Motor imagery is the mental simulation of movement in the absence of physical execution (Jeannerod, 2001). Motor imagery spans across various disciplines, from fundamental neuroscience to sports science and rehabilitation (Guerra et al., 2017; Guillot et al., 2021; Ladda et al., 2021; Paravlic et al., 2018). The ability to perform motor imagery can vary between individuals (Floridou et al., 2022; Guillot et al., 2008), but this ability is challenging to measure because of the covert nature of multiple cognitive processes occurring during motor imagery (Cumming & Eaves, 2018).

The Hand Laterality Judgement Task (HLJT) has been proposed as a tool for measuring imagery ability (Heremans et al., 2013; McAvinue & Robertson, 2008) and has also been considered an “implicit” measure of mental representation of the body in action, as opposed to paradigms that require greater action monitoring (Brusa et al., 2023; Scarpina et al., 2019). In this task, the individual has to determine whether a picture of a rotated hand corresponds to the right or left side (Parsons, 1987). The task has been used as the paradigmatic example to argue that motor imagery is implicitly used in body recognition tasks. This model posits that the individual will use the mental representation of their own hand to decide the laterality (Parsons, 1987; Sekiyama, 1982).

Behavioural evidence has repeatedly found a phenomenon which differentiates the HLJT from mental rotation of other objects, namely the ‘biomechanical constraints’ effect (Parsons, 1994). The effect illustrates how anatomical restrictions can influence the mental rotation process: when the hand image is rotated in directions that are more easily physically achieved (i.e., towards the midline or medial direction), the time needed to process the stimulus decreases, compared to images rotated towards anatomically awkward angles (i.e., away from the midline or lateral direction). This phenomenon has been consistently replicated across differing HLJT paradigms (Bek et al., 2022; Conson et al., 2021; Ionta et al., 2007; Meng et al., 2016), and has not been found in mental rotation tasks using letters or alphanumerical symbols as stimuli (Bek et al., 2022; ter Horst et al., 2012). These findings have been used to claim that the biomechanical constraints effect represents a hallmark of the use of motor imagery to complete the task (Bek et al., 2022; Hoyek et al., 2014; Meng et al., 2016; Vannuscorps & Caramazza, 2016).

Critical knowledge gaps are still present regarding how stimuli are processed during the HLJT. For example, previous research has shown an interaction between the degree of rotation and the view of the hand that is presented, yet the nature of this interaction is not well understood. There is behavioural (Conson et al., 2020, 2021; Hoyek et al., 2014) and neurophysiological (Meng et al., 2016; Zapparoli et al., 2014) work suggesting that hand images in a palmar view (i.e., the palm facing up) are processed in different ways than hand images in a dorsal view (i.e., the dorsum facing up). Some authors (Brady et al., 2011; Nagashima et al., 2019) have suggested that the difference may lie in the fact that they might elicit different processing strategies, where processing based on spatial/visual aspects largely occurs for dorsal views (i.e., using an allocentric/third-person reference frame) and processing based on anatomical/motor aspects occurs for palmar views (i.e., using an egocentric/first-person reference frame). In fact, previous work has shown that when hand images display an orientation more plausible from a third-person perspective than from a first-person perspective, as in a rotation angle of 180° in the dorsal view, a mirror-like mapping is automatically generated in the brain (Shmuelof & Zohary, 2008). In other words, dorsal views might more easily trigger an allocentric strategy when stimuli are rotated to a greater extent and closer to a mirror-like position, whereas palmar views would more easily trigger an egocentric strategy regardless of the rotational angle (Bek et al., 2022). However, there are no studies that have formally investigated this hypothesis in the HLJT. Here, using a combination of ‘forced response’ behavioral measurements and computational modelling, we provide the first empirical evidence to support this proposal. Our results indicate that palmar stimuli presented at more extreme rotations are processed in a fundamentally different manner from other hand stimuli, which we attribute to differences between separable allocentric and egocentric modes of processing.

The possibility of differing processing strategies in the palmar and the dorsal view is supported by compelling evidence suggesting that the strength (or even the presence) of the ‘biomechanical constraints’ effect can depend largely on the view of the hand that is presented (Bek et al., 2022; Conson et al., 2021; Hoyek et al., 2014; Mibu et al., 2020; Nagashima et al., 2019). Indeed, hand images in the palmar view typically trigger this effect, whereas hand images in the dorsal view typically show a much weaker effect or do not produce it at all. This could be indirect evidence of motor-based processing occurring only for the palm of the hand (Conson et al., 2021). Understanding the underlying cognitive subprocesses of this medial-to-lateral advantage is useful to determine which specific stimuli are more likely to elicit the different processing strategies, a knowledge that would be more complete if the information processing time-course is studied as a whole, instead of only studying its end-point through classical reaction time paradigms. In our study, we approached this experimentally by modifying the traditional HLJT into a ‘forced response’ protocol, which allowed us to reconstruct the time-course of the ‘biomechanical constraints’ effect. This information is crucial to use the HLJT as a potential motor imagery ability index and increase our understanding of the possible use of motor imagery in this task, which in turn will have an impact on its use in applied contexts.

As indicated above, the traditional HLJT paradigm effectively considers two separate but inherently linked dependent variables (i.e., reaction time and accuracy), and therefore it has a limited ability to study how the stimulus is processed, as speed-accuracy trade-offs can interfere with the dynamics of information processing (Wickelgren, 1977). Moreover, simple reaction time paradigms, although straightforward, only provide data from the endpoint of the processing time-course. Several methods have been proposed to overcome these limitations, both experimentally (Katsimpokis et al., 2020) and statistically (Davidson & Martin, 2013), but other behavioural paradigms could also be useful to provide insight into the time-course of processing during this task. In this scenario, ‘forced response’ paradigms, which ‘force’ individuals to respond after a specific amount of time (i.e., effectively manipulating the duration of stimulus presentation as an independent variable) could help to unveil behavioural signatures typically observed in the HLJT. Forced response paradigms provide a valuable means to reconstruct the time-course of information processing during a task, assessing accuracy as a function of the time allowed to process the stimulus. The forced response approach therefore allows researchers to study the time-course of behaviour during stimulus processing with fine-grained precision (Haith et al., 2016). As an example, forced response protocols enable researchers to understand the evolution of ‘basic’ effects in the HLJT, like the effect of stimulus rotation, which is still not fully understood. Many studies in the literature from general mental rotation paradigms show this effect linearly increases reaction time (i.e., there is a monotonic and proportional relationship between rotation angle and reaction time), and is the main driver of information processing in these tasks. Nonetheless, the HLJT is a particular paradigm where differences in the effect of rotation angle have been reported. For instance, previous studies have found non-linear increases in reaction times for this task (which in some cases are reflective of the ‘biomechanical constraints’ effect, but in other cases are not) (Bek et al., 2022; Conson et al., 2020; Ionta et al., 2007). However, it is challenging with traditional reaction time paradigms to ascertain whether this particular difference is due to actual differences in processing, or artifacts caused by simple shifts in the speed-accuracy trade-off (i.e., a participant may respond more slowly to maintain accuracy, or respond faster at the cost of making more errors).

This paper studied the time-course of information processing during the HLJT. We used a ‘forced response’ paradigm to analyse how factors such as stimulus rotation or hand view affect performance during the task, how they interact with each other, and their influence on the ‘biomechanical constraints’ effect. We hypothesised that information processing is primarily driven by stimulus rotation, and that a strong interaction between rotational angle and hand view would be evidenced. Additionally, we predicted that a medial-to-lateral advantage (consistent with a ‘biomechanical constraints effect’) would only be evident for palmar views, as they are more likely to produce an egocentric (i.e., motor-based) strategy. Using computational modelling, we further characterised different response profiles for hand views and rotation angles, which is key to understand how information processing strategies might be used in this task. Overall, our study stepped beyond current limitations of behavioural paradigms to increase our understanding of information processing in the HLJT.