lecture 3- visual search Flashcards
describe Treisman & Gelade’s (1980) classic search experiment
- participants have to find target among distractors
- factors varied- number of distractors (size of array) and presence/absence of targets (positive/negative trials)
- feature (target distinct by colour) or conjoined search (colour and orientation are two conjoined features- combined)
what did Treisman & Gelade (1980) find?
- single feature: immediate, size of array doesn’t have impact when target is positive, but RT does increase with array when target negative
- conjoined features: larger array increased RT for both positive and negative
- both positive and negative = conjoined takes longer
how are Treisman & Gelade’s (1980) findings represented using an equation?
y = mx + c
y = reaction time
m = no. milliseconds that each item adds to search (different for different types of stimuli)
x = no. of items
c = time it takes to do initial processing and then to produce a response (same across all)
steeper slope = longer the search by item
(adds most to RT) conjuctive -ve > conjunctive +ve > single -ve > single +ve (adds least)
explain feature integration theory
two processing stages-
1. initial stage feature detection- parallel, fast, efficient, automatic, pre-attentive (single feature ‘pop out’
2. feature integration- happens with conjoined, gluing features together (takes attentional focus), serial processing needed, slow and inefficient (each element one by one- speed determined by how many elements in array)
what should all search data look like and what did Wolfe (1998) find?
- should see pattern of slopes between present/absent/conjunction/feature
- they did not find expected pattern- target absent slopes higher overall
what is Wolfe’s model of guided search?
- said Treisman’s model was too simple- only based one side of attention (bottom up model)
- brain also tells us where to look
- bottom-up info is influenced by top-down commands
- stimulus is filtered through broadly tunes ‘categorical’ channels (black/vertical)
- top-down commands to feature maps activate locations possessing specific contegorical attributes (activate ‘black’ lines)
- the output produces feature maps with activation based on local differences (bottom-up) and task demands (top-down)
- a weighted sum of these activations forms the activation map- attention deploys limited capacity in order of decreasing activation
what factors guide search?
bottom-up: salience (between targets and distractors), attributes (elements that capture deployment of attention)
top-down: scene properties, values (what’s important in scene to you)
what is the the attentional engagement theory and who proposed it?
duncan & humphreys 1989
- efficiency of search (slope on graph) is based on aspects of task
- target/non-target similarity and non-taget/non-target similarity
- more distinct = shorter (unlike treisman- says all should be efficient)
- heterogeneous non-targets (all different) = more inefficient
give some examples of guiding attributes and their likeliness to cause fast automatic searches
definitely: colour, motion, orientation, size
probably: depth, luminance, closure, shape
maybe: lighting (shading), luster, number, apect ratio
doubtful: novelty, letter identity
non-feature: conjunctions, categroy, identity, 3D shape, threat
how does shading and luster affect visual search?
shinier (difference between target and non-target) = better search
how does a scene (top-down information) affect search?
- bread on plain background -> bread doesn’t pop out -> serial search
- kitchen background makes search faster as brain recognises scene and directs attention to counter to look for bread
- brain takes gist of scene
what are the two pathways in wolfe’s model?
selective- info comes up through cognitive system (slow and effortful, how info from environment is processed- leads to recognition/identification of stimuli elements)
non-selective- extracts basic semantic information from scene quickly, provides guidance (directs attentional process) to selective pathway (constrains)
e.g non-selective directs to look at bottom half of image of forest, to then count chickens which is then selective
what study was done on the effect of value on visual search?
anderson 2011- used training
- p’s had to say what the orientation of a line was inside a red or green circle
- were rewarded more highly for red circles than for green (or vice versa)
- completed 1008 trials to build up association between colours and chance of reward
- test phase: p’s look for unique shape among 6 objects in coloured circles, half the time there was high or low value distractor
- high value distractor increased RT
- individual differences mattered: lower WM and higher impulsivity both = higher distraction
