Language and Attention

Language and communication problems are key characteristics of autism spectrum
disorders. According to several theories, language problems in autism arise
from a bias towards detailed information, resulting in a failure to process
higher levels of language such as semantics, syntax and pragmatics. Yet, recent
results suggest that although individuals with autism are biased to attend to
details, they are actually able to process information for its global meaning
when instructed to do so. It therefore is assumed that autistic information
processing is caused by superior local information processing, rather than a
deficit in global information processing.
However, from our point of view, the drive to process language for its details
might very well stem from a deficit. We expect the focus on details to be a
necessary adaptation of the autistic individual. We propose that during normal
language development, children become increasingly automatised in language
processing, enabling them to attend to increasingly higher levels of language
(e.g. semantics, syntax), without jeopardizing lower level processing (e.g.
orthography). In individuals with autism, however, lower level processing
remains relatively attention demanding, leading to performance deterioration at
this lower level, whenever attention is directed at a higher level. This is in
contrast with current theories about information processing in autism, which
assume superior processing of local aspects of language regardless of
performance at other levels of language.

The study has three main objectives:

a) to investigate whether autistics can achieve normalized levels of language
processing when receiving explicit attentional instructions.

b) to study if (and if yes, how) an instructed attentional change towards
higher levels of language affects the simultaneous processing of lower language
levels.

c) to examine whether differences in attention needed for language processing
can be observed not only on a behavioural, but also on a neuropsychological
level, measuring Event Related Potentials.

The proposed theory will be studied in a series of experiments during which
autistic and typically developing participants will be tested on lower levels
(orthography) and higher levels (semantics, syntax) of visual language
processing.
During the experiments, the participants will be reading sentences from a
computer screen. Each participant will be tested on three conditions; a) a
low-level condition in which participants are instructed to focus only on
lower-level information, b) a high-level condition in which participants are
instructed to focus only on higher-level information, c) a dual-level condition
in which participants are instructed to analyze both lower-level and
higher-level information.
It will be examined how the instructed attentional focus affects the
proficiency of processing on various levels of language. The experiments only
differ in the high-level aspect of language that is included. The experiments
will be conducted twice with different participants; once measuring behavioral
parameters and once measuring Event Related Potentials.

The burden associated with participation is minimal. Autistic participants are
free to choose how many experiments they would like to participate in. They can
decide to participate in one or more experiments (45-60 minutes per
experiment). In addition, autistic participants are offered the possibility to
participate in multiple experiments at one day. Only when psychological
background information about the autistic participant is absent or outdated,
psychological tests will be administered to the participant under supervision
of a GZ-psychologist working at the institution. The non-autistic participants
will participate in only one of the experiments, and they have to perform only
one psychological test.
Furthermore, we expect the risks associated with participation to be minimal.
The study does not involve a clinical trial. We do not intervene on any aspect
of a participant*s life. We do not administer a drug or other forms of
treatment to the participants. The tasks we use to observe language processing
are similar to tasks that have to be performed at school (reading sentences).
Duration of the experiments is less than one hour, and the participants decide
for themselves how many experiments they will be participating in. We therefore
do not expect the study to lead to significant disabilities, incapacities or
mental health problems.