MEG) as well as more invasive recordings, is related to the dynamics of cognitive performance (Box 1). There is increasing evidence that this relationship is revealing, and an increasing theoreti- cal understanding of how it might come about based on computational neural models [8–10]. Here I review recent data and two provocative models from the large literature linking EEG (and MEG) recordings of the large-scale oscillatory activity of the brain with the dynamics of the fundamental cognitive processes of memory, attention, and consciousness. The goal is to make a case for serious consideration of such data and models in the effort to understand the origins and nature of cognition. There are, of course, limitations to what this approach can tell us about cognition. These include the relatively poor spatial resolution of EEG and MEG (although MEG spatial resolution can approach that of fMRI), the fact that the dendritic field potentials of the cortical pyramidal neurons recorded by EEG and MEG constitutes only part of the brain’s relevant dynamics, the correlative nature of the associations reported which beg questions of causality, and various more specialized technical problems such as volume conduction (EEG) and noise filtering (MEG). In addition, it is early days in this endeavor, so that models are not complete; they are simply illustrative of what can be accomplished within the dynamical approach. EEG oscillations and cognitive processes The EEG varies with activity, both in humans and other animals, and particularly with the sleep-wakefulness cycle. Moreover, spectral power at various frequencies (Box 2) changes with age; alpha power increases as children mature whereas theta and delta power decrease. These changes are linked to the more general increase in cognitive competence with maturation, whereas the reverse changes signal declining mental abilities in old age [11]. Alpha waves have been apparent in EEG recordings ever since electroencephalography was invented by Hans Berger in the 1930s. Classically, because alpha power was larger with eyes closed than with eyes open, it was thought that alpha reflected a relaxed, unoccupied brain. An overall decrease in alpha power has been linked to increasing demands of attention, alertness, and task load in general [11]. Theta power, by contrast, tends to increase in memory tasks, especially during encoding [8,11,12]. These complementary effects have been thought to reflect different cognitive operations occurring in cortico-thalamic circuits: theta for encoding and alpha for search and retrieval [11]. In what follows I will discuss more recent data and models that are Corresponding author: Lawrence M. Ward (lward@psych.ubc.ca). Review TRENDS in Cognitive Sciences Vol.7 No.12 December 2003 553 http://tics.trends.com 1364-6613/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.tics.2003.10.012