Early research investigated whether microsaccades could reflect external attention shifts. Hafed and Clark (2002) first investigated this topic, conducting experiments where participants were instructed to maintain central fixation while attending to four peripheral target locations. A sequence of visual cues presented randomly between the fixation and a peripheral target, inducing covert attention shifts. Subsequent to an unpredictable number of cues, the most recently cued target location displayed a color after a variable cue-target interval, and participants identified the displayed color. The results revealed an early microsaccades bias toward the target location (170 to 340 ms post-cue). Furthermore, this bias correlated with an enhanced target discrimination performance in trials characterized by that particular cue-target interval. A second experiment used anti-cues, wherein the cue consistently indicated a direction opposite to the actual target, and revealed that microsaccades were initially attracted to the cue side, then shifted toward the target side. Notably, in this scenario, optimal target discrimination was achieved with longer cue-target intervals. These findings provide compelling evidence for a direct relationship between the directionality of microsaccades and real-time attention shifts. Subsequent studies examined four key characteristics that link microsaccade directionality with attentional shifts: (1) Influence of attentional cues: Both endogenous and exogenous cues are known to precipitate shifts in attention. Endogenous cues, however, induce these shifts more gradually and with a subtler impact compared to exogenous cues (e.g., Müller and Rabbitt, 1989). The direction of microsaccade aligns well with such difference in attention shifts: exogenous cues caused a fast and strong microsaccade direction bias toward the cue, whereas endogenous cues only caused a late and weak bias (Engbert and Kliegl, 2003; Laubrock et al., 2005). These results suggest that microsaccade direction can distinguish different types of attention shifts, strengthening a direct correlation between attention shifts and microsaccade directionality. (2) Enhancements in early visual cortex: Meyberg et al. (2015) found that attention shifts correlate with changes in the early visual cortex, as indicated by a pronounced P1/N1 microsaccade-related potential component at occipital scalp sites. Notably, these changes are ipsilateral to the direction of microsaccades and emerge subsequent to the bias in microsaccade direction, suggesting a temporal and spatial link between microsaccade direction and neural activity associated with attention shifts. (3) Inhibition of Return (IOR) phenomenon: The IOR effect, where attention is less likely to return to a previously attended location (Posner and Cohen, 1984), also influences microsaccade direction. In scenarios involving invalid cues and the occurrence of IOR, microsaccades are more frequently oriented in the direction opposite to the cue (Galfano et al., 2004; Betta et al., 2007). This indicates that the direction of microsaccades is not only a response to immediate attentional cues but also to the overall pattern of attention shifts. (4) Microsaccade direction and attention-related performance: Evidence particularly from nonhuman primate studies underscores a direct relationship between microsaccade direction and attentional performance. Hafed et al. (2011) observed that when microsaccades were directed toward a cued location soon after a target appeared, the subjects exhibited a higher likelihood of target discrimination. Conversely, when microsaccades were oriented away from the cue, the likelihood of successful discrimination decreased. This underscores the functional significance of microsaccade direction in tasks that require attention.

Researchers further investigated whether microsaccades could directly influence or even cause attentional activity. Yuval-Greenberg et al. (2014) conducted the first study from the perspective of how microsaccades influence covert attention allocation. After real-time detecting participants’ spontaneous microsaccades during fixation, probe items were briefly presented surrounding the initial fixation (Experiment 1) or the actual eye position (Experiment 2). Participants were required to discriminate the orientation of the target from probe items according to a postcue that either pointed to the direction of the microsaccade or opposite to it. In both Experiments, they found that participants exhibited significantly better recognition performance for stimuli presented in the direction of the microsaccades compared to the opposite direction. This result revealed a direct impact of microsaccades on the allocation of visual attention, although this effect only occurred for binocular microsaccades (Zhang et al., 2016). Lowet et al. (2018) provided further neurophysiological evidence. They recorded the influence of microsaccades on attention-related modulation of neuronal activity in macaque V4 and the inferior temporal cortex, and observed enhanced neural processing occurred only following microsaccades directed toward the attended location. However, a number of empirical investigations have yielded consistently contrary findings to the aforementioned perspective. At the behavioral level, researchers found that spatial perception is altered prior to microsaccade onset (Hafed, 2013), and there is an improvement in visual discrimination at the upcoming microsaccade location as well as an impairment at the opposite location prior to microsaccade onset (Shelchkova and Poletti, 2020). At the neural level, neurons in the superior colliculus and frontal eye fields exhibit attention-related modulation prior to microsaccades onset (Chen et al., 2015), even in the absence of microsaccades (Yu et al., 2022). Given the accumulating body of evidence suggesting an earlier emergence of attention-related activities compared to microsaccade onset, the current consensus does not support the notion that microsaccades directly cause external attentional activity.

Researchers continue to broaden our understanding of how microsaccades reflect external attention shifts from various perspectives. For example, a simple model including microsaccade generation and peri-microsaccadic changes in vision sufficiently accounted for attention capture and IOR (Tian et al., 2016). Dynamic changes in microsaccade amplitudes differentiated attention shifts under different cue validities (Lv et al., 2022). Attention shifts induced by emotional stimuli modulated microsaccadic activity (Kashihara et al., 2014). In more ecologically valid environments, some studies replicated the link between attention and microsaccade direction (e.g., Barnhart et al., 2019; Xue et al., 2020), while others found that this relationship disappears with increasing environmental complexity (Willett and Mayo, 2023).

Overall, current results suggest that although microsaccades may not be deterministic for external attention activities, the direction of them still can be regarded as a useful, yet not perfect, marker of external attention shifts. Specifically, express microsaccades occurring 60–100 ms after cue presentation (Tian et al., 2018), microsaccades occurring 200–400 ms after cue presentation, the first microsaccade occurring after cue presentation (Laubrock et al., 2010), and microsaccades occurring during periods of extremely low microsaccade rate (Pastukhov and Braun, 2010) are believed to provide a more accurate reflection of attention shifts.

In recent years, there has been growing interest in using eyetracking to reflect internal attention shifts. van Ede et al. (2019) conducted a pioneering study investigating the relationship between gaze shifts and attention shifts in working memory. They required participants firstly to memorize the orientations of two colored bars and then to reproduce the orientation of one bar according to the color of the fixation (see Figure 2A), while analyzing the precise gaze position during the maintenance period. The results revealed that after the onset of colored retro-cue and it guided attention to a particular representation in working memory, the gaze position involuntarily shifted, in line with microsaccades’ magnitudes, slightly toward the location where that representation was once presented (see Figure 2B). Moreover, when there were two retro-cues and the first can be either valid or neutral (see Experiment 2 in van Ede et al., 2019), the pattern of gaze shifts did not emerge until the valid retro-cue presented (see Figure 2C). These findings support the notion that microsaccades can serve as an indicator of internal attention shifts, rather than sustained attention biases. Follow-up research further demonstrated that such reflection owned a high temporal precision (van Ede et al., 2021), and could be achieved via both voluntary and involuntary attention selection (van Ede et al., 2020).

Similar to research on external attention shifts, researchers also examined whether microsaccades can directly influence or cause internal attention shifts. Behavioral studies suggested that microsaccades do not cause internal attention shifts, since that an involuntary gaze-shift manipulation did not bring an item into the focus of attention so as to benefit subsequent performance (van Ede et al., 2019). This perspective is further corroborated by neural research, notably by Liu et al. (2022). They used lateralized alpha EEG as an objective reference for internal attention (see van Ede, 2018 for a review) and revealed a close relationship between microsaccade direction and alpha lateralization: alpha lateralization is stronger in trials with microsaccades toward versus away from the memorized location, and occurs earlier when microsaccades orient to the target earlier. Intriguingly, alpha lateralization was still observable even in the absence of microsaccades, supporting that internal attention shifts can occur independently of microsaccades. However, Liu et al. (2023) also noted that microsaccades can elicit stronger alpha activity when they oriented toward a specific direction, facilitating transient lateralized alpha activity. It implied that microsaccades may modulate internal attention to some extent. Given these insights, it is imperative for future research to further explore the influence of microsaccades on internal attention processes and to elucidate the underlying mechanisms through which microsaccades reflect shifts in internal attention.