Pupillary Manifolds and Eye Tracking

The paper “Pupillary manifolds: Uncovering the latent geometrical structures behind phasic changes in pupil size” by Blini, Arrighi, and Anobile explores the underlying physiological structures governing changes in pupil size.

Generally, a pupillary manifold refers to the underlying geometric structure that captures how the pupil changes size rapidly—known as phasic changes—across various cognitive and physiological conditions. A pupillary manifold is supposed to serve as a low-dimensional representation, like a simplified map, that reveals consistent patterns in pupil responses regardless of the specific task or stimulus. Note, “pupillary” pertains to the pupil and its behavior, while “manifold” is a mathematical term for a smooth, often curved surface that simplifies complex, high-dimensional data into a more manageable form.

The core research question in the Blini, Arrighi, and Anobile study revolves around whether functionally distinct pupillary signatures can be mapped onto a low-dimensional space, and if this space reflects task-specific functions or rather unspecific physiological constraints. This inquiry addresses two significant limitations in current pupillometry research: the lack of analytical consensus and the challenge of discerning distinct sources of observed effects from an integrated pupillometry readout.

Pupillometry Eye-Tracking Methodology

The authors utilized dimensionality reduction techniques, primarily Principal Components Analysis (PCA) and promax-rotated PCA (rPCA), to unveil these latent processes. They conducted three experiments: mapping pupillary responses to luminance changes (Pupillary Light Reflex – PLR), mapping responses to varying Working Memory Load (WML), and a combined task manipulating both luminance and memory load.

The study employed a remote infrared-based eye-tracker, the SR Research EyeLink 1000, as a crucial tool in data collection. The EyeLink 1000 was set to continuously monitor participants’ pupil size at a 500 Hz sampling rate, ensuring high-fidelity data acquisition. This precision was vital for capturing the subtle, dynamic changes in pupil size that formed the basis of the research. Furthermore, the system allowed the researchers to identify and discard pupil size measurements during blinks and saccades.

Phasic Changes in Pupil Size are Low-Dimension

The key finding is that phasic changes in pupil size are inherently low-dimensional, with modes that are highly consistent across behavioral tasks of very different natures. This suggests that these changes occur along “pupillary manifolds” that are highly constrained by underlying physiological structures, such as the relative balance between sympathetic and parasympathetic activity, rather than specific functions. 

With promax-rotated PCA, three remarkably similar components were recovered across all three tasks, accounting for over 87% of the total phasic pupillary dynamics. Critically, these components were found to track both light reflexes and mental effort, indicating a common underlying structure. One component (RC3) was specific to luminance changes but not cognitive load, loading on the earliest time points and possibly reflecting a parasympathetic component. The consistency of these latent structures across diverse tasks points to fundamental physiological constraints shaping pupillary responses.

 The robust and accurate pupillometric data provided by the EyeLink system was critical in enabling the researchers to effectively apply sophisticated dimensionality reduction techniques, ultimately uncovering the low-dimensional pupillary manifolds and advancing the understanding of how pupil size reflects physiological and cognitive processes. 

For information regarding how eye tracking can help your research, check out our solutions and product pages or contact us. We are happy to help!