Eye movements disrupt EEG alpha-band coding of behaviorally relevant and irrelevant spatial locations held in working memory

Abstract

Oscillations in the alpha frequency band (∼8-12 Hz) of the human electroencephalogram play an important role in supporting selective attention to visual items and maintaining their spatial locations in working memory (WM). Recent findings suggest that spatial information maintained in alpha is modulated by interruptions to continuous visual input, such that attention shifts, eye closure, and backward masking of the encoded item cause reconstructed representations of remembered locations to become degraded. Here, we investigated how another common visual disruption-eye movements-modulates reconstructions of behaviorally relevant and irrelevant item locations held in WM. Participants completed a delayed estimation task, where they encoded and recalled either the location or color of an object after a brief retention period. During retention, participants either fixated at the center or executed a sequence of eye movements. Electroencephalography (EEG) was recorded at the scalp and eye position was monitored with an eye tracker. Inverted encoding modeling (IEM) was applied to reconstruct location-selective responses across multiple frequency bands during encoding and retention. Location-selective responses were successfully reconstructed from alpha activity during retention where participants fixated at the center, but these reconstructions were disrupted during eye movements. Recall performance decreased during eye-movements conditions but remained largely intact, and further analyses revealed that under specific task conditions, it was possible to reconstruct retained location information from lower frequency bands (1-4 Hz) during eye movements. These results suggest that eye movements disrupt maintained spatial information in alpha in a manner consistent with other acute interruptions to continuous visual input, but this information may be represented in other frequency bands.NEW & NOTEWORTHY Neural oscillations in the alpha frequency band support selective attention to visual items and maintenance of their spatial locations in human working memory. Here, we investigate how eye movements disrupt representations of item locations held in working memory. Although it was not possible to recover item locations from alpha during eye movements, retained location information could be recovered from select lower frequency bands. This suggests that during eye movements, stored spatial information may be represented in other frequencies.

Keywords: alpha; eye movements; inverted encoding model; spatial representations; working memory.

Graphical abstract

Figure 1.

Delayed estimation task (experiment 1). Schematic examples of the trial procedure for each of the four conditions. Each trial began with the participant aligning a green dot reflecting their current eye position with the dark gray fixation dot and pressing the space bar once fixation was achieved. The green dot then disappeared and after a brief delay (jittered between 0.6 and 1.5 s) the probe stimulus was presented (0.25 s). The participant was required to retain either the location (“spatial” conditions) or the color (“color” conditions) of the probe throughout the 1.75 s retention period. In the “fixate” conditions, the participant maintained gaze at the center throughout the entire retention period. In the “move” conditions, the participant was required to track the position of the dot as it moved to the first location (eye-movements cue 1, EM1 Cue), the second location (eye-movements cue 2, EM2 Cue) and then returned to fixation (RF Cue). At the end of the retention period, the participant either reported the recalled probe spatial location or color by clicking on a response wheel. Conditions were blocked and order was fully counterbalanced across participants.

Figure 2.

Delayed estimation task (experiment 2). Schematic examples of the trial procedure for each of the four conditions. The task stimuli and procedure were identical to experiment 1 except during the retention period participants were just required to make one eye movements away from fixation [eye-movements cue 1 (EM1) Cue] and then return gaze to fixation [return to fixation (RF) Cue].

Figure 3.

Behavior (experiment 1). Precision (SD in degrees from true location) (A) and guess rate (proportion of guess trials, i.e., where there was no memory; B) are shown for all conditions. In this and subsequent figures, the conditions are labeled as follows: S/F, Spatial/Fixate; C/F, Color/Fixate; S/M, Spatial/Move; C/M, Color/Move. Error bars = means ± SE.

Figure 4.

Alpha lateralization (experiment 1). A: topographic plots depict the distribution of alpha power across the scalp, averaged across trials and participants. Columns represent the four conditions. Rows represent the final 50 ms of stimulus presentation (0.2–0.25 s) and the retention period divided into three 0.5-s time windows that correspond to the three eye-movements windows in the “move” conditions [eye-movements cue 1 (EM1), eye-movements cue 2 (EM2), and return to fixation]. Note that the “fixate” conditions are divided into the same time windows so that alpha topography and lateralization can be compared between “fixate” and “move” conditions. The position of each head plot in the ring of plots reflects the mean activity across all trials where the stimulus was presented in the corresponding angular location bin. B: bar plots depict alpha lateralization indices for all conditions and time windows. Each row corresponds to the time window and topographic plots denoted in A. Higher values reflect greater lateralization. *Lateralization index significantly different to zero (Bayes factors, BF > 3). Error bars = means ±SE. S/F, Spatial/Fixate; C/F, Color/Fixate; S/M, Spatial/Move; C/M, Color/Move.

Figure 5.

Modeling total alpha power (experiment 1). A: estimated channel response functions (CRFs) presented as heatmaps. B: CRF slopes, computed by folding each CRF at the peak channel offset at each time point and computing linear slope. Horizontal bars at base of plot indicate time points where there is moderate evidence for a statistical difference between each real CRF and its permuted baseline (Bayes factors, BF > 3). C: actual BF values for tests comparing each real CRF and its permuted baseline. D: BF ANOVA model results comparing the “real” slope data across conditions (horizontal lower and upper dashed lines indicate BF > 3 and BF > 10, respectively). E: BF pairwise comparisons comparing the “real” slope data (horizontal lower and upper dashed lines indicate BF > 3 and BF > 10, respectively). Dashed vertical lines in all plots indicate stimulus onset (On) and offset (Off) and the three eye-movements cues [eye-movements cue 1 (EM1), eye-movements cue 2 (EM2), and return to fixation (RF)].

Figure 6.

Temporal generalization of CRFs (experiment 1). The linear slopes of the channel response functions (CRFs) are plotted for each training and testing time-point, with greater slopes represented as hotter colors. Time-points where there is no statistical difference between real and permuted slope (Bayes factors, BF < 3) are plotted as uniform dark blue. Training and testing within each time-point [as per the previous inverted encoding modeling (IEM) analysis] is represented by activation along the diagonal stretching from the top-left to the bottom-right corners of each plot. Off-diagonal activations represent encoding endurance both backward in time (activations to the left of the diagonal) and forwards in time (activations to the right of the diagonal). Dashed horizontal and vertical lines indicate stimulus onset (On) and offset (Off) and the three eye-movements cues [eye-movements cue 1 (EM1), eye-movements cue 2 (EM2), and return to fixation (RF)].

Figure 7.

Tracking location-selective representations across a broad range of frequency bands for total (A) and evoked power (experiment 1) (B). Channel response functions (CRFs) slopes are plotted for each time sample across frequency bands 1–30 Hz. For each sample the slope of the real CRF is compared with the permuted baseline and if the resulting BF < 3 then the point is represented in uniform dark blue. Dashed vertical lines indicate stimulus onset (On) and offset (Off) and the three eye-movements cues [eye-movements cue 1 (EM1), eye-movements cue 2 (EM2), and return to fixation (RF)].

Figure 8.

Eye-movements analysis (experiment 1). A: Euclidian error (minimum distance between eye-gaze cue and actual eye gaze) for the first and second eye-movements windows [eye-movements cue 1 (EM1) and eye-movements cue 2 (EM2), respectively] for both “move” conditions. B: time during each eye-movements window where actual eye gaze was closest to gaze cue. Error bars = means ±SE.

Figure 9.

Behavior (experiment 2). Precision (SD in degrees from true location) (A) and guess rate (B, proportion of guess trials, i.e., where there was no memory) are shown for all conditions. In this and subsequent figures, the conditions are labeled as follows: S/F, Spatial/Fixate; C/F, Color/Fixate; S/M, Spatial/Move; C/M, Color/Move. Error bars = means ± SE.

Figure 10.

Alpha lateralization (experiment 2). A: topographic plots depict the distribution of alpha power across the scalp, averaged across trials and participants. Columns represent the four conditions. Rows represent the final 50 ms of stimulus presentation (0.2–0.25 s) and the retention period divided into the Eye-Movements Window 1 (0.5–1.5 s) and the Return to Fixation window (1.5–2.0 s). Note that the “fixate” conditions are divided into the same time windows so that alpha topography and lateralization can be compared between “fixate” and “move” conditions. The position of each head plot in the ring of plots reflects the mean activity across all trials where the stimulus was presented in the corresponding angular location bin. B: bar plots depict α lateralization indices for all conditions and time windows. Each row corresponds to the time window and topographic plots denoted in A. Higher values reflect greater lateralization. *Lateralization index significantly different to zero (Bayes factors, BF > 3). S/F, Spatial/Fixate; C/F, Color/Fixate; S/M, Spatial/Move; C/M, Color/Move. Error bars = means ± SE.

Figure 11.

Modeling total alpha power (experiment 2). A: estimated channel response functions (CRFs) presented as heatmaps. B: CRF slopes, computed by folding each CRF at the peak channel offset at each time point and computing linear slope. Horizontal bars at base of plot indicate time points where there is moderate evidence for a statistical difference between each real CRF and its permuted baseline (Bayes factors, BF > 3). C: actual BF values for tests comparing each real CRF and its permuted baseline. D: BF ANOVA model results comparing the “real” slope data across conditions (horizontal dashed lines indicate BF > 3 and BF > 10, respectively). E: BF pairwise comparisons comparing the “real” slope data (horizontal lower and upper dashed lines indicate BF > 3 and BF > 10, respectively). Dashed vertical lines in all plots indicate stimulus onset (On) and offset (Off) and the two eye-movements cues [eye-movements cue 1 (EM1) and return to fixation (RF)].

Figure 12.

Temporal generalization of channel response functions (CRFs) (experiment 2). The linear slopes of the CRFs are plotted for each training and testing time-point, with greater slopes represented as hotter colors. Time-points where there is no statistical difference between real and permuted slope (Bayes factors, BF < 3) are plotted as uniform dark blue. Training and testing within each time-point [as per the previous inverted encoding modeling (IEM) analysis] is represented by activation along the diagonal stretching from the top-left to the bottom-right corners of each plot. Off-diagonal activations represent encoding endurance both backward in time (activations to the left of the diagonal) and forward in time (activations to the right of the diagonal). Dashed horizontal and vertical lines indicate stimulus onset (On) and offset (Off) and the two eye-movements cues [EM and return to fixation (RF)].

Figure 13.

Tracking location-selective representations across a broad range of frequency bands for total (A) and evoked power (experiment 2) (B). Channel response functions (CRFs) slopes are plotted for each time sample across frequency bands 1–30 Hz. For each sample, the slope of the real CRF is compared with the permuted baseline and if the resulting Bayes factors (BF) < 3 then the point is represented in uniform dark blue. Dashed vertical lines indicate stimulus onset (On) and offset (Off) and the three eye-movements cues [eye-movements cue 1 (EM1), eye-movements cue 2 (EM2), and return to fixation (RF)].

Figure 14.

Eye-movements analysis (experiment 2). Euclidian error (minimum distance between eye-gaze cue and actual eye gaze) throughout the trial for both “move” conditions. Error bars = means ± SE. EM, eye movement; RF, return to fixation.