4 - Vision & Attention
• Looking but not seeing o__ O
• Lots of drivingexamples→ e.g.,less motorcycle-car accident in the city eventhough there's more
motorcycle in the city; perhaps they're more used to it.
From last class
• Visual selection :
○ Multiplevisual input → sensory buffer →attention selection & visual meaning analysis
→long term memory → attended object
○ A stage by stage copy of the auditory selection
• For reading,you can present words in 2 colours & have P read one colour, with the different
colours in (1) alternating words, or (2) alternating lines
○ Similarto dichotic listening task (different messagein each ear → P tend to attend to one &
• What did subjects rememberabout the unattended message?
○ Physical properties: colour, lowercase font
○ Salient words: (low threshold stimuli)own name, words of interest
• What mighteye movementtracking tell us about selectivereading?
○ (Tracking where eyesare looking towards)
○ P are reading lines they're asked to attend to, but occasionally flickeron other lines (esp.
○ More precise way of seeing what people are paying attention to
○ (more on eye movements later inthe course)
• What ifwe reduced the need for eye movement?
○ From motivation to control for eye movements → can put the stimuli inthe same location
II. Looking but not seeing-- overlapping stimuli
• Language modulein audio (what if you don't understand the language?)complication! Get rid of
it! → geometric shapes (something simple)
• If you could see in front and behind you at the sametime, how clear would it be?
○ Mirror apparatus to test this o__O (it looks like a horrifying helmet)with half-silvermirror
(mirrorlike but still transparent). To be fair, this wasn't an actual experiment, he just built it.
○ Chameleonscan do it, but they also walksuper slow(trade off?) → perhaps ifP had it from
○ e.g.,window at night, you can see your own reflection as well as outside → watch your
neighbour's car get trashed vs. focus on the chainsaw murdererbehind you
• Overlapping words of differentcolours (Filth vs. Thief)
• Neisser-- developed overlapping videomethods
○ Btw: he wrote "CognitivePsychology", which gavethe nameof this field
○ Like old school TV when the signalsfor the channels overlap
○ (1) hand slapping gamevideo overlapping with (2) ball passing gamevideo
When asked to click the button everytime there's a successful (1) hand slap or (2) ball When asked to click the button everytime there's a successful (1) hand slap or (2) ball
pass, P does very well in ignoringother scene and accurately count
When asked to count both, P missdetails
○ When asked to attend (1) with unexpected event in (2) (they put ball down and "air pass"),
most P don't notice
Or attend to (2) whilepeople shook hands in (1)
○ Intentional limitation(likeevolvedto focus on what's important)
○ Watch overlapping ball video and have them focus on (1) how manytimes white team pass
When a lady with an umbrellawalks across the screen, P don't notice
○ Watch the people play ball passing game(no overlap)and a gorillaor person with umbrella
walksacross the screen → similarfindings(P don't notice gorillaor person with umbrella)
• Overlapping nonsense figures
○ Taking meaning analysisout of the equation
○ "Rate how pleasing the red figureis." → "recognize which shapes you saw" (force choice
recognition: "of A or B, which one did you see")
Rememberedthe one them attended, but no the unattended one
○ Critic: [limitedattention or limitedmemory]perhaps P did process the unattended, but over
time of study (~1 min)P forgot it → then it's a memory problem
However, this wasn't the case for shapes that were familiar(e.g.,house, tree)
○ Do we notice the gaps in unattended figure?(a linearshape with holes)
Later, ask what they rememberabout the unattended one, the shape with no holes
was preferred → they didn't actually process the unattended figure(details)
• What isa head-updisplay? → information isdisplayed on a surface (e.g.,windshield)(e.g.,on a
plane, in high end Mercedes)
○ Supposed to makevisual processes more effective b/c pilot won't need to look down on
○ Not so bad when you're flying the plane (not likethere's other planes turning left that you
have to look out for)
○ But when landing?Tested on 747 (militarytraining/ experience) → many P crashed into
○ → they're not as safe as they were supposed to be! What about all the littlegadgets in cars?
○ Overconfidence at how much information we can take in
III. Negativeeffect of priming
• [Positive]Primingwith words: response to the first givenword will affect response to the second
○ Lexical decisiontask: "is this a word or not a word" task
○ IS CAT A WORD? WHAT ABOUT FNE?
○ Faster if the current word is semanticallyrelated to a word on a previous trial
e.g.,hot dog in (1), hamburger in (2)
• No memoryof unattended shapes = early selection?
○ P don't rememberunattended audio message'smeaning, sort of replicated in visual tests
○ This would be consistent with early selection
Sensory buffer → attentional selection → object recognition → long term memory
Where attentional selection comes before recognition • Tipper -- negativepriming