Hazard anticipation of young novice drivers - SWOV

when they occur infrequently in time and space. This can be sudden sounds,
a flash of light or something with a bright colour in a grey environment.
Objects or events can also be instinctively salient (e.g. something with the
shape of a snake) or can have become salient by learning. The well learned
salient stimuli can be the content aspects and the context aspects of the road
and traffic situations experienced drivers perceive and of which based on gut
feeling, they immediately 'know' there is a latent hazard (see Section 3.8.1).
This subconscious knowledge is stored in neural representations (schemata).
Activation of neural representations (schemata) is not only the result of
stimuli in the road and traffic situation, but also the result of knowledge
stored in long term memory, feelings (internal state of the driver), the actions
she or he can take (motor) and information from other senses than the eye
(sensory). In terms of the theory presented in Chapter 3, the bottom-up
process of salience filtering can be seen as an element of contention
scheduling (the autonomous process of activation and inhibition of
supporting and conflicting schemata) and the neural representations as the
activated schemata at a particular moment in time. Note that bottom-up
salience filtering in Knudsen's functional model of attention is not synonym
with bottom-up fixations. A top-down fixation (e.g. when a driver looks in a
direction where nothing can be seen, but something is expected) can be the
result of bottom-up salience filtering for experienced drivers. The neural
representations (or activated schemata) are not only determined by salience
filtering, but also by sensitivity control. This is the process of the energizing
of certain schemata of the group of schemata that already have been selected
by the salience filters, comparable with what SAS does in Norman &
Shallice's model on willed and automatic control of behaviour (1986). The
black arrows in Figure 4.1 represent top-down attention and the grey arrows
bottom-up attention. The third process in Knudsen's functional model of
attention is 'competitive selection'. This means that signal strengths of
activated neural representations (schemata) are compared. The weighing of
the signal strengths of neural representations (schemata) leads to the
selection of a dominant neural representation (dominant schema). Note that
the dominant neural representation does not have to be processed in
working memory in order to elicit an eye movement. If for instance in case of
driving, the selected dominant schema is about a routine hazard (for an
experienced driver), gaze control will be bottom-up. This is represented in
Figure 4.1 by the gray arrow from 'competitive selection' to 'gaze control'.
When information is not processed in working memory, the attentional
processes remain in the realm of the contention scheduler. Also note that in
the bottom-up process new schemata can be energized (sensitivity control)
128