Striatal response to negative feedback in a stop signal task operates as a multi-value learning signal

Area of Science:

• Neuroscience
• Cognitive Psychology
• Functional Magnetic Resonance Imaging (fMRI)

Background:

• Error-driven learning is crucial for adapting behavior based on mistakes.
• The stop signal task is a standard paradigm for studying response inhibition.
• Previous research has explored neural correlates of learning, but a robust characterization is needed.

Purpose of the Study:

• To investigate error-driven learning using fMRI data from 217 subjects performing a stop signal task.
• To characterize the relationship between behavioral learning measures and neural signals.
• To determine if striatal activity encodes specific learning signals within the stop signal task.

Main Methods:

• Utilized functional magnetic resonance imaging (fMRI) to measure brain activity (BOLD signal) in 217 participants.
• Employed a stop signal task requiring participants to inhibit responses.
• Compared brain activity and behavioral responses following successful versus failed inhibition trials.

Main Results:

• Found significantly greater bilateral striatal activity after correct stop trials compared to failed stop trials, particularly in the putamen.
• Correlated median striatal activity with learning signals for stop signal probability and latency.
• Mixed-effects models revealed that changes in reaction time and stop trial probability significantly predicted striatal activity.

Conclusions:

• Striatal activity appears to function as a learning signal encoding changes in reaction time and the probability of stop signal occurrence.
• This finding extends the understanding of the striatum's role in encoding reward prediction error signals.
• Demonstrates that striatal signals encode both changes in stop signal probability and delay within the stop signal task.