Limits...
Familiar Face Detection in 180 ms.

Visconti di Oleggio Castello M, Gobbini MI - PLoS ONE (2015)

Bottom Line: We investigated the effect of personal familiarity on the earliest stages of face processing by using a saccadic-choice task to measure how fast familiar face detection can happen.By contrast, accuracy of saccades to unfamiliar faces with familiar faces as distractors did not exceed chance.Saccades to faces with object distractors were even faster (110 to 120 ms) and equivalent for familiar and unfamiliar faces, indicating that familiarity does not affect ultra-rapid saccades.

View Article: PubMed Central - PubMed

Affiliation: Department of Psychological and Brain Sciences, Dartmouth College, Hanover, New Hampshire, United States of America.

ABSTRACT
The visual system is tuned for rapid detection of faces, with the fastest choice saccade to a face at 100 ms. Familiar faces have a more robust representation than do unfamiliar faces, and are detected faster in the absence of awareness and with reduced attentional resources. Faces of family and close friends become familiar over a protracted period involving learning the unique visual appearance, including a view-invariant representation, as well as person knowledge. We investigated the effect of personal familiarity on the earliest stages of face processing by using a saccadic-choice task to measure how fast familiar face detection can happen. Subjects made correct and reliable saccades to familiar faces when unfamiliar faces were distractors at 180 ms--very rapid saccades that are 30 to 70 ms earlier than the earliest evoked potential modulated by familiarity. By contrast, accuracy of saccades to unfamiliar faces with familiar faces as distractors did not exceed chance. Saccades to faces with object distractors were even faster (110 to 120 ms) and equivalent for familiar and unfamiliar faces, indicating that familiarity does not affect ultra-rapid saccades. We propose that detectors of diagnostic facial features for familiar faces develop in visual cortices through learning and allow rapid detection that precedes explicit recognition of identity.

No MeSH data available.


Related in: MedlinePlus

Parameter estimates for Task in the Logit Mixed-Effects Models obtained by changing reference level for target position. Error bars represent 95% bootstrapped confidence intervals.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4549263&req=5

pone.0136548.g003: Parameter estimates for Task in the Logit Mixed-Effects Models obtained by changing reference level for target position. Error bars represent 95% bootstrapped confidence intervals.

Mentions: Table 1 reports the accuracies for each condition, and Fig 3 shows the parameter estimates (log Odds) of the model for the Task variable with 95% confidence intervals computed through parametric bootstrapping with 10,000 repetitions. Since the model was computed without intercept, all parameter estimates are contrasted against 0, or chance level. We thus determined that a task was significantly above chance if its confidence interval did not contain 0. Moreover, confidence intervals can be used to judge whether parameter estimates for two tasks are likely to be different.


Familiar Face Detection in 180 ms.

Visconti di Oleggio Castello M, Gobbini MI - PLoS ONE (2015)

Parameter estimates for Task in the Logit Mixed-Effects Models obtained by changing reference level for target position. Error bars represent 95% bootstrapped confidence intervals.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4549263&req=5

pone.0136548.g003: Parameter estimates for Task in the Logit Mixed-Effects Models obtained by changing reference level for target position. Error bars represent 95% bootstrapped confidence intervals.
Mentions: Table 1 reports the accuracies for each condition, and Fig 3 shows the parameter estimates (log Odds) of the model for the Task variable with 95% confidence intervals computed through parametric bootstrapping with 10,000 repetitions. Since the model was computed without intercept, all parameter estimates are contrasted against 0, or chance level. We thus determined that a task was significantly above chance if its confidence interval did not contain 0. Moreover, confidence intervals can be used to judge whether parameter estimates for two tasks are likely to be different.

Bottom Line: We investigated the effect of personal familiarity on the earliest stages of face processing by using a saccadic-choice task to measure how fast familiar face detection can happen.By contrast, accuracy of saccades to unfamiliar faces with familiar faces as distractors did not exceed chance.Saccades to faces with object distractors were even faster (110 to 120 ms) and equivalent for familiar and unfamiliar faces, indicating that familiarity does not affect ultra-rapid saccades.

View Article: PubMed Central - PubMed

Affiliation: Department of Psychological and Brain Sciences, Dartmouth College, Hanover, New Hampshire, United States of America.

ABSTRACT
The visual system is tuned for rapid detection of faces, with the fastest choice saccade to a face at 100 ms. Familiar faces have a more robust representation than do unfamiliar faces, and are detected faster in the absence of awareness and with reduced attentional resources. Faces of family and close friends become familiar over a protracted period involving learning the unique visual appearance, including a view-invariant representation, as well as person knowledge. We investigated the effect of personal familiarity on the earliest stages of face processing by using a saccadic-choice task to measure how fast familiar face detection can happen. Subjects made correct and reliable saccades to familiar faces when unfamiliar faces were distractors at 180 ms--very rapid saccades that are 30 to 70 ms earlier than the earliest evoked potential modulated by familiarity. By contrast, accuracy of saccades to unfamiliar faces with familiar faces as distractors did not exceed chance. Saccades to faces with object distractors were even faster (110 to 120 ms) and equivalent for familiar and unfamiliar faces, indicating that familiarity does not affect ultra-rapid saccades. We propose that detectors of diagnostic facial features for familiar faces develop in visual cortices through learning and allow rapid detection that precedes explicit recognition of identity.

No MeSH data available.


Related in: MedlinePlus