Lateralized power spectra of the EEG as an index of visuospatial attention

Abstract

The electroencephalogram (EEG) was measured in an endogenous orienting paradigm where symbolic cues indicated the likely side of to-be-discriminated targets. Combined results of event-related lateralizations (ERLs) and a newly derived measure from wavelet analyses that we applied on the raw EEG and individual event-related potentials (ERPs), the lateralized power spectra (LPS) and the LPS-ERP, respectively, confirmed the common view that endogenous orienting operates by anterior processes, probably originating from the frontal eye fields, modulating processing in parietal and occipital areas. The LPS data indicated that modulation takes place by increased inhibition of the irrelevant visual field and/or disinhibition of the relevant to-be-attended visual field. Combined use of ERLs, the LPS, and the LPS-ERP indicated that most of the involved processes can be characterized as externally evoked, either or not with clear individual differences as some evoked effects were only visible in the LPS-ERERP, whereas few processes seemed to have an internally induced nature. Use of the LPS and the LPS-ERP may be advantageous as it enables to determine the involvement of internally generated lateralized processes that are not strictly bound to an event like stimulus onset.

Keywords: ERLs; ERPs; inhibition; visuospatial attention; wavelets.

Figure 1.

The sequence of events in a trial. Four types of targets were used; they had either low or high spatial frequencies (LSF or HSF) and had a vertical or horizontal orientation. Left or right button presses depended on target orientation.

Figure 2.

Grand average event-related lateralizations (filtered for display purposes with a low pass filter of 12 Hz, 24 dB/oct) and cue-direction specific hEOG as observed during the cue-target interval at parietal, occipito-parietal, and occipito-temporal sites. In our labeling of the channels, we indicated the locations on the left hemisphere but observed potentials are the result of double subtractions (see the Method section). The time windows in which significant effects were observed after application of the Benjamini and Hochberg (1995) procedure are indicated in gray boxes. The early directing attention negativity (EDAN) was significant at O1 (see Table 2), which is not displayed here. The time interval in which this occurred is indicated for PO7 with the light gray box. LDLDAP = late directing attention positivity.

Figure 3.

Topographical maps of the event-related lateralizations in 20-ms windows from 200 to 1,000 ms after cue onset based on interpolation of spherical splines (fourth order). In the left hemisphere, the contra-ipsilateral difference map is displayed, whereas an inverted ipsi-contralateral difference map is presented for the right hemisphere.

Figure 4.

Results of source analysis with BESA on the event-related lateralizations observed in the cue-target interval. Estimated activity of the posterior (green) and the anterior source (blue) is indicated in the left panel.

Figure 5.

The lateralized power spectra for the θ₂ and the θ₃ bands. Positive values mean increased ipsilateral relative to contralateral power. The initial frontal effect (see Table 3) and the later posterior effects thus both reflect increased ipsilateral as compared to contralateral power. Significant effects (at least two successive windows p < .01) for the θ₂ and the θ₃ bands are indicated in light and dark gray boxes, respectively.

Figure 6.

Topographical maps for the θ₂ and the θ₃ bands for time windows in which significant effects were observed, in the left and right panel, respectively. The left hemisphere reflects the contra-ipsilateral power difference, whereas the right hemisphere displays the ipsi-contralateral power difference. Positive values in the right hemisphere thus mean increased ipsilateral as compared to contra- lateral power. LPS = lateralized power spectra.

Figure 7.

The lateralized power spectra for the α₁ and the α₂ bands. Positive values mean increased ipsilateral relative to contralateral power. The occipital effect (see Table 3) for the α₁ and the α₂ band thus both reflect increased ipsilateral as compared to contralateral power. Significant effects (at least two successive windows p < .01) for the α₁ and the α₂ bands are indicated in light and dark gray boxes, respectively.

Figure 8.

Topographical maps for the α₁ and the α₂ bands for time windows in which significant effects were observed, in the upper and lower panel, respectively. For further descriptions, see Figure 6. LPS = lateralized power spectra.

Figure 9.

The lateralized power spectra for the β₁ and the β₂ bands. Positive values mean increased ipsilateral relative to contralateral power. Significant effects (at least two successive windows p < .01) for the β₁ and the β₂ bands are indicated in light and dark gray boxes, respectively.

Figure 10.

Topographical maps for the β₁ and the β₂ bands for time windows in which significant effects were observed. For further descriptions, see Figure 6. LPS = lateralized power spectra.

Figure 11.

Topographical maps for the θ₁, the θ₂, and the θ₃ bands for time windows in which significant effects were observed (see Table 4) after performing the lateralized power spectra analysis on the event related potentials (LPS-ERP). Ipsi-contralateral estimates are projected on the right hemisphere.

Figure 12.

Topographical maps for the α₁, the α₂, the β₁, and the the β₂ band for time windows in which significant effects were observed (see Table 4) after performing the lateralized power spectra analysis on the event related potentials (LPS-ERP). Ipsi-contralateral estimates are projected on the right hemisphere.