unit 1: intro, history/methods/philosophy, perception/attention

long-term learning vs short-term performance

performance: immediate, short-term ability; observable during practice

learning: long-term ability that endures; not observable during practice

Simon & Bjork key press experiment

two trials

1. blocked practice

2. interleaved practice

overall findings

• performance during practice was better for blocked trials

• long-term retention was better for interleaved trials

desirable difficulties

• methods that are most effective for long-term learning tend to introduce difficulties that make short term performance worse

intuitions about what study strategies are most effective are often wrong - evidence

short term performance is worse during study strategies that cause better long-term learning, giving the illusion that they are not as effective

highlighting/underlining strategy (2 reasons highlighting is not effective)

1. highlighting causes focus on isolated facts, ignoring connections

2. difficulty distinguishing between central and peripheral ideas

Fowler and Baker study and results

• highlighting did not help during trials that tested learning for groups that read an article, highlighted an article, and read a highlighted article

rereading (3 problems)

1. diminishing returns compared to other methods

2. does not improve performance on inference-based questions

3. gives a mistaken illusion of mastery (recall vs recognition!)

Rothkopf study and results

students were given 2 passages with missing key words, and were asked to fill in the missing words

• trials: 1) never seen the passages, 2) read both passages once, 3) read both passages twice, 4) read both passages four times

findings

• reading once leads to big improvement compared to no reading, but after the 2nd time, there is no significant difference onwards

generating explanations - Pressley study and results

idea; make connections between well-understood ideas and develop an understanding of why an observation is true

trials:

1. no explanation why

2. given an explanation why

3. generated an explanation why

results

• same performance for first two trials, with a significant improvement for the 3rd generated explanation group

interleaving (blocked vs interleaved practice)

interleaved: switch between topics while studying, building confidence across multiple areas

VS
blocked: study on topic until confident and studying the next…etc

Rohrer & Taylor study and results

studying interleaved vs blocked study methods, using tactics to memorize the volume of geometric shapes

results

• better practice performance for the blocked trial

• BUT improved exam performance for the interleaved group

distributed practice and spacing

spacing out studying enhances learning!

final retention tests see that more spacing = better memory

Bahrick study and results

group 1: all 6 sessions back to back on one day (cramming)

group 2: all 6 sessions one day apart

group 3: all 6 sessions one full month apart

Results

• more proportion correct for crammed group in the beginning, but on the final test group 3 say the most proportion correct

testing & utility

testing on concepts improves memory because it requires recall and generating an explanation

Butler study and results

trials:

• all groups read text passages but one group tested on half of the passages while the other group re-read the passages

results

• testing on studied passage saw improved accuracy compared to solely re-reading the passages

learning vs performance - desirable difficulties (Bjock/Bjock reading)

assessing easy forms of learning that improve recognition in the beginning ≠ long-term learning

• many difficulties are undesirable in the beginning → desirable because they trigger encoding and retrieval processes that support learning, comprehension, and remembering

varying the conditions of practice (Bjock/Bjock reading)

-spaced vs massed studying: spacing is more effective (aka blocking)

• massing supports short-term, but does not improve overall long-term learning

• massing does not center on learning more difficult concepts

• little will be recalled over the long run, despite cramming

-interleaving vs blocking

• blocked improved more rapidly, but saw a decrease in long-term learning

• blocking does not create a ‘desirable’ difficulty

• blocked misinterpreted successful short term learning as ‘successful’

generation effects (Bjock/Bjock reading)

repeated testing vs repeated study

• looking up answers and getting answer directly prevents learning and troubleshooting; power of tests in learning

• how familiarity can be misleading

• metacognitive effects of tests

  • limited judgment in deciding whether content from a textbook will be on an exam

  • more time should be spent testing/practicing!

strategies to optimize learning (Putnam reading)

1. starting the semester

2. preparing for class

3. during class

4. after class

5. preparing for tests

6. final exam

introspectionism (Wundt) problems

observing one’s thoughts, experiences, and feelings

• developed theories on how the mind works based on just observations, impacted by expectations

• private mental lives cannot be verified by others - problem of the minds (indirect observation)

problems

1. difficult to verify

2. private events, not public

3. end product, not the product itself

behaviorism (Thorndike and Skinner) + experiments

emphasis on what can be directly observed such as:

• stimulus

• responses

• reinforcements/rewards

• IGNORING THE MIND!

cat in a box experiments, observing ‘how long to escape’ to press a lever

law of effect

if a behavior is followed by something positive, it will be strengthened

if followed by something negative, it will be weakened

methodological vs radical behaviorism

Skinner

• animal’s behavior operates on the environment to make something happen

• all animal and human behavior can be explained by simple rules of learning

• core beliefs of radical behaviorism

  • mental activity does not exist

  • mental activity is epiphenomenal (no causal impact)

instrumental conditioning

• animal’s actions are instrumental in producing some kind of outcome

methodological behaviorism

• NO introspection

downfall of behaviorism

• inability to explain human language

  • people can create new sentences and understand them even without prior experience

  • limiting science to strictly the observable

epiphenomenal mind

mental activity has no causal impact, as proposed by radical behaviorists

cognitivism and computational view of the mind

mainstream underlying assumption: the mind is a computer program

• signal detection theory

• information theory

• symbolic logic (true, false operations)

• Turig’s thesis (AI): anything can be computed in a mechanical manner

interpreting graphs (main effects and interactions)

main effect: two variables are parallel → there is a main effect

• presence of an effect of one IV across levels on the DV

interaction: if two variables intersect → there is an interaction

• when the effect of an IV on the DV depends on the level of another IV

mental chronometry

Donders (study of the time course of mental processes)

• mental processes take time and we can measure that time

information processing stages

stimulus → processing → response

choice reaction time

stimulus → detection → decision → response

Donders’ subtraction method and problems

subtracting the time between a choice reaction tsimple reaction time (stimulus → detection → response) to calculate the duration of the decision stage

problems

1. assumption of pure insertion

   1. all stages remain the same when the new one is added; adding the decision stage may influence another stage

2. assumption of additivity

   1. duration of all stages add together to yield the reaction time

3. assumption of omniscience

   1. we ‘know’ the information processing stages (dont really)

sternberg reading: structuralism (wundt, titchener), functionalism, pragmatism (james, dewey), behaviorism (watson, skinner)

structuralism:

• goal: understand the structures of the mind by studying the mind in its constituent parts

• introspection (delving into the private mind)

functionalism:

• what do people do, and why do they do it?

• how mental processes allow one to adapt, survive, and flourish

pragmatism:

• knowledge is validated by usefulness

• consciousness enables people to adapt to the environment

behaviorism:

• focuses on associations between the environment and behavior

• little impact of the mind

• emphasis on animal research

perception

• properties of the world and receptors

process of taking bits of energy that come into your brain and actively constructing a mental representation (percept)

transduce physical energy (light, sound) into electrochemical energy (AP, neurotransmitters)

signal detection theory

addresses how we make detection decisions and how expectations/biases impact perception

signal detect matrix

hits: correctly identify a signal

misses: miss a signal

correct rejections: correctly reject that there is no signal

false alarms: detect a false signal that does not exist

sensitivity biases

sensitivity: how easy or difficult it is to discriminate signal from noise

bias: your tendency to say yes or no, determined by expectations

signal vs noise

signal: something in the environment you are trying to detect

noise: things in the environment other than the signal

stimuli types

distal: external image

proximal: light and image that hits the retina

percept/representation: representation inside our mind that develops

perception as construction

illusions: convert energy out there to a model in there inside our head (constructive process where internal model does not match external model)

ambiguous figures:

lack of correspondence

when the percept does NOT correspond to the distal stimulus

example: perceptual illusions

paradoxical correspondence

when the proximal stimulus does not correspond to the distal stimulus

• the percept does!

example: moving objects, moving eyes, constancies

perceptual constancies

our perception of an object’s features remains constant even when viewpoint and proximal stimulus changes

• size doesn’t change with distance

• perception of color/lightness doesn’t change with light

• perception of shape doesn’t change with angle

direct perception (stimulus theory)

environment provides all necessary cues

stimulus info is always unambiguous

constructivist theory

perception uses data from the world and our prior knowledge/expectations

• sensory info is often ambiguous

stimulus-driven

processing that is driven by the external stimulus NOT internal knowledge

• direct perception

goal driven processing

processing that is driven by knowledge and expectation

• constructivism: stimulus driven AND goal driven

Sacks reading

Dr. P’s patient deficits and preserved capabilities:

• issue with ventral stream in the brain (temporal)

• remembered key events of people, but not their faces

• not impaired visual imagination

agnosia: unable to identify objects with visual pathways in the eye left intact

relationship to dorsal/ventral distinction: independent of each other, where impairment in one area does not cause impairment in others

receptor properties: retina, rods/cones, bipolar cells, ganglion cells

cornea: transparent front layer

iris: colored, light-controlling diaphragm

lens: focusing structure

bipolar cells: signal integrators and relations

• lateral inhibition

ganglion cell layer: output

photoreceptor layer: light detectors

rods vs cones

• rods detect brightness

  • many

  • located in periphery

  • low acuity

• cones are concentrated in the fovea, detecting colors light blue, green, and red

  • low light sensitivity

  • few

  • high acuity

neuron parts

dendrites → soma → axon (myelin sheath)

types of potentials

resting: at -70mV

action: spike in potential, causing an increase positive concentration to 30 mV

threshold: potential must get above a level for a neuron to fire

all-or-none principle: action potentials always have same strength (either they fire or stay at resting potential)

propagation: once past threshold, active process of ion pumping propagates action potentials down axon

refractory period: short period after firing before neuron can fire again

neurotransmitters, synapses, presynaptic/postsynaptic cells, electrochemical transmission, summation

chemicals that are send between small gaps of cells (synapses) = neurotransmitters

electrochemical transmission

• electrical AP within cells

• chemical neurotransmitter between cells

summation: if combined effects at all synapses take AP across axon above threshold, then neuron will fire an action potential

receptive field of a neuron

• area of external world in which stimulation causes a neuron to respond

• Ganglion cell: points of lights of about 0.1-1+ degree of visual angle

magnocellular vs parvocellular layers, lateral geniculate nucleus

magnocellular: movement/location; transient respond; large receptive field

parvocellular: patterns/color/form; small receptive field; sustained response

spots of light → edges

• RGCs care about spots of light

• connected to a single V1 cell through LGN cell connection

• detection of spots of light in a specific arrangement (edge) by V1 cells

cerebral cortex and lobes, primary visual cortex, simple/complex cells

simple cells: bar of light; specific orientation; specific retinal position

complex cells: edges, movement

PET scanning

mental and neural activity → blood flood → more radioactive tracer → more positrons emitted

PET Study:

• spatial where task should activate occipital and parietal regions

• object what tasks should activate occipital and temporal regions

Kohler experiment

Kohler experiment:

• spatial task: locations same or different? (different)

• objects: same or different (same)

• RESULTS: areas of activation

  • spatial: parietal/dorsal (area 39)

  • object: temporal/ventral (area 37)

patient DF

condition: visual object agnosia (deficit in shape perception) after damage to the temporal lobe

• cannot identify objects

• cannot copy objects

• cannot recognize own drawing

performs poorly on perceptual matching task, but performs well on a posting taks

• intact visually guided action (dorsal pathway)!

Mishkin experiment

what task

• lesions in inferior temporal cortex in the ventral stream product deficits in shape discrimination

where task

• lesions in the parietal cortex along the dorsal stream produce deficits in landmark where tasks

population coding

• no grandmother cell that is specific to one object/orientation

• needs a pattern of activation across a population of cells to recognize objects

  • preserves object recognition in case of cellular death/damage

McCloskey Reading - Patient AH

modality-specific localization impairment

• impairment selective to vision

• did not have a spatial deficit affecting localization

identification vs localization

• able to correctly identify objects, but could not locate them properly

• location and identity are processed separately

visual experience and indirectness

• could correctly identify objects from vision itself, but could not localize and perceive them correctly

• objects of one’s awareness are NOT things in the world, but rather representations of those constructed by visual pathways

parsing and filling in

parsing: breaking the complex scene up into individual objects (edge detection, filling in and grouping!)

• objects are occluded by other others → need to fill in the details to recognize objects

gestalt principles

main theory: see things as a whole, rather than the sum of their parts

similarity: similar things should be grouped together

good continuation: group things together if they follow a smooth curve

Pragnanz: simplest possible interpretation is preferred

template theories

having a template for which objects we see and identify

problems:

-transformations (angled E is not the same as a perfectly vertical E)

-exemplar variation (different fonts of E are not identical)

-occluded objects(the object is not the same as the template → requires parsing of features that are visible)

feature theory, evidence for features, visual search, problems

• feature decomposition: cells in temporal lobe are not tuned to specific stimuli (no grandmother cell)

problems

• lack of relationship considered

• different arrangements of the same features produce different objects

EVIDENCE
- physiology (on area and off areas)

• stabilized retinal images

• visual search

pandemonium model, caricatures

• feature based theory of pattern recognition

• humans identify objects by breaking down into individual components

• chaotic noisy and parallel nature of process (metaphor of demons)

• brain has specialized neurons that fire in response to specific, simple features

• patterns with similar features (p and r) will excite similar feature demons → recognition errors

structural description theories, RBC (geons, non-accidental properties), evidence + problems

RBC:

three keys

1. objects are composed of a number of distinct geons

2. recognize shapes and objects are composed of because of non-accidental properties

3. any object can be specified by its structural description (components and spatial relations)

matching process

1. detect elementary features

2. find non-accidental properties

3. determine component geons

4. match to memory

Evidence

• partial or degraded objects - can still identify based on non-accidental properties

  • deletion of accidental properties still enables recognition, not with deleting non-accidental

• unusual orientations

problems

• facial recognition — faces are composed of the same geons, yet they aren’t the same face

• lack of brain evidence (geon-selective brain areas)

face perception - holistic

holistic vs feature perception

• holistic: faces are largely perceived and stored holistically

  • minimal decomposition into parts

  • relationships among parts = key

  • perception of inverted faces is different

neural evidence for face-specific processing

face-selective intracranial ERP responses can be recorded from ventral occipito-temporal cortex

prosopagnosia, expertise vs face for FFA

prosopagnosia: damage to the ventral stream can cause selective deficit in face perception

• typically use other cues to identify people

FFA: face-selective region

• found by Kanwisher et al study which showed periods of faces or objects

• FFA is involved in face-specific processing

faces or expertise?

• FFA is strongly activated by greebles for greeble experts but not for non-experts

• prosopagnosics who are experts lose the ability to identify previous expertise features → brain has a general system for the expert metric subordinate-level visual classification, NOT a face-specific system

level of recognition (subordinate, basic, superordinate)

levels:

superordinate: general categories (animal)

basic-level: typical balance of general and specific (dog)

subordinate level: highly specific (golden retriever)

depth: paradox of depth perception (aka paradoxical correspondece)

distal system (the world) = 3D

proximal stimulus on retina = 3D

perceptual experience is 3D (paradoxical correspondence)

monocular vs binocular depth cues

monocular

• linear perspective (parallel line

• shape or texture cues (less distinct as distance increases)

• relative size: when two objects are similar in size, smaller retinal image of further one

• interposition: blocking and occluded object = farther

• shadows

• motion parallax: as we move, nearby objects move quicker

• accommodation: lens changes shape to focus on objects at different distances

binocular

• retinal disparity: eyes are horizontally separated, so they receive subtly different images

• convergence: nearby objects causes the eyes to rotate inward and converge

stroop task, controlled vs automatic processing

stroop task

• identifying processing speed in reaction time when naming the color of a word that is incongruent with semantic meaning

• shows automaticity where human brains read words faster than identifying colors

controlled - conjunctions

automatic - features

auditory attention, dichotic listening, basis for selection, processing of unattended info

dichotic listening: participants hear two different audio streams

• early: only know perceptual features BUT NOT meaning

• issues: hard to pin down exactly when selection happens

early vs late selection/filtering, attenuation, perceptual load theory

early filtering:

• filtering information after detection but before recognition

• issues: GSR showed the effect of conditioned words even when unaware of them

• similar concepts as attenuation (filtering before recognition)

late filtering:

• filtering information after recognition

overt vs covert attention, Posner’s spatial cueing task, spotlight/zoom-lens model

overt: eye movements

covert: shifting attention without using your eyes

Posner’s spatial cueing task

• when attention is shirting using an arrow

  • costs and benefits of either slow or fast reaction times

spotlight metaphor

• specifically focusing on partical aspects in visual field

visual attention - Treisman

• conjunction: shape/color/object

• illusory conjunctions: incorrectly perceiving features of two objects into one

  • occurring only at unattended locations

1. attended locations → detection and recognition is faster

2. visual search is fast IF target is different from nontargets in a simple feature

3. attended locations → features are correctly bound together

4. arrangements of colored shapes into textures is much easier if diff parts of a scene differ in simple features

integrative spotlight

visual scene parsed into simple visual features

master map of locations codes where visual info is, but not what

attention is direction randomly around the master map

when a location is attended, features from diff dimensions become glued

blinding problem of perception

failure to notice unexpected yet fully visible stimuli due to attention being focused elsewhere

• perception is limited by attention, not just vision

feature integration theory: know the model

how perceive visual objects by identifying simple visual features (color, shape, motion) and how they are automatically processed

• require focused attention to be glued into unified objects

feature vs conjuction search; illusory conjuctions, texture segmentation

• illusory conjunctions: incorrectly perceiving features of two objects into one

  • occurring only at unattended locations

feature search

• automatic

• fast, pop-out

• independent of number of distractions

conjunction search

• controlled

• slow, effortful, needs glue

• depends on number of distractions

space vs object vs feature based attention

space

• specific region or location

• operates like a spotlight enhancing processing of all info that falls within the attended spatial area

object

• unified perceptual object or chunk of visual info

• once an object is selected, attention spreads throughout entire boundary, enhancing processing of features and locations

features

• non-spatial visual property such as color, motion, or orientation

• enhances processing of a target feature across the entire visual field

Driver Reading - cocktail party problem, selective shadowing, non-shadowed message

cocktail party problem

• in many situations, many sounds enter ears → how do we pick which sounds are revelant?

selective shadowing:

• most efficient with clear differences between two messages, such as differences pitch

• only notice simple physical properties such as pitch changes in non-shadowed message

nonshadowed message

• limited processing of information due to superficial processing

• perceived as having lower subjective loudness compared to shadowed message

• unattended input

Driver Reading - Broadbent’s filtering model, methodological problem of assessing unattended info, falsifications

Broadbent filtering model:

• fueled less by cocktail parties than by problems of radar operators

• two main empirical questions described selection attention and selective shadowing (efficient with clear physical differences between two concurrent messages)

• two successive stages of perceptual processing: 1. physical properties (pitch of sounds) is extracted for stimuli in a parallel manner; 2. more complex psychological properties that go beyond physical characteristics would be extracted

• limited capacity channel; bottleneck theory (filter prevents overloading the brain, where unattended info is lost)

unattended info

• sometimes unattended information can be processed more deeply

• limited awareness of unattended stimuli → rejection from entry into memory or into control of deliberate messages

late selection (deutsch and deutsch)

deep processing was the rule rather than the exception

• limited awareness of unattended stimuli is due to rejection from entry into memory or into control

• unattended info may undergo full processing, yet without conscious recollection

attenuation (treisman) and evidence (roles of partial info and priming)

unattended info were attenuated rather than completely filtered out

• weaker inputs from unattended such as in abstract properties like word identity

• key roles that partial information (attenuated inputs) and priming have psychological processes

evidence

• priming from an attended message becomes larger as one extends related context → lowers recognition threshold

• partial info: attenuator does not completely block info, allowing recognition of low threshold strongly primed words

parallels in visual attention - Lavie, early/late reconciliation and perceptual load theory

results favoring late selection → obtained in situations of low perceptual load (single target and single distractor)

• distractor interference greater in low-load

• perceptual processing = automatic

• if target is higher in load, may exhaust perceptual capacity so less distractor processing occurs

• low load = late selection (only portion of capacity is used, causing spare to process distractors)

• high load = early selection (difficult = all capacity used to process)

feature integration theory, simple features in parallel, attention to bind features together AND studies

feature integration

• not effective for oriented elements which make a shape

• spotlight model of attention

• special role allocated to location (Tsal & Lavie)

• does not take grouping processes sufficiently into account (Duncan and Humpreys)

neglect: extinction, processing of neglected info

reaction times to stimuli on non neglected side can be influenced by neglected stimuli on the contralesional (left) side

• left extinction: detect single event, regardless of where it appears

• when two events shown at the same time, patient will miss the contralesional event they usually could detect, focusing only on the right event

• attention biased to the right

• losses of awareness for single features; greater in cnojunction tasks than in feature detection tasks

attention modulates (alters) sensory neural responses

sensory responses to external stimuli → modulated by animal’s attentional state

substantial effects of selective attention on sensory neural response could be obtained at many brain sites

main principles:

• selective attention can modulate perceptual coding from early processing

• most cellular evidence shows attenuation

visuospatial neglect/extinction (disengage deficit, unbalanced competition between hemispheres)

spatial vs object centered neglect

Balint’s syndrome

change blindness, inattentional blindness (plus studies), link between attention and awareness