Mechanisms of Working Memory Prioritization: Retro-cues vs. Saccadic preparation

Selective internal attention allows working memory (WM) contents to be prioritized according to task demands, a process typically investigated with the retro-cue paradigm. In this paradigm, a cue presented during WM maintenance indicates which item will be tested, enabling voluntary allocation of attentional resources and typically enhancing memory performance. Recent findings suggest, however, that saccade preparation toward the location of a WM item can also produce an involuntary prioritization effect. In the present study, we compared behavioral performance and neural oscillatory dynamics associated with involuntary prioritization induced by saccade preparation and voluntary prioritization elicited by spatial and color retro- cues. Behaviorally, saccade-based prioritization produced only a modest benefit relative to incongruent items, whereas both types of retro-cues yielded substantially larger improvements. At the neural level, alpha-band lateralization showed a similar spatial-attention pattern across all prioritization conditions, supporting a shared covert attentional mechanism. In contrast, bilateral posterior alpha power was reduced following retro-cues compared to saccadic preparation, consistent with a release of WM load when non-cued items are deprioritized. Overall, our results indicate that voluntary and involuntary mechanisms of WM prioritization rely on overlapping spatial-attention processes, but differ in strategic control and their consequences for load management. Importantly, top-down prioritization through retro-cues produces a stronger behavioral enhancement than involuntary prioritization driven by saccade preparation.

Working memory (WM) encompasses processes responsible for the retention and temporal manipulation of information that is relevant to future behavior (Gazzaley & Nobre, 2012). Internal selective attention facilitates flexible behavior by prioritizing WM representations that are imminently relevant for the upcoming action (Itti & Koch, 2000). Most of the empirical evidence supporting flexible WM representations has been obtained using the retro-cue paradigm (Griffin & Nobre, 2003). In these experiments, the relevance of the maintained content is biased by providing an informative cue once the perceptual array is no longer present. Retrospective cues indicate which content is relevant for the test stage, enabling participants to strategically use this information to optimize their performance. Although memory benefits have been observed at the item and the dimensional level (Hajonides et al., 2020; Heuer & Schub¨o, 2016), most of the research has used retro-cues linked to object representations. There are numerous ways in which the relevance of an item can be retrospectively cued. For example, by providing an arrow spatial cue that explicitly indicates the position of the object to be probed, or by using a feature cue, such as color or shape, which acts as an indicator of the relevant item that matches the feature of the cue (Heuer et al., 2016; Heuer & Rolfs, 2021; Kuo et al., 2009; Panichello & Buschman, 2021; Van Ede, 2020). A robust neurophysiological correlate of item-based prioritization is the retinotopic alpha modulation in occipito-parietal regions, which has been extensively observed following spatial cues, and to a lesser extent, after feature retro-cues (Fu et al., 2022; Poch et al., 1 Posted on 18 Dec 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.176604315.58106628/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. 2017; Ursino, 2025; Zhou et al., 2025). This neural signature emphasizes the role of spatial attention in modulating internal representations, even when the cue itself is not spatial (Liu et al., 2024; Myers et al., 2017; Pertzov et al., 2013; Snyder & Foxe, 2010; Van Diepen et al., 2016). According to the premotor theory, item prioritization would be achieved through the sensory biases that the oculomotor system exerts on internal representations, similar to the perceptual benefit for features of upcoming saccade targets in the perceptual domain (Moore & Egeth, 1998; Moore et al., 2003; Van Ede et al., 2019). In fact, a recent study found a gaze bias associated with the retro-cued representation, evidencing the implication of the oculomotor system in WM prioritization (Van Ede et al., 2019). Furthermore, several studies show that, in the absence of explicit retro-cues, eye movements programmed during the delay period can determine the contents of visual WM, even when the saccade cue lacked predictive value (Gersch et al., 2008; Ohl & Rolfs, 2017; Rajsic et al., 2017). It has been proposed that it is the oculomotor selection of task-relevant locations, rather than task relevance itself, that impacts WM maintenance. A clever experiment found dissociated effects of task relevance, oculomotor selection and saccade execution in WM performance (Hanning et al., 2016). The cited study observed a behavioral benefit for the memory item that matched the saccade preparation target, but not for the executed saccade target itself nor for the item that was relevant to perform the task but was not a saccade target (Hanning & Deubel, 2018). These findings emphasize the overlapping mechanisms that operate in spatial WM and saccade selection and support the proposal that the selection of WM maintained representations is mediated through oculomotor selection. In contrast to external retro-cues, which provide task-relevant information that enables a voluntary allocati- on of attentional resources to enhance performance, the memory bias produced by saccade preparation arises from oculomotor selection and constitutes an involuntary form of prioritization. Although it has been ques- tioned whether these forms of prioritization are functionally equivalent (Heuer et al., 2020), to our knowledge no study has directly compared voluntary and involuntary sources of WM selection. In the present work, we directly contrasted behavioral performance and oscillatory neural correlates associated with voluntary and involuntary prioritization in WM. Participants performed two voluntary retro-cue tasks—one spatial and one feature-based—and a saccadic preparation task. The spatial retro-cue was included to closely match the perceptual properties of the cue in the saccade condition, enabling a comparison of alpha power modulations. However, because arrow cues may also induce exogenous attentional shifts due to their perceptual salience, potentially confounding alpha responses, we additionally included a color retro-cue, which is considered a purely endogenous cue (Hommel et al., 2001). Although both spatial and feature retro-cues can enhance memory and modulate alpha oscillations, a direct comparison of their neural signatures—and how these relate to saccade-based prioritization—had not yet been conducted.

Participants The present experiment was conducted with a sample of 36 adult participants (mean age, 24.33 years; standard deviation, 4.76; range, 18-38 years), consisting of 20 females and 16 males. All were right-handed according to the Edinburgh Handedness Inventory (Oldfield, 1971). The sample size was determined using G*Power, aiming to detect an effect size of 0.25 and to achieve a statistical power of 0.9. This indicated a minimum sample size of 30 participants (Faul et al., 2007; Macedo-Pascual et al., 2023). Participants signed an informed consent form in accordance with the guidelines of the Declaration of Helsinki. Exclusion criteria included a history of neurological or psychiatric disorders, family history of epilepsy, and use of psychotropic drugs prior to the experiment. Experimental task The experimental tasks are illustrated in Figure 1. Each of the three different conditions was presented in randomized blocks. The structure of the three conditions, based on a retro-cue paradigm, was the same, but differed in the instructions given between the saccadic and the spatial retro-cue conditions, and the physical appearance of the retro-cue between the color and the spatial retro-cue conditions. In the saccadic condition, participants were instructed to prepare but not execute a saccade in the direction of the cue during the 2 Posted on 18 Dec 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.176604315.58106628/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. delay period. In both the spatial and the color retro-cue conditions, participants were told that the retro-cue predicted the probe item. The experiment consisted of a total of 420 trials, distributed in three blocks of 140 trials corresponding to each experimental condition. At the beginning of each block, participants were informed about the characteristics of the stimuli and the task to be performed. The visual stimuli were presented on a screen located at a distance of 60 cm from the participants, ensuring uniformity in the size of the perceived stimuli. The programming and execution of the experiment were carried out using the open- source Psychtoolbox-3 library in MATLAB (Kleiner et al., 2007). Each trial began with the presentation of a fixation point for 1200 ms. Next, a memory set was displayed for 250 ms, consisting of four rectangles with specific orientations that participants had to memorize (0 °, 45°, 90°, 135°) and located within a visual angle of 3.8° (Macedo-Pascual et al., 2023). After a maintenance period of 1200 ms, a cue was presented for 200 ms indicating the location/color of the test stimulus in the retro-cue blocks or the direction towards which they had to prepare an eye movement in the saccadic preparation blocks. After a 1200 ms delay period participants were asked to make an orientation judgement on one of the four rectangles in the saccadic conditions, and on the cued item in the retro-cue conditions by pressing a key on the keyboard. The test stimulus was presented for 2000 ms, after which participants had 2000 ms to perform a saccade in the saccadic condition. While in the retro-cue conditions the cue was 100% predictive, in the saccadic conditions the saccadic cue was only congruent with the test item in 25% of the trials. Prioritization comparison between tasks was performed taking into account only congruent trials. In all conditions, the orientation of the test rectangle matched that of the sample rectangle in 50% of the trials, while in the remaining 50% the orientation was randomized. In addition, there was a 25% probability that the retro-cue would point in any of the directions of the presented rectangles. Fig. 1. Schematic illustration of the experimental task. EEG recording The brain electrical signal was recorded using an ActiChamp system (BrainProducts), equipped with 64 active channels. In addition, electrooculographic (EOG) electrodes were placed to record eye activity, with one horizontal and one vertical electrode, and an external reference electrode was used at the tip of the nose. The data was initially sampled at a rate of 1000 Hz, and a low-pass filter with a cutoff frequency of 100 Hz was applied during acquisition. Subsequently, the data was referenced offline to the tip of the nose and the sampling rate was reduced to 250 Hz using MATLAB and the FieldTrip toolbox (Oostenveld et al., 2011; www.fieldtriptoolbox.org). Both preprocessing and subsequent analyses of the electrophysiological signals were carried out using the same FieldTrip toolbox. Preprocessing and time-frequency analysis EEG data were preprocessed before analysis, working exclusively with artefact-free data. First, the data were 3 Posted on 18 Dec 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.176604315.58106628/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. epoched in segments from -2.65 to 2.75 seconds around retro-cue presentation. For the removal of artefacts related to blinks and eye movements (vertical and horizontal), an Independent Component Analysis (ICA) was applied using EEGLAB’s ‘runica’ algorithm, implemented in FieldTrip. Subsequently, individual epochs were visually inspected, discarding those containing remaining artefacts. Channels with excessive noise were identified and interpolated. Oscillatory activity in the alpha band was obtained for each trial by Hilbert transform. Initially, the preprocessed data were filtered in the 8-14 Hz range, and then the time course of the spectral amplitude of the signal was extracted by calculating the absolute value of the Hilbert transform. Lateralized alpha activity was determined by combining the electrodes from the left retro-cue condition with the mirrored version of the electrodes from the right condition. In this way, contralateral activity was represented at the right electrodes by averaging the right electrodes from the left condition with the left electrodes from the right condition. Similarly, ipsilateral activity was plotted at the left electrodes. Statistical analysis Working memory accuracy was assessed using repeated-measures one way ANOVA taking the total percent- age of hits on a standardized scale from 0 to 1. Prioritization differences were assessed between retro-cues conditions and saccadic cue condition taking into account only congruent saccadic trials. In this way, we compared performance on the valid cue trials. To evaluate a prioritization effect within the saccadic cue condition, we ran a t-test between congruent and incongruent saccadic trials. Experimental conditions were compared with a statistical significance threshold set at α = 0.05. Statistical analysis of alpha oscillatory activity was performed using a non-parametric cluster-based permutation analysis implemented in FieldTrip (Maris & Oostenveld, 2007), designed to control for type I error arising from multiple comparisons. Ini- tially, a parametric statistical test was applied to each electrode-time pair. Subsequently, clusters formed by electrode-time pairs significantly adjacent in temporal or spatial dimensions were identified. For each identified cluster, a cluster statistic was calculated by summing the parametric statistical values within the cluster. Statistical significance was assessed by comparing this statistic with a null distribution, drawn from 10,000 permutations. During each permutation, data were randomly assigned to two subsets, recalculating the maximum cluster statistic for each iteration. The analysis was performed in a 1400 ms time window following cue presentation. At the final stage, the cluster p-value was determined by constructing a his- togram of the cluster statistics obtained from the 10,000 permutations, and calculating the proportion of permutations whose statistics exceeded the value observed in the original data.

Memory performance Memory accuracy was modulated by prioritization condition (F (2,35) = 33.34, p < .001; η² = 0.488). While item prioritization did not behaviorally differ between the two voluntary prioritization conditions, spatial and color retro-cue conditions, ( t (35) = -0.44, p = 0.663; Mdiff = -0.004), they both differed from the congruent saccadic condition (t (35) = 6.35, p < .001; Mdiff = 0.110; t (35) = 6.01, p < .001; Mdiff = 0.106 for spatial and color retro-cue, respectively) (Figure 2A). Although performance in congruent saccadic trials and retro-cue trials was not equivalent, we did find a subtle memory advantage for congruent compared to incongruent trials in the saccadic condition ( t (35) = 2.14, p = .04; d’ = 0.35) (Figure 2B). The same pattern of results was obtained for reaction times (RT). RT differed across conditions (F (2, 35) = 141.48, p < .001; η² = 0.802) (Figure 2C). No significant differences in RT were observed between the spatial and color retro-cue conditions ( t (35) = -1.58, p = 0.124; Mdiff = -0.027). However, both conditions were performed significantly faster than the saccadic condition (t (35) = -12.32, p < .001; Mdiff = -0.327; t (35) = -13.98, p < .001; Mdiff = -0.354 for the spatial and color retro-cue conditions, respectively). In the saccadic condition we obtained a robust difference between saccadic congruent trials compared to incongruent trials (t (35) = -5.05, p < .001; d’ = -0.84) (Figure 2D). 4 Posted on 18 Dec 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.176604315.58106628/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. Fig. 2. Raincloud plots depicting behavioral performance and reaction times across conditions. Each subplot displays individual data points (cloud of dots), a boxplot, and a one-sided violin plot to illustrate data distribution. A- Mean accuracy for color, spatial, and congruent saccadic trials. B- Mean accuracy for congruent and incongruent saccadic trials. C- Mean reaction times (ms) for color, spatial, and congruent saccadic trials.D- Mean reaction times (ms) for congruent and incongruent saccadic trials. EEG results After retro-cue presentation, alpha oscillatory activity was significantly lateralized in posterior electrodes in the three conditions. In the arrow retro-cue condition, alpha was significantly lateralized from 250 ms to 920 ms (p < .001), in the arrow saccadic condition from 250 to 910 (p .05) (Figure 3). Additionally, bilateral alpha power was compared between the two arrow conditions. After the arrow pre- sentation, alpha oscillatory activity was significantly higher in the saccadic condition (p < .001) than in the retro-cue condition. This effect was found over the whole scalp, although more pronounced in posterior- central regions, from 470 ms to 1100 ms (Figure 4). 5 Posted on 18 Dec 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.176604315.58106628/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. Fig. 3. Neural dynamics of alpha activity across experimental conditions. The top left display displays the time course of lateralized alpha activity for each condition. Pink shaded brackets highlight time windows with statistically significant lateralized alpha activity. The remaining panels show alpha lateralization as depicted by topo plots, computed as contralateral minus ipsilateral activity. Colored circles indicate electrodes forming the significant cluster. Fig. 4. Bilateral alpha activity. On the left: Time course of bilateral alpha activity from the electrodes comprising the significant cluster. The pink shaded area highlights the time interval in which statistically significant differences were observed between the saccadic and the spatial retro-cue task. On the right: Topo plot illustrating the difference in bilateral alpha activity between the saccadic and spatial cue conditions. Alpha activity was averaged within the time window corresponding to the significant cluster.

Prioritization of working-memory (WM) contents can arise from voluntary, strategic control of internal at- tention, but also from involuntary processes such as automatic biases induced by saccade planning. In the present study, we compared behavioral performance and alpha oscillatory activity associated with voluntary and involuntary WM prioritization, elicited by two types of retro-cues and a saccadic cue, respectively. Be- haviorally, voluntary prioritization yielded similar benefits in the color and spatial retro-cue conditions, and both produced a markedly stronger enhancement than the congruent trials in the saccadic preparation condi- tion. Nonetheless, within the saccadic condition, items congruent with the planned saccade showed a modest performance advantage over incongruent items. At the neural level, alpha-band markers of covert attention revealed a comparable lateralization pattern across all three prioritization conditions. However, bilateral posterior alpha power was reduced for the spatial retro-cue relative to the saccade preparation condition. In the following, we will argue that these findings suggest that voluntary and involuntary prioritization may rely on shared covert attentional mechanisms, while voluntary selection likely recruits additional strategic resources that can down-regulate non-cued representations (Pertzov et al., 2013; Souza et al., 2014). In this study, we provide further evidence for a bias toward the representation located at the saccade target (congruent location), relative to the incongruent location (Ohl et al., 2024; Ohl & Rolfs, 2018). Previous research has suggested that saccade-based prioritization may rely on mechanisms distinct from those underlying retro-cue benefits. This proposal is motivated by findings showing that, when memory load or delay is manipulated in saccadic paradigms, the resulting effects deviate from those expected under retro-cue conditions (Ohl & Rolfs, 2017, 2018, 2020). Our data contribute to this literature by showing that the magnitude of facilitation is not comparable between voluntary and involuntary prioritization. A number of factors inherent to the task structure and demands could underlie the behavioral discrepancy between 6 Posted on 18 Dec 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.176604315.58106628/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. saccade-based prioritization and retro-cue conditions. First, the comparatively smaller behavioral benefit in the saccadic-cue condition may be related to dual-task demands. Although we did not include a retro-cue neutral baseline condition, due to time limitations, we have repeatedly tested this paradigm with the same set of stimuli and structure obtaining a robust retro-cue benefit greater than 10% in all cases (Macedo- Pascual et al., 2022, 2023; Poch et al., 2014). In contrast, here the average benefit between congruent and incongruent targets was only about 4%, comparable to the saccadic benefit reported by Hanning and Deubel (2018). In a previous study we used the same structure and stimuli in a dual retro-cue task, in which after the retro-cue participants had to do detect a high contrast or a low contrast stimulus. We found that the high demanding detection condition impaired performance in the neutral condition, but did not affect the retro-cued item. The dual demands of both experiments could be considered similar as they recruit visual attention resources not related to the memory task itself. Based on those findings, if the voluntary and involuntary prioritization were based on similar mechanisms and the only factor accounting for behavioral differences was the dual task demand, we would not have found a worsened performance in the congruent saccadic trials compared to the retro-cued items. Second, one might argue that participants were not sufficiently encouraged to prepare the saccade until immediately before execution and thus may not have actively prepared it. In such a case, the facilitatory effect could reflect the exogenous component of the cue or memory of its location rather than genuine oculomotor preparation. However, in a similar paradigm, Hanning et al (2016) found that benefits emerged only in blocks where the cue indicated a saccade target, but not when participants were required to maintain fixation— ruling out bottom-up explanations—or during antisaccade blocks, suggesting that merely maintaining the location was insufficient to interact with WM representations. In our study, the benefit is of a comparable magnitude as the one found in Hanning et al (2016), and, as we will discuss later, we found neural markers of preparatory attention following the saccadic cue, which would support the notion that participants were indeed preparing the saccade. Third, and perhaps most importantly, we believe that cue validity is the key factor to interpreting the present findings. The retro-cue was valid on 100% of trials, enabling participants to adopt a voluntary, strategic allocation of resources to maximize performance, whereas the saccade target carried no predictive value. Cue reliability has previously been shown to modulate retro-cueing (Gunseli et al., 2015; Liu et al., 2024). Although retro-cue benefits can emerge even when validity is low (Berryhill et al., 2012; Dube et al., 2019; Gunseli et al., 2015), high-reliability cues generally produce stronger enhancement of the cued item. This enhancement appears to be driven by validity-related costs: when reliability is high, non-cued items are dropped from active maintenance, whereas under low reliability they are still preserved. Thus, the retro- cue effect likely reflects multiple cognitive mechanisms that vary depending on the strategic allocation of resources. In line with this, our saccadic cue was non-predictive, and participants likely maintained all items voluntarily. Although oculomotor selection automatically prioritized the saccade-congruent item relative to incongruent items, the concurrent maintenance of all representations attenuated the prioritization effect compared to the retro-cue conditions. Although there are reasons to believe that saccades influence working-memory content through mechanisms distinct from endogenous cueing, our results provide evidence that saccade-based prioritization may rely on attentional networks similar to those supporting voluntary prioritization. First, covert and overt attention appear to rely on closely related neural mechanisms. Preparatory saccade activity modulates visual process- ing across early visual cortex (e.g., T. Moore et al., 1998; Saber et al., 2015), enhancing perceptual sensitivity at the movement target even before eye displacement occurs—an effect thought to underlie presaccadic at- tention shifts (Rolfs & Carrasco, 2012). Likewise, covert attention can operate within representational space to select relevant items maintained in WM. In this regard, alpha-band lateralization is a well-established neural signature of spatial attention allocation in both perception and memory (Capilla et al., 2014; Poch et al., 2014; Thut et al., 2006; Worden et al., 2000), and has been linked to oculomotor control (Popov et al., 2021). Consistent with this view, covert shifts within mnemonic space appear to be driven by the oculomotor system (Van Ede et al., 2019), yielding a gaze-bias toward prioritized items. Conversely, oculo- motor selection during WM maintenance can bias performance even when task relevance is absent (Hanning 7 Posted on 18 Dec 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.176604315.58106628/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. & Deubel, 2018). In our data, we observed alpha lateralization indexing covert attention shifts in both retro-cue conditions—associated with selection of the cued representation—and in the saccadic condition, where it was aligned with the saccade target. This convergence suggests that both voluntary and involuntary prioritization recruit covert spatial attention and supports the notion that saccade preparation necessarily involves a preceding covert attentional shift. We also found increased alpha power over posterior-central regions in the saccade condition relative to retro- cueing. One possibility is that this modulation reflects processes specific to saccade preparation. However, prior work generally reports alpha desynchronization during the delay between cue and eye movement (Brig- nani et al., 2007; Van Noordt et al., 2017), which makes an increase in alpha during saccade preparation somewhat counterintuitive. An alternative interpretation is that alpha scaling reflects differences in memory load or maintenance demands. Alpha delay activity is known to scale with WM load (Heinz & Johnson, 2017; Macedo-Pascual et al., 2019; Poch et al., 2018; Schroeder et al., 2018; Tuladhar et al., 2007) and has been associated with internally directed attention or increased cognitive demands (Benedek et al., 2014; Cooper et al., 2006; Palva & Palva, 2007; Poch et al., 2017a; Ray & Cole, 1985). As discussed above, the high cue reliability in the retro-cue conditions likely encouraged participants to drop non-relevant items from active maintenance, whereas the non-predictive nature of the saccade target required maintaining all items, potentially explaining the higher alpha power in the saccadic condition (Macedo-Pascual et al., 2023; Poch et al., 2018). Having discussed saccade-related prioritization and its overlap with voluntary attentional mechanisms, we now turn to a different question: whether spatial and color retro-cues enhance memory through the same processes. The key difference between these two types of cues lies in the bottom-up properties of the arrow cue that can additionally induce automatic shifts of spatial attention, while color or feature cues rely on purely endogenous processes. Although there is previous evidence that a feature retro-cue can lead to an item memory advantage similar to the more commonly used spatial retro-cue (Lepsien & Nobre, 2007; Li & Saiki, 2015; Pertzov et al., 2013), some authors have found null effects after other than arrow retro-cues (Berryhill et al., 2012), or more subtle retro-cue benefits after a feature retro-cue (Heuer et al., 2016). These differential effects are believed to have arisen due to methodological aspects such as the short delay between the encoding and the retro-cue. It has been hypothesized that feature retro-cues, such as shape and color, might require additional processing demands taking a longer time to be equally effective, as it has been demonstrated in the perceptual domain (T. Liu et al., 2007). Additional processing might be related to the spatial recoding of the cue in order to apply an internal spatial attentional mechanism as in the spatial retro-cue condition (Pertzov et al., 2013; Poch et al., 2017b). In our study, we used a one-second interval between the encoding and the retro-cue, which has been proven long enough to allow for a spatial recoding of the cue, leading to equivalent behavioral benefits. Regarding the neural mechanisms triggered by arrow and color retro-cues, both induced a similar shift of covert spatial attention; however, alpha lateralization modulation began around 100 ms later in the color condition, providing further evidence that additional resources are involved in the spatial recoding of the color cue (Keefe & St ¨ormer, 2021; Pertzov et al., 2013; Poch et al., 2017b; Souza & Oberauer, 2016). In sum, our findings indicate that both voluntary and involuntary prioritization can bias working-memory representations, although to different behavioral extents. While retro-cues produced a robust enhancement of the cued item, saccade-based prioritization yielded a smaller benefit—likely due to differences in cue reliability and strategic resource allocation. Nevertheless, converging neural evidence suggests that both forms of prioritization recruit overlapping attentional mechanisms. Alpha-band lateralization emerged in both saccadic and retro-cue conditions, supporting the view that saccade preparation and retro-cued attention both rely on covert shifts of spatial attention. Differences in posterior alpha power further point to additional memory-maintenance demands in the saccade condition, where all items had to be preserved due to the non- predictive nature of the cue. Finally, our comparison of spatial and feature-based retro-cues reveals that both can lead to comparable behavioral enhancement when sufficient time is allowed for spatial recoding, although feature cues require longer processing, as reflected in delayed alpha-lateralization onset. Together, these results support the idea that distinct prioritization cues differ in their behavioral expression depending 8 Posted on 18 Dec 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.176604315.58106628/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. on task structure and strategic factors, yet converge at the neural level on a common spatial-attention mechanism that enhances prioritized representations in working memory.

Benedek, M., Schickel, R. J., Jauk, E., Fink, A., & Neubauer, A. C. (2014). Alpha power increa- ses in right parietal cortex reflects focused internal attention. Neuropsychologia , 56 , 393-400. htt- ps://doi.org/10.1016/j.neuropsychologia.2014.02.010Berryhill, M. E., Richmond, L. L., Shay, C. S., & Olson, I. R. (2012). Shifting Attention among Working Memory Representations: Testing Cue Type, Awareness, and Strategic Control. Quarterly Journal of Experimental Psychology ,65 (3), 426-438. htt- ps://doi.org/10.1080/17470218.2011.604786Brignani, D., Maioli, C., Maria Rossini, P., & Miniussi, C. (2007). Event-related power modulations of brain activity preceding visually guided saccades. Brain Re- search , 1136 , 122-131. https://doi.org/10.1016/j.brainres.2006.12.018Capilla, A., Schoffelen, J.-M., Pa- terson, G., Thut, G., & Gross, J. (2014). Dissociated α-Band Modulations in the Dorsal and Ven- tral Visual Pathways in Visuospatial Attention and Perception. Cerebral Cortex , 24 (2), 550-561. https://doi.org/10.1093/cercor/bhs343Cooper, N. R., Burgess, A. P., Croft, R. J., & Gruzelier, J. H. (2006). Investigating evoked and induced electroencephalogram activity in task-related alpha power increases dur- ing an internally directed attention task. Neuroreport , 17 (2), 205-208.Dube, B., Rak, S., Iannucci, L., & Al-Aidroos, N. (2019). Too little too late: No flexible control of memory by retro-cues. Journal of Vision , 19 (10), 310c. https://doi.org/10.1167/19.10.310cFaul, F., Erdfelder, E., Lang, A.-G., & Buch- ner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences.Behavior Research Methods , 39 (2), 175-191. https://doi.org/10.3758/BF03193146Fu, X., Ye, C., Hu, Z., Li, Z., Liang, T., & Liu, Q. (2022). The impact of retro-cue validity on work- ing memory representation: Evidence from electroencephalograms. Biological Psychology , 170 , 108320. https://doi.org/10.1016/j.biopsycho.2022.108320Gazzaley, A., & Nobre, A. C. (2012). Top-down modu- lation: Bridging selective attention and working memory. Trends in Cognitive Sciences , 16 (2), 129- 135. https://doi.org/10.1016/j.tics.2011.11.014Gersch, T. M., Kowler, E., Schnitzer, B. S., & Dosher, B. A. (2008). Visual memory during pauses between successive saccades. Journal of Vision ,8 (16), 15-15. https://doi.org/10.1167/8.16.15Griffin, I. C., & Nobre, A. C. (2003). Orienting Atten- tion to Locations in Internal Representations. Journal of Cognitive Neuroscience , 15 (8), 1176-1194. https://doi.org/10.1162/089892903322598139Gunseli, E., Fahrenfort, J., Daoultzis, K., Meeter, M., & Oliv- ers, C. (2015). The Loss of Information from Visual Working Memory depends on Retro-Cue Reliability. Journal of Vision , 15 (12), 91. https://doi.org/10.1167/15.12.91Hajonides, J. E., Van Ede, F., Stokes, M. G., & Nobre, A. C. (2020). Comparing the prioritization of items and feature-dimensions in visual working memory. Journal of Vision ,20 (8), 25. https://doi.org/10.1167/jov.20.8.25Hanning, N. M., & Deubel, H. (2018). Independent Effects of Eye and Hand Movements on Visual Working Memory. Frontiers in Systems Neuroscience ,12 , 37. https://doi.org/10.3389/fnsys.2018.00037Hanning, N. M., Jonikaitis, D., Deubel, H., & Szinte, M. (2016). Oculomotor selection underlies feature retention in visual working memory. Journal of Neurophysiology , 115 (2), 1071-1076. https://doi.org/10.1152/jn.00927.2015Heinz, A. J., & Johnson, J. S. (2017). Load-Dependent Increases in Delay-Period Alpha-Band Power Track the Gating of Task-Irrelevant Inputs to Working Memory. Frontiers in Human Neuroscience , 11 . https://doi.org/10.3389/fnhum.2017.00250Heuer, A., Ohl, S., & Rolfs, M. (2020). Memory for ac- tion: A functional view of selection in visual working memory. Visual Cognition , 28 (5-8), 388- 400. https://doi.org/10.1080/13506285.2020.1764156Heuer, A., & Rolfs, M. (2021). Incidental en- coding of visual information in temporal reference frames in working memory. Cognition , 207 , 104526. https://doi.org/10.1016/j.cognition.2020.104526Heuer, A., & Schub ¨o, A. (2016). Feature-based and spatial attentional selection in visual working memory. Memory & Cognition , 44 (4), 621-632. https://doi.org/10.3758/s13421-015-0584-5Heuer, A., Schub¨o, A., & Crawford, J. D. (2016). Different Corti- cal Mechanisms for Spatial vs. Feature-Based Attentional Selection in Visual Working Memory. Frontiers in Human Neuroscience , 10 . https://doi.org/10.3389/fnhum.2016.00415Hommel, B., Pratt, J., Colza- to, L., & Godijn, R. (2001). Symbolic Control of Visual Attention. Psychological Science , 12 (5), 360-365. https://doi.org/10.1111/1467-9280.00367Itti, L., & Koch, C. (2000). A saliency-based search 9 Posted on 18 Dec 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.176604315.58106628/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. mechanism for overt and covert shifts of visual attention. Vision Research , 40 (10-12), 1489-1506. https://doi.org/10.1016/S0042-6989(99)00163-7Keefe, J. M., & St¨ormer, V. S. (2021). Lateralized alpha acti- vity and slow potential shifts over visual cortex track the time course of both endogenous and exogenous ori- enting of attention. NeuroImage , 225 , 117495. https://doi.org/10.1016/j.neuroimage.2020.117495Kleiner, M., Brainard, D., & Pelli, D. (2007). What’s new in Psychtoolbox-3? Kuo, B.-C., Rao, A., Lepsien, J., & Nobre, A. C. (2009). Searching for Targets within the Spatial Layout of Visual Short-Term Memory. Journal of Neuroscience , 29 (25), 8032-8038. https://doi.org/10.1523/JNEUROSCI.0952-09.2009Lepsien, J., & Nobre, A. C. (2007). Attentional Modulation of Object Representations in Working Memory. Ce- rebral Cortex , 17 (9), 2072-2083. https://doi.org/10.1093/cercor/bhl116Li, Q., & Saiki, J. (2015). Dif- ferent effects of color-based and location-based selection on visual working memory. Attention, Per- ception, & Psychophysics ,77 (2), 450-463. https://doi.org/10.3758/s13414-014-0775-3Liu, Q., Guo, L., Wang, J., Nie, D., & Ye, C. (2024). The modulatory mechanism of spatial distance on the role of in- ternal attention in unattended working memory representations. Visual Cognition , 32 (8), 638-656. htt- ps://doi.org/10.1080/13506285.2025.2471454Liu, T., Stevens, S. T., & Carrasco, M. (2007). Comparing the time course and efficacy of spatial and feature-based attention. Vision Research ,47 (1), 108-113. https://doi.org/10.1016/j.visres.2006.09.017Macedo-Pascual, J., Campo, P., Capilla, A., Hinojosa, J. A., & Poch, C. (2022). Expectance of low contrast retro-cues does not modulate anticipatory alpha power.Biological Psychology , 173 , 108400. https://doi.org/10.1016/j.biopsycho.2022.108400Macedo-Pascual, J., Capilla, A., Campo, P., Hinojosa, J. A., & Poch, C. (2023). Selection within working memory impairs percep- tual detection. Psychonomic Bulletin & Review , 30 (4), 1442-1451. https://doi.org/10.3758/s13423-022- 02238-2Macedo-Pascual, J., Hinojosa, J. A., & Poch, C. (2019). Oscillatory correlates of prioritizati- on of emotional stimuli in WM: The interaction between bottom-up and top-down processes. Biological Psychology , 145 , 167-173. https://doi.org/10.1016/j.biopsycho.2019.05.005Maris, E., & Oostenveld, R. (2007). Nonparametric statistical testing of EEG- and MEG-data. Journal of Neuroscience Methods , 164 (1), 177-190. https://doi.org/10.1016/j.jneumeth.2007.03.024Moore, C. M., & Egeth, H. (1998). How does feature-based attention affect visual processing? Journal of Experimental Psychology: Human Perception and Performance , 24 (4), 1296-1310. https://doi.org/10.1037/0096-1523.24.4.1296Moore, T., Armstrong, K. M., & Fallah, M. (2003). Visuomotor Origins of Covert Spatial Attention. Neuron , 40 (4), 671-683. https://doi.org/10.1016/S0896-6273(03)00716-5Moore, T., Tolias, A. S., & Schiller, P. H. (1998). Visual re- presentations during saccadic eye movements. Proceedings of the National Academy of Sciences ,95 (15), 8981-8984. https://doi.org/10.1073/pnas.95.15.8981Myers, N. E., Stokes, M. G., & Nobre, A. C. (2017). Prioritizing Information during Working Memory: Beyond Sustained Internal Attention. Trends in Cogniti- ve Sciences , 21 (6), 449-461. https://doi.org/10.1016/j.tics.2017.03.010Ohl, S., Kroell, L. M., & Rolfs, M. (2024). Saccadic selection in visual working memory is robust across the visual field and linked to saccade metrics: Evidence from nine experiments and more than 100,000 trials. Journal of Experimental Psycholo- gy: General , 153 (2), 544-563. https://doi.org/10.1037/xge0001520Ohl, S., & Rolfs, M. (2017). Saccadic eye movements impose a natural bottleneck on visual short-term memory. Journal of Experimental Psy- chology: Learning, Memory, and Cognition , 43 (5), 736-748. https://doi.org/10.1037/xlm0000338Ohl, S., & Rolfs, M. (2018). Saccadic selection of stabilized items in visuospatial working memory. Consciousness and Cognition , 64 , 32-44. https://doi.org/10.1016/j.concog.2018.06.016Ohl, S., & Rolfs, M. (2020). Bold moves: Inevitable saccadic selection in visual short-term memory. Journal of Vision , 20 (2), 11. htt- ps://doi.org/10.1167/jov.20.2.11Oldfield, R. C. (1971). The assessment and analysis of handedness: The Edinburgh inventory.Neuropsychologia , 9 (1), 97-113. https://doi.org/10.1016/0028-3932(71)90067-4Palva, S., & Palva, J. M. (2007). New vistas for α-frequency band oscillations. Trends in Neurosciences , 30 (4), 150-158. https://doi.org/10.1016/j.tins.2007.02.001Panichello, M. F., & Buschman, T. J. (2021). Shared mechanisms underlie the control of working memory and attention. Nature , 592 (7855), 601- 605. https://doi.org/10.1038/s41586-021-03390-wPertzov, Y., Bays, P. M., Joseph, S., & Husain, M. (2013). Rapid forgetting prevented by retrospective attention cues. Journal of Experimental Psychol- ogy: Human Perception and Performance , 39 (5), 1224-1231. https://doi.org/10.1037/a0030947Poch, C., Campo, P., & Barnes, G. R. (2014). Modulation of alpha and gamma oscillations related to retro- spectively orienting attention within working memory. European Journal of Neuroscience , 40 (2), 2399- 10 Posted on 18 Dec 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.176604315.58106628/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. 2405. https://doi.org/10.1111/ejn.12589Poch, C., Capilla, A., Hinojosa, J. A., & Campo, P. (2017). Selection within working memory based on a color retro-cue modulates alpha oscillations. Neuropsy- chologia ,106 , 133-137. https://doi.org/10.1016/j.neuropsychologia.2017.09.027Poch, C., Valdivia, M., Capilla, A., Hinojosa, J. A., & Campo, P. (2018). Suppression of no-longer relevant information in Working Memory: An alpha-power related mechanism? Biological Psychology , 135 , 112-116. https://doi.org/10.1016/j.biopsycho.2018.03.009Popov, T., Miller, G. A., Rockstroh, B., Jensen, O., & Langer, N. (2021).Alpha oscillations link action to cognition: An oculomotor account of the brain ’s dominant rhythm . Neuroscience. https://doi.org/10.1101/2021.09.24.461634Rajsic, J., Liu, H., & Pratt, J. (2017). Eye movements can cause item-specific visual recognition advantages. Visual Cognition , 25 (9-10), 903- 912. https://doi.org/10.1080/13506285.2017.1352639Ray, W. J., & Cole, H. W. (1985). EEG activity during cognitive processing: Influence of attentional factors. International Journal of Psychophysiology ,3 (1), 43- 48. https://doi.org/10.1016/0167-8760(85)90018-2Rolfs, M., & Carrasco, M. (2012). Rapid Simultaneous Enhancement of Visual Sensitivity and Perceived Contrast during Saccade Preparation. The Journal of Neu- roscience , 32 (40), 13744-13752a. https://doi.org/10.1523/JNEUROSCI.2676-12.2012Saber, G. T., Pestilli, F., & Curtis, C. E. (2015). Saccade Planning Evokes Topographically Specific Activity in the Dorsal and Ven- tral Streams. The Journal of Neuroscience , 35 (1), 245-252. https://doi.org/10.1523/JNEUROSCI.1687- 14.2015Schroeder, S. C. Y., Ball, F., & Busch, N. A. (2018). The role of alpha oscillations in dis- tractor inhibition during memory retention. European Journal of Neuroscience , 48 (7), 2516-2526. https://doi.org/10.1111/ejn.13852Snyder, A. C., & Foxe, J. J. (2010). Anticipatory Attentional Suppression of Visual Features Indexed by Oscillatory Alpha-Band Power Increases:A High-Density Electrical Mapping Study. The Journal of Neuroscience , 30 (11), 4024-4032. https://doi.org/10.1523/JNEUROSCI.5684- 09.2010Souza, A. S., & Oberauer, K. (2016). In search of the focus of attention in working mem- ory: 13 years of the retro-cue effect. Attention, Perception, & Psychophysics , 78 (7), 1839-1860. https://doi.org/10.3758/s13414-016-1108-5Souza, A. S., Rerko, L., & Oberauer, K. (2014). Unloading and reloading working memory: Attending to one item frees capacity. Journal of Experimental Psychol- ogy: Human Perception and Performance , 40 (3), 1237-1256. https://doi.org/10.1037/a0036331Thut, G., Nietzel, A., Brandt, S. A., & Pascual-Leone, A. (2006). α-Band Electroencephalographic Activity over Occipital Cortex Indexes Visuospatial Attention Bias and Predicts Visual Target Detection. The Journal of Neuroscience , 26 (37), 9494-9502. https://doi.org/10.1523/jneurosci.0875-06.2006Tuladhar, A. M., Huurne, N. T., Schoffelen, J., Maris, E., Oostenveld, R., & Jensen, O. (2007). Parieto-occipital sources account for the increase in alpha activity with working memory load. Human Brain Mapping ,28 (8), 785-792. https://doi.org/10.1002/hbm.20306Ursino, M. (2025). Modeling the Role of the Alpha Rhythm in Attentional Processing during Distractor Suppression. Journal of Cognitive Neuroscience ,37 (12), 2469-2492. https://doi.org/10.1162/JOCN.a.65Van Diepen, R. M., Miller, L. M., Mazaheri, A., & Geng, J. J. (2016). The Role of Alpha Activity in Spatial and Feature-Based Attention. Eneuro ,3 (5), ENEURO.0204-16.2016. https://doi.org/10.1523/ENEURO.0204-16.2016Van Ede, F. (2020). Visual work- ing memory and action: Functional links and bi-directional influences. Visual Cognition , 28 (5-8), 401-413. https://doi.org/10.1080/13506285.2020.1759744Van Ede, F., Chekroud, S. R., Stokes, M. G., & Nobre, A. C. (2019). Concurrent visual and motor selection during visual working memory guided action. Nature Neuroscience , 22 (3), 477-483. https://doi.org/10.1038/s41593-018-0335-6Van Noordt, S. J. R., Desjardins, J. A., Gogo, C. E. T., Tekok-Kilic, A., & Segalowitz, S. J. (2017). Cognitive control in the eye of the beholder: Electrocortical theta and alpha modulation during response preparation in a cued saccade task. NeuroImage , 145 , 82-95. https://doi.org/10.1016/j.neuroimage.2016.09.054Worden, M. S., Foxe, J. J., Wang, N., & Simpson, G. V. (2000). Anticipatory Biasing of Visuospatial Attention Indexed by Retinotopi- cally Specific α-Bank Electroencephalography Increases over Occipital Cortex. The Journal of Neuroscience , 20 (6), RC63-RC63. https://doi.org/10.1523/JNEUROSCI.20-06-j0002.2000Zhou, Y. J., Van Es, M. W. J., & Haegens, S. (2025). Distinct alpha networks modulate different aspects of perceptual decision-making. PLOS Biology ,23 (10), e3003461. https://doi.org/10.1371/journal.pbio.3003461 Declarations Availability of data and materials: The datasets generated during and/or analyzed during the current study 11 Posted on 18 Dec 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.176604315.58106628/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. are available from the corresponding upon request. Author contributions: CRC: Investigation, Formal Analysis, Writing; JSS: Investigation, Formal Analy- sis, Writing; JMP: Investigation, Methodology; NC: Conceptualization, Writing; AC: Conceptualization, Writing, Funding acquisition; CP: Conceptualization, Formal Analysis, Funding acquisition, Project Admin- istration, Writing. Funding: This work was funded by the Ministerio de Ciencia, Innovaci´ on y Universidades under grant PID2022-143111NB-I00 and PID2021-125841NB-I00, and by the Comunidad de Madrid under Grant PIPF- 2022/SAL-GL-25279. Conflicts of interest/Competing interests: The authors have no relevant financial or non-financial interests to disclose. Ethics approval: This study was performed in line with the principles of the Declaration of Helsinki. The study was approved by the Ethics Committee of the Universidad Nebrija. Consent to participate: Informed consent was obtained from all individual participants included in the study. 12 Posted on 18 Dec 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.176604315.58106628/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. 13