CH6 Flashcards
Task-set switching
efficient, coordinated performance of complex tasks often
depends on the ability to switch from one task or task component to another.
Advance reconfiguration can help with this: they’re like processes that facilitate this
transition. Switching costs are increases in RT and error rates on the task after the
switch and reflect the time needed to adopt the appropriate task set. Task-set
switching depends on two components: a top-down, control component and a
bottom-up component in which the stimulus triggers the appropriate set. It does not
seem to be the case that task-set switching is always all-or-none, and repeating the
same task has benefits over task switching regardless of foreknowledge.
Advance reconfiguration
intentionally prepare for
task switch.
Switching costs
increases in RT and error rates, time needed to adopt. Poorer performance for task-switch trials than for consecutive same-task trials
Task-set switching and executive control:
it may be that executive control is
responsible for determining which task will be performed, but that readiness (and RT)
to perform the task depends on more automatic processes of inhibition and activation
from preceding trials. Also, there might be a general need to reinstate the stimulus-
response mapping (aka retrieve it from memory), so switching seems to contain
restart costs and may also reflect negative transfer.
De Jong: model of residual switching costs
based on limitations in the ability to fully
prepare for a new task before the presentation of the first stimulus and on the failure
to take advantage of opportunities for advance preparation. He proposes that
residual switch costs rest largely on a “failure to engage” in advance preparation for
an upcoming task. If this is true, residual switch costs should be observed only on
trials for which performers fail to prepare for the new task.
Residual switch costs
failure to fully discard or inhibit
previous task set.
task-set switching Common assumptions
- task set depends on configuration of processing
pathways/modules in brain through which some operations are facilitated and others inhibited, degree
of readiness, - processing system stays in state of readiness until switched again. But: even
well-prepared switches show residual switching costs.
James: performance becomes automatic after
practice: it becomes independent of
attentional control. There are two sides to this: reductions in capacity demands, and
independence of voluntary control. According to resource views of attention, two tasks can
be performed together only to the extent that sufficient attentional resources are available
Both single- and multiple-resource views imply
that if one of two tasks is automatised, it is
possible to perform the two tasks simultaneously with little or no cost. Studies found this to
be true, but it could be that really the subjects just got better at switching attention between
the two tasks. Although switching takes time, it is plausible to suppose that responses to one
task can be buffered, and this buffering provides the participant with the chance to select
information or responses on the other task. Research has also shown that automatised
processes are not entirely immune to interference, but with practice, performance improves.
Stimulus onset asynchrony
Stimulus-onset asynchrony is a measure used in experimental psychology. SOA denotes the amount of time between the start of one stimulus, S1, and the start of another stimulus, S2. In this respect, a stimulus may consist of, e.g., a presented image, sound or printed word.
Rather, interference occurs
at moments when response
selection must occur in both tasks
The psychological refractory period effect (PRP)
dual-task performance may suffer
because both stimuli cannot be fixated at the same time, and to avoid this, the stimuli
for the two tasks are often presented in different sensory modalities and the
responses are made with different effectors. An important finding is that the RT to the
second stimulus is slowed relative to when that stimulus is presented alone, and this
is called the psychological refractory period (PRP) effect. This shows that at least
some processes may be carried out for only one task at a time.
Graded capacity allocation
a characteristic of the response-selection bottleneck model is that response-selection processing is all-or-none. However, capacity- sharing models can account for most of the data from the PRP paradigm just as well as the bottleneck models, so that’d be a good alternative.
Effects of practice on the PRP effect
the PRP effect was reduced considerably when
extensive practice with one SOA was given, but there was little transfer of this benefit
to shorter or longer SOAs. There is evidence that the reduction of the PRP effect with
practice is primarily due to a decrease in the duration of the bottleneck stage for task
1.
Alternatives to bottleneck models:
Their explanation of the PRP effect is based on an
information-processing architecture, the executive-process interactive control (EPIC)
framework. Here, the psychological refractory period effect is attributed to a task
strategy, specifically one of response deferment for task 2. However, this model is not
very strong. The main difference is whether the sequential processing is a “built-in:
limitation of the basic information-processing architecture or a strategy that subjects
adopt in order to comply with the task instructions.
