Fig. 3: Displaying the relationships between subclinical symptoms and task readouts in Experiment 1.

A shows the relationship between average block-wise score and paranoia scores, including standard error. B shows the relationship between average block-wise performance error and paranoia scores, including standard error. C shows the relationship between average block-wise performance error and delusion ideation scores. Specifically, results shown for the noise (blue) and volatility (brown) blocks demonstrating the interaction found for individuals scoring at the upper and lower end of the delusions symptom scale. D shows the the relationship between the averaged block-wise beam widths and paranoia scores in high and low noise conditions. Results show that individuals with paranoia use a significantly wider beam in low noise, and a narrower beam in high noise. E shows the relationship between learning rates and psychometric schizotypy for each block. Significance indicated for planned contrast noise block vs volatility block, showing that high and low schizotypy significantly differ in their learning rate use between these blocks. F Left: shows the learning rates in stable and volatile blocks. Each block dot is one person, and their data is joined between conditions with a straight black line. The mean for each condition is indicated in red. Right: As with FA but the learning rate difference is the difference between noisy trials and volatile trials. G shows the learning rate difference between volatile and stable blocks for each individual. The learning rate difference is then correlated with psychosis-trait symptom scores paranoia (red), delusional ideation (light blue), and psychometric schizotypy (purple). H As with (G), this plot shows the learning rate difference (volatile − noise) and symptom scores paranoia (red), delusional ideation (light blue), and psychometric schizotypy (purple).