Limits...
Rapid gist perception of meaningful real-life scenes: Exploring individual and gender differences in multiple categorization tasks.

Vanmarcke S, Wagemans J - Iperception (2015)

Bottom Line: Since this pioneering work, follow-up studies consistently reported population-level reaction time differences on different categorization tasks, indicating a superordinate advantage (animal versus dog) and effects of perceptual similarity (animals versus vehicles) and object category size (natural versus animal versus dog).In this study, we replicated and extended these separate findings by using a systematic collection of different categorization tasks (varying in presentation time, task demands, and stimuli) and focusing on individual differences in terms of e.g., gender and intelligence.In addition to replicating the main findings from the literature, we find subtle, yet consistent gender differences (women faster than men).

View Article: PubMed Central - PubMed

Affiliation: Laboratory of Experimental Psychology, University of Leuven (KU Leuven), Leuven, Belgium, e-mail: steven.vanmarcke@ppw.kuleuven.be.

ABSTRACT
In everyday life, we are generally able to dynamically understand and adapt to socially (ir)elevant encounters, and to make appropriate decisions about these. All of this requires an impressive ability to directly filter and obtain the most informative aspects of a complex visual scene. Such rapid gist perception can be assessed in multiple ways. In the ultrafast categorization paradigm developed by Simon Thorpe et al. (1996), participants get a clear categorization task in advance and succeed at detecting the target object of interest (animal) almost perfectly (even with 20 ms exposures). Since this pioneering work, follow-up studies consistently reported population-level reaction time differences on different categorization tasks, indicating a superordinate advantage (animal versus dog) and effects of perceptual similarity (animals versus vehicles) and object category size (natural versus animal versus dog). In this study, we replicated and extended these separate findings by using a systematic collection of different categorization tasks (varying in presentation time, task demands, and stimuli) and focusing on individual differences in terms of e.g., gender and intelligence. In addition to replicating the main findings from the literature, we find subtle, yet consistent gender differences (women faster than men).

No MeSH data available.


Related in: MedlinePlus

Overview of reaction time (A) and accuracy (C) outcomes in the ultrarapid categorization animal/vehicle task. The data are represented as the mean performance across participants, with error bars depicting the standard error of the mean (SEM). For accuracy, mean and SEM were calculated based on the logistic transformation of the values and then retransformed into percentage correct (%) data. A similar overview is provided for the reaction time (B) and accuracy (D) outcomes in the ultrarapid categorization social task. Men are always depicted in blue, women in red. The dotted lines depict the mean expected RT (A, B) or accuracy (C, D) as fitted by the final model for both men (light blue) and women (light red) based on (1) a random intercepts regression analysis for RT and (2) a random intercepts logistic regression analysis for accuracy. The latter fit was averaged (for both RT and accuracy) over time.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4441019&req=5

Figure 3: Overview of reaction time (A) and accuracy (C) outcomes in the ultrarapid categorization animal/vehicle task. The data are represented as the mean performance across participants, with error bars depicting the standard error of the mean (SEM). For accuracy, mean and SEM were calculated based on the logistic transformation of the values and then retransformed into percentage correct (%) data. A similar overview is provided for the reaction time (B) and accuracy (D) outcomes in the ultrarapid categorization social task. Men are always depicted in blue, women in red. The dotted lines depict the mean expected RT (A, B) or accuracy (C, D) as fitted by the final model for both men (light blue) and women (light red) based on (1) a random intercepts regression analysis for RT and (2) a random intercepts logistic regression analysis for accuracy. The latter fit was averaged (for both RT and accuracy) over time.

Mentions: The final model (see Table 2 for parameter estimates and 95% confidence intervals) for both RT and accuracy (see Figure 3A and C) provided a significant random intercept (RT: t59.64 = 44.21; p < .001 / Accuracy: Z = 15.24; p < .001). The analysis furthermore provided a clear replication of the superordinate effect [Aim 1] (e.g., Macé et al., 2009; Praβ et al., 2014) due to the significant impact of the fixed (within-subjects) factor Level of Categorization (RT: t47.18 = −10.27; p < .001 / Accuracy: Z = 12.72; p < .001) on outcome prediction. It also yielded an interesting finding regarding the processing of scene gist perception versus object perception, namely a significantly faster and more accurate categorization of “Object” versus “Scene” information. The latter was exemplified by the significant effect of the fixed (within-subjects) factor Goal on predicting task RT and accuracy (RT: t46.45 = 8.47; p < .001 / Accuracy: Z = −6.60; p < .001). The conducted random intercept regression analysis also indicated that people were faster and more accurate at detecting Inanimate versus Animate information as exemplified by the significant effect of the fixed (within-subjects) factor Animacy (RT: t45.96 = 3.61; p < .001 / Accuracy: Z = −3.19; p < .01) on outcome prediction. The moment of testing (tested by adding the fixed (within-subject) factor Time (RT: t26.96 = 1.12; p = .27 / Accuracy: Z = −1.70; p = .09) did not lead to a significantly better prediction of the dependent variables. This indicated that the randomized sequence of test blocks did not elicit any detrimental/learning effects on RT or accuracy.


Rapid gist perception of meaningful real-life scenes: Exploring individual and gender differences in multiple categorization tasks.

Vanmarcke S, Wagemans J - Iperception (2015)

Overview of reaction time (A) and accuracy (C) outcomes in the ultrarapid categorization animal/vehicle task. The data are represented as the mean performance across participants, with error bars depicting the standard error of the mean (SEM). For accuracy, mean and SEM were calculated based on the logistic transformation of the values and then retransformed into percentage correct (%) data. A similar overview is provided for the reaction time (B) and accuracy (D) outcomes in the ultrarapid categorization social task. Men are always depicted in blue, women in red. The dotted lines depict the mean expected RT (A, B) or accuracy (C, D) as fitted by the final model for both men (light blue) and women (light red) based on (1) a random intercepts regression analysis for RT and (2) a random intercepts logistic regression analysis for accuracy. The latter fit was averaged (for both RT and accuracy) over time.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4441019&req=5

Figure 3: Overview of reaction time (A) and accuracy (C) outcomes in the ultrarapid categorization animal/vehicle task. The data are represented as the mean performance across participants, with error bars depicting the standard error of the mean (SEM). For accuracy, mean and SEM were calculated based on the logistic transformation of the values and then retransformed into percentage correct (%) data. A similar overview is provided for the reaction time (B) and accuracy (D) outcomes in the ultrarapid categorization social task. Men are always depicted in blue, women in red. The dotted lines depict the mean expected RT (A, B) or accuracy (C, D) as fitted by the final model for both men (light blue) and women (light red) based on (1) a random intercepts regression analysis for RT and (2) a random intercepts logistic regression analysis for accuracy. The latter fit was averaged (for both RT and accuracy) over time.
Mentions: The final model (see Table 2 for parameter estimates and 95% confidence intervals) for both RT and accuracy (see Figure 3A and C) provided a significant random intercept (RT: t59.64 = 44.21; p < .001 / Accuracy: Z = 15.24; p < .001). The analysis furthermore provided a clear replication of the superordinate effect [Aim 1] (e.g., Macé et al., 2009; Praβ et al., 2014) due to the significant impact of the fixed (within-subjects) factor Level of Categorization (RT: t47.18 = −10.27; p < .001 / Accuracy: Z = 12.72; p < .001) on outcome prediction. It also yielded an interesting finding regarding the processing of scene gist perception versus object perception, namely a significantly faster and more accurate categorization of “Object” versus “Scene” information. The latter was exemplified by the significant effect of the fixed (within-subjects) factor Goal on predicting task RT and accuracy (RT: t46.45 = 8.47; p < .001 / Accuracy: Z = −6.60; p < .001). The conducted random intercept regression analysis also indicated that people were faster and more accurate at detecting Inanimate versus Animate information as exemplified by the significant effect of the fixed (within-subjects) factor Animacy (RT: t45.96 = 3.61; p < .001 / Accuracy: Z = −3.19; p < .01) on outcome prediction. The moment of testing (tested by adding the fixed (within-subject) factor Time (RT: t26.96 = 1.12; p = .27 / Accuracy: Z = −1.70; p = .09) did not lead to a significantly better prediction of the dependent variables. This indicated that the randomized sequence of test blocks did not elicit any detrimental/learning effects on RT or accuracy.

Bottom Line: Since this pioneering work, follow-up studies consistently reported population-level reaction time differences on different categorization tasks, indicating a superordinate advantage (animal versus dog) and effects of perceptual similarity (animals versus vehicles) and object category size (natural versus animal versus dog).In this study, we replicated and extended these separate findings by using a systematic collection of different categorization tasks (varying in presentation time, task demands, and stimuli) and focusing on individual differences in terms of e.g., gender and intelligence.In addition to replicating the main findings from the literature, we find subtle, yet consistent gender differences (women faster than men).

View Article: PubMed Central - PubMed

Affiliation: Laboratory of Experimental Psychology, University of Leuven (KU Leuven), Leuven, Belgium, e-mail: steven.vanmarcke@ppw.kuleuven.be.

ABSTRACT
In everyday life, we are generally able to dynamically understand and adapt to socially (ir)elevant encounters, and to make appropriate decisions about these. All of this requires an impressive ability to directly filter and obtain the most informative aspects of a complex visual scene. Such rapid gist perception can be assessed in multiple ways. In the ultrafast categorization paradigm developed by Simon Thorpe et al. (1996), participants get a clear categorization task in advance and succeed at detecting the target object of interest (animal) almost perfectly (even with 20 ms exposures). Since this pioneering work, follow-up studies consistently reported population-level reaction time differences on different categorization tasks, indicating a superordinate advantage (animal versus dog) and effects of perceptual similarity (animals versus vehicles) and object category size (natural versus animal versus dog). In this study, we replicated and extended these separate findings by using a systematic collection of different categorization tasks (varying in presentation time, task demands, and stimuli) and focusing on individual differences in terms of e.g., gender and intelligence. In addition to replicating the main findings from the literature, we find subtle, yet consistent gender differences (women faster than men).

No MeSH data available.


Related in: MedlinePlus