Bridging the gap between experimental control and ecological validity in human sensorimotor science

Human sensorimotor science faces a fundamental trade-off between precise experimental control and real-world applicability. Laboratory experiments, though excellent in experimental control, produce findings with limited generalizability to diverse populations and ecological environments; conversely, real-world experiments, though effective at eliciting movements relevant to daily life, often lack the experimental control necessary for consistent replication and detailed mechanistic understanding. In this opinion piece, we illustrate the need for more complementary methods, comparative studies and emerging technologies to bridge the gap between experimental control and ecological validity. Human sensorimotor science is the study of how the human nervous system uses sensory information to effectively move and interact with the environment. This field spans a wide range of topics, from the neural correlates of sensorimotor control to the computational algorithms of de novo skill acquisition. This field also encompasses a wide range of methods, from behavioural psychophysics to transcranial magnetic stimulation. As with other scientific domains, human sensorimotor science confronts a fundamental trade-off between experimental control and real-world translatability (Fig. 1A) (Cisek & Green, 2024; Fooken et al., 2023). On one hand, laboratory experiments often involve manipulations requiring precise spatiotemporal control (Fig. 1A, bottom right). Yet these experimental contexts tend to diverge sharply from those of daily life, involving elementary movements, artificial sensory feedback and brief training sessions. On the other hand, real-world experiments often involve natural movements essential to day-to-day life (Fig. 1A, top left). Yet, these experiments tend to be unconstrained, making it challenging to assess sensorimotor behaviour, isolate underlying mechanisms and replicate previous phenomena. We offer three recommendations to address the control–translatability trade-off (Fig. 1B). First, we recommend the use of complementary methods. Take, for instance, motor adaptation experiments involving the upper limb. Although many laboratory studies have significantly advanced our mechanistic understanding of motor adaptation, the process of reducing motor errors through feedback and practice, their limited sample sizes and participant diversity often prevent them from effectively addressing questions about how individual differences impact sensorimotor performance (Bastian, 2008; Ranganathan et al., 2022). To tackle these challenges, researchers have turned to a crowdsourcing approach, which enables the collection of larger datasets involving participants from diverse backgrounds (Clode et al., 2024; Listman et al., 2021; Malone et al., 2023; Tsay et al., 2021). Strikingly, these datasets have not only replicated classic findings but also revealed unappreciated individual differences underlying motor adaptation (Tsay et al., 2024). With ongoing advances in markerless pose tracking and virtual reality, we anticipate that these large-scale motor experiments will expand to include full-body movements. In sum, by complementing fine-grain laboratory studies with large-scale real-world experiments, we can broaden the scope of inquiries in human sensorimotor science. Second, we recommend the use of comparative studies to evaluate how sensorimotor behaviour differs between in-lab and real-world settings. For instance, despite being biomechanically simpler, learning to throw in the lab is often significantly slower than in natural, real-world environments (Zhang & Sternad, 2021). Similarly, learning in the lab often involves the progression of a single learning strategy, whereas learning in real-world settings often encompasses a mixture of strategies (Nardi et al., 2024). Moreover, the rate of motor learning in the lab is often modulated by motor variability only in task space (Wu et al., 2014), whereas the rate of learning in real-world settings is often modulated by multiple sources of variability (Haar et al., 2020). In sum, by conducting more comparative studies, we can deepen our understanding into how robustly sensorimotor mechanisms generalize between in-lab and real-world settings. Third, we recommend the use of emerging technologies to mitigate the control–translation trade-off. For example, wearable technologies – such as portable electroencephalography, optically pumped magnetometers, chronic electrode implants for real-time adaptive deep-brain stimulation, epidural electrograms and high-density electromyography – can now monitor neural and muscle activity as participants move in naturalistic environments. Furthermore, virtual reality (VR) can now be used to provide realistic feedback while maintaining tight experimental control (Cesanek et al., 2024); for example, by manipulating the timing and trajectory of a virtual basketball. However, VR is not without its limitations. For instance, while manipulating virtual visual and auditory feedback is straightforward, providing virtual touch and proprioception is challenging; interactions with virtual objects are also mediated by game controllers, which may diminish the level of immersion and embodiment, potentially reducing the amount of skill transfer from VR to real-world environments. To address these challenges, researchers have been developing embodied virtual reality (EVR) environments that enable virtual objects to be physically sensed and directly manipulated, paving the way for a more immersive and realistic sensorimotor experience (Haar et al., 2021; Nardi et al., 2024). In conclusion, we propose that leveraging complementary methods, conducting comparative studies between lab-based and real-world tasks, and using emerging technologies for task presentation and evaluation will bridge the gap between experimental control and ecological validity. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. None declared. All authors contributed to conceptualizing, reviewing, and editing. J.T., N.S., and S.H. investigated and wrote the original draft. S.H. supervised the work. None.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.