PSYCH LEC 7 (DRIVER)

Cocktail party problem

• Picking a specific stimulus to focus one, with many auditory stimulus occurring (noisy room full of people)

Selective shadowing

• I.e. picking out a relevant conversation at a party

• Most efficient with clear PHYSICAL differences between two messages

  • diff loc and voice freq.

Non-shadowed message: what can/can’t be reported

• I.e. ignored conversations at party that enter in the ear

• Can report

  • Substantial change in pitch of voice (female to make)

  • Sudden inserts of loud tone

  • Message endings

• Broadbent’s filtering model

• 1st Stage: Physical properties (pitch + sound location)

  • Processes them in parallel AKA pre-attentive stage

• 2cd Stage: Psychological properties (meaning of spoken words)

  • Filters to prevent overload AKA attentive stage

Methodological problem of assessing unattended info

1. Retrospective questioning; Where they have surprise questions abt non-shadow message at end of selective shadowing

2. Asking directly shifts attention away from selective shadowing

Falsifications

• Filter theory says unattended information does not bypass filter

  • Used Galvinic Skin Response, zapped ppl when word is said and caused their ANS to activate before zap= unattended message is present in ear during shadowing

  • Hearing your name

Late selection (Deutsch and Deutsch)

• Proposed limited awareness of unattended stimuli not abt rejection in full perceptual processing, MORE abt rejecting it from entering memory and control responses

• Unattended stimuli actually processed, refutes early filtration theory that it’s removed in 1st stage

Attenuation (Treisman) and evidence

• Deep unattended stimuli processing was ‘attenuated’ than filtered out

• Unattended enters 2cd stage but is much weaker than attended

• Stressed key roles of priming and partial info can effect psychological processing

Key roles of partial information and priming

• Partial input can exceed identification thresholds if item is primed to detect name, so less work to ID

Parallels in visual attention

• Found that they do not process unattended shapes using surprise memory test at the end of exp.

• Slower response times (negative priming) when superimposed green + red drawings into 2 different visual objects were presented and had to speed name object in one specified color

Early/late reconciliation and perceptual load theory

RESEARCHER LAVIE

Perceptual processing is automatic until reaches capacity

• late selection

  • Results obtained in low perceptual load (single stimulus) + high distractor interference = harder tasks

• early selection

  • Results obtained in higher perceptual load (more stimulus) + low distractor interference = hard to block distractors

Feature integration theory: simple features in parallel

• different visual features extracted preattentively in parallel w/o needing prior authorization

  • color and orientation

• Results in producing multidimensional objects

Attention to "bind" features together

Supporting studies

Neglect

"Extinction"

Processing of neglected info

Attention modulates sensory neural responses