429
TABLE I
SELECTIVE ATTENTION: TEN EMPIRICAL
GENERALIZATIONS REFLECTING THE LITERATURE
UPTO 1986
W.A. Johnston and V.J. Dark, Ann. Rev. PsychoLi 1986, 37: 43—75.
1. All levels of stimulus analysis can be primed for
particular stimuli.
2. Selection based on sensory cues is usually superior to
selection based on semantic cues.
3. Irrelevant stimuli sometimes undergo semantic
analysis.
4. Spatial cues are especially effective cues.
5. Attention is independent of eye fixation and can assume
the characteristics of an adjustable-beam spotlight.
6. Stimuli outside the spatial focus of attention undergo
little or no semantic processing — is restricted mainly to
simple physical features.
7. Overlapping objects can be selectively processed.
8. Non-Selected objects within the spatial foci of attention
undergo little or no semantic processing.
9. Selective processing is sometimes performed passively
and sometimes actively.
10. Selective attention can be guided by active schemata.
Mirsky and Orren (1977) described 3 such spheres
— consciousness, sleep-wakefulness and orienta-
tion-habituation. Together they strongly influence
arousal, vigilance, attention and adaptive interac-
tions of the organism with the environment.
Cutting across these schemata are numerous at-
tempts to describe the functional mode of attentional
mechanisms. These account for numerous observa-
tions in terms of categories of characteristics that in
turn allow predictions for more detailed experiment.
They are thus heuristically useful but intellectually
limited by the constraint of the dichotomy they erect.
Examples include serial vs. parallel processing,
active∕passive or automatic∕controlled processes,
stimulus-∕response-set, sensory-∕concept-driven,
open-∕closed-loop or exogenous/endogenous con-
trol of processing (Straube and Oades 1992 and
references therein). Suffice it to say, in this short
overview, that one can imagine a concept-(experi-
ence)-driven selection proceeding ‘automatically’
and, depending on the demands of the situation,
involving either serial or parallel processing.
For someone setting out to investigate biologi-
to objects
Fig. 2. Kahneman’s scheme (1973, with permission) illus-
trating interactions of perceptual mechanisms and attentional
processes.
cal measures associated with selective attention the
model elaborated now 20 years ago by Kahneman
(1973) is still useful (Fig. 2). Some advances in
ERP research can be usefully placed against the
background of this model even though it is biased
toward serial processing schemes. Thus, for ex-
ample, we can see that dimensions of a stimulus
may be processed hierarchically according to per-
ceptual difficulty, as was demonstrated with meas-
ures of processing negativity (PN), where locus in
a dichotic paradigm is processed faster than pitch
(Hansen and Hillyard 1983). But fundamental to
the design of most modern studies of selective
processes is the emphasis on the adaptive rather
than the salient aspect of the stimulus: the ability
to discriminate a stimulus that is relevant to the