Table 1.

Studies on aberrant salience.

Study	Study Type	Sample Sizes	Clinical Sample(s) Characteristics	Experimental Task	Aberrant Salience Behavioural Results	Main Neuroimaging Results—Group Comparisons	Additional Neuroimaging Results
Esslinger et al., 2012 [59]	A comparative study of FEP and HC; an fMRI study	FEP = 27 HC = 27	Convenience sample from an admissions centre for a mental health hospital; never medicated	Implicit Salience Attribution Task (famous and non-famous faces)	-No significant effect reported for face x group interaction -“Three way interaction indicated that patients profited less from famousness once the stimuli were colourful (control task condition, colourful vs dull)”—Esslinger et al., 2012, p. 117.	-“There were no significant differences in activation between patients and controls regarding the contrasts of interest in the whole brain or ROI analyses”—Esslinger et al., 2012, p. 118	-A correlation in the bilateral ventral striatum ROI between activation in two tasks contrasts (monetary incentive vs control in monetary incentive task) vs (famous vs non-famous face presentation). Correlations for peak voxels revealed significant relationships between contrasts in FEP, right ventral striatum: r = 0.58, left ventral striatum: r = 0.48. Correlations in the HC were not significant. Additionally, correlation coefficients were not significantly different between groups.
Katthagen et al., 2018 [55]	A comparative study of SCH and HC; an fMRI study	SCH = 42 HC = 42	Convenience sample from inpatient and outpatient units; all participants had antipsychotic medication	Implicit Salience Paradigm	-SCH showed increased implicit aberrant salience compared to HC (d = 0.39) -SCH showed greater irrelevance bias (described in the paper) compared to HC (d = 0.55)	Not applicable	-relevance weighted prediction error (described in the paper) correlated with increased response in the anterior cingulate cortex in all participants -relevance weighted absolute prediction errors (described in the paper) correlated with decreased response in the L-hippocampus in all participants -irrelevance bias correlated with decreased relevance weighted prediction errors in bilateral nucleus accumbens response
Pankow et al., 2016 [31]	A comparative study of SCH, subclinical delusions and HC; an fMRI study	SCH = 29 subclinical delusions = 24 HC = 50	SCH: a convenience sample from a hospital department of psychiatry and psychotherapy. Most of the patients were medicated; subclinical delusions: people with results in 4Q of PDI from a large internet sample	Salience Attribution Test	-Implicit Aberrant Salience: SCH showed significantly greater IAbS in comparison to HC (d = 0.68), but not subclinical delusions group (d = 0.56 -Explicit Aberrant Salience: no significant differences -Implicit Adaptive Salience: subclinical delusions group showed decreased IAdS salience in comparison to HC (d = −0.85), but not SCH (d = −0.14); difference between SCH and HC was also not significant (d = −0.52) -Explicit Adaptive Salience: SCH showed significantly decreased EAdS in comparison to HC (d = −1.40) and subclinical delusions group (d = −1.24)	Not applicable	-A significant relationship between aberrant salience score in SCH related to self-referential task activation in the ventromedial prefrontal cortex (r = −0.60) -No such correlation was observed for subclinical delusions group and HC
Walter et al., 2016 [36]	A comparative study of SSD with lower and higher levels of positive symptoms; an fMRI study	SSD with higher positive symptoms = 21 SSD with lower positive symptoms = 21	A convenience sample of patients with FEP and chronic SCH patients with a history of violent offense; most of the patients were medicated	Salience Attribution Test	-Implicit Aberrant Salience: no significant differences -Explicit Aberrant Salience: no significant differences -Implicit Adaptive Salience: no data -Explicit Adaptive Salience: no data	-Aberrant salience was associated with increased BOLD response in left insula in the higher positive symptoms group	-Aberrant reward prediction signals in the left posterior insula correlated negatively with the chlorpromazine-equivalent doses (r = −0.31)
Smieskova et al., 2015 [56]	A comparative study of ARMS, FEP and HC; an fMRI study	FEP with medication = 12 FEP without medication = 17 ARMS = 34 HC = 19	A convenience sample of patients with FEP (medicated and unmedicated) and ARMS patients	Salience Attribution Test	-Implicit Aberrant Salience: no significant differences -Explicit Aberrant Salience: a trend result (p = 0.096) for group effect (FEP with medication > HC > FEP without medication > ARMS) (d for FEP with medication > ARMS = 0.88) -Implicit Adaptive Salience: no significant differences -Explicit Adaptive Salience: no significant differences	-In aberrant salience, FEP without medication showed higher BOLD response in comparison to HC in the R-cuneus and R-middle occipital gyrus -In aberrant salience, HC showed higher BOLD response in comparison to ARMS and FEP in L-inferior parietal lobule -In adaptive salience, ARMS showed lower BOLD response in comparison to HC in R-supramarginal gyrus and R-inferior parietal lobule -In adaptive salience, FEP without medication showed lower BOLD response in comparison to HC in dorsal part of L-anterior cingulate gyrus, L-middle frontal gyrus and L-precentral gyrus -In adaptive salience, FEP with medication showed lower BOLD response in comparison to HC in R-insula and R-precentral gyrus, R-paracingulate gyrus, and R-anterior cingulate gyrus -In adaptive salience, combined FEP groups showed lower BOLD response in comparison to HC in R-precentral gyrus and R-insula -In adaptive salience, FEP with medication showed lower BOLD response in comparison to ARMS in bilateral paracingulate gyri and L-anterior cingulate gyrus	-Negative correlation between R-insular adaptive salience activation and hallucinations in FEP without medication (r = −0.64)
Roiser et al., 2013 [33]	Comparative study of UHR and HC; an fMRI and PET study	UHR = 18 HC = 18	A convenience sample; 2 participants with antipsychotic medication (authors checked that data from these participants did not alter the results)	Salience Attribution Test	-Implicit Aberrant Salience: no significant differences -Explicit Aberrant Salience: UHR showed greater EAbS in comparison to HC (d = 0.88) -Implicit Adaptive Salience: no significant differences -Explicit Adaptive Salience: no significant differences	-No significant group differences	-In both groups collapsed, there was a positive relationship between aberrant reward prediction response in the ventral striatum and explicit aberrant salience, “however, the slope of the regression line was significantly flatter and nonsignificant in UHR” (Roiser et al., 2013, p. 1331) -Aberrant reward prediction in the hippocampus was positively correlated with dorsal striatal dopamine synthesis capacity in HC (r = 0.65) and negatively in the UHR (r = −0.52) in a peak voxel -Adaptive reward prediction response in the caudate was positively correlated with ventral striatum dopamine levels in HC (r = 0.63) and negatively in UHR (r = −0.63) in a peak voxel
Schmidt et al., 2017 [34]	A longitudinal study of UHR and HC; an fMRI study	UHR = 23 HC = 13	A convenience sample from a clinical service for people at high risk of psychosis	Salience Attribution Test	-Implicit Aberrant Salience: in both timepoints UHR showed significantly higher IAbS than HC (d = 0.93); at follow-up, the UHR group showed significantly higher IAbS (d = 1.25) (drop of IAbS in HC), but not at baseline -Explicit Aberrant Salience: at baseline, UHR showed significantly higher EAbS (d = 0.78), but not at follow-up -Implicit Adaptive Salience: in both timepoints, UHR showed lower IAdS (d = −1.21); at baseline significantly lower IAdS (d = −1.33) and at follow-up at trend level (p = 0.062, d = −0.69) -Explicit Adaptive Salience: UHR showed lower EAdS (d = −0.84) in both timepoints; significantly lower EAdS (d = −1.07) at baseline, but not at follow-up (d = −0.42)	-No significant effects of group or time, and no group x time interactions, significant group differences in aberrant reward prediction at either baseline or follow-up -No significant group × time interactions for adaptive reward prediction -In both timepoints, UHR showed weaker BOLD response than HC in the ventral striatum, calcarine sulcus and midbrain bilaterally and in the left cuneus and middle temporal gyrus in adaptive reward prediction -At baseline UHR showed significantly less activation in the ventral striatum bilaterally and the left parahippocampal and middle temporal gyrus, and cerebellum during adaptive reward prediction -At follow-up the UHR showed significantly less activation in the ventral striatum bilaterally during adaptive reward prediction	-No significant relationships between changes in clinical features and longitudinal changes in brain activation during aberrant reward prediction -Improvement in abnormal beliefs over time was associated with the increase in activation during adaptive reward prediction in the right ventral striatum and in the supplementary motor cortex bilaterally (peak-voxel correlation)
Winton-Brown et al., 2017 [27]	A comparative study of UHR and HC; an fMRI study	UHR = 29 HC = 32	A convenience sample from a clinical service for people at high risk of psychosis	Salience Integration Task	-No significant differences for reward, novelty and aversion conditions	-UHR showed significantly greater BOLD response to reward-predicting cues than HC in the L-ventral pallidum and L-midbrain; no areas where HC showed greater activation than UHR- No significant differences for novelty condition -No significant differences for aversion condition	-Reward-induced modulation of ventral striatum/pallidum to midbrain connectivity was significantly greater in UHR than in HC -Described above modulation was correlated with CAARMS unusual thought content in UHR (r = 0.50), but not other positive symptoms subscales -No significant correlations between group differences in activation related to reward processing and CAARMS positive symptoms

FEP—first episode of psychosis group, HC—healthy controls group, SCH—schizophrenia or schizophrenia spectrum disorder group, ARMS—at risk mental states group, UHR—clinical ultra-high risk of psychosis group, IAbS—implicit aberrant salience condition, EAbS—explicit aberrant salience condition, IAdS—implicit adaptive salience condition, EAdS—explicit adaptive salience condition, CAARMS—Comprehensive Assessment of at Risk Mental States.