disrupting face biases in visual attention anna s. law, liverpool john moores university stephen r....

1
Disrupting face biases in visual attention Anna S. Law, Liverpool John Moores University Stephen R. H. Langton, University of Stirling Introduction Method The primary (dot probe) task was a replication of Bindemann et al. (2007). The secondary task was identical to that of Lavie & de Fockert (2005). Undergraduate participants (N=24) did 2 blocks of trials in counterbalanced order: 1 single task (dot probe only) and 1 dual task (dot probe + digit rehearsal). Results Fig 3: Mean RT at 500ms SOA, collapsed across demand level + 03124 + Fig 1: “Face-valid” dual task trial The stimulus onset asynchrony (SOA) between the picture cues and the target was either 100ms or 500ms Participants localised target with “left” and “right” button press responses (vertical mapping of keys) Participants had to recall the number that came next in the earlier sequence (in this case, 1) Fig 2: Mean RT at 100ms SOA, collapsed across demand level Discussion In this preliminary study, a demanding secondary task had little impact on the face bias. This emphasises the strong influence that faces exert on attention. However, there was some evidence that the face bias was reduced for those who attempted the task initially under dual task load (at least at the longer SOA). Results are therefore more consistent with prediction 2 above than prediction 1. Endogenous mechanisms could be involved in the creation of an attentional set for faces, which once adopted is hard to overcome. Dual task demand may make it less likely that such an attentional set is created. Future work will investigate this possibility. References Bindemann, M., Burton, A. M., Langton, S. R. H., Schweinberger, S. R., & Doherty, M. J. (2007). The control of attention to faces. Journal of Vision, 7(10):15, 1-8 Langton, S. R. H., Law, A. S., Burton, A. M., & Schweinberger, S. R. (2008). Attentional capture by faces. Cognition, 107, 330-342. Lavie, N., & de Fockert, J. (2005). The role of working memory load in attention capture. Psychonomic Bulletin & Review, 12, 669-674. Accuracy in the digit task: Type of trial in the dot probe task made no difference error rate in the memory task. Overall error rate was 11.46%. 3 Target appeared in location of face or object 300 350 400 450 500 550 600 650 700 Face Object Reation Tim e (m s) Single first Dual first 300 350 400 450 500 550 600 650 700 Face Object Reaction Tim e (m s) Single first Dual first At 100ms SOA there was a marginal face bias for “single first” and “dual first” groups, and no interaction. At 500ms SOA there was a face bias for the “single first” group but this was eliminated for the “dual first” group. 100ms SOA 500ms SOA Face Object Face Object Single Task 501 (73) 503 (62) 489 (71) 497 (71) Dual Task 589 (111) 602 (107) 537 (149) 541 (115) Table 1: Mean (and SD) of RTs to targets appearing at face and object locations, at both SOAs under single and dual task conditions Results cont. Overall, people were faster on face trials then object trials, F(1,22)=4.19, p=.053. But an important finding was that the order in which participants attempted the single and dual task blocks interacted with both soa and the face effect, F(1,22)=12.78, p<.01. To explore this 3-way interaction, data were collapsed across single and dual task blocks, and analysed separately for each SOA. At 100ms SOA, there were marginal effects of face, F(1,22) = 4.05, p=.057, and order, F(1,22) = 3.81, p=.064 but no interaction. At 500ms SOA, there was no main effect of face but there was a significant interaction, F(1,22) = 6.61, p<.05. Analysis of simple main effects showed that there was a significant effect of face for those who did the single- task block first, F(1,44) = 15.56, p<.001, but not for those who did dual first. 1) Control processes are normally involved by suppressing (to some extent) the inherently distracting influence of faces. Therefore, dual-task load will lead to an increase in the face bias. 2) Control processes are normally involved in creating an attentional set that results in a face bias during the task. Therefore, dual-task load will lead to a decrease in the face bias. Faster response if target then appears here Slower response if target then appears here Faces attract visual attention (e.g., Langton et al., 2008). Bindemann et al. (2007) demonstrated this with a dot probe task. They also demonstrated that the face bias was largely under voluntary or endogenous control. The present study explored the nature of this endogenous control, using a dual task procedure. There are 2 possible predictions: Contact: [email protected] .uk

Upload: reynold-garrison

Post on 18-Jan-2018

217 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Disrupting face biases in visual attention Anna S. Law, Liverpool John Moores University Stephen R. H. Langton, University of Stirling Introduction Method

Disrupting face biases in visual attention

Anna S. Law, Liverpool John Moores UniversityStephen R. H. Langton, University of Stirling

Introduction

MethodThe primary (dot probe) task was a replication of Bindemann et al. (2007). The secondary task was identical to that of Lavie & de Fockert (2005). Undergraduate participants (N=24) did 2 blocks of trials in counterbalanced order: 1 single task (dot probe only) and 1 dual task (dot probe + digit rehearsal).

Results

Fig 3: Mean RT at 500ms SOA, collapsed across demand level+

03124

+

Fig 1: “Face-valid” dual task trial

The stimulus onset asynchrony (SOA) between the picture cues and the target was either 100ms or 500ms

Participants localised target with “left” and “right” button press responses (vertical mapping of keys)

Participants had to recall the number that came next in the earlier sequence (in this case, 1)

Fig 2: Mean RT at 100ms SOA, collapsed across demand level

DiscussionIn this preliminary study, a demanding secondary task had little impact on the face bias. This emphasises the strong influence that faces exert on attention.

However, there was some evidence that the face bias was reduced for those who attempted the task initially under dual task load (at least at the longer SOA). Results are therefore more consistent with prediction 2 above than prediction 1.

Endogenous mechanisms could be involved in the creation of an attentional set for faces, which once adopted is hard to overcome. Dual task demand may make it less likely that such an attentional set is created. Future work will investigate this possibility.

References

Bindemann, M., Burton, A. M., Langton, S. R. H., Schweinberger, S. R., & Doherty, M. J. (2007). The control of attention to faces. Journal of Vision, 7(10):15, 1-8

Langton, S. R. H., Law, A. S., Burton, A. M., & Schweinberger, S. R. (2008). Attentional capture by faces. Cognition, 107, 330-342.

Lavie, N., & de Fockert, J. (2005). The role of working memory load in attention capture. Psychonomic Bulletin & Review, 12, 669-674.

Accuracy in the digit task: Type of trial in the dot probe task made no difference error rate in the memory task. Overall error rate was 11.46%.

3

Target appeared in location of face or object

300

350

400

450

500

550

600

650

700

Face Object

Rea

tion

Tim

e (m

s)

Single f irst

Dual f irst

300

350

400

450

500

550

600

650

700

Face Object

Rea

ctio

n Ti

me

(ms)

Single f irst

Dual f irst

At 100ms SOA there was a marginal face bias for “single first” and “dual first” groups, and no interaction.

At 500ms SOA there was a face bias for the “single first” group but this was eliminated for the “dual first” group.

100ms SOA 500ms SOA

Face Object Face Object

Single Task 501 (73) 503 (62) 489 (71) 497 (71)

Dual Task 589 (111) 602 (107) 537 (149) 541 (115)

Table 1: Mean (and SD) of RTs to targets appearing at face and object locations, at both SOAs under single and dual task conditions

Results cont.Overall, people were faster on face trials then object trials, F(1,22)=4.19, p=.053.

But an important finding was that the order in which participants attempted the single and dual task blocks interacted with both soa and the face effect, F(1,22)=12.78, p<.01.

To explore this 3-way interaction, data were collapsed across single and dual task blocks, and analysed separately for each SOA. At 100ms SOA, there were marginal effects of face, F(1,22) = 4.05, p=.057, and order, F(1,22) = 3.81, p=.064 but no interaction.

At 500ms SOA, there was no main effect of face but there was a significant interaction, F(1,22) = 6.61, p<.05. Analysis of simple main effects showed that there was a significant effect of face for those who did the single-task block first, F(1,44) = 15.56, p<.001, but not for those who did dual first.

1) Control processes are normally involved by suppressing (to some extent) the inherently distracting influence of faces. Therefore, dual-task load will lead to an increase in the face bias.

2) Control processes are normally involved in creating an attentional set that results in a face bias during the task. Therefore, dual-task load will lead to a decrease in the face bias.

Faster response if target then

appears here

Slower response if target then

appears here

Faces attract visual attention (e.g., Langton et al., 2008).

Bindemann et al. (2007) demonstrated this with a dot probe task.

They also demonstrated that the face bias was largely under voluntary or endogenous control.

The present study explored the nature of this endogenous control, using a dual task procedure.

There are 2 possible predictions:

Contact: [email protected]