Exploring biological measures to facilitate the discovery and development of new treatments for social and cognitive deficits in Alzheimer*s disease, schizophrenia, and major depression: replication and generalisability of the Psychiatric Ratings using Intermediate Stratified Markers (PRISM)1 study

The development of treatments for neuropsychiatric conditions forms one of the
major challenges of our time in public health and in pharmaceutical industry.
According to the World Health Organisation (WHO) report on priority medicines
for Europe and the World (Update Report, 2013), neuropsychiatric conditions
rank third on mortality for the EU (5.4% of total) and rank second on burden of
disease as measured in disability, adjusted for life years (DALYs) (19.6% of
total DALYs).
The current pharmacopeia for the treatment of neuropsychiatric disorders
amounts to over 100 compounds. However, most of them were discovered by
serendipity and have the same mechanisms of action. There is still a high,
unmet clinical need because of issues including insufficient efficacy, limited
tolerability, as well as the existence of domains that have no single approved
therapeutic yet. Since the 1950*s, very few new therapeutic concepts and
Mechanisms of Action (MoAs) have been identified, and no new relevant treatment
options have emerged for either neuropsychiatric or neurodegenerative disorders.
One reason for this lack of meaningful progress may be that the diagnostic
classification system of neuropsychiatric disorders (be it International
Classification of Diseases-11 (ICD-11) or Diagnostic and Statistical Manual of
Mental Disorders, 4th Edition (DSM-IV)) separates such disorders into
non-overlapping diagnostic categories, such as schizophrenia (SZ), Major
Depressive Disorder (MDD), or Alzheimer*s disease (AD). This separation is not
based on any underlying aetiologies, but on convention-based, qualitative
clustering of clinical symptoms. While these current diagnostic categories are
helpful to provide a basis for general clinical management, they fail to
consider the underlying neurobiology that often gives rise to variations in
symptoms between patients. The ability to link symptoms to the underlying
neurobiology would be expected to facilitate the development of better
(personalised) treatments and could also allow physicians to provide patients
with a better understanding of the complexities and management of their illness.
The original PRISM (Psychiatric Ratings Using Intermediate Stratified Markers)
project, conducted between 2016 and 2020, sought to determine whether similar
symptomatologies, that are assumed to result from different pathological
processes, could be associated using quantitative parameters. The PRISM project
selected social dysfunction as the *symptom* of interest. AD and SZ, two
disorders of different origin (one being neurodegenerative, the other
neurodevelopmental) and with a different age-profile, but both partly
characterised by social dysfunction, were used as the prototype disorders.
Choosing social dysfunction as the phenotypic transnosological domain for PRISM
was motivated by the fact that this provides a major burden to patients and
their families, and this area has no approved approach for clinical
intervention yet.
As anticipated, PRISM has generated a wide variety of interesting new data and
insights, including links between social dysfunction and functional integrity
of the Default Mode Network (DMN). The DMN is a network of brain regions,
characterised by high functional connectivity, that generally exhibits higher
activity during rest than during performance of many attention-demanding tasks.
The PRISM 2 refined test battery will include only the most relevant paradigms
to test for the reproducibility and generalizability of the identified
relationships between social dysfunction and DMN integrity, including
functional magnetic resonance imaging (fMRI) and electroencephalography (EEG))
and connectional (diffusion tensor imaging (DTI) & resting state-fMRI
(RS-fMRI)) paradigms, as well as social dysfunction measures (total social
functioning scale (SFS) score and BeHapp smartphone data). This refined and
optimized test battery will be deployed in SZ/AD patients and age-matched
healthy controls. In addition, to further attest the generalisability and
therefore transdiagnostic utilty of DMN social-dysfunction interactions, we
will add MDD patients to our sample, given the substantial social deficits
documented in this patient population that previous research has linked to
altered DMN integrity. In addition to the effort to determine the
reproducibility and generalizability of the identified quantitative,
transdiagnostic, biological parameters which have significant relationships
with social dysfunction, PRISM 2 will also include the assessment of novel
biomarkers. It is well understood that there is a neurobiological basis to SZ,
MDD and AZ which is present at the DNA, RNA, and epigenome level (Khavari &
Cairns, 2020; Lozupone et al., 2019; Qazi et al., 2018; Smigielski et al.,
2020). Therefore, a blood draw and subsequent analysis will also allow the
exploration of biological and epigenetic biomarkers. Analysis of peripheral
blood samples with subsequent transcriptome and proteome profiling may aid the
discovery of novel biomarkers that predict social functioning outcomes or
identify new therapeutic targets. In addition, spoken language contains several
measurable biomarkers that can indicate various aspects of cognitive health and
social skills.

The overall goal of the PRISM programme of research is to develop a
quantitative, transdiagnostic, neurobiological approach to the understanding of
neuropsychiatric disorders in order to accelerate the discovery and development
of better treatments for patients with those disorders.
The main goal of PRISM 2 is to show the reproducibility and generalizability of
quantitative biological parameters which were identified in the original PRISM
clinical study (Protocol number ABR59359) as having significant relationships
with social dysfunction, in a transdiagnostic manner. Namely,
1. SFS total score and rostromedial prefrontal cortex (PFC) activity
2. SFS and BeHapp composite score
3. BeHapp composite score and DTI fractional anisotropy (FA)
4. BeHapp composite score and EEG
5. SFS and connectivity between the medial PFC (mPFC) and amygdala

We aim to work towards this goal in the current study by having a set of
distinct objectives. Objectives are defined as either replication objectives,
generalisation objectives, or discovery objectives, with clearly defined
end-points

Replication Objectives:
1. To replicate, in a separate cohort, the finding that DMN functional
connectivity in the rostromedial prefrontal cortex (rmPFC) on rsfMRI is
negatively associated with social functioning in patients with schizophrenia,
Alzheimer*s Disease and in Healthy Control participants.
2. To replicate in a separate cohort, the correlation of SFS self-report social
functioning scales with the objective BeHapp measurement of social
functioning.
3. To replicate in a separate cohort, the finding that high BeHapp social
functioning scores are associated with high DTI fractional anisotropy (FA) in
the Inferior Frontotemporal Fasciculus (IFF) and forceps minor (FM)
4. To replicate in a separate cohort, the finding that high BeHapp social
functioning scores are associated with high resting state EEG connectivity
index scores in the DMN of patients with AD, SCZ, and in HC, independent of
their diagnostic labels.
5. To replicate in a separate cohort, the finding that reduced connectivity in
response to emotional valent faces between the mPFC and amygdala is associated
with lower social functioning in patients with SZ, AD, and in HC independent of
diagnostic labels.

Generalisation Objectives:
1. To demonstrate that the relationship of comparatively lower DMN functional
connectivity in the rmPFC predicts worse social functioning in SZ, AD, and HC
applies to patients with Major Depressive Disorder (MDD).
2. To demonstrate that the finding of high BeHapp social functioning scores
predicts high DTI FA in the Inferior Frontotemporal Fasciculus (IFF) and
forceps minor (FM) in SZ, AD, and HC applies to patients with MDD.
3. To demonstrate that the finding of higher BeHapp social functioning scores
predicts higher resting state EEG functional connectivity in the DMN applies to
patients with MDD.
4. To demonstrate that reduced connectivity in response to emotional valent
faces between the mPFC and amygdala is associated with lower social functioning
in patients with MDD.

Discovery Objectives:
1. To determine whether speech-based endpoints measured from a variety of
speech elicitation tasks are related to social dysfunction in participants with
SZ, AD, MDD and HC participants. The speech-based endpoints measure various
aspects of cognitive-linguistics and speech motor control.
2. To determine whether EEG DMN connectivity index scores (in response to
emotional valent faces) are associated with social dysfunction in patients with
SZ, AD, MDD and in HC participants independent of diagnostic labels.
3. To discover multimodal biomarkers that predict social functioning outcomes
using data-driven machine learning
4. To explore population subtyping and the relationships between these subtypes
and outcomes through application of unsupervised clustering techniques.
5. To determine multivariate relationships between biomarkers and demographic
characteristics that predict social functioning outcomes.
6. To discover candidate biomarkers informed by diagnostic labels or symptom
severity.
7. To discover biologically interpretable biotypes described by the epigenome
and/or the transcriptome through the application of unsupervised methods
8. To identify candidate and discovery inflammation, synaptic integrity, and
neurodegeneration gene expression in peripheral blood samples as biomarkers of
disease using transcriptome and proteome profiling
9. To support verification of the identified genes and pathways using more
focused approaches such as RT-qPCR, mass spectrometry, and immunoassay.
10. To explore the correlation between the identified peripheral gene
expression and social functioning in AD, SZ, MDD, and HC.

The study comprises a naturalistic, cross-sectional study of four groups. Three
patient cohorts, 1) patients with probable AD, 2) patients with SZ, and 3)
patients with MDD will be studied, together with 4) a healthy control group,
approximately matched in age distribution and gender proportion to the SZ, AD,
and MDD groups.
Approximately 40 participants will be recruited into each of the AD, SZ, and
MDD groups, together with approximately 60 healthy controls, resulting in a
total of approximately 180 participants.
First, a pre-screening is performed. The purpose of the pre-screening email
correspondence or telephone call is to discuss the study and answer questions
that potential participants might have, and to ascertain the likelihood that
participants will be eligible (i.e. satisfying inclusion/exclusion criteria).
By performing a telephone or email pre-screening we intend to reduce the burden
for potential candidates, by preventing them for as much as possible to visit
the study center only to find out they are not eligible for participation, for
example, determining any obvious MRI contraindications, such as claustrophobia.
Typically, participants will attend the study centre for a single assessment
day. This assessment day includes screening, the collection of questionnaire
measures, behavioural testing, a blood draw, the installation of BeHapp on
participants* phones, and a short MRI and EEG neuroimaging session.
In some cases, participants may indicate fatigue during assessments days, or
due to restrictions in their availability may not be able to complete all the
required activities for a given assessment day. In these cases, the participant
will be invited, if they prefer, to attend a greater number of assessment
visits (e.g., a separate visit for the EEG and MRI scan) to complete all the
study activities. The number of assessment visits required to complete all the
study activities will be determined by the participant*s capacity and left to
the judgement of the clinician. All study activities will preferably be
collected within a single week. No assessment day will have a duration of
greater than 6 hours, and, as far as possible, the order of study events will
be held constant across participants.
Follow-up, during which time remote data collection with BeHapp will be
collected, will continue for a maximum total follow-up period of up to 42 days
(6 weeks) from the screening/assessment day. Participants will be encouraged to
engage in follow-up for at least 28 days (4 weeks) to allow for the collection
sufficient data for meaningful analysis, but they can choose to discontinue
BeHapp data collection at any time whilst continuing with other aspects of the
study. Therefore, the maximum duration of the study, if a participant consents
to BeHapp installation, is 42 days.

Potential benefits to participating in the study are low on the short term.
Participants may value contributing to research that may lead to new treatment
approaches in the future. However, risks associated with participating in the
study are also low. Some risk is associated with the blood-draw procedure, such
as bruising. Risks associated with MRI scanning are minimized by screening of
research participants prior to each scan. Although there are no associated
risks with the EEG, it may cause minor discomfort.