Please enable JavaScript.

Coggle requires JavaScript to display documents.

CogPsych (L1 (Cognition involves: (perception, attention, memory, decision…

- - - - creates(REPRESENTATIONS)
        
        of(world)
        
        purpose(ACT within it)
        
        to(achieve goals)
  - - - categorise entities
      - infer relationships between them
  - - - If AI's linguistic competence satisfies us, this indicates the machine has a mind
  - - - creates(CONCEPTUAL STRUCTURE)
        
        RECOGNISE(objects & events
        
        to(RESPOND) based on(Knowledge)
  - - - Sensorimotor,
      - Mental Images
      - Symbolic representations
    - - sensing
      - acting
      - thinking
  - - - Symbolic code for relationship among concepts
      - Propositions composed of (predicate) and (arguments / semantic elements).
        
        Relationship between elements
  - - - with(Representations directly analogous to sensory experience)
      - time taken to confirm object pairs increased proportional to angular rotational difference between objects
    - - Classical view of semantic knowledge
      - (Collins and Quillian’s model) = hierarchical
  - - - questions a system’s computational goals
      - Behaviour and relationship w/ environment
      - logic of the strategy
    - - Questions representations and algorithms to achieve these goals
      - Representation(for the input and output)
      - Algorithm (for the transformation)
        
        CAREFUL [In notes, this level also mentions COMPONANT operations / COMPONENTS and their operations?? meaning components of the algorithm
    - - questions THE WAY algorithms/representations are physically realised
      - component parts where operations are localised
    - - Error setting the text for the following input: = mechanism’s component operations as well as their functional organization.
      - ratherThan(implementational = component parts and their structural organisation)
  - - - Real-time cognition(continuously changing pattern of neuronal activity) rather than(sequence of logical operations on discrete symbols)
      - Support(eye-tracking and computer-mouse-tracking)
      - Mental activity is being conducted in between those seemingly discrete thoughts
      - Cognition as a (continuously dynamic biological process)
        
        rather than (staccato series of abstract computer-like symbols)
        
        MDthoughts(Think of a if then operation of a swtich (single neuron level), vs a neural net with the overal probability/function producting the congition)
  - - - establishing(relevant goals and affect)
      - producing(supportive bodily and motor states)
      - predicting(likely outcomes)
- - - - Reflexive system is (faster) but more (transient)
      - Voluntary is (slower) but more (sustained)
  - - - (Spatial Cuing Paradigm) {Posner}
        
        Attract attention to A - present stimulus at A or B - compare performance
        
        May take time to move the spotlight (switch attention) from one place to another (echoes filter theory)
        
        (SOA = Stimulus onset time) = Time between the cue and the presentation of stimulus
        
        (RT = reaction time) = Time between the stimulus and response
        
        (Valid cue) - The cue (arrow) predicts of stimulus (80% of the time)
        
        Miscued (invalid) trial: stimulus occurs at other (low probability of occurrence)
        
        Uninformative cue (arrow points both ways) Baseline to compare valid and invalid
        
        Findings
        
        Valid cue = (Faster) RT
        
        Invalid cue = (Slower) RT
        
        Relevant to baseline, there are (symmetrical) benefits and costs with valid vs invalid cuing
        
        Surprising findings because expected no attention engagement on such a simple task
        
        Flexible findings - Replicates with (RT / accuracy / any stimuli)
        
        2 Explanations of Causes of Cuing Effects
        
        1 - Spotlight account = (Switching time) - Time to move the spotlight
        
        Costs of disengaging from wrong location and to move spotlight to other location
        
        Benefit from engaging correct location before stimulus
        
        2 - Capacity theory account (Unequal Capacity Allocation)
        
        RT inversely proportional to the amount of capacity allocated to a location
        
        Allocate capacity in response the the cue
        
        Neutral = capacity spread (across) locations
        
        Focused = capacity (concentrated on one) location
    - - Clinical patients show deficits of both kinds = failure to (focus attention) or failure to (disengage attention)
    - - Endogenous (voluntary) = Cognitive (need to interpret)
      - Exogenous (reflexive) = Direct / Spatial (no need to interpret)
    - - Different time course of central and peripheral cuing
        
        Peripheral effect (Fast & transient - AKA peaks rapidly but then falls off very rapidly))
        
        Central effect (Slow and sustained - AKA peaks slowly but once it gets there, remains constant, as have shifted it there and are waiting for something to happen)
        
        How they found this:
        
        An arrow = semantic cue - so Cognitive and Slow - need to interpret then make decision
        
        Flash is (reflexive)
        
        SOA is the time between (cue) and (stimulus)
        
        SOA also time taken to reorientate attention?
        
        Cuing effect is the difference between RT on valid and invalid trails
        
        The magnitude of the cuing effect as a function of the SOA between the cue and the target
      - Different effects of load (Jonides)
        
        (Memory) task & (Attention) task
        
        Voluntary orienting slowed by (memory allocated)
        
        Reflexive orienting is (not sowed)
        
        Inhibition of Return - Found only with (peripheral) cues, not with (central) cues
        
        Ecological argument: Allows efficient (search) of complex environment.
        
        Maintaining a (mental map of searched locations) would be costly
        
        IOR gives you this for free by preventing repeated search of same location
      - Eye Gaze Directional Cues (fast orienting to central cue) may have it's own (cortical pathway)
- - - - Dear Aunt Jane(w/ Meaningful material - preferred recall follows semantic context - not ear presentation)
      - Own Name(detected on unattended channel)
    - - Definitions
        
        Early selection: explains by attenuating filter
        
        Late selection: all stimuli activate semantic representations in LTM, but need to be selected by pertinence to get into consciousness
      - Consistent with:
        
        Early
        
        Consistent with partial, but reduced processing of semantic targets (“tap'”)
        
        Late
        
        Consistent with indirect measures of semantic processing on the unattended channel
      - Evidence
        
        Early
        
        Split span experiment - preferred recall ear by ear over temporal{time} order)
        
        explanation(more switches needed)
        
        Treisman's Evidence for Early(Tapping, %correct higher on attended channel with non-0 unattended indicating attenuation rather than channel blocking)
        
        Late
        
        Ambiguous stimuli(recognitions biased by previous shadowing task - eg Stones towards the bank)
        
        Semantic activation in absence of attention shown by Classically conditioned GSR for words in same category in absence of attention
      - Filter Disagreement
        
        Location
        
        Early(Sensory analysis - > Filter -> Semantic analysis{in LTM})
        
        Late(Sensory analysis - > Semantic analysis{in LTM} -> Filter)
        
        Properties
        
        Early(Attenuation - Treisman)
        
        Information gets through at different amounts, with weaker signals causing partial semantic activation in LTM
        
        Late Selection
        
        Bottom up + top down selection mechanisms both activated to get through filter
        
        BU(stimulus driven)
        
        TD(Selection by pertinence{relevance to task})
        
        Semantic LTM activation
        
        Early(LTM activation = conscious awareness)
        
        Late (All stimuli access LTM, but not sufficient for awareness {pass filter=awareness})
        
        Criticisms
        
        Early
        
        Complex filter must respond to semantic context
  - - - Disagreement about location and properties of filter
        
        Early Selection theory (Attenuation Model - Treisman)
        
        Sensory analysis - > Filter -> Semantic analysis in LTM
        
        Rather than blocking unattended material, the channel is attenuated (turning down the volume instead of blocking)
        
        Semantic analysis
        
        Late Selection (Deutsch & Deutsch / Norman)
        
        Sensory analysis - > Semantic analysis in LTM -> Filter
- - - - Evidence(reduced detection accuracy on unattended channel {Treisman})
    - - Evidence(Semantic activation on unattended channel {McKay})
    - - ES (Weak) semantic activation VS LS (brief) semantic activation (on unattended channel)
  - - - Sel (Selective - monitor one channel only)
      - XOR (Exclusive OR - A or B but not both)
      - IOR (Monitor both channels, AND simultaneous targets possible)
    - - OR<SEL(Divided attention < selective attention) (ATTENUATION W/ DIVIDED ATTENTION)
      - AND<OR (Simultaneous detection < single channel detection) (AND DEFICIT)
    - - Predicts OR<SEL, (lower signal when 2 channels go attenuated by the filter, faster times when signal stronger from one channel)
        
        AKA(attenuation w/ divided attention)
      - NOT AND<OR (need to pass two channels through the filter in both cases)
    - - Predicts AND<OR (AND = two targets causing semantic activation, both need to fit through filter) (SELECTION COMPETITION)
      - NOT OR<SEL (Same amount of semantic activation that needs to get through the filter) (NO COMPETITION = NO DEFICIT)
  - - - Capacity (the ability to activate neural structures)
        
        Capacity is limited, finite, but (flexibly allocated) to simultaneous tasks
        
        Deficit in divided attention tasks caused by(Reduction of capacity)
        
        Pros(Put study Emphasis on divided attention & flexibility of attentional control)
        
        Can theorise mathematically using (Signal Detection Theory)
        
        Cons(vagueness of theory made it difficult to falsify)
      - Interfering Effects of Divided Attention
        
        CellPhone while Driving(More red lights missed & slower reaction time)2
      - Dual task performance
        
        GIST PROCESSING task (<capacity) than PHASE DISCRIMINATION task
        
        Shows:
        
        Different tasks (load capacity) differently
        
        Difficult task is much more affected by (central load)
      - Intentional blindness
        
        (Monkey business) illusion
        
        {In lab} Focus on flashing (arms of a cross) - Clearly visible (square) not detected
      - Systematic study of Capacity
        
        Allocate fractional amounts of capacity to two tasks
        
        Expect (graceful degradation) of performance as available capacity is reduced, not (catastrophic failure)
        
        Theoretical curve of trade off curve is called an AOC (Attention Operating Characteristic)
      - Visual and Auditory Dual tasks
        
        Detection vs Discrimination(increase/decrease in intensity)
        
        Discrimination task curve (consistent with capacity tradeoff)
        
        So Discrimination looks like a capacity demanding task
        
        Detection (no performance decrease when doing two tasks simultaneously)
        
        Task does not seem to require capacity
        
        Therefore, Detection processed (pre attentively) outside the (limited capacity system)
        
        Explanation(Detection by onset transient used for attention orienting)