3. Top-down Processing Flashcards
constructivism and context
- contextual cues allow us to decode stimulus
- can facilitate or hinder processing (appropriate vs inappropriate)
illusions as experience
- gestalt principles were believed to be responsible for the illusion
- we look for regularities in the env.
cross-cultural experience
refer to slide 7 of powerpoint
constructivism and world recognition
- one of the most researched areas in experimental psychology is language
- most early studies focused on elementary features in perception (identification and recognition) and memory (serial recall)
1. frequency effects
2. similarity effects
3. semantic contexts effects
constructivism and world recognition: frequency effects
high freq words produce better ACC and fast RT than low freq words
constructivism and world recognition: similarity effects
- words exist in context of other words
- words with many neighbours (book, look, took) produce poorer results than those with fewer neighbours
- issues of distinctiveness/competition (recall uncertainty in info)
constructivism and world recognition: semantic context effects
- performance increases for word recognition if a word is presented with semantically similar words (e.g pen paper ink) or in a sentence
- allows deeper meaning of the word
word superiority effect
- when people are more accurate in identifying a letter when it is part of the word
- RECALL discussion of info theory and decrease of uncertainty
- more likely to recall/detect stimulation when presented in context of words due to LTM
- refer to slide 11-12 for example
interactions between bottom up and top down mechanisms
- studies suggest that there is an interaction between bottom up and top-down mechanisms which process info in parallel
- the IAM (interactive activation model) is a pandemonium-like system that includes both BU and TD processing
the IAM
the study:
1. features, letters and words were built from the alphabet (slide 14)
2. the system was equipped with 1179 letter words
the IAM elements
1. perceptual processing occurs within a system in which there are multiple levels of processing
- each level is concerned with forming a representation of the input at a different level of abstraction (features, letters, words)
- they use output of one system as their input
2. visual perception involves parallel processing
3. visual processing occurs at several levels at the same time
the IAM: representational assumptions
- there is a node for each word and each letter (in each letter position)
- the nodes are organized into levels
- the nodes are connected to all other nodes within levels or between adjacent levels
- connections may be excitatory or inhibitory
refer to slide 20 of powerpoint
the IAM: operational assumptions
- when stimulus is represneted, a set of featural inputs = made
- initially, all nodes are in a quiescent state
- features activate the letter s, which in turn activate the words, which finally send back activation to the letters
- the answer of the system is the letter with the most activation at a given time
refer to slides 22-23 of powerpoint
the IAM: conclusions
- with a limited nb of clearly defined postulates, the IAM can account for the word superiority effect
- this model is a direct precursor of contemporary connectionist models
speech perception as categorical perception
- speech perception examines how we identify speciial kinds of auditory stimuli
- speech perception examines how we identify special kinds of auditory stimuli
- in most types of perception people perceive stimli along a continuum - does this hold for speech stimulus
- speech errors (and typos) seem to indicate a nb of interesting aspects of our language
speech perception: bottom-up and top-down
when listening to a speaker, we are required to perform numerous operations
speech perception: phonemes
- auditory cues are used to differentiate speech sounds
- phonemes: represent smallest unit of sound that provides meaningful info in terms of the meaning of the word
1. stop consonants: b/d/g
2. voicing (unvoiced or voiced): p/b - voice-onset-time: delay between
1. release of articulators
2. vibration of the vocal folds
auditory vs speech perception
- when presented with stimuli, we would expect that the proportion of ‘loud’ responses would increase with actual increase in loudness
- looks like psychophysics
- allows to detect different stimuli (slide 28)
- people have access to acoustic properties of the stimuli, then group them into phonemes they have stored in LTM
What should we expect to see in speech sounds auditory stimuli?
- a systematic decrease in the nb of responses reported as a given phoneme
- despite differences, neighbouring stimuli within a category are perceived as the same
- we observe a discontinuity or category boundary
speech perception: crosscultural findings
- categorical perception is a orbust phenomena, however, we must recognize that it is a function of individual experience
- there is some element of learning regarding different category boundaries
- certain lab techniques can be used to add a category boundary (e.g. non-native phonemic category can be perceived)
- prior linguistic knowledge can impair phonemes sometimes (being overridden by our knowledge)
what is the trigger to an affective response?
- initially encoding a stimlus, which starts a cascade of processing that leads to an affective response
- refer to diagram in notebook page 20
multimodal perception
- most perception in our everyday lives is multimodal wherein we integrate different types of info
- McGurk effect
- multimodal and multidimensional cues also provide the basis for illusions (e.g. vestibular and visual = crazy kitch)
multimodal perception: McGurk effect
- observed when we integrate info from visual and auditory (e.g. b vs g) modalities
- results in a perceived qualitative difference (e.g. d)
