In this eye-tracking study, we explored the visual-exploration patterns of participants involved in a free-to-look-at task where ten unfamiliar graspable objects extracted from the NOUN database (Horst & Hout, 2016) were presented. Specifically, we analysed the participants’ temporal allocation of visual-spatial attention to objects’ manipulation (i.e., the area of the object aimed at manipulating it) and functional (i.e., the part of the object through which one may gather its identity and function) AOIs. Also, to investigate how increasing stimulus familiarity impacts on the way participants look at unfamiliar objects, we administered three repetitions. Results highlighted peculiar visual-exploration patterns, with participants looking at functional and manipulations AOIs in a reversed-U way. Indeed, we found that unfamiliar-object visual exploration was fully described by quadratic curves, with participants looking at functional AOIs in such a way as to give rise to an almost-convex parabolic trend. Instead, participants showed an almost-concave parabolic trend of fixations to objects' manipulation AOIs. During the first 750ms of visual exploration, participants increased their visual-spatial attention to the objects’ functional AOI (Fig. 4C). Conversely, participants' visual-spatial attention to manipulation AOIs (Fig. 4D) decreased during the same time interval. After the 750ms peak, participants reversed their visual exploration trend, hence reducing their visual-spatial attention to the functional AOIs while relatively increasing their fixations toward the manipulation AOIs. Crucially, we found that familiarization effects related to stimuli repetitions reverberated only in the way participants visually explored functional AOIs. Indeed, during the third stimuli repetition, participants significantly reduced their visual-spatial attention to the functional AOIs (Fig. 4A and Fig. 4C) while maintaining unchanged their fixations to manipulation AOIs (Fig. 4B and Fig. 4D).
According to the findings we report here, the visual exploration of unfamiliar graspable objects seems to be characterised by an explorative gaze behaviour firstly aimed at identifying the object's function and identity by looking at its functional area (Natraj et al., 2015; Van Der Linden et al., 2015; Tamaki et al., 2020). After such a preliminary semantic-driven visual exploration of the stimulus, we found that participants oriented their visual-spatial attention toward the object’s manipulation areas, as if they were to prepare themselves for action (Riddoch et al., 2003; Handy et al., 2003). Such a perspective is in line with the perception-for-action theoretical framework (e.g., Milner & Goodale, 2008) while emphasizing at the same time the relevance of high-level cognitive processes involved in signification, action understanding and technical reasoning (Federico & Brandimonte; 2020). Critically, in line with such a hybrid approach, we found that familiarization effects produced by stimuli repetitions did not impact the manipulation-related visual exploration of objects, thus influencing only the functional-related gaze behaviour. Indeed, one may reasonably assume that imagining how to grasp unfamiliar objects should not produce so many fluctuations in visually exploring their manipulation areas over time, whereas identifying and recognizing the function of an object might be a process susceptible to the stimulus’ exposure time (Federico & Brandimonte, 2019).
The visual exploration of unfamiliar but graspable objects appears to reflect the interactions between affordance-based (e.g., Humphreys et al., 2013; Masson et al., 2011) and higher-level cognitive processing (e.g., Federico et al., 2021; Wurm & Caramazza, 2019; Bar et al., 2006). Additionally, as we mentioned above, the exploration of objects’ functional areas might also be associated with technical-knowledge processing through which observers may reason about how objects can be used with other objects mechanically (e.g., looking at the head of a hammer, thus focusing on the action-performing area that will hit a nail; Tamaki et al., 2020; Goldenberg & Spatt, 2009). In this sense, our results align with multiple studies that have shown how participants may concentrate on the action’s goal component more than on its manipulation component (e.g., Massen & Prinz, 2007; Osiurak & Badets, 2014). These goal-related patterns have also been traced in observational investigations where observers looked at an actor using an object (e.g., Decroix & Kalénine, 2019; Naish, Reader, Houston-Price, Bremner, & Holmes, 2013; Nicholson, Roser, & Bach, 2017; van Elk, van Schie, & Bekkering, 2008). Such studies investigated the action’s goal component regarding objects’ functional AOIs, thus implicitly referring to the semantic/technical knowledge retrievable by looking at objects. However, those results are typically interpreted only in terms of manipulation/sensorimotor knowledge (e.g., Thill et al., 2013), possibly because of the absence of alternative theoretical frameworks (Osiurak & Federico, 2020).
The polymorphic interactions between distinct kinds of knowledge we summarised above give space to the idea of a cognitive functioning oriented towards integrating multiple information modalities through which humans may endow reality with meaning and exploit the environment for action. Such a kind of hybrid cognitive mechanism involved in the way humans integrate distinct kinds of information in order to generate representations that may be used in everyday life has been recently labelled as “action reappraisal” in the field of human tool use (Federico & Brandimonte, 2019). The results presented here provide further experimental support for the action reappraisal mechanism and interrogate about the possible neurocognitive systems supporting such a multifaceted phenomenon. Indeed, brain areas underlining the identification, recognition and use of objects comprise an extensive and multifunctional fronto-temporo-parietal network (e.g., Rizzolatti & Matelli, 2003; Goldenberg & Spatt, 2009; Almeida, Fintzi, & Mahon, 2013; Ishibashi et al., 2016; Lesourd et al., 2021). Consequently, specific hypotheses have been generated about the involvement of such fronto-temporo-parietal areas in the context of the action-reappraisal approach (e.g., Federico & Brandimonte, 2020). Although the scientific debate about the neural correlates of the action reappraisal mechanism is very far from being concluded, increasing and converging support to the action reappraisal idea comes from studies that highlighted the involvement of specific brain networks involved in integrating information about action and objects across different modalities (e.g., Chen, Garcea, Jacobs & Mahon, 2018; De Bellis et al. 2020; Lambon Ralph, Jefferies, Patterson & Rogers, 2017; Wurm & Caramazza 2019; Lesourd et al., 2021; Pupíková, Šimko, Gajdoš, Rektorová, 2021). Significantly, a most recent fMRI/tDCS study (Pupíková et al., 2021) demonstrated how stimulating the fronto-parietal network with twenty minutes of 2mA anodal tDCS increased the recognition performance of participants involved in a yes-no object recognition task which was similar to the one developed by Federico and Brandimonte (2020), hence providing the first solid causal evidence for the action reappraisal mechanism.
By following the tripartite neural-stream division introduced by Rizzolatti and Matelli (2003), the central neurocognitive systems involved in object processing and tool use are those related to the motor-control system (i.e., the dorso-dorsal system), the mechanical/technical-knowledge system (i.e., the dorso-ventral system), and the semantic system (i.e., the ventral system; e.g., Goldenberg & Spatt, 2009; Almeida, Fintzi, & Mahon, 2013; Osiurak et al., 2017; Ishibashi et al., 2016). Intriguingly, the recent Three-Action System model (Osiurak et al., 2017) identifies a part of the ventro-dorsal system, namely the left inferior parietal lobe and specifically its related cytoarchitectonic area PF (Caspers et al., 2006), as a kind of bridge between the semantic system related to objects’ identity (i.e., the ventral system), and the motor-control system (i.e., the dorso-dorsal stream). Additionally, most recent neuroscientific and neuropsychological evidence highlighted the inferior parietal and middle temporal brain areas as the ones pertaining to an integrative cognitive layer related to the generation of object-related action multi-modal representations (e.g., Chen et al., 2018; De Bellis et al. 2018; Lambon Ralph et al., 2017; Wurm & Caramazza 2019; Lesourd et al., 2021; see also Humphreys, Lambon Ralph, & Simons, in press). Also, when considering the frontal and prefrontal cortex involvement in object processing, it appears that these areas might be easily related to high-level executive functions, motor timing, sequencing and simulation (e.g., Koechlin & Summerfield, 2007; Bortoletto & Cunnington, 2010). However, whereas the above-specified integrated processes involving temporo-parietal brain areas have been increasingly investigated, the frontal and prefrontal areas involvement did not get the same popularity. Notwithstanding, these areas might actively take part in the action reappraisal mechanism as they may signal a specific cognitive functioning through which an observer, from the multiple environment-available information and action possibilities, may select only those that are consistent with their intentions (for a discussion, see Federico et al., 2021).