The open field is a standard experimental tool used to assess locomotion and anxiety-like behavior in animal models of depression. It involves placing a mouse into a barren, over-lit box and analyzing their movement during this stressful experience. Despite its continued use, open field has been criticized for being too contrived, and yielding limited, anthropomorphized readouts. Recent advances in machine vision tools, such as DeepLabCut or MoSeq, promise to extract richer insights from recordings of freely behaving animals and to bypass the need for biased human interpretation.

In this artistic research project, I aimed to explore and subtly subvert these tools through a performative self-experiment. Over several days, I recorded myself in my home under three conditions: (1) no intervention (naturalistic behavior), (2) while behaving according to randomly generated activity commands, and (3) real-time commands derived from a previous day’s movement pattern analysis. I annotated my activity via an unsupervised machine readout using DeepLabCut pose estimation and MoSeq sequencing and compared it to auto-phenomenographical notes.

Rather than seeking empirical conclusions, this project aims to interrogate agency and the gap between lived experience and behavior observed from the outside. By placing myself in the role of both subject and observer, I explored what is captured by machine vision, what is left out. Echoing Harun Farocki’s critique of instrumental seeing and Sophie Calle’s intimate surveillance, this work inhabits the boundary between data analysis and embodied knowledge.

Authors:
Terezie Sedlinska, Okinawa Institute of Science and Technology, Japan

About the Presenter(s)
Dr.med. Terezie Sedlinska is an architect, medical doctor, trans-disciplinary researcher and currently neuroscience PhD candidate at the Okinawa Institute of Science and Technology.

See this presentation on the full schedule – Tuesday Schedule

Bookmark