Analyzing the Time Course of Pupillometric Data.

Area of Science:

• Cognitive Psychology
• Psycholinguistics
• Data Analysis

Background:

• Pupil dilation is a popular measure for tracing language processing in psychological and psycholinguistic research.
• Current analysis methods often focus on extracted features, lacking consensus and potentially losing information from continuous pupil dilation signals.
• Recent advancements suggest analyzing pupil dilation trajectories directly for a more comprehensive understanding.

Purpose of the Study:

• To provide a tutorial for analyzing pupillometric data by focusing on continuous pupil dilation trajectories.
• To demonstrate the application of nonlinear regression and generalized additive mixed modeling for analyzing the full-time course of pupil dilation signals.
• To address challenges in time series analysis, specifically autocorrelation in pupillary signals.

Main Methods:

• Application of nonlinear regression analysis.
• Utilizing generalized additive mixed modeling (GAMM) to analyze the full-time course of pupil dilation.
• Inclusion of nonlinear random effects to account for participant and item variation.
• Addressing autocorrelation in residuals through simulations and experimental data analysis.

Main Results:

• Generalized additive mixed models effectively incorporate complex nonlinear interactions between stimulus properties or participant characteristics and pupil dilation.
• The proposed methods allow for a more coherent interpretation of pupillary data compared to traditional feature-based techniques.
• Potential causes of extreme autocorrelation in pupillary signals are explained, and methods to mitigate their adverse effects are demonstrated.

Conclusions:

• Direct analysis of pupil dilation trajectories using nonlinear regression and GAMM offers a more informative approach in psychological and psycholinguistic research.
• This methodology enhances the interpretation of pupillometric data by utilizing the full continuous signal.
• Addressing autocorrelation is crucial for robust analysis and reliable conclusions from pupillary data.