Rosenholtz’s re-examination of visual attention stems from a clear understanding of the capabilities and limits of mature peripheral vision. This marks the starting point to uncover and challenge anomalies in attention-related phenomena. Despite this, such an understanding of peripheral function is still partial over early development. Rosenholtz’s position is therefore not integrated into infant research paradigms examining visual attention. It is our view that development is critical for understanding mature attentional mechanisms. Foundations of attentional networks, such as alerting and orienting, develop in infancy, and ground mature and more complex attentional and cognitive processes (Hendry, Johnson, & Holmboe, Reference Hendry, Johnson and Holmboe2019). Development can also help us determine how perceptual capacities and limitations impact attention. As highlighted by Rosenholtz, perceptual mechanisms tend to be included under the same umbrella definition of attention. A better understanding of early perceptual development can provide essential information on the complexity of mature phenomena.

Currently, relatively few studies have investigated peripheral processing during development. What we know is that visual fields gradually expand, particularly over the first postnatal year (Lewis & Maurer, Reference Lewis and Maurer1992). At birth, the peripheral retina is relatively mature compared to the central/foveal retina (Hendrickson & Drucker, Reference Hendrickson and Drucker1992). Accordingly, a peripheral advantage for movement detection is observable at birth (Johnson et al., Reference Johnson, Dziurawiec, Ellis and Morton1991) and during the third trimester of gestation (Reid et al., Reference Reid, Dunn, Young, Amu, Donovan and Reissland2017). Of note, stimuli with evolutionary significance (e.g. face-like stimuli) can be detected at higher eccentricities compared to other stimuli (Capparini, To, & Reid, Reference Capparini, To and Reid2022a). Stimulus features such as size and distance also play a role in early peripheral functioning (de Schonen et al., Reference de Schonen, McKenzie, Maury and Bresson1978). Nevertheless, there is currently no comprehensive mapping of peripheral processing during different developmental stages. Most research focused on the first postnatal weeks, with very little investigations beyond 6–7 months. However, we know that peripheral vision is still developing at 9 months (Capparini, To, & Reid, Reference Capparini, To and Reid2022b). It is therefore unsurprising that only hints are present in the relationship between peripheral information processing and attention during infancy (Johnson, Reference Johnson1990; Johnson & de Haan, Reference Johnson and de Haan2015).

In stark contrast to peripheral work, infancy research abounds with visual attention studies. Examples include experiments on covert and overt attention, with tasks such as the fixation shift paradigm, the gap/overlap paradigm, or other adaptations of the spatial cueing paradigm. In preverbal participants, covert attention is often revealed by physiological variations (Finlay & Ivinskis, Reference Finlay and Ivinskis1984) or changes in saccadic behaviour (Clohessy et al., Reference Clohessy, Posner, Rothbart and Vecera1991) following the brief presentation of a peripheral cue. Further, brain activation in the anterior cingulate cortex, as indexed by the Negative component (Nc), has been proposed to mediate early attentional orienting and to distinguish periods of focused vs. unfocused processing (Richards, Reference Richards2003). All the above paradigms involve the presentation of peripheral stimuli, but the limits of the developing visual field are often not considered. In parallel with Rosenholtz’s reasoning, we argue that performance on attentional tasks may be explained in terms of a developing peripheral vision rather than attentional processes, and that the contributions from the perceptual and attentional systems both need to be considered. Future work is needed to clarify whether complex behaviour during development, such as rapid orientation to some emotions over others, is led by bottom-up perceptual and/or top-down attentional mechanisms.

Visual processing in the peripheral field is qualitatively different from processing in central regions. Crowding is a phenomenon that is primarily present in the periphery and impacts the detectability of objects outside fixation (Levi, Reference Levi2008). Farzin, Rivera, & Whitney, (Reference Farzin, Rivera and Whitney2010) demonstrated that visual experience during infancy cannot be simply predicted by developing acuity and contrast sensitivity, rather infants seem to have a coarser and more diffuse window of visual awareness than adults due to crowding. Crowding can also impact saccadic orienting behaviour towards a peripheral target and, hence, play a role in the above-mentioned visual attention studies. The effect of perceptual mechanisms such as crowding on complex attentional behaviour needs to be further explored in infancy. Infants and their developing peripheral capacities can therefore allow us to test the limits on task complexity raised by Rosenholtz and identify the nature of such a limit.

As highlighted in the target paper, a non-attentional pathway accounting for statistics of low-level visual features may explain attentional phenomena. In infancy, low-level processing has proposed cascading roles for attentional and cognitive mechanisms. An example is the relation of low-level contextual processing indicated by a slow wave ERP effect for congruent and incongruent actions (Michel et al., Reference Michel, Kaduk, Ní Choisdealbha and Reid2017). This then leads to an Nc attentional effect at 7 months and then an N400 semantic effect at 9 months (Reid et al., Reference Reid, Hoehl, Grigutsch, Groendahl, Parise and Striano2009; Kaduk et al., Reference Kaduk, Bakker, Juvrud, Gredebäck, Westermann, Lunn and Reid2016). Overall, summary statistics models have had limited applications to developing populations (see Balas, Reference Balas2017, with children). Whether summary statistics can explain infant behaviours, such as the preference for high-contrast information or delayed orienting responses, is still open to investigation.

On top of conceptually rethinking early attentional processes, resolving the vision-attention imbalance would be helpful in addressing newly emerging trends in developmental research. Notably, there has been increasing interest in implementing more naturalistic setups that can capture the child’s naturalistic visual environment, for instance, using head-mounted cameras (Anderson, Seemiller, & Smith, Reference Anderson, Seemiller and Smith2022; Jayaraman, Fausey, & Smith, Reference Jayaraman, Fausey and Smith2015). Such investigations are essential for generalizing laboratory findings to everyday environments. They currently operate with incomplete information about the child’s visual capacities during development. How much visual information is available across the child’s visual field, and how is this affected by peripheral limitations or crowding? These issues cannot be addressed using a simplified camera view. Another area that can benefit from a better mapping of perceptual limits over time is computational models of cognitive development (Mattern et al., Reference Mattern, Schumacher, López, Raabe, Ernst, Aubret and Triesch2024). Implementing immersive visual environments that capture visual capabilities and limitations across different developmental stages can therefore affect further processing and decision making.

It is our view that a comprehensive mapping of peripheral vision throughout early development could lead to a better understanding of what can be defined as attentional rather than perceptual. Beyond a theoretical rethinking of attention, perceptual capabilities and limits need to be carefully considered in the developmental sciences.