Orienting auditory attention in time: Lateralized alpha power reflects spatio-temporal filtering

Abstract

The deployment of neural alpha (8-12 Hz) lateralization in service of spatial attention is well-established: Alpha power increases in the cortical hemisphere ipsilateral to the attended hemifield, and decreases in the contralateral hemisphere, respectively. Much less is known about humans' ability to deploy such alpha lateralization in time, and to thus exploit alpha power as a spatio-temporal filter. Here we show that spatially lateralized alpha power does signify - beyond the direction of spatial attention - the distribution of attention in time and thereby qualifies as a spatio-temporal attentional filter. Participants (N = 20) selectively listened to spoken numbers presented on one side (left vs right), while competing numbers were presented on the other side. Key to our hypothesis, temporal foreknowledge was manipulated via a visual cue, which was either instructive and indicated the to-be-probed number position (70% valid) or neutral. Temporal foreknowledge did guide participants' attention, as they recognized numbers from the to-be-attended side more accurately following valid cues. In the magnetoencephalogram (MEG), spatial attention to the left versus right side induced lateralization of alpha power in all temporal cueing conditions. Modulation of alpha lateralization at the 0.8 Hz presentation rate of spoken numbers was stronger following instructive compared to neutral temporal cues. Critically, we found stronger modulation of lateralized alpha power specifically at the onsets of temporally cued numbers. These results suggest that the precisely timed hemispheric lateralization of alpha power qualifies as a spatio-temporal attentional filter mechanism susceptible to top-down behavioural goals.

Fig. 1

(A) Trial design. After a visually presented temporal cue to indicate the to-be-probed number position (valid, invalid, neutral), an auditorily presented spatial cue indicated whether spoken numbers on the left versus right side were to be attended. Participants indicated via button press which one of two numbers appeared on the to-be-attended side. (B) Bars and error bars respectively show average ±1 SEM proportion of correct responses for the three temporal cues. Thin lines show data of N = 20 individual participants. Statistical comparisons were performed on logit-transformed proportion correct scores. *** p < 0.001; n.s. not significant. (C) Same as B for individual probed number positions (1–5). (D) Same as B for response speed, quantified as 1/response time (RT).

Fig. 2

(A) Bottom: Grand-average time-frequency representations of inter-trial phase coherence (ITPC), averaged across all (102) combined gradiometer sensors. Dashed vertical lines indicate onsets of number positions 1–5. Top: Topographic map and brain surfaces show low-frequency (1–5 Hz) ITPC averaged across the entire trial duration (0–7.5 s). LH: left hemisphere; RH: right hemisphere. (B) Bottom: Grand-average time-frequency representations of oscillatory power (relative change with respect to –1 to 0 s), averaged across all (102) combined gradiometer sensors. Top: Topographic map and brain surfaces show grand-average alpha lateralization (averaged across trials with neutral and instructive temporal cues), calculated for 8–12 Hz alpha power (Pow) according to the formula: (Pow_attend-left – Pow_attend-right) / (Pow_attend-left + Pow_attend-right). The colour bar applies to both, sensor- and source-level data. (C) The 45-degree plot shows the hemispheric difference in alpha lateralization (LH – RH) for trials with neutral temporal cues (green, x-axis) versus instructive temporal cues (orange, y-axis). Grey dots correspond to N = 20 participants. Bars and error bars in the inset show average ±1 between-subject SEM of the hemispheric difference in alpha lateralization for neutral and instructive temporal cue trials. ** p < 0.01; *** p < 0.001; n.s. not significant.

Fig. 3

(A) Grand-average 1–5 Hz inter-trial phase coherence (ITPC), averaged across all (102) gradiometer sensors and N = 20 participants, for trials with neutral (green) and instructive temporal cues (orange). Black triangles indicate onsets of individual spoken numbers. The inset shows single-subject 0.8-Hz phase angles and resultant vectors of ITPC during number presentation (1.5–7.5 s). (B) Average spectral amplitude of ITPC during number presentation. Shaded areas show ±1 between-subject SEM. (C) Source localization of the amplitude of 0.8-Hz ITPC modulation, averaged across neutral and instructive temporal cue conditions. (D) The 45-degree plot shows 0.8-Hz ITPC amplitude (averaged across frequencies 0.76–0.84 Hz) for trials with neutral (green, x-axis) versus instructive temporal cues (orange, y-axis). Bars and error bars in the inset show average ±1 between-subject SEM of 0.8-Hz ITPC amplitude. n.s. not significant. (E, F, H) Same as (A, B, D) for alpha lateralization, calculated for 8–12 Hz alpha power (Pow) at sensors ipsilateral (ipsi) versus contralateral (contra) relative to the spatial focus of attention, using the formula: (Pow_ipsi – Pow_contra) / (Pow_ipsi + Pow_contra). * p < 0.05. Horizontal black bars in E indicate time points for which overall alpha lateralization (averaged across neutral and instructive temporal cue trials) differed significantly from zero (multiple t-tests; p < 0.05; uncorrected). (G) Source localization of the 0.8-Hz modulated amplitude of alpha lateralization, for instructive versus neutral temporal cue trials (z-values masked in the range: –1.96 < z < 1.96). Since the source analysis of rhythmically modulated alpha lateralization results in symmetric results in the left and right hemisphere (see Materials and Methods for details), only the left hemisphere is shown here from the outside (top) and from the inside (bottom).

Fig. 4

Scatterplots show relation of 0.8-Hz amplitude of alpha lateralization for instructive versus neutral temporal cue conditions (x-axis) to rated temporal cue benefit (left panel; r_Spearman = 0.456; p_permutation = 0.044) and temporal cue usage (right panel; r_Spearman = 0.425; p_permutation = 0.065).

Fig. 5

(A) In each trial with an instructive temporal cue, there were four temporally uncued and one cued number positions. (B) In each trial with a neutral temporal cue, all five number positions can be considered neutral. Note that the first and last number positions are shaded in A&B, since these number positions were not considered for the present analysis (see Materials and Methods for details). (C) Top: Temporal alpha lateralization index for cued number positions (blue) and uncued number positions (pink; calculated by sub-sampling the same quantity of uncued MEG epochs as there were cued epochs available for each participant, number position, and side of attention; see Materials and Methods for details). Shaded areas show the 2.5^th and 97.5^th percentiles of alpha lateralization across 10,000 sub-samples of uncued number positions. Since no sub-sampling was required for cued number positions (see Materials and Methods for details), no error bars are shown. Bottom: Solid line and shaded area show the average across single-subject Z-values to quantify alpha lateralization for cued versus uncued number positions and 95-% bootstrap confidence interval, respectively. P-values (indicated in color) were derived by testing Z-values of N=20 participants against zero using multiple non-parametric permutation tests. (D) Same as C but for the comparison of cued versus neutral number positions.