Neural representation of prediction error signals in substance users

Addict Biol. 2021 May;26(3):e12976. doi: 10.1111/adb.12976. Epub 2020 Nov 24.


Abnormal decision making can result in detrimental outcomes of clinical importance, and decision making is strongly linked to neural prediction error signalling. Activation likelihood estimation (ALE) meta-analyses were used to examine the neural correlates of prediction error signals of individuals taking different types of substances and healthy controls with contrast and conjunction analyses. Twenty-eight studies were included in the meta-analysis, representing 424 substance users' individuals and 834 healthy control individuals. Robust brain activity associated with prediction error signals in substance users was found for the bilateral striatum and insula. Healthy control subjects also activated bilateral striatum, midbrain, right insula and right medial-inferior frontal gyrus. Compared with healthy controls, substance users showed blunted activity in the bilateral putamen, right medial-inferior frontal gyrus and insula. The current meta-analysis of cross-sectional findings investigated neural prediction error signals in substance users. PE abnormalities in substance users might be related to poor decision making. In conclusion, the present study helps identify the pathophysiological underpinnings of maladaptive decision making in substance users.

Keywords: addiction; fMRI; meta-analysis; prediction error signals; striatum; substance use.

Publication types

  • Meta-Analysis
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Brain / physiopathology*
  • Case-Control Studies
  • Decision Making / physiology
  • Female
  • Humans
  • Linear Models
  • Magnetic Resonance Imaging
  • Male
  • Substance-Related Disorders / physiopathology*
  • Young Adult