Obstacle Avoidance and Visual Field Deficits

In everyday life, we are faced with a complex environment that might contain
many possibilities for action. Depending on our intention we plan an action
toward a particular object, i.e. if we want to sip our drink an action is
planned toward our tea cup. Other objects that are in proximity of the target
object can act as obstacles in that they obstruct the movement of the hand and
arm. Therefore, a movement around an obstacle and toward the target needs to be
programmed. Our daily success at this task seems to indicate that this is a
trivial problem for our brain to solve. Evidence from case studies seems to
point towards the dorsal stream (Milner & Goodale, 2008) as a necessary neural
substrate for this ability (Schindler et al., 2004). The dorsal stream is a
brain network that stretches from the primary visual cortex (V1) in the
occipital lobe to several areas of the posterior parietal lobe (AIP, LIP, VIP,
etc.) and from there on to frontal areas such as the frontal eye fields,
supplementary motor and motor areas (M1). The dorsal stream is primarily
involved in the processing of visual information in order to guide
goal-directed actions.

It has been noted that the dorsal stream is cognitively impenetrable (Schenk &
McIntosh, 2010). This means that awareness of the obstacles does not seem to be
a prerequisite for a visually guided action such as planning a movement around
an obstacle (McIntosh et al., 2004). In bold strokes, planning an avoidance
movement goes as follows: visual or eye-centered information about the target
and obstacles enters the primary visual areas and is processed along the dorsal
stream which feeds into (pre)motor areas where this information is combined
with body-centered information about the current (start) posture in order to
program a movement that goes from the start posture to a comfortable
end-posture near the target while omitting limb configurations that occupy the
same space as the obstacle (Rosenbaum, Meulenbroek, & Vaughan, 2001).

Since the dorsal stream is necessary for avoidance behavior, obstacle avoidance
research appears to be driven in large part by manipulations of visual
information that is processed by the dorsal stream. Several features that are
processed by the dorsal stream have been manipulated and their effect has been
described: size, orientation, and location of obstacles have been proven to
affect hand movement trajectories. Conversely, a feature such as object
identity, a ventral stream property, has only been investigated in limited
quantity. The ventral stream is associated with object recognition and form
representation and runs from the successive visual areas in the occipital lobe
(V1, V2, and V4) to the inferior temporal lobe (PIT, CIT, and AIT). The ventral
stream enables us to consciously experience the world.

In order to be perceived as obstacles non-target objects first have to capture
attention (Castiello & Tresilian, 1999). We showed that the attentional capture
by obstacles can be increased and that the strength of attentional capture
determined the avoidance response (Menger, Van der Stigchel, & Dijkerman,
2012)). We did this by manipulating features of the target and obstacle that
were not picked-up by the dorsal stream, i.e. the similarity of the
target-obstacle pair. Indeed, we found that participants responded to ventral
stream information: participants deviated more from dissimilar obstacles than
from similar obstacles. Because similarity properties were never investigated
in participants who had no access to visual awareness of the scene and ventral
stream information we can (still) ask what mechanism drives this effect: is it
the saliency of a particular obstacle that affects dorsal stream processing of
the obstacle? Or is it its identity as a similar or dissimilar obstacle that
affects ventral stream processing of the obstacle?

The mechanisms underlying the interference of a dissimilar obstacle with hand
movements toward a target are currently unknown. In the current research, the
question will be answered whether target-obstacle similarity interferes with
processing when an obstacle is presented at a location in the blind field, i.e.
when it is not consciously perceived. This will reveal whether the ability to
consciously perceive an obstacle is a crucial factor for target-obstacle
similarity to interfere with hand movements.

The above mentioned research question will be answered by studying patients
with visual deficits. Specifically, we will try to recruit participants with
visual form agnosia, or hemianopia. Visual form agnosia is a rare, pathological
condition characterized by the inability to consciously perceive object shape.
Some of them however, show the ability to accurately grasp objects according to
their form. Simply put, the dorsal stream is intact while the ventral stream is
damaged. Hemianopia is a more common deficit where people suffer from blindness
in one half of their visual field. In a subsample of hemianopic patients,
residual processing of visual input in the blind hemifield has been observed,
without it entering awareness. This has been called blindsight (Weiskrantz,
1989, 2009). Thus, patients have the ability to direct behavior accurately
towards stimuli in their blind field while reporting no conscious experience of
anything present in the blind field. *Blindsight* is determined experimentally.

If a similar or dissimilar obstacle is presented at a location where it does
not enter visual awareness (either in the blind field for hemianopics or at all
for visual form agnosia patients), and it still influences behavior, this will
have interesting consequences for the dual stream hypothesis by Milner &
Goodale. The separation between *what* and *how* might then be less distinct
than assumed in their model. In that case, either the properties of the dorsal
stream need to be expanded or there is more interaction between the two visual
streams than originally had been assumed.

In this study, we will systematically investigate the influence of attentional
capture by obstacles in patients with visual field deficits. These patients are
impaired in the conscious perception of part of their environment. It will be
investigated whether an obstacle differentially interferes with the hand
movement when it is either similar or dissimilar to the target and when it
cannot capture attention. This will reveal whether the saliency of a particular
obstacle affects dorsal stream processing of the obstacle or whether its
identity as a similar or dissimilar obstacle affects ventral stream processing
of the obstacle.

A group (n=12) of adult participants with visual field deficits are invited to
participate in a behavioral study to investigate the mechanisms underlying the
attentional capture of non-target objects. A control group (n=12), matched in
age, will be tested using the same experimental set-up.

This is a non-therapeutic study in which the risks are negligible and the
burden is minimal. The experiments are non-invasive, so there is no need for
special preparation for the participants. There are no health risks doing the
tests. The data are primarily used for research purposes.