Dyadic Interaction Platform: A novel tool to study transparent social interactions

Reviewer #1 (Public review):

Summary:

In this manuscript, the authors aim to address significant limitations of existing experimental paradigms used to study dyadic social interactions by introducing a novel experimental setup - the Dyadic Interaction Platform (DIP). The DIP uniquely allows participants to interact dynamically, face-to-face, with simultaneous access to both social cues and task-related stimuli. The authors demonstrate the versatility and utility of this platform across several exemplary scenarios, notably highlighting cases of significant behavioral differences in conditions involving direct visibility of a partner.

Major strengths include comprehensive descriptions of previous paradigms, detailed explanations of the DIP's technical features, and clear illustrations of multimodal data integration. These elements greatly enhance the reproducibility of the methods and clarify the potential applications across various research domains and species. Particularly compelling is the authors' demonstration of behavioral impacts related to transparency in interactions, as evidenced by the macaque-human experiments using the Bach-or-Stravinsky game scenario.

Strengths:

The DIP represents a methodological advance in the study of social cognition. Its transparent, touch-sensitive display elegantly solves the problem of enabling participants to attend to both their social partner and task stimuli simultaneously without requiring attention switching. This paper marks a notable step forward toward more options for naturalistic yet still lab-based studies of social decision-making, an area where the field is actively moving, especially given recent research highlighting significant differences in neural activity depending upon the context in which an action is performed. The DIP offers researchers a valuable tool to bridge the gap between tightly controlled laboratory paradigms and the dynamic, bidirectional nature of real-world social interactions.

The authors do well to provide comprehensive documentation of the technical specifications for the four different implementations of the platform, allowing other researchers to adapt and build upon their work. The detailed information about hardware configurations demonstrates careful attention to practical implementation details. They also highlight numerous options for integration with other tools and software, further demonstrating the versatility of this apparatus and the variety of research questions to which it could be applied.

The historical review of dyadic experimental paradigms is thorough and effectively positions the DIP as addressing a critical gap in existing methodologies. The authors convincingly argue that studying continuous, dynamic social interactions is essential for understanding real-world social cognition, and that existing paradigms often force unnatural attention-splitting or turn-taking behaviors that don't reflect naturalistic interaction patterns.

The four example applications showcase the DIP's versatility across diverse research questions. The Bach-or-Stravinsky economic game example is particularly compelling, demonstrating how continuous access to partners' actions substantially changes coordination strategies in non-human primates. This highlights a key strength of the DIP, which is that it removes a level of abstraction that can make tasks more difficult for non-human primates to learn. By being able to see their partner and actions directly, rather than having to understand that a cursor on a screen represents a partner, the platform makes the task more accessible to non-human primates and possibly children as well. This opens up important avenues for enhanced cross-species investigations of cognition, allowing researchers to study social dynamics in a setting that remains naturalistic yet controlled across different populations.

Weaknesses:

Some of the experimental applications would benefit from stronger evidence demonstrating the unique advantages of the transparent setup. For instance, in the dyadic foraging example, it's not entirely clear how participants' behavior differs from what might be observed when simply tracking each other's cursor movements in a non-transparent setup. More evidence showing how direct visibility of the partner, beyond simply being able to track the position of the partner's cursor, influences behavior would strengthen this example. Similarly, in the continuous perceptual report (CPR) task, the subjects could perform this task and see feedback from their partners' actions without having to see their partner through the transparent screen. Evidence showing that 1) subjects do indeed look at their partner during the task and 2) viewing their partner influences their performance on the task would significantly strengthen the claim that the ability to view the partner brings in a new dimension to this task. These additions would better demonstrate the specific value added by the transparent nature of the DIP beyond what could be achieved with standard cursor-tracking paradigms.

A significant limitation that is inadequately addressed relates to neural investigations. While the authors position the platform's ability to merge attention to social stimuli and task stimuli as a key advantage, they don't sufficiently acknowledge the challenges this creates for dissociating neural signals attributed to social cues versus task-based stimuli. More traditional lab-based experiments intentionally separate components like task-stimulus perception, social perception, and decision-making periods so that researchers can isolate the neural signals associated with each process. This deliberate separation, which the authors frame as a weakness, actually serves an important functional purpose in neural investigations. The paper would be strengthened by explicitly discussing this limitation and offering potential approaches to address it in experimental design or data analysis. For instance, the authors could suggest methodological innovations or analytical techniques that might help disentangle the overlapping neural signals that would inevitably arise from the integrated presentation of social and task stimuli in the DIP setup.

Furthermore, the authors' suggestion to arrange task stimuli around the periphery of the screen to maintain a clear middle area for viewing the partner appears to contradict their own critique of traditional paradigms. This recommended arrangement would seemingly reintroduce the very problem of attentional switching between task stimuli and social partners that the authors identified as a limitation of previous approaches. The paper would be strengthened by discussing the potential trade-offs associated with their suggested stimulus arrangement. Additionally, offering potential approaches to address these limitations in experimental design or data analysis would enhance the paper's contribution to the field.

Reviewer #2 (Public review):

Summary:

This work proposes a new platform to study social cognition in a more naturalistic setting. The authors give an overview of previous work that extends from static unidirectional paradigms (i.e., subject is presented with social stimuli such as still images or faces), to more dynamic unidirectional paradigms (i.e., the subject is presented with movies, or another individual's behavior) to dyadic interactions in a laboratory setting or in real life (i.e., interacting with a real person). Overall, this literature demonstrates that findings from realistic social situations can differ dramatically from unidirectional laboratory settings. Moreover, current and previous work are put in the perspective of an experimental framework that has tightly controlled experimental set-ups and low ecological validity on one end, and high ecological validity, naturalistic, without any experimental constraints on the other end, and all that is in between. The authors frame previous work along a spectrum, ranging from highly controlled, low-ecological-validity experiments to naturalistic, unconstrained approaches with high ecological validity, situating their current work within this continuum. They focus on a specific sub-domain of social interactions, i.e., goal-directed contexts in which interactions are purposeful for solving joint tasks or obtaining rewards. This new dyadic interaction platform claims to embed tight experimental control in a naturalistic face-to-face social interaction with the goal of investigating social information processing in bidirectional, dynamic social interactions.

Strengths:

The proposed dyadic interaction platform (DIP) is highly flexible, accommodating diverse visual displays, interactive components, and recording devices, making it suitable for various experiments.

The manuscript does a good job of highlighting the strengths and weaknesses of the various display options. This clarity allows readers to easily assess which display best suits their specific experimental setup and objectives.

One of the platform's key strengths is its versatility, allowing the same experimental setup to be used across multiple species and developmental stages, and enabling NHPs and humans to be studied as subjects within the same paradigm. Highlighting this capability could further underscore the platform's broad applicability.

Weaknesses:

The manuscript emphasizes the importance of ecological validity alongside tight experimental control, a significant challenge in naturalistic neuroscience. While the platform achieves tight control, the ecological validity of such a set-up remains questionable and warrants further testing and validation. For example, while the platform is designed to be more naturalistic in principle, its application to NHPs is still complex and may be comparably constrained as traditional NHP research. To realize its full potential for animal studies, the platform should be combined with complementary methodologies - such as wireless electrophysiology and freely moving paradigms - to truly achieve a balance between ecological validity and experimental control. Further validation in this direction could significantly enhance its utility.

The manuscript is somewhat lengthy and occasionally reads more like a review paper, which slightly shifts the focus away from the primary emphasis on the innovative technological advancement and the considerable effort invested in optimizing this new platform. Streamlining the presentation to more directly highlight these key contributions could enhance clarity and impact.

Overall, there is compelling evidence supporting the feasibility and value of DIP for investigating specific types of social interactions, particularly in contexts where individuals share a workspace and have full transparency regarding their opponent's actions. While I believe that DIP has the potential to significantly impact the field, which is supported by preliminary data, its broader applicability remains an open question. This platform aligns well with recent initiatives aimed at enhancing ecological validity in neuroscience research across both human and animal models. To maximize its impact, it would be beneficial to more explicitly situate this work within that broader movement, emphasizing its relevance and potential to advance ecologically valid approaches in the field.