Model-based pupil deconvolution reveals latent attentional dynamics

This dataset contains behavioural and pupillometry data collected during an experiment investigating the temporal deployment of attention in response to rapid visual stimulation and its modulation by the emotional content of successive targets. The experiment used a rapid serial visual presentation (RSVP) paradigm in which participants viewed a sequence of emotional faces containing two task-relevant targets (T1 and T2). The temporal separation between the two targets was manipulated across different lags, and the emotional content of the targets was varied. Participants reported the emotional category of the targets after each trial. Pupil responses were recorded during the task. The dataset provides the processed behavioural and pupil data used to examine how temporal separation and emotional content influence attentional processing. Participants viewed a rapid sequence of face stimuli presented using an RSVP paradigm. Each stimulus was presented for 100 ms, followed by a 50 ms blank interval, resulting in a stimulus onset asynchrony (SOA) of 150 ms. Two task-relevant targets, T1 and T2, were embedded within the RSVP stream. T1 appeared at serial positions 6–8, while T2 appeared at one of three temporal lags following T1: Lag 1, 2, 3 and 5. The target faces were either happy or angry, producing four possible target-pair conditions: HH: Happy T1 – Happy T2, HA: Happy T1 – Angry T2, AH: Angry T1 – Happy T2 and AA: Angry T1 – Angry T2. The remaining stimuli in the RSVP stream were neutral faces. Trials included both two-target and single-target conditions. The dataset contains data from 37 participants, out of which 32 were included in the final analysis. Participants were between 18 and 24 years of age. Participant-level identifiers in the dataset are pseudonymised and do not contain directly identifying information. Pupil data were recorded using an EyeLink 1000 eye tracker at a sampling rate of 500 Hz, using the right eye. The recorded pupil signal was processed before analysis. Processing included blink handling, interpolation of missing samples, downsampling, and baseline correction. The resulting pupil time series were aligned to the RSVP onset and used for subsequent analysis. File Format: The data are provided in DataMatrix (.dm) format. DataMatrix is a Python-based data structure and file format designed for behavioural and experimental research data. The .dm files can be read using the DataMatrix Python package. The dataset is provided in its original .dm format to preserve the structure used during data preprocessing and analysis. Here is a direct link to DataMatrix Package