Abstract
Background
Schizophrenia spectrum disorders and depression have been associated with reductions in brain activation during reward anticipation. It is not known whether brain signals associated with reward anticipation relate to psychopathology dimensions of depression or schizophrenia in childhood prior to adolescence.
Method
We examined whether fMRI brain correlates of reward anticipation related to psychotic-like experiences and symptoms of depression, in 2129 children from the ABCD study aged 9-10 years.Psychotic-like experiences and depression were assessed using the Prodromal Questionnaire Brief Child version and the K-SADS. We fused regional MRI summary statistics for reward anticipation activation in the ABCD study data release 1.0 (contrast of expected large reward versus neutral expectation). Relations between brain activation and psychopathology were assessed using linear regressions in R for 82 brain regions, corrected for multiple comparisons for the number of regions using false discovery rate.
Results
From several regressions, there was an isolated unilateral association between right parsorbitalis activation and psychotic-like experiences, but no other significant associations between brain activation and psychopathology.
Conclusions
In 9-10 year old children, reward anticipation is not strongly related to psychotic-like experiences or depression. As previous evidence links depression and schizophrenia to reduced reward anticipation in adults and older adolescents, it appears likely that such associations develop over the adolescent period: this can be tested in follow-up studies of the ABCD cohort.
Introduction
Dysfunction of the neural circuits underpinning reward processing has been hypothesised to relate to the pathogenesis of the symptoms of schizophrenia \cite{Ziauddeen_2010}\cite{Heinz2002}\cite{Kring2014}. Studies in adult patients with schizophrenia have indicated reduced brain activation during reward anticipation, especially in the ventral striatum \cite{Juckel2006}\cite{Hägele2015}\cite{Arrondo2015}\cite{Radua2015}. The Monetary Incentive Delay (MID) task \cite{Knutson2000} is the functional MRI paradigm that has been used most extensively in both health and mental disorder to probe the neural basis of reward anticipation. Within patients with schizophrenia, inter-individual variability in the degree of reduced striatal activation on the MID has been associated with the severity of negative symptoms \cite{Radua2015}, consistent with the possibility that ventral striatal dysfunction in schizophrenia compromises reward anticipation, leading to real world deficits in motivation and/or enjoyment in this disorder. A recent study, however, found no evidence of ventral striatal dysfunction in help-seeking young adults with prodromal symptoms of psychosis \cite{Michielse_2017}. However, reduced brain reward anticipation in the MID has also been shown in in other disorders, and may relate to severity of depressive symptoms in a transdiagnostic fashion \cite{Hägele2015}\cite{Arrondo2015}.
Thus far, the majority of clinical studies of reward anticipation using fMRI have examined older adolescents and adults. However, the largest study to date of the MID has been the IMAGEN study of around 2000 14 year olds \cite{Jia2016}. This study showed the MID task robustly activated a large network of brain regions, including the ventral stratum and medial prefrontal cortex, in addition to several other regions, as has been shown previously in adults. In IMAGEN, current clinical depression (n=22) or sub threshold depression (n=101) at age 14 was associated with reduced ventral striatal activation compared to a sample of 123 healthy subjects matched in age, sex handedness and imaging site \cite{Stringaris2015}; furthermore, lower ventral striatal activation at age 14 in healthy adolescents was associated with an increased risk of transition to sub threshold or clinical depression at age 16 in this sample. Whilst anhedonia was associated with reduced ventral striatal activation, accentuated in the presence of both anhedonia and low mood, the presence of low mood without anhedonia was associated with normal ventral striatal reward anticipation. A recent analysis of IMAGEN data showed that 149 individuals with high levels of psychotic like experiences at age 19 had reduced reward related brain activity at age 14 on the MID \cite{Papanastasiou_2018}. Taken together, these findings indicate that reduced reward anticipation on the MID task is associated with depression and schizophrenia in adults, and is related to anhedonia, clinical depression status and possibly depression or psychotic-like symptom pathogenesis in mid-adolescence. Whilst the aforementioned studies have shown associations between lower striatal activation during reward anticipation and psychopathology, other studies have suggested that increased activation during reward anticipation may be associated with psychopathology. For example, a recent study \cite{Lahat2018} showed that greater striatal (caudate) activation during reward anticipation was associated with greater social anxiety in a sample of 40 children age 10-13.
However, it is not yet known at what stage in psychotic or depressive illness do the neural signals of reward anticipation become compromised. No previous large study has used the adult MID task investigate neural reward anticipation signals in early or middle childhood, although "child-friendly" modified versions of the MID task have been shown to engage similar brain regions to the adults MID task \cite{Helfinstein2013}. Furthermore, no large scale study has examined the relation of reward anticipation to psychopathology in the pre-adolescent stage. A complete understanding of the relationships between reward anticipation and the emergence of psychopathology and mental illness will require examination of precision in the relationships between neutral processing of reward anticipation and the the timeframe of the development of psychopathology in adolescence and young adulthood. A necessary step is to show that the MID engages the classic reward processing network in pre-adolecents and whether or not it relates to psychopathology (especially depression and psychotic-like experiences) before adolescence. Preliminary analysis of the reward receipt phase of the MID task in the large Adolescent Brain Cognitive Development (ABCD) study \cite{Jernigan2018} suggested it produced similar task activations to the those previously seen in adults \cite{Casey2018}. We therefore used the publicly available ABCD dataset to investigate the following questions:
1) Does the MID task as used in the ABCD study engage similar brain regions during reward anticipation in the pre-adolescent period as previously reported in adolescence and adulthood?
2)Is reduced brain activation in reward anticipation associated with severity or presence of psychotic-like experiences in the pre-adolescent stage?
3) Is reduced brain activation in reward anticipation association with the severity or presence of depressive symptoms in the pre-adolescent stage?
4) Is reduced brain activation in reward anticipation associated with anhedonia in the pre-adolescent stage?
Methods
Subjects
3923 children aged 9-11 years from the ABCD dataset (Data Release 1.0) were initially included. These subjects were drawn from 21 centres throughout the US, with participants largely recruited through the school system. Sampling plans and recruitment procedures based on considerations of age, gender, race, socio-economic status and urbanicity were designed to reflect the sociodemographics of the US. Details of recruitment and study design are described elsewhere \cite{Garavan_2018}. Details of demographic, physical and mental health assessments are described elsewhere \citep{Barch_2018}. Institutional review board approval was obtained for each site before data collection. All parents provided written informed consent and all children provided assent. Task imaging data was available for 3179 subjects. Subjects with a diagnosis of ADHD, autism spectrum disorder, schizophrenia and intellectual disability, or who were under-weight as defined by CDC, were excluded from analysis, as were subjects who had poor task performance (see below). A total of 2129 subjects were ultimately included in the analysis. A breakdown of demographic data is described in Table \ref{421683}.