Lecture 4: The World in your Head

Memory, STM, LTM, sensation, perception

Perception

experience of the world with interpretation and processing

combining sensation and expectations of how the world works

successful perception involves educated, unconscious guesses about the world

Sensation

See world as it really is free of bias

true experience (light, sound, touch)

not interpreting

Never really experience it because we are always perceiving (see faces of your family and they are __your__ family in your head)

Behaviourist views on perception/ attention/ memory/ sensation

think they are made-up constructs

but, by agreeing these exist, then behaviouralism is refuted

Dualist views on perception/ attention/ memory/ sensation

turned mental life into physical process, so dualism is refuted

also, shows that there is a physical world out there (“I think therefore I am”)

ie light reflects off retinas, neurons fire, brain processes info = not consistent with dualism

Piaget claims babies dont perceive. Is this true or false?

False. babies perceive because otherwise would be blur of colours. they see objects as objects

Naive realism

our senses capture the world as it really is

not exactly true because we can be wrong (ie illusions)

Psychophysics

science of how sensations relate to the stimuli that produce them

Weber’s Law

humans think/distinguish sensations in proportions, not absolute

ie if buying a house for $1mil then check later to see its $1mil and $100, doesnt matter much, BUT if buy coffee usually for $5, then they change price to $105, it is ridiculous

Sensation puzzle

get different sensations from seeing a rainbow or hearing a baby giggle

Occipital lobe takes the input and turns into colour & light

parietal lobe takes input and turns into taste and smell

but what if reversed?

taste a rainbow?

= synesthesia

Sensation is __(easy/hard)__ to access.

Hard

ie dont really know what languages you know sound like (to others who dont speak it)

Wrong view of perception

looking at the world through camera in eyes

this doesnt solve the problem, just pushes it back

Right way to view perception

have the eye and out from it comes neural firings

from the pattern of neural firings = recognize the world

* eye gives number outputs, but brain has to recognize objects & people

Why perception is hard

Need to infer a 3D world from 2D (eye) inputs

if lighting is right, cant tell the difference

How do we solve the problem of perception?

unconscious assumptions about how the world works

educated guesses

* this is how optical illusions arise

MIT robot story

wanted machine to interface with/perceive the world

computer looks through camera so it could describe what it saw

Demonstrates how simple the problem of perception seems and how hard it actually is

Instinct blindness (Tooby, Cosmides)

processes that give rise to perception are automatic and effortless, so cant appreciate their complexity

how eyes work

light hits off eyes, light passes through corneas, goes through lenses (which use muscles to see far and close), hits off back of eyes, which are filled with photoreceptors (120 mil rods (highly sensitive, good for low light), 6 mil cones (mostly in center & do colour vision)

from rods & cones, get neural firing that goes to back of head

infers 3D from 2D

successful percetion involves coordination of 2 types of info

1. input to the visual system: neuronal firing from light hitting retina (bottom-up info)
2. assumption about the world: some might be wired in, part of visual system itself, from memories, etc. (top-down info)

ex. you are in a dark room and a circle of white light is expanding, see it as approaching, unless we are walking towards it, then it is seen as stable

3 case studies of perception

1. Colour
2. Objects
3. Depth

Colour: how to get a spectrum of it

neurons fire more when brighter

neurons fire less when darker

But, this system is confounded by objects around it (snowball in darkness may be same colour as something else), SO

simple assumption: shadows make surfaces darker (the object doesnt change colour, the shadows made it)

Objects: getting 3D from 2D

We have to interpret 2D info into objects (know that the podium is separate from the professor)

Simplifying principles:

* Proximity: closer together = likely to be same thing
* Similarity: separate patterns usually = separate things
* Closure: square is a closed shape so infer it is on top of circle (picture)
* Good continuation: follow routes that cross (assume simplicity)
* Common movement: move together = likely same object
* Good form: natural forms (+ vs 2 rectangles at 90\* angles)

binocular disparity

use differences between close-up and far-away perception to calculate a rough sense of how distant objects are in space

Depth: object differentiation

know if objects are in front or behind others

characteristics of the stimulus influence our perception

* its more reasonable that a person who seems tiny is just far away

but, can be confounded (mueller-lyer line illusions)

3 general assumption of depth

1. typical size: if a woman approaches you and looks bigger than a house, then likely she is just close (based on typical sizes of women and houses)
2. interposition: if outline of the woman is complete and blocks the house, then likely she is in front
3. motion parallax: objects farther away seem like moving slower then objects closer (use this to infer distance)

Phonemic restoration effect

mental filling in of gaps

if hear a sentence (“it was found that the wheel was on the axle”) and replace a phoneme with a cough and hear it as “it was found that the \[cough\]eel was on the axle”. Your brain fills in the missing info with what is most plausible

this is why proofreading is difficult

efficiency gains outweigh cost of occasionally getting it wrong

Attention

world input that gets focused on

spotlight

moves info from sensory memory to working memory

Sensory memory

unattended information that is quickly lost

Working memory

aka short-term memory

unrehearsed information is quickly lost

Flow of memory

world/sensory input -~~attention~~>working memory -~~encode it~~> long-term memory

(called retrieval if moving from LTM to STM/working memory)

Limitations of attention (3)

* Stroop effect: hard to read colours of letters of words if the words are different colours
* attention is limited (flip back and forth between images)
* Change blindness: don’t attend to most of the world (basketball with gorilla)

Autobiographical memory

memory of personal experiences

Procedural memory

how to do things (read a book, etc.)

5 memory distinctions

1. Sensory vs short-term/working vs long-term memory
2. Implicit vs explicit
3. Semantic vs episodic
4. Encoding vs storage vs retrieval
5. Recall vs recognition

Sensory vs ST/WM vs LTM

Sensory memory: resides in senses for fraction of a moment (like writing with the sparklers)

ST/WM: conscious memory

LTM: library, every word, whatever you remember

Implicit vs explicit

Implicit: dont know you know (distinguish faces, riding a bike, etc.)

Explicit: know you have it

Semantic vs episodic

Semantic: library (facts)

Episodic: episodes, memories (ie first kiss)

Encoding vs storage vs retrieval

Encoding: getting information

storage: keeping

retrieval: getting it out

Recall vs recognition

Recall: ask you a question (short answer)

recognition: is it x or y or z (multiple choice, fill in blank)

Storage difference (LTM and STM)

LTM: virtually unlimited, in theory might fill up, just havent

STM: smaller storage, Miller estimates 5-9 chunks

Maintenance rehearsal

repeating something over and over in head to try to remember

Primacy/recency effect

Primacy: remember better stuff at beginning (ie list, movie, etc)

Recency: remember better stuff at end/most recent

Expertise effects

remember more about what you know more about

How to get STM to LTM (5 ways)

* rehearsall is not enough


1. depth-of-processing: more you think about it, more likely to remember (think of meaning = remember more)
2. mnemonics (foer): memory tricks
3. understanding: knowing and understanding things
4. make it interesting/vivid
5. consolidation: sleep helps to keep memory into the brain

How to get info out of LTM (3)

1. Retrieval cues (ie need to go to dentist and remember when you brush your teeth)
2. Compatibility principle: Relationship between encoding and retrieval: tend to remember stuff better if you are tested (retrieval) in the place you learned (encoded) \[ie scuba experiment\]
3. Searching strategies: in head but cant get it out (yearbook: go through brain, searching for stuff)

Failure of memory/normal forgetting (3 reasons)

1. Decay: atrophy bc fleshy meat
2. Interference: forget stream of numbers quicker if given another stream of numbers
3. Change of retrieval cues: learning things in this context, the feeling in your seat makes it easier, but taking an exam in another hall = harder to remember

Puzzling case of childhood amnesia (Freud)

* forget first few years of life, not decay bc phenomenon happens with 10yr olds
* could be bc the brain rewires

Memory is sensitive to plausibility. T or F?

True.

able to create false memories.

ex. given a string of words that are sleep-related, most people will say they remember ‘sleep’ even though it was never said

Failure of memory/ amnesia

retrograde amnesia: loss of memory for some period period

anterograde amnesia: (korsokoff’s syndrome) lose ability to form new memories, perpetual present

* new explicit cant be formed, but implicit can be
* ie tracing shapes, get better and better although forgets

Clive Wearing (implicit damage)

* lost temporal lobes and hippocampus
* perception of condition is like waking up for the first time
* habituated to condition, writes down everything

Failure of memory/ false memories

* Sam Stone: came in, said hi and left
* asked same questions over and over
* implanted memories in kids: not only did they change answer to please experimenter, but believed changed memory

_________________ questions are more likely to form false memories.

Leading

ex. “did you see __the__ yield sign?” instead of “did you see __a__ yield sign?”

Wrong way to think of memory

memory is recorded like a phone, get at it through hypnosis or ask the right questions

Right way to think of memory

a lot gets lost from the world, retrieval is not the problem, but reconstruction

Why cant trust memory

* filling in the blanks: sometimes fill in missing information with plausible info
* witness testimony: if asked specific question (did you see __the__ school bus? vs __a__ school bus?)
* Implant memories into kids
* Hypnosis: just makes super cooperative (if you know it, youll say it, but might not know)
* Repressed memories: traumatic memory is purposefully forgotten and can be retrieved

TedX

* train self to remember
* mental castles
* more bizarre = more likely to remember

Baker/baker paradox (elaborative encoding)

when told to remember a guy who is a baker and a guy who’s name is Baker. More likely to remember job bc its more relatable/emotional hooks/relations.