Introduction

Spatial working memory—the maintenance and manipulation of spatial information over brief delays—is essential for goal-directed behaviour. Neuroimaging studies using positron emission tomography (PET) have identified a network of prefrontal and parietal regions activated during spatial working memory tasks, but the temporal dynamics of this network at the millisecond timescale remain unknown. Electroencephalography (EEG) provides the temporal resolution necessary to examine dynamic patterns of neural oscillation during working memory maintenance.

Recent evidence suggests that oscillatory activity in specific frequency bands may index distinct cognitive operations. For example, theta-band (4–7 Hz) oscillations have been associated with attentional gating in dorsolateral prefrontal cortex, whilst alpha-band (8–12 Hz) activity may reflect the active maintenance of sensory information in parietal cortex. However, few studies have directly examined frequency-specific oscillations during spatial working memory, particularly across the range of memory loads employed in classic psychological experiments.

Method

Participants

Twenty-eight right-handed volunteers (mean age 22.3 years, SD 2.1; 14 female) gave written informed consent and received course credit. Participants had normal or corrected-to-normal vision and no history of neurological or psychiatric disorder.

Procedure

Participants performed a delayed spatial localization task while EEG was recorded from 64 scalp electrodes. On each trial, a sample stimulus (a small square) appeared at a random location within a 10° × 10° virtual grid for 300 ms. After a delay period of 1, 2, or 4 seconds, a probe stimulus appeared, and participants judged whether the probe matched the sample location. Feedback was provided auditorily after each response. Each memory load condition (load 1, 2, or 4 locations memorised concurrently) was tested in separate blocks of 32 trials; blocks were presented in counterbalanced order. Continuous EEG was digitised at 250 Hz with 0.01–100 Hz bandpass and average-referenced offline. Artefacts (eye movements, muscle activity) were manually rejected.

Results

Mean accuracy increased with shorter delays and decreased with higher memory load (range 78%–92% correct). Response times were analysed via one-way repeated-measures ANOVA with load as the within-subjects factor; a main effect of load was observed, F(2, 54) = 11.8, p < .001, with longer RTs for higher load. Wavelet power in the 8–12 Hz band during the delay period showed a significant load × region interaction, F(4, 108) = 3.2, p < .02. Post-hoc tests revealed that load effects were maximal over central parietal sites, with alpha power increasing from low to high load. Theta-band (4–7 Hz) activity was lateralised to right prefrontal electrodes and showed a main effect of load, F(2, 54) = 6.1, p < .01.

Power spectral density in posterior alpha was positively correlated with behavioural accuracy across participants (r = .44, p < .05), suggesting that individual differences in parietal oscillatory activity predicted working memory performance.

Discussion

This study provides evidence that spatial working memory is subserved by task-dependent oscillatory activity in dissociable frequency bands. The load-dependent increase in alpha-band activity over parietal regions aligns with prevailing models of visual-spatial attention and is consistent with the proposal that alpha oscillations reflect the active allocation of processing resources in sensory cortex. The concurrent engagement of theta-band activity in prefrontal sites suggests that dorsolateral prefrontal cortex may coordinate the maintenance and updating of spatial information, consistent with previous PET work.

Future investigations should employ source-localisation techniques to determine the anatomical origins of these oscillatory components and examine their relationship to fMRI-defined regions of interest. The correlation between individual alpha power and task performance suggests that oscillatory biomarkers may index fundamental constraints on working memory capacity, with implications for understanding both normal cognition and disorders of memory and attention.

References

  • Brouwer, G. J., & Heeger, D. J. (in press). Cross-orientation suppression in human visual cortex. Journal of Neurophysiology.
  • Gevins, A., Smith, M. E., McEvoy, L., & Yu, D. (1997). High-resolution EEG mapping of cortical activation related to working memory: Effects of task difficulty, type of processing, and practice. Cerebral Cortex, 7(4), 374–385.
  • Klimesch, W., Doppelmayr, M., Russegger, H., Pachinger, T., & Schwaiger, J. (1998). Induced alpha band power changes in the human EEG and attention. Neuroscience Letters, 244(2), 73–76.
  • Posner, M. I., & Petersen, S. E. (1990). The attention system of the human brain. Annual Review of Neuroscience, 13, 25–42.
  • Tallon-Baudry, C., & Bertrand, O. (1999). Oscillatory gamma activity in humans and its role in object representation. Trends in Cognitive Sciences, 3(4), 151–162.
  • Ungerleider, L. G., & Mishkin, M. (1982). Two cortical visual systems. In D. J. Ingle, M. A. Goodale, & R. J. W. Mansfield (Eds.), Analysis of visual behaviour (pp. 549–586). MIT Press.
  • Wechsler, D. (1997). Wechsler Adult Intelligence Scale-III. The Psychological Corporation.