Impaired sensory processing in Autism Spectrum Disorder: Towards better phenotyping and cellular models*

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder with a >1%
prevalence characterized by impairments in social skills and flexibility and by
a complex aetiological heterogeneity. The neurobiological mechanisms underlying
ASD are still poorly understood and the aetiological heterogeneity complicates
their elucidation. Recently, sensory processing sensitivities, such as extreme
sensitivity to light, sound, or touch, were added to the ASD diagnostic
criteria (DSM5) as negative and prevalent symptoms (reported up to 87%). These
symptoms are now considered *a critical cornerstone for characterizing and
understanding ASD*. A new network theory explains ASD by a disturbed
excitation/inhibition (E/I) balance in synaptic networks in the brain. We aim
to test this theory using behavioral, neuropsychological, europhysiological
and synaptic analyses of sensory processing deficits in ASD.

We will first exploit our Netherlands Autism Register to pre-select extreme
cases of high and low sensory sensitivity among ASD participants (currently
683) and control subjects and examine the relationship between sensory
sensitivity and other clinical characteristics and daily functioning skills.
Secondly, we will assess 80 pre-selected participants further, using customized
sensory processing tasks and EEG, which we have previously optimized to analyze
sensory sensitivity and E/I balance. Thirdly, we will utilize our unique
expertise to generate small, standardized neuronal networks in vitro, made of
human neurons derived from induced pluripotent stem cells (iPSCs) and to record
differences in synaptic and network properties. The sensory sensitivity and E/I
balance tests will be exploited to generate these standardized in vitro
networks of the most extreme cases. Finally, it has been proposed that the
(unknown) neurobiological mechanisms underlying sensory processing deficits are
similar to those underlying higher order social function deficits. Therefore,
we will test whether the outcomes of our neuropsychological and
neurophysiological tests are predictive for the reported higher order function
deficits.

This project brings together unique expertise to bridge across behavioral,
neuropsychological, neurophysiological and cellular research domains of ASD,
focusing on a symptom that is only recently recognized as a central symptom,
also in relation to other indicators of ASD, but for which integrated studies
across research domains are still lacking. Whereas higher order social
functions are difficult to measure quantitatively, sensory processing can be
easily quantified in a manner that is less biased for age, gender, or cultural
background. This quantification is expected to lead to a more objective,
unbiased and homogeneous ASD subtyping which is again expected to improve ASD
diagnosis and facilitate cellular studies with neurons derived from ASD
participants. Such iPSC-derived neuronal models, in contrast to current
(monogenic) animal- or cell-models, present the full polygenic complexity of
ASD. Given the expected subtle cellular effects and the heterogeneity in ASD,
current IPSCs-studies are often underpowered. We will exploit the precise
classification of the disease subtypes to overcome this problem and arrive at
homogenously phenotyped ASD subgroups. Together, this project provides a unique
opportunity to bring together information from these different research
domains, enhance our understanding of the complexity of ASD, and provide
intervention targets to alleviate the burden of ASD.

The main aim of this proposal is to characterize the neurobiological basis of
sensory processing deficits in ASD.

The first key objective is to exploit our Netherlands Autism Register to
pre-select extreme cases of high- and low sensory sensitivity among ASD- and
control subjects and examine the relationship between sensory sensitivity and
social and daily functioning skills.

The second key objective is to perform a detailed assessment of 80 pre-selected
ASD- and 80 control participants using sensory processing tasks and EEG,
specifically targeting sensory sensitivity and E/I balance.

The third key objective is to generate small, standardized neuronal networks in
vitro, made of human neurons derived from induced pluripotent stem cells
(iPSCs) and to record differences in synaptic and network properties related to
E/I balance.

We have defined two projects to reach these objectives. Project A will target
objectives 1 and 2 and provides the classification in ASD subtypes using our
two-step selection principle (pre-selection in key objective 1 and detailed
assessment using sensory processing tasks and EEG in key objective 2). In
addition, this project addresses to what extent altered sensory processing can
be used as a proxy for other aspects of ASD. This project will provide new fast
and robust diagnostic tools for ASD based on sensory processing tasks and EEG.
Project B will target objective 3 and characterize the neurobiological basis of
sensory processing deficits in standardized neuronal networks, with a focus on
the excitatory/inhibitory balance. This project provides insight into the
cellular basis of sensory processing deficits in ASD and new cellular models
for ASD. Both objectives contribute to the awareness and recognition of ASD as
a heterogeneous condition with sensory processing deficits as a defining
criterion.

a) The NAR: a large, unique ASD sample
We use one of the largest clinical ASD samples to date with sensory processing
data. All participants are recruited via the Netherlands Autism Register (NAR)
(www.nederlandsautismeregister.nl) which is a longitudinal cohort of 2000
participants with ASD (1000 adults; IRB approved). Since 2013, NAR participants
complete yearly online questionnaires that assess an extensive range of data,
among which clinical severity, treatment and medication, comorbid psychiatric
problems, education and wellbeing (Figure 2A). The NAR has added in 2016 a
validated self-report questionnaire on sensory processing (Tavassoli et al.,
2014). We will use the first results of this self-report to conduct the
preselection of low and high sensitive individuals in NAR (see Figure 1). In
addition we will utilize the extensive NAR data collection to explore how
sensory processing sensitivities in adults correlates with (predicts) ASD
severity.

b) Functional biomarkers for ASD: EEG
We will collaborate with NBT Analytics (www.nbt-analytics.com), a spin-off
company from the Linkenkaer-Hansen lab in Amsterdam, to produce an ASD
diagnosis tool, based on EEG measurements. This tool will be used for the
second step in our selection procedure for extreme
sensory sensitivity, depending on the outcome of subsequent synaptic analyses
in project B, further optimized as a diagnostic tool for ASD. NBT Analytics has
developed a unique approach to classify brain disease phenotypes based on EEG
biomarkers (see project A below), which has been successfully applied for
patient classification in other disease domains before (Figure 5). In addition,
they have developed a novel computational method using Detrended Fluctuation
Analysis (DFA) to quantify E/I ratio from brain oscillations measured with EEG
(Figure 3). We will apply this method to investigate the differences in E/I
balance in high- and low sensitivity in ASD. After finalizing the EEG study,
participants will be approached to participate in a DNA study, based on saliva
collection via mail.

c) The use of IPSCs in ASD research
In ASDs about 10% of the syndromes are monogenic, with reasonably good animal
models available (Nelson and Valakh, 2015). However, it is not clear to what
extent these animal models are relevant for the other 90% syndromes in ASD with
highly polygenic origin, where many disease genes contribute only a small
fraction to the total disease risk. Therefore, we chose to study these forms of
ASD using micro-networks of human neurons derived from IPSCs from selected
extreme cases of sensory processing.

see protocol