Schizophrenia as a disorder of social cognition: an fMRI study in adolescents

Psychosis manifests itself in social interactions. This is most evident in the
core symptoms of psychosis, especially paranoid delusions, which are
characterised by a fundamental lack of trust in others. Trust is a necessary
component of successful human interactions, yet appears too complex to probe
experimentally. However, the contemporary development of *neuroeconomics* has
shown that complex social interactions, such as trust, can be operationalised
in economic exchange games. Recent studies in healthy individuals measuring
cortical responses during such games have shown that being engaged in a real
interaction involves motivational and emotional processes, that are crucial to
the development of trust between interacting subjects (King-Casas et al.,
2005). This approach has highlighted activation in brain areas linking sensory
representations with their motivational significance (e.g., striatum, amygdala,
orbitofrontal cortex) (King-Casas et al., 2005; Rilling et al., 2002; Singer et
al., 2004), in addition to the set of brain regions traditionally associated
with social cognitive tasks (i.e., the posterior superior temporal sulcus, the
temporal poles, the inferior frontal cortex, and the medial prefrontal cortex)
(Gallagher & Frith, 2003; Frith & Frith, 2003; Gallese, Keysers & Rizzolatti,
2004). This points to a possible mechanism underlying disturbed social
interactions in psychosis, bringing into play contemporary theories of dopamine
(DA) function. Mesolimbic DA has a central role in reward, learning and
motivation (Schultz, 2002), and is also thought to be crucial to the
pathophysiology of psychotic symptoms (Davis, Kahn, Ko, & Davidson, 1991;
Seeman & Kapur, 2000). Through linking recent insights in the mechanisms
underlying social interactions, with contemporary theories on the role of DA in
psychosis (Kapur, 2003), it is possible to develop the hypothesis that an
aberrant reward processing mechanism underlies disturbed social interactions
which in turn may lead to the formation of paranoid delusions.
Paranoid delusions and hallucinations can also be triggered by the mental state
of the person and his interaction with his environment. To elucidate this
proces, we added the Experience sampling method. This method alows us to
register fluctuations in daily activity, emotions, mental states and stress.
Correlations with other tasks will be investigated.
Recently, studies investigating psychosis include induviduals with high risk of
psychosis. Siblings and relatives are one group, by their shared genetic
vulnerability. Another group is the *Ultra High Risk* (UHR) group, individuals
who show some characteristics of psychosis, but are not (yet) psychotic. These
individuals are usually medication-free, this means that there is one
confounding factor less. This makes outcomes more reliable. By including a
Ultra High Risk group, the difference between consequences of the disorder and
precursors/markers of it, can be distinguished: If the UHR group shows the same
characteristics as the patients, these characteristics will be markers of
psychosis; characteristics only found in the patient group will be due to the
disorder, and consequences of it.
For research in the field of psychosis, especially the early detection and
intervention, discovering markers is vital.

The aim of this project is to elucidate the mechanisms underlying one of the
most incapacitating psychiatric disorders, psychosis. Despite antipsychotic
treatment, up to a third of patients with psychotic illness continue to
experience significant distressing and disabling psychotic symptoms. Acquiring
an understanding of the cognitive and emotional mechanisms that have led to the
psychotic beliefs is crucial. The current project applies novel approaches
derived from behavioural game theory to further understanding of the mechanisms
underlying symptom formation. These approaches allow the study of the symptoms
as they are expressed in social interactions.

This is an experimental study which consists of two separate sessions: During
the first session, participants will be screened, assessed on the
questionnaires and trained on the trust game task. The second session consists
of a one hour fMRI-scan, during which participants will perform three trust
games (i.e. each with a different partner).
The task in this study represents a slightly modified version of the multiround
game implemented by King-Casas et al. (2005). Participants in the study play
the role of the investor throughout the whole game, and thus always make the
first move. Both investor and trustee start with the same amount of money
(10¤). The investor decides which part of this amount, i.e. between 1¤ and 10¤,
he wants to share. Shared money will be tripled. The partner, i.e. the trustee,
can either repay part of the received money back to the investor or keep all
the money. Functional MRI data will be obtained throughout the entire game
session. Participants will play against the computer, but are told they are
playing against two different human partners and a computer with a random
repayment strategy. Specifically, they will be informed that their human
partners are tested in a different location and they play together via the
internet. The computer algorithm will consist of two versions, reflecting a
cooperative and a deceptive style of playing, all programmed in a probabilistic
way. Each experiment starts with a few practice trials in order to ensure that
participants fully understand the instructions. The entire experiment consists
of 3 games with 40 trials that include 20 control trials per game.
Besides the trust game, in the scanner the Social Mindfulness task (SoMi) will
be performed. In this paradigm, choices are made in favour of oneself or
keeping in mind the interest of another person. 4 Identical (but for one detail
as colour or shape) items are shown, food and non-food. The items are shown in
a ratio 2:2 (2 red and 2 green apples) or 3:1 (3 blue pencils and 1 green
pencil). The 2:2 ratio is the controlcondition, the choice does not influence
the options the second (virtual) player has. In the ratio 3:1 the choice does
influence the options for the second player: if the subject chooses the single
item, the second player has no choice; does he choose one of 3, the second
player still has the choice between 2 different items. In short: does the
subject choose for himself (unmindful) or does he consider the options of the
other and mae a mindful choice?
Every item is presented in the choice (3:1) and in the control (2:2) condition,
making analysis possible as to what he really likes and if he chooses that
option in the choice-condition or not (mindful). Thus we can control for
personal preferences (Van Doesum & Van Lange, submitted).

In the week prior to the testing/scanning, participants receive an IPod touch
from the VU University. This device contains a questionnaire that participants
are requested to fill in 10 times a day, for 7 consecutive days, at randomly
chosen moments between 7.30 and 22.30. The IPod will beep (give a short alarm)
and participants are to fill in the questionnaire directly after that.
Questions are asked about activities, environment, mood, stress and consumption
(including drugs and alcohol). Psychosis-related symptoms will also be
assessed.

The burden and risks associated with participation in this study are minimal.
Subjects will be exposed to a high static magnetic field of 3 Tesla. Ample
experience with such field strengths yields no indication of adverse effects on
humans. In order to protect the subjects* hearing, protective headphones will
be provided.