Disrupted reinforcement learning during post-error slowing in ADHD | JoVE Visualize

Area of Science:

Neuroscience
Cognitive Science
Developmental Psychology

Background:

Attention-Deficit/Hyperactivity Disorder (ADHD) is linked to altered dopamine-regulated reinforcement learning, particularly concerning prediction errors.
While neuroimaging studies have detailed normal reinforcement learning, specific neural mechanisms, like ventral striatum's thresholding influence on substantia nigra and dorsal striatum, remain unidentified in ADHD.

Purpose of the Study:

To identify error magnitude-related thresholding influences in the ventral striatum during reinforcement learning in adolescents with and without ADHD.
To differentiate error detection from post-error adjustment processes using functional magnetic resonance imaging (fMRI) in the stop-signal task.

Main Methods:

Utilized fMRI to examine brain activity during an inhibition task (stop-signal task) in typically developing (TD) adolescents and those with ADHD.
Separated neural activity related to error detection from activity associated with post-error adjustments (e.g., post-error slowing).
Analyzed correlations between activity in the dorsal striatum, ventral striatum, amygdala, and other neurotransmitter nuclei (locus coeruleus, raphe nucleus, medial septal nuclei).

Main Results:

In TD adolescents, deactivation in the dorsal striatum during error detection correlated with response-phase activity, and ventral striatum deactivation during post-error slowing correlated with dorsal striatum activation, consistent with dopamine-regulated learning.
In adolescents with ADHD, ventral striatum activity correlated with heightened amygdala responses, suggesting impaired reinforcement learning.
Altered correlations were observed between neurotransmitter nuclei (locus coeruleus, raphe nucleus, medial septal nuclei) and the substantia nigra in ADHD, further indicating disrupted learning mechanisms.

Conclusions:

Findings in TD individuals support dopamine-regulated reinforcement learning models for post-error adjustments.
In ADHD, heightened amygdala activity and altered neurotransmitter systems appear to impede normal reinforcement learning processes.
The study successfully identified distinct neural correlates of error detection and post-error adjustments, highlighting differences between TD and ADHD adolescents.