Saliency from the decision perspective [Elektronische Ressource] : inferring the processing architecture of pre-attentive vision with mental chronometry / Michael Zehetleitner

Michael Zehetleitner

Sujets

• Psychologie

Informations

• Publié par: ludwig-maximilians-universitat_munchen
• Publié le: 01 janvier 2007
• Nombre de lectures: 11
• Langue: English

Extrait

Saliency from the decision
perspective:
Inferring the processing architecture
of pre-attentive vision with mental
chronometry
Michael Zehetleitner
Munc¨ hen 2007Saliency from the decision
perspective:
Inferring the processing architecture
of pre-attentive vision with mental
chronometry
Michael Zehetleitner
Dissertation
an der Fakult¨at fur¨ Psychologie und P¨adagogik
der Ludwig–Maximilians–Universit¨at
Munc¨ hen
vorgelegt von
Michael Zehetleitner
aus Kempten im Allg¨au
Munc¨ hen, den 2. November 2007Erstgutachter: Prof. Dr. Hermann J. Muller¨
Zweitgutachter: Prof. Dr. Joseph Krummenacher
Tag der mundlic¨ hen Prufung¨ : 20. Dezember 2007To Norbert Bischof, without whom not.vi
ACKNOWLEDGMENTS
Norbert Bischof, in his lectures and in personal communication, ignited my cu-
riosity for animal and human psychology and provided me with a psychological and
philosophical framework as footholds for investigating such issues. Additionally his
cybernetic, supported by Felix Tretter’s system theoretical approach influenced my
thinking intensly.
In particular, I would like to thank my mentor and supervisor of this thesis, Her-
mann Muller,¨ for focussing my mind to the art of psychological experimentation and
for providing me with the freedom to pursue very interesting questions answerable
with mental chronometry. I thank Joseph Krummenacher for introducing me to the
redundant-signals paradigm and being my contact person for experimental, writing,
or statistical questions from the beginning of my studies of neuro-cognitive psychol-
ogy. Being part of a large unit of colleagues also was of great help for developing and
discussing experiments, as well as for gaining new insights. Specifically, I would like
to mention Thomas Geyer, Thomas T¨ollner, Zhuanghua Shi, and Dragan Rangelov
for valuable discussions and input for this thesis. I had the chance to be supported
by Henning Bumann, Yvonne Schiller, Michael Hegenloh, and Frieder Wormser, who
carried out the experiments and collected the data.
Allworkrelatedsupportwouldbenothingwithoutmyfamily: mywifeIlona,oursons
Franz and Peter, as well as our parents. Finally, I thank the Deutsche Forschungsge-
meinschaft for funding the CoTeSys Excellence Cluster, which financed my research.vii
TABLE OF CONTENTS
Page
ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Salience map models . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Modulation of salience . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 The redundant-signals effect . . . . . . . . . . . . . . . . . . 3
1.2.2 Dimension switch costs and cueing benefits . . . . . . . . . . 4
1.3 Alternative processing architectures . . . . . . . . . . . . . . . . . . 6
1.4 Scope of the present study . . . . . . . . . . . . . . . . . . . . . . . 8
1.5 Summary of findings . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.5.1 Chapter 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.5.2 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.5.3 Chapter 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5.4 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.6 Conclusion and outlook . . . . . . . . . . . . . . . . . . . . . . . . . 14
2 WHAT THE REDUNDANT-SIGNALS PARADIGM CAN REVEAL . . 17
2.1 The redundant-signals paradigm . . . . . . . . . . . . . . . . . . . . 18
2.2 The question of architecture: parallel, co−active, or serial? . . . . . 22
2.3 Is integration spatially selective? . . . . . . . . . . . . . . . . . . . . 27
2.4 Is integration feature−based or dimension−based? . . . . . . . . . . 32
2.5 Weighting or priming? . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.6 Implementationofsaliencymapsanddimensionalweightinginthebrain 38
2.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3 CO-ACTIVATION VS. SERIAL AND PARALLEL MODELS . . . . . . 45
3.1 Experiment 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.1.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.1.2 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.1.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.1.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.2 Experiment 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.2.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.2.2 Results and Discussion . . . . . . . . . . . . . . . . . . . . . 65
3.3 Experiment 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.3.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.3.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74viii
Page
3.3.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.4 General Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4 INTENTION AND TRIAL HISTORY IN LOCALIZATION . . . . . . . 89
4.1 Experiment 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
4.1.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.1.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
4.1.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.2 Experiment 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
4.2.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.2.2 Design and Procedure . . . . . . . . . . . . . . . . . . . . . 106
4.2.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
4.2.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
4.3 General Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
4.3.1 Summary of findings . . . . . . . . . . . . . . . . . . . . . . 110
4.3.2 Relations to post-selective accounts . . . . . . . . . . . . . . 112
4.3.3 Relation to further studies . . . . . . . . . . . . . . . . . . . 114
4.3.4 Summary and conclusion . . . . . . . . . . . . . . . . . . . . 118
5 DECISION PERSPECTIVE ON SALIENCY . . . . . . . . . . . . . . . 121
5.1 The RSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.1.1 Manipulation of feature contrast . . . . . . . . . . . . . . . . 133
5.1.2 Maion of response bias . . . . . . . . . . . . . . . . . 134
5.1.3 The effect of spatial attention . . . . . . . . . . . . . . . . . 136
5.2 Benefits from Dimensional Cues . . . . . . . . . . . . . . . . . . . . 136
5.3 Dimension Switch Costs . . . . . . . . . . . . . . . . . . . . . . . . 137
5.4 Experiment 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
5.4.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
5.4.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
5.4.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
5.5 Experiment 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
5.5.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
5.5.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
5.5.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
5.6 General Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
5.6.1 Relation to previous studies . . . . . . . . . . . . . . . . . . 156
5.6.2 Theoretical Implications . . . . . . . . . . . . . . . . . . . . 164
5.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
VITA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178ix
ABSTRACT
There are several cognitive and neuro-scientific models of early, pre-attentive vi-
sual processing, with saliency map models being the particular dominant ones. Al-
though they are very specific about how feature contrast and salience are being com-
puted(includingstimulus-andobserver-driveninfluences),thereusuallyisatheoreti-
cal gap between models dealing with visual analysis (such as the dimension weighting
account, Muller,¨ Heller, & Ziegler, 1995; Found & Muller,¨ 1996) and models de-
scribing decision and response selection processes (e.g. the Ratcliff Diffusion Model,
Ratcliff, 1978). Consequently, I propose that investigating saliency from a decision
perspective that is by applying mathematical theories of decisions to several tasks
that can be performed supposedly based on a salience map (e.g. detection, localiza-
tion, attentional selection), explanatory power is increased and new hypothesis can
be generated.
Further, several issues of visual pre-attentive processing are currently still under de-
bate. Specifically, the exact nature of the pre-attentive architecture is disputed with
regardtotrialsequence, intention, andredundancyeffects. Thepresentstudytargets
at shedding new light on the question of pre-attentive processing architecture (serial,
parallelindependentorinteractive,co-active),top-downpenetrabilityofpre-attentive
vision, and pre-attentive vs. post-selective locus of dimensional inter-trial effects.
In summary, the fi