Mind wandering at the fingertips: automatic parsing of subjective states based on response time variability - Sorbonne Université Accéder directement au contenu
Article Dans Une Revue Frontiers in Psychology Année : 2013

Mind wandering at the fingertips: automatic parsing of subjective states based on response time variability

Résumé

Research from the last decade has successfully used two kinds of thought reports in order to assess whether the mind is wandering: random thought-probes and spontaneous reports. However, none of these two methods allows any assessment of the subjective state of the participant between two reports. In this paper, we present a step by step elaboration and testing of a continuous index, based on response time variability within Sustained Attention to Response Tasks (N = 106, for a total of 10 conditions). We first show that increased response time variability predicts mind wandering. We then compute a continuous index of response time variability throughout full experiments and show that the temporal position of a probe relative to the nearest local peak of the continuous index is predictive of mind wandering. This suggests that our index carries information about the subjective state of the subject even when he or she is not probed, and opens the way for on-line tracking of mind wandering. Finally we proceed a step further and infer the internal attentional states on the basis of the variability of response times. To this end we use the Hidden Markov Model framework, which allows us to estimate the durations of on-task and off-task episodes.
Fichier principal
Vignette du fichier
fpsyg-04-00573.pdf (1.7 Mo) Télécharger le fichier
Origine : Publication financée par une institution
Loading...

Dates et versions

hal-01537208 , version 1 (12-06-2017)

Licence

Paternité

Identifiants

Citer

Mikaël Bastian, Jérôme Sackur. Mind wandering at the fingertips: automatic parsing of subjective states based on response time variability. Frontiers in Psychology, 2013, 4, pp.573. ⟨10.3389/fpsyg.2013.00573⟩. ⟨hal-01537208⟩
177 Consultations
166 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More