Computational markers show specific deficits for dyslexia and ADHD in complex learning settings

Area of Science:

• Cognitive Science
• Neuroscience
• Computational Psychology

Background:

• Neurodevelopmental disorders, including ADHD and dyslexia, present challenges in learning and cognitive processing.
• Understanding how these conditions affect learning in complex, uncertain environments is crucial for developing targeted interventions.
• Previous research often focuses on specific deficits, but less is known about real-world learning complexities.

Purpose of the Study:

• To investigate learning in multidimensional environments for individuals with ADHD and dyslexia.
• To compare performance on complex and simple reinforcement learning tasks.
• To identify distinct computational mechanisms underlying learning differences in neurodevelopmental disorders.

Main Methods:

• Participants with ADHD and dyslexia, and controls, completed complex and simple reinforcement learning tasks.
• The complex task required discovering a hidden reward-predictive dimension.
• Computational modeling (Bayesian inference, decay rates) was used to analyze learning strategies.

Main Results:

• Individuals with ADHD and dyslexia performed worse than controls on both tasks.
• ADHD group showed reduced use of optimal Bayesian inference strategies.
• Dyslexia group exhibited higher decay rates, discounting recent information faster.

Conclusions:

• ADHD and dyslexia are associated with distinct computational impairments in learning.
• Reduced Bayesian inference in ADHD and increased decay in dyslexia offer insight into neurodevelopmental learning challenges.
• These findings highlight unique computational markers for neurodevelopmental disorders in naturalistic learning contexts.