Eye-gaze to memory

Autistic disorder has three core symptom domains: deficits in communication,
abnormal social interactions, and restrictive and/or repetitive interests and
behaviours. The last 20 years have seen a drastic increase in our understanding
of the behavioural and neural basis of the social impairment in autism.
Autistic individuals differ in the way they process faces. One main feature
seems to be an avoidance of the eye region.
A delay in the development of visual joint attention is also reported, this
delay could follow from a lack of interest in the information conveyed by the
eye. There seems to be a lack in this fundamental ability to gaze at what
another person is looking at, and this deficit appears to be correlated with
their level of social impairment.
Psychological studies on gaze following have demonstrated that another*s
individual gaze triggers a shift of attention automatically, and is even
reported when the gaze cannot be perceived consciously.
Gaze is special in triggering this shift of visual-spatial attention.
The effect of eye-gaze is not limited to a shift in visual-spatial attention.
Firstly, objects that are gazed at by someone else later seem more familiar,
and are given a positive value. Secondly, direct gaze also influences access to
categorical knowledge related to the person and the later recognition of the
face. Actually, eye-gaze perception is probably a very important factor to
developing social cognition.
The hypothesis was advanced that, in autism, an early impairment at processing
eye-gaze could lead to more drastic social impairments later on. However,
evidence for a deficit of eye-gaze processing in autism has been quite
inconclusive so far. Only 2 out of some 10 studies showed abnormal orienting to
eye-gaze in autism and one study demonstrated that autistic fail to show the
same advantage as control for detecting direct- over averted-gaze in an
odd-ball paradigm. The paucity of positive findings is surprising as, in
naturalistic situations, impairments of mutual gaze is documented in the
majority of children.
Previous experiments on eye-gaze in adults with autism typically involve
testing the presence of a shift of spatial attention in a situation where a
schematic face or a photograph of a face looks left or right before the
apparition of a target on the congruent or the incongruent side of the screen.
Sometimes, one condition also includes arrow cues instead of faces as control.
The gaze is always consciously perceived.
The lack of positive findings in previous studies could have two main reasons:
(1) the gaze was available to consciousness, opening the door to compensation
strategies in the autistic group, and (2) the use of schematic faces and arrow
cues hamper the ecological validity of the experiment.
One key finding demonstrating that autistic are actually impaired at automatic
processing of eye-gaze information is that they fail to show the same
difference between gaze and arrow cue as control subjects when the cue is
counter-predictive. This suggests that eye-gaze is processed as a non-social
stimulus by autistic individuals in these experiments.

In the present study we want to show that automatic processing of eye-gaze is
impaired in autism, and that this deficit influences the way autistic
individuals perceive the world. We will test this hypothesis using a masked
priming procedure to avoid any compensatory strategy in the autistic group. The
eye-gaze of the participants will be recorded in order to obtain a measure of
the shift of attention. In order to assess the influence of eye-gaze on memory,
we will test the participant*s recognition of objects that were or were not
gazed at during the encoding phase of the experiment. The recognition memory
test will use a remember/know procedure in which the participants have to
indicate whether they can remember something about when they saw the item the
first time (e.g., what they thought, how they were feeling). This will provide
an information about the type of memory process that is influenced by social
contact (i.e. eye-gaze).
In one condition, we will test the effect of a straight gaze on eye movements
and memory performance. Straight gaze is not processed normally in autism. In
typically developing (TD) individuals, straight gaze facilitates semantic
categorization, and face recognition. We hypothesize that autistic will not
show any effect of straight gaze, while in controls, we expect straight gaze to
delay the shift of attention towards the cups, and therefore impede memory
performance.
In order to test the hypothesis that autistic individuals use compensatory
strategies when processing eye-gaze in an explicit visual-spatial orientation
task, we will also test our participants on such task.
Finally, females have better social skills than males, and autistic are
sometimes considered as having an extreme male brain. In order to test the
hypothesis that females are more sensitive to eye-gaze than males, and males
than autistic, we include a group of healthy female participants next to the
autistic and the male controls.

Given that an IQ in the normal range is an inclusion criterion, the IQ test
will be administered before the main experiment.

For the present study, we designed a set of stimuli consisting of 80
photographs of a pair of cups (160 cups), with one cup on each side of a table,
and another 320 photographs of a single cup (160 old and 160 new cups) for
recognition.
There will also be a picture of a face with eyes gazing straight, down, to the
left, or to the right (at the location were the cups supposed to be). The
pictures of the face were cropped around the face to remove the context.
In addition one picture was used to create a mask of the same luminance and
color as the face by scrambling the picture.
Although it is known that a face presented for 25 ms and masked cannot be
perceived consciously, we designed a pre-test to ensure that this is the case
with our stimuli. In this test, 15 typically-developing volunteers were
presented the faces between the 2 masks, which were displayed for 80 ms, and
the face for 26.7 ms. In one task, the participant had to decide whether there
was a face between the 2 masks or nothing. Ten out of 15 participants answered
below chance level. In a second task, the participants had to decide whether
the face was looking left, right, or straight at them. None of the participants
scored above chance level. On this basis, we can be sure that eye-gaze cannot
be consciously detected with our procedure.

1 Memory experiment
In the main experiment, participants will not be told that a face is flashed
before the stimulus. The stimuli of the memory task consist of 320 different
cups. 160 cups are presented during the encoding part of the memory task. The
cups are presented by pairs in one of 4 priming conditions: face (1) gazing
down, (2) gazing at the camera, (3) gazing at the left object, (4) gazing at
the right object. The number of items is constant across conditions (n = 20).
The items are randomly assigned to the 4 conditions for each participant in
order to prevent any object-specific effect. The prime is flashed for 26.7 ms
between the 2 masks which are displayed for 80 ms. Each pair of cups is
presented for two seconds immediately after the second mask. The participants
eye-gaze is monitored during encoding using a camera controlled by a second
computer synchronized with the stimulus presentation computer. Participants*
head remains stable thanks to a head-holder attached to the table. In the
second phase of the experiment, 320 cups (160 old, and 160 new cups) are
presented at a central location on the screen, and the participant has to
decide if s/he confidently saw the cups before. The response is given by
pressing one of two keys (*yes I am sure* vs. *no I am not sure*). When the
participant is sure a specific cup was seen at encoding, s/he will be asked
whether s/he just know s/he saw it, or s/he can remember something about when
s/he saw the cup. Specific episodic memories will be manually recorded by the
experimenter. The total duration of the experiment is about 30 minutes

2 Detection post-tests
Although the timing parameters of the priming experiment were determined
according to a pilot experiment, we will also test the ability of every
participant to consciously detect the face, and the direction of eye-gaze
(left, right, straight).
The detection post-tests shall take place after the main experiment, and before
the explicit orienting task. Every participant will perform two tests. In the
first test, a face or a scrambled picture of the face is presented between 2
masks. The participant must decide if there was a face between the two masks
(yes / no). There are 120 items, half with a face, and one fourth at each
exposure duration. In the second test, the participant must decide if the face
between the 2 masks is looking straight, to the right, or to the left. There
are 180 items in total; 60 items in every condition (straight, left, or right).
The total duration of the detection post-tests is 15 minutes.

3 Explicit orienting task
In this test we use the same faces. This time, the face will be displayed for
250 ms, and followed by a square appearing on the left or on the right of the
screen at the same place as the cup used to be. The asynchrony between the face
and the apparition of the square is either 100 or 300 ms. The task is to press
the left or right key according to the position of the target square. Half of
the squares appear on the left and half on the right. The eye-gaze does not
predict the position of the square (50% congruent trials). The participant will
be told at the beginning that the gaze is non-predictive. There are 160 items,
80 at each onset asynchrony. The explicit orienting test will be completed in
about 10 minutes. With this post-test, we hope to show that our group of
autistic participants orient towards the direction of gaze in the context of an
explicit task.

The experiment will not entail more than minimal risk to the participants. The
study is not intended to benefit the subjects directly. However, the data
collected during this study could improve our understanding of autism and human
cognition at large.