May

Interacting Minds – A Workshop on the Neural Basis of Real-Time Social Interaction

5 May 2026 09:00 to 6 May 2026 15:00 | Workshop

Illustration. Two stick figures dancing with each other. In the background are two floating brains that are connected to each other by a data cable.

Registration is now closed

Social interaction lies at the heart of human cognition, yet much of cognitive neuroscience has traditionally examined individuals in isolation. Approaches such as hyperscanning—simultaneous recording of neural activity from two or more interacting individuals—now allow researchers to study the neural dynamics underlying real-time social behaviour.

Organised within the Natural and Artificial Cognition (NAC) profile area at Lund University, this workshop brings together researchers interested in the neural basis of interactive behaviour. Through invited talks, discussions, and demonstrations, the event will highlight how hyperscanning approaches can advance understanding of communication, cooperation, and other forms of social interaction across NAC research areas.

Time: May 5-6, 2026

Venue: The Pufendorf Institute for Advanced Studies, Biskopsgatan 3, Lund, Sweden

Primary audience researchers (estimate: 30-40 attendants).
The number of seats are limited, so we will have to prioritise who we can receive.

Registration: Registration is now closed

Programme - 5 May

Full programme with abstracts below.

09:00 – Fika

09:20 – Welcome and introduction

10:30 – Break

11:00 – Talk 2: An Introduction to fNIRS Hyperscanning: From Methods to Neural Synchrony in Children
Speaker: Verena Schäfer, University of Vienna

11:45 – Lunch

14:40 – Break

15.30 – Talk 4: Deep Language as a Cognitive Model for the Development and Processing of Natural Language
Speaker: Uri Hasson, Professor of Psychology and Neuroscience, Princeton University

15:15 – Fika

Programme - 6 May

9:00 – Fika

09:30 – Panel discussion

10:30 – Break

10:45 – Meet the speakers

12:00 – Lunch

13:00-15:00 – Demos by fNIRS and EEG companies:

ANT Neuro, Akash Basu
mBrainTrain, Filip Rakas
Artinis-NIRx, Dalila Burin and David Zijderveld

Programme 5 May with abstracts

09:00 – Fika

09:20 – Welcome and introduction

09:30 – Talk 1: The neurocognition of social interaction
Speaker: Antonia Hamilton, Professor of Social Neuroscience and leader of the Social Neuroscience group at the Institute of Cognitive Neuroscience (UCL)
Abstract: Interpersonal neural synchrony has been described in a variety of contexts where brain activity is recorded from two or more people simultaneously. However, there has been controversy around how to interpret this signal. Here, I will present a mutual prediction model of hyperscanning effects and argue that these studies can be informative, even if there is no direct cross-brain causal mechanism. I will share new data from studies of conversation and learning to illustrate how hyperscanning can be used to understand social coordination between people in a variety of contexts.

10:30 – Break

11:00 – Talk 2: An Introduction to fNIRS Hyperscanning: From Methods to Neural Synchrony in Children
Speaker: Verena Schäfer, University of Vienna
Abstract: Hyperscanning has become an important approach for investigating real-time social interaction by simultaneously recording neural activity from multiple individuals. This talk provides an introduction to hyperscanning research with a focus on functional nearinfrared spectroscopy (fNIRS), covering basic principles to assessing interpersonal neural synchrony. Particular attention is given to practical challenges in conducting hyperscanning experiments, including issues related to experimental control, data quality, and preprocessing of synchrony measures. These challenges are discussed with a specific emphasis on developmental research. The perspective highlights practical considerations and lessons learned from an early career researcher. The second part of the talk presents a practical example of a hyperscanning study investigating interpersonal neural synchrony in children. The study examines how shared laughter relates to neural synchrony and subsequent social outcomes such as liking and bonding. Overall, the talk combines a methodological introduction with an applied research example, aiming to provide a practical and accessible entry point into hyperscanning research.

11:45 – Lunch

13:30 – Talk 3: EEG hyperscanning for measuring rhythms of social interaction: Insights & practical lessons
Speaker: Anna Zamm, Associate Professor at School of Communication and Culture, at Aarhus University.
Abstract: Social interaction in humans is often facilitated by rhythmic structures such as turn-taking, which provide a clear framework for interpersonal coordination. Understanding how social partners use rhythm to predict and coordinate actions requires methods with high temporal resolution. EEG hyperscanning is increasingly used for this purpose, due to the excellent temporal resolution of EEG and the increasing mobility of modern EEG systems. However, EEG hyperscanning has received criticism highlighting theoretical and methodological inconsistencies across studies. In this talk, I outline guidelines for avoiding these pitfalls when designing and implementing EEG hyperscanning studies, and present empirical findings highlighting the utility of EEG hyperscanning for revealing key neural dynamics underlying rhythmic social coordination.

14:40 – Break

15.30 – Talk 4: Deep Language as a Cognitive Model for the Development and Processing of Natural Language
Speaker: Uri Hasson, Professor of Psychology and Neuroscience, Princeton University

Abstract: Understanding human learning and cognition requires models that can handle the complexity of real-world experience while remaining grounded in biological plausibility. Traditional computational models generalize poorly beyond controlled laboratory conditions. Modern deep learning systems, despite their success on naturalistic tasks, face the opposite problem: they are data-hungry, developmentally implausible, and largely disconnected from how humans actually learn. We aim to bridge this divide by building cognitively plausible learning agents grounded in children's natural environments. Central to this effort is the First 1000 Days (1kD) dataset—a longitudinal, audiovisual record of children's daily lives from birth to age 3, captured with unprecedented detail of the individual developmental conditions that shape learning. Using 1kD recordings, we are developing child-centered learning agents that acquire language in socially grounded, multimodal contexts, directly mirroring the conditions of infant development. In parallel, we align AI model representations with human neural data by comparing them to invasive electrocorticography (ECoG) recordings of natural language processing — grounding our models not just behaviorally, but neurally. Together, these efforts define a unified framework for ecologically valid cognitive neuroscience: one that combines longitudinal developmental data, embodied learning agents, and neural alignment to illuminate the core computational principles underlying human cognition.